rust: run a clippy pass with the latest stable version

Our current version of clippy is older than the latest stable.
The newest version has new lints that are moslty good advice, so let's apply
them ahead of time. This has the added benefit of reducing the noise for
developpers like myself that use clippy as an IDE helper, as well as being
more prepared for a future clippy upgrade.

rust/hg-core/src/config/layer.rs

@@ -1,349 +1,349 b''
 // layer.rs
 //
 // Copyright 2020
 //      Valentin Gatien-Baron,
 //      Raphaël Gomès <[e-mail]>
 //
 // This software may be used and distributed according to the terms of the
 // GNU General Public License version 2 or any later version.
 
 use crate::errors::HgError;
 use crate::exit_codes::CONFIG_PARSE_ERROR_ABORT;
 use crate::utils::files::{get_bytes_from_path, get_path_from_bytes};
 use format_bytes::{format_bytes, write_bytes, DisplayBytes};
 use lazy_static::lazy_static;
 use regex::bytes::Regex;
 use std::collections::HashMap;
 use std::path::{Path, PathBuf};
 
 lazy_static! {
     static ref SECTION_RE: Regex = make_regex(r"^\[([^\[]+)\]");
     static ref ITEM_RE: Regex = make_regex(r"^([^=\s][^=]*?)\s*=\s*((.*\S)?)");
     /// Continuation whitespace
     static ref CONT_RE: Regex = make_regex(r"^\s+(\S|\S.*\S)\s*$");
     static ref EMPTY_RE: Regex = make_regex(r"^(;|#|\s*$)");
     static ref COMMENT_RE: Regex = make_regex(r"^(;|#)");
     /// A directive that allows for removing previous entries
     static ref UNSET_RE: Regex = make_regex(r"^%unset\s+(\S+)");
     /// A directive that allows for including other config files
     static ref INCLUDE_RE: Regex = make_regex(r"^%include\s+(\S|\S.*\S)\s*$");
 }
 
 /// All config values separated by layers of precedence.
 /// Each config source may be split in multiple layers if `%include` directives
 /// are used.
 /// TODO detail the general precedence
 #[derive(Clone)]
 pub struct ConfigLayer {
     /// Mapping of the sections to their items
     sections: HashMap<Vec<u8>, ConfigItem>,
     /// All sections (and their items/values) in a layer share the same origin
     pub origin: ConfigOrigin,
     /// Whether this layer comes from a trusted user or group
     pub trusted: bool,
 }
 
 impl ConfigLayer {
     pub fn new(origin: ConfigOrigin) -> Self {
         ConfigLayer {
             sections: HashMap::new(),
             trusted: true, // TODO check
             origin,
         }
     }
 
     /// Parse `--config` CLI arguments and return a layer if there’s any
     pub(crate) fn parse_cli_args(
         cli_config_args: impl IntoIterator<Item = impl AsRef<[u8]>>,
     ) -> Result<Option<Self>, ConfigError> {
         fn parse_one(arg: &[u8]) -> Option<(Vec<u8>, Vec<u8>, Vec<u8>)> {
             use crate::utils::SliceExt;
 
             let (section_and_item, value) = arg.split_2(b'=')?;
             let (section, item) = section_and_item.trim().split_2(b'.')?;
             Some((
                 section.to_owned(),
                 item.to_owned(),
                 value.trim().to_owned(),
             ))
         }
 
         let mut layer = Self::new(ConfigOrigin::CommandLine);
         for arg in cli_config_args {
             let arg = arg.as_ref();
             if let Some((section, item, value)) = parse_one(arg) {
                 layer.add(section, item, value, None);
             } else {
                 Err(HgError::abort(
                     format!(
                         "abort: malformed --config option: '{}' \
                     (use --config section.name=value)",
                         String::from_utf8_lossy(arg),
                     ),
                     CONFIG_PARSE_ERROR_ABORT,
                     None,
                 ))?
             }
         }
         if layer.sections.is_empty() {
             Ok(None)
         } else {
             Ok(Some(layer))
         }
     }
 
     /// Returns whether this layer comes from `--config` CLI arguments
     pub(crate) fn is_from_command_line(&self) -> bool {
         matches!(self.origin, ConfigOrigin::CommandLine)
     }
 
     /// Add an entry to the config, overwriting the old one if already present.
     pub fn add(
         &mut self,
         section: Vec<u8>,
         item: Vec<u8>,
         value: Vec<u8>,
         line: Option<usize>,
     ) {
         self.sections
             .entry(section)
-            .or_insert_with(HashMap::new)
+            .or_default()
             .insert(item, ConfigValue { bytes: value, line });
     }
 
     /// Returns the config value in `<section>.<item>` if it exists
     pub fn get(&self, section: &[u8], item: &[u8]) -> Option<&ConfigValue> {
         self.sections.get(section)?.get(item)
     }
 
     /// Returns the keys defined in the given section
     pub fn iter_keys(&self, section: &[u8]) -> impl Iterator<Item = &[u8]> {
         self.sections
             .get(section)
             .into_iter()
             .flat_map(|section| section.keys().map(|vec| &**vec))
     }
 
     /// Returns the (key, value) pairs defined in the given section
     pub fn iter_section<'layer>(
         &'layer self,
         section: &[u8],
     ) -> impl Iterator<Item = (&'layer [u8], &'layer [u8])> {
         self.sections
             .get(section)
             .into_iter()
             .flat_map(|section| section.iter().map(|(k, v)| (&**k, &*v.bytes)))
     }
 
     /// Returns whether any key is defined in the given section
     pub fn has_non_empty_section(&self, section: &[u8]) -> bool {
         self.sections
             .get(section)
             .map_or(false, |section| !section.is_empty())
     }
 
     pub fn is_empty(&self) -> bool {
         self.sections.is_empty()
     }
 
     /// Returns a `Vec` of layers in order of precedence (so, in read order),
     /// recursively parsing the `%include` directives if any.
     pub fn parse(src: &Path, data: &[u8]) -> Result<Vec<Self>, ConfigError> {
         let mut layers = vec![];
 
         // Discard byte order mark if any
         let data = if data.starts_with(b"\xef\xbb\xbf") {
             &data[3..]
         } else {
             data
         };
 
         // TODO check if it's trusted
         let mut current_layer = Self::new(ConfigOrigin::File(src.to_owned()));
 
         let mut lines_iter =
             data.split(|b| *b == b'\n').enumerate().peekable();
         let mut section = b"".to_vec();
 
         while let Some((index, bytes)) = lines_iter.next() {
             let line = Some(index + 1);
             if let Some(m) = INCLUDE_RE.captures(bytes) {
                 let filename_bytes = &m[1];
                 let filename_bytes = crate::utils::expand_vars(filename_bytes);
                 // `Path::parent` only fails for the root directory,
                 // which `src` can’t be since we’ve managed to open it as a
                 // file.
                 let dir = src
                     .parent()
                     .expect("Path::parent fail on a file we’ve read");
                 // `Path::join` with an absolute argument correctly ignores the
                 // base path
-                let filename = dir.join(&get_path_from_bytes(&filename_bytes));
+                let filename = dir.join(get_path_from_bytes(&filename_bytes));
                 match std::fs::read(&filename) {
                     Ok(data) => {
                         layers.push(current_layer);
                         layers.extend(Self::parse(&filename, &data)?);
                         current_layer =
                             Self::new(ConfigOrigin::File(src.to_owned()));
                     }
                     Err(error) => {
                         if error.kind() != std::io::ErrorKind::NotFound {
                             return Err(ConfigParseError {
                                 origin: ConfigOrigin::File(src.to_owned()),
                                 line,
                                 message: format_bytes!(
                                     b"cannot include {} ({})",
                                     filename_bytes,
                                     format_bytes::Utf8(error)
                                 ),
                             }
                             .into());
                         }
                     }
                 }
             } else if EMPTY_RE.captures(bytes).is_some() {
             } else if let Some(m) = SECTION_RE.captures(bytes) {
                 section = m[1].to_vec();
             } else if let Some(m) = ITEM_RE.captures(bytes) {
                 let item = m[1].to_vec();
                 let mut value = m[2].to_vec();
                 loop {
                     match lines_iter.peek() {
                         None => break,
                         Some((_, v)) => {
                             if COMMENT_RE.captures(v).is_some() {
                             } else if CONT_RE.captures(v).is_some() {
                                 value.extend(b"\n");
                                 value.extend(&m[1]);
                             } else {
                                 break;
                             }
                         }
                     };
                     lines_iter.next();
                 }
                 current_layer.add(section.clone(), item, value, line);
             } else if let Some(m) = UNSET_RE.captures(bytes) {
                 if let Some(map) = current_layer.sections.get_mut(&section) {
                     map.remove(&m[1]);
                 }
             } else {
                 let message = if bytes.starts_with(b" ") {
                     format_bytes!(b"unexpected leading whitespace: {}", bytes)
                 } else {
                     bytes.to_owned()
                 };
                 return Err(ConfigParseError {
                     origin: ConfigOrigin::File(src.to_owned()),
                     line,
                     message,
                 }
                 .into());
             }
         }
         if !current_layer.is_empty() {
             layers.push(current_layer);
         }
         Ok(layers)
     }
 }
 
 impl DisplayBytes for ConfigLayer {
     fn display_bytes(
         &self,
         out: &mut dyn std::io::Write,
     ) -> std::io::Result<()> {
         let mut sections: Vec<_> = self.sections.iter().collect();
         sections.sort_by(|e0, e1| e0.0.cmp(e1.0));
 
         for (section, items) in sections.into_iter() {
             let mut items: Vec<_> = items.iter().collect();
             items.sort_by(|e0, e1| e0.0.cmp(e1.0));
 
             for (item, config_entry) in items {
                 write_bytes!(
                     out,
                     b"{}.{}={} # {}\n",
                     section,
                     item,
                     &config_entry.bytes,
                     &self.origin,
                 )?
             }
         }
         Ok(())
     }
 }
 
 /// Mapping of section item to value.
 /// In the following:
 /// ```text
 /// [ui]
 /// paginate=no
 /// ```
 /// "paginate" is the section item and "no" the value.
 pub type ConfigItem = HashMap<Vec<u8>, ConfigValue>;
 
 #[derive(Clone, Debug, PartialEq)]
 pub struct ConfigValue {
     /// The raw bytes of the value (be it from the CLI, env or from a file)
     pub bytes: Vec<u8>,
     /// Only present if the value comes from a file, 1-indexed.
     pub line: Option<usize>,
 }
 
 #[derive(Clone, Debug, PartialEq, Eq)]
 pub enum ConfigOrigin {
     /// From a configuration file
     File(PathBuf),
     /// From [ui.tweakdefaults]
     Tweakdefaults,
     /// From a `--config` CLI argument
     CommandLine,
     /// From a `--color` CLI argument
     CommandLineColor,
     /// From environment variables like `$PAGER` or `$EDITOR`
     Environment(Vec<u8>),
     /// From configitems.toml
     Defaults,
     /* TODO extensions
      * TODO Python resources?
      * Others? */
 }
 
 impl DisplayBytes for ConfigOrigin {
     fn display_bytes(
         &self,
         out: &mut dyn std::io::Write,
     ) -> std::io::Result<()> {
         match self {
             ConfigOrigin::File(p) => out.write_all(&get_bytes_from_path(p)),
             ConfigOrigin::CommandLine => out.write_all(b"--config"),
             ConfigOrigin::CommandLineColor => out.write_all(b"--color"),
             ConfigOrigin::Environment(e) => write_bytes!(out, b"${}", e),
             ConfigOrigin::Tweakdefaults => {
                 write_bytes!(out, b"ui.tweakdefaults")
             }
             ConfigOrigin::Defaults => {
                 write_bytes!(out, b"configitems.toml")
             }
         }
     }
 }
 
 #[derive(Debug)]
 pub struct ConfigParseError {
     pub origin: ConfigOrigin,
     pub line: Option<usize>,
     pub message: Vec<u8>,
 }
 
 #[derive(Debug, derive_more::From)]
 pub enum ConfigError {
     Parse(ConfigParseError),
     Other(HgError),
 }
 
 fn make_regex(pattern: &'static str) -> Regex {
     Regex::new(pattern).expect("expected a valid regex")
 }

rust/hg-core/src/config/mod.rs

@@ -1,810 +1,810 b''
 // config.rs
 //
 // Copyright 2020
 //      Valentin Gatien-Baron,
 //      Raphaël Gomès <[e-mail]>
 //
 // This software may be used and distributed according to the terms of the
 // GNU General Public License version 2 or any later version.
 
 //! Mercurial config parsing and interfaces.
 
 pub mod config_items;
 mod layer;
 mod plain_info;
 mod values;
 pub use layer::{ConfigError, ConfigOrigin, ConfigParseError};
 use lazy_static::lazy_static;
 pub use plain_info::PlainInfo;
 
 use self::config_items::DefaultConfig;
 use self::config_items::DefaultConfigItem;
 use self::layer::ConfigLayer;
 use self::layer::ConfigValue;
 use crate::errors::HgError;
 use crate::errors::{HgResultExt, IoResultExt};
 use crate::utils::files::get_bytes_from_os_str;
 use format_bytes::{write_bytes, DisplayBytes};
 use std::collections::HashSet;
 use std::env;
 use std::fmt;
 use std::path::{Path, PathBuf};
 use std::str;
 
 lazy_static! {
     static ref DEFAULT_CONFIG: Result<DefaultConfig, HgError> = {
         DefaultConfig::from_contents(include_str!(
             "../../../../mercurial/configitems.toml"
         ))
     };
 }
 
 /// Holds the config values for the current repository
 /// TODO update this docstring once we support more sources
 #[derive(Clone)]
 pub struct Config {
     layers: Vec<layer::ConfigLayer>,
     plain: PlainInfo,
 }
 
 impl DisplayBytes for Config {
     fn display_bytes(
         &self,
         out: &mut dyn std::io::Write,
     ) -> std::io::Result<()> {
         for (index, layer) in self.layers.iter().rev().enumerate() {
             write_bytes!(
                 out,
                 b"==== Layer {} (trusted: {}) ====\n{}",
                 index,
                 if layer.trusted {
                     &b"yes"[..]
                 } else {
                     &b"no"[..]
                 },
                 layer
             )?;
         }
         Ok(())
     }
 }
 
 pub enum ConfigSource {
     /// Absolute path to a config file
     AbsPath(PathBuf),
     /// Already parsed (from the CLI, env, Python resources, etc.)
     Parsed(layer::ConfigLayer),
 }
 
 #[derive(Debug)]
 pub struct ConfigValueParseErrorDetails {
     pub origin: ConfigOrigin,
     pub line: Option<usize>,
     pub section: Vec<u8>,
     pub item: Vec<u8>,
     pub value: Vec<u8>,
     pub expected_type: &'static str,
 }
 
 // boxed to avoid very large Result types
 pub type ConfigValueParseError = Box<ConfigValueParseErrorDetails>;
 
 impl fmt::Display for ConfigValueParseError {
     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
         // TODO: add origin and line number information, here and in
         // corresponding python code
         write!(
             f,
             "config error: {}.{} is not a {} ('{}')",
             String::from_utf8_lossy(&self.section),
             String::from_utf8_lossy(&self.item),
             self.expected_type,
             String::from_utf8_lossy(&self.value)
         )
     }
 }
 
 /// Returns true if the config item is disabled by PLAIN or PLAINEXCEPT
 fn should_ignore(plain: &PlainInfo, section: &[u8], item: &[u8]) -> bool {
     // duplication with [_applyconfig] in [ui.py],
     if !plain.is_plain() {
         return false;
     }
     if section == b"alias" {
         return plain.plainalias();
     }
     if section == b"revsetalias" {
         return plain.plainrevsetalias();
     }
     if section == b"templatealias" {
         return plain.plaintemplatealias();
     }
     if section == b"ui" {
         let to_delete: &[&[u8]] = &[
             b"debug",
             b"fallbackencoding",
             b"quiet",
             b"slash",
             b"logtemplate",
             b"message-output",
             b"statuscopies",
             b"style",
             b"traceback",
             b"verbose",
         ];
         return to_delete.contains(&item);
     }
     let sections_to_delete: &[&[u8]] =
         &[b"defaults", b"commands", b"command-templates"];
     sections_to_delete.contains(&section)
 }
 
 impl Config {
     /// The configuration to use when printing configuration-loading errors
     pub fn empty() -> Self {
         Self {
             layers: Vec::new(),
             plain: PlainInfo::empty(),
         }
     }
 
     /// Load system and user configuration from various files.
     ///
     /// This is also affected by some environment variables.
     pub fn load_non_repo() -> Result<Self, ConfigError> {
         let mut config = Self::empty();
         let opt_rc_path = env::var_os("HGRCPATH");
         // HGRCPATH replaces system config
         if opt_rc_path.is_none() {
             config.add_system_config()?
         }
 
         config.add_for_environment_variable("EDITOR", b"ui", b"editor");
         config.add_for_environment_variable("VISUAL", b"ui", b"editor");
         config.add_for_environment_variable("PAGER", b"pager", b"pager");
 
         // These are set by `run-tests.py --rhg` to enable fallback for the
         // entire test suite. Alternatives would be setting configuration
         // through `$HGRCPATH` but some tests override that, or changing the
         // `hg` shell alias to include `--config` but that disrupts tests that
         // print command lines and check expected output.
         config.add_for_environment_variable(
             "RHG_ON_UNSUPPORTED",
             b"rhg",
             b"on-unsupported",
         );
         config.add_for_environment_variable(
             "RHG_FALLBACK_EXECUTABLE",
             b"rhg",
             b"fallback-executable",
         );
 
         // HGRCPATH replaces user config
         if opt_rc_path.is_none() {
             config.add_user_config()?
         }
         if let Some(rc_path) = &opt_rc_path {
             for path in env::split_paths(rc_path) {
                 if !path.as_os_str().is_empty() {
                     if path.is_dir() {
                         config.add_trusted_dir(&path)?
                     } else {
                         config.add_trusted_file(&path)?
                     }
                 }
             }
         }
         Ok(config)
     }
 
     pub fn load_cli_args(
         &mut self,
         cli_config_args: impl IntoIterator<Item = impl AsRef<[u8]>>,
         color_arg: Option<Vec<u8>>,
     ) -> Result<(), ConfigError> {
         if let Some(layer) = ConfigLayer::parse_cli_args(cli_config_args)? {
             self.layers.push(layer)
         }
         if let Some(arg) = color_arg {
             let mut layer = ConfigLayer::new(ConfigOrigin::CommandLineColor);
             layer.add(b"ui"[..].into(), b"color"[..].into(), arg, None);
             self.layers.push(layer)
         }
         Ok(())
     }
 
     fn add_trusted_dir(&mut self, path: &Path) -> Result<(), ConfigError> {
         if let Some(entries) = std::fs::read_dir(path)
             .when_reading_file(path)
             .io_not_found_as_none()?
         {
             let mut file_paths = entries
                 .map(|result| {
                     result.when_reading_file(path).map(|entry| entry.path())
                 })
                 .collect::<Result<Vec<_>, _>>()?;
             file_paths.sort();
             for file_path in &file_paths {
                 if file_path.extension() == Some(std::ffi::OsStr::new("rc")) {
                     self.add_trusted_file(file_path)?
                 }
             }
         }
         Ok(())
     }
 
     fn add_trusted_file(&mut self, path: &Path) -> Result<(), ConfigError> {
         if let Some(data) = std::fs::read(path)
             .when_reading_file(path)
             .io_not_found_as_none()?
         {
             self.layers.extend(ConfigLayer::parse(path, &data)?)
         }
         Ok(())
     }
 
     fn add_for_environment_variable(
         &mut self,
         var: &str,
         section: &[u8],
         key: &[u8],
     ) {
         if let Some(value) = env::var_os(var) {
             let origin = layer::ConfigOrigin::Environment(var.into());
             let mut layer = ConfigLayer::new(origin);
             layer.add(
                 section.to_owned(),
                 key.to_owned(),
                 get_bytes_from_os_str(value),
                 None,
             );
             self.layers.push(layer)
         }
     }
 
     #[cfg(unix)] // TODO: other platforms
     fn add_system_config(&mut self) -> Result<(), ConfigError> {
         let mut add_for_prefix = |prefix: &Path| -> Result<(), ConfigError> {
             let etc = prefix.join("etc").join("mercurial");
             self.add_trusted_file(&etc.join("hgrc"))?;
             self.add_trusted_dir(&etc.join("hgrc.d"))
         };
         let root = Path::new("/");
         // TODO: use `std::env::args_os().next().unwrap()` a.k.a. argv[0]
         // instead? TODO: can this be a relative path?
         let hg = crate::utils::current_exe()?;
         // TODO: this order (per-installation then per-system) matches
         // `systemrcpath()` in `mercurial/scmposix.py`, but
         // `mercurial/helptext/config.txt` suggests it should be reversed
         if let Some(installation_prefix) = hg.parent().and_then(Path::parent) {
             if installation_prefix != root {
                 add_for_prefix(installation_prefix)?
             }
         }
         add_for_prefix(root)?;
         Ok(())
     }
 
     #[cfg(unix)] // TODO: other plateforms
     fn add_user_config(&mut self) -> Result<(), ConfigError> {
         let opt_home = home::home_dir();
         if let Some(home) = &opt_home {
             self.add_trusted_file(&home.join(".hgrc"))?
         }
         let darwin = cfg!(any(target_os = "macos", target_os = "ios"));
         if !darwin {
             if let Some(config_home) = env::var_os("XDG_CONFIG_HOME")
                 .map(PathBuf::from)
                 .or_else(|| opt_home.map(|home| home.join(".config")))
             {
                 self.add_trusted_file(&config_home.join("hg").join("hgrc"))?
             }
         }
         Ok(())
     }
 
     /// Loads in order, which means that the precedence is the same
     /// as the order of `sources`.
     pub fn load_from_explicit_sources(
         sources: Vec<ConfigSource>,
     ) -> Result<Self, ConfigError> {
         let mut layers = vec![];
 
         for source in sources.into_iter() {
             match source {
                 ConfigSource::Parsed(c) => layers.push(c),
                 ConfigSource::AbsPath(c) => {
                     // TODO check if it should be trusted
                     // mercurial/ui.py:427
                     let data = match std::fs::read(&c) {
                         Err(_) => continue, // same as the python code
                         Ok(data) => data,
                     };
                     layers.extend(ConfigLayer::parse(&c, &data)?)
                 }
             }
         }
 
         Ok(Config {
             layers,
             plain: PlainInfo::empty(),
         })
     }
 
     /// Loads the per-repository config into a new `Config` which is combined
     /// with `self`.
     pub(crate) fn combine_with_repo(
         &self,
         repo_config_files: &[PathBuf],
     ) -> Result<Self, ConfigError> {
         let (cli_layers, other_layers) = self
             .layers
             .iter()
             .cloned()
             .partition(ConfigLayer::is_from_command_line);
 
         let mut repo_config = Self {
             layers: other_layers,
             plain: PlainInfo::empty(),
         };
         for path in repo_config_files {
             // TODO: check if this file should be trusted:
             // `mercurial/ui.py:427`
             repo_config.add_trusted_file(path)?;
         }
         repo_config.layers.extend(cli_layers);
         Ok(repo_config)
     }
 
     pub fn apply_plain(&mut self, plain: PlainInfo) {
         self.plain = plain;
     }
 
     /// Returns the default value for the given config item, if any.
     pub fn get_default(
         &self,
         section: &[u8],
         item: &[u8],
     ) -> Result<Option<&DefaultConfigItem>, HgError> {
         let default_config = DEFAULT_CONFIG.as_ref().map_err(|e| {
             HgError::abort(
                 e.to_string(),
                 crate::exit_codes::ABORT,
                 Some("`mercurial/configitems.toml` is not valid".into()),
             )
         })?;
         let default_opt = default_config.get(section, item);
         Ok(default_opt.filter(|default| {
             default
                 .in_core_extension()
                 .map(|extension| {
                     // Only return the default for an in-core extension item
                     // if said extension is enabled
                     self.is_extension_enabled(extension.as_bytes())
                 })
                 .unwrap_or(true)
         }))
     }
 
     /// Return the config item that corresponds to a section + item, a function
     /// to parse from the raw bytes to the expected type (which is passed as
     /// a string only to make debugging easier).
     /// Used by higher-level methods like `get_bool`.
     ///
     /// `fallback_to_default` controls whether the default value (if any) is
     /// returned if nothing is found.
     fn get_parse<'config, T: 'config>(
         &'config self,
         section: &[u8],
         item: &[u8],
         expected_type: &'static str,
         parse: impl Fn(&'config [u8]) -> Option<T>,
         fallback_to_default: bool,
     ) -> Result<Option<T>, HgError>
     where
         Option<T>: TryFrom<&'config DefaultConfigItem, Error = HgError>,
     {
         match self.get_inner(section, item) {
             Some((layer, v)) => match parse(&v.bytes) {
                 Some(b) => Ok(Some(b)),
                 None => Err(Box::new(ConfigValueParseErrorDetails {
                     origin: layer.origin.to_owned(),
                     line: v.line,
                     value: v.bytes.to_owned(),
                     section: section.to_owned(),
                     item: item.to_owned(),
                     expected_type,
                 })
                 .into()),
             },
             None => {
                 if !fallback_to_default {
                     return Ok(None);
                 }
                 match self.get_default(section, item)? {
                     Some(default) => {
                         // Defaults are TOML values, so they're not in the same
                         // shape as in the config files.
                         // First try to convert directly to the expected type
                         let as_t = default.try_into();
                         match as_t {
                             Ok(t) => Ok(t),
                             Err(e) => {
                                 // If it fails, it means that...
                                 let as_bytes: Result<Option<&[u8]>, _> =
                                     default.try_into();
                                 match as_bytes {
                                     Ok(bytes_opt) => {
                                         if let Some(bytes) = bytes_opt {
                                             // ...we should be able to parse it
                                             return Ok(parse(bytes));
                                         }
                                         Err(e)
                                     }
                                     Err(_) => Err(e),
                                 }
                             }
                         }
                     }
                     None => {
                         self.print_devel_warning(section, item)?;
                         Ok(None)
                     }
                 }
             }
         }
     }
 
     fn print_devel_warning(
         &self,
         section: &[u8],
         item: &[u8],
     ) -> Result<(), HgError> {
         let warn_all = self.get_bool(b"devel", b"all-warnings")?;
         let warn_specific = self.get_bool(b"devel", b"warn-config-unknown")?;
         if !warn_all || !warn_specific {
             // We technically shouldn't print anything here since it's not
             // the concern of `hg-core`.
             //
             // We're printing directly to stderr since development warnings
             // are not on by default and surfacing this to consumer crates
             // (like `rhg`) would be more difficult, probably requiring
             // something à la `log` crate.
             //
             // TODO maybe figure out a way of exposing a "warnings" channel
             // that consumer crates can hook into. It would be useful for
             // all other warnings that `hg-core` could expose.
             eprintln!(
                 "devel-warn: accessing unregistered config item: '{}.{}'",
                 String::from_utf8_lossy(section),
                 String::from_utf8_lossy(item),
             );
         }
         Ok(())
     }
 
     /// Returns an `Err` if the first value found is not a valid UTF-8 string.
     /// Otherwise, returns an `Ok(value)` if found, or `None`.
     pub fn get_str(
         &self,
         section: &[u8],
         item: &[u8],
     ) -> Result<Option<&str>, HgError> {
         self.get_parse(
             section,
             item,
             "ASCII or UTF-8 string",
             |value| str::from_utf8(value).ok(),
             true,
         )
     }
 
     /// Same as `get_str`, but doesn't fall back to the default `configitem`
     /// if not defined in the user config.
     pub fn get_str_no_default(
         &self,
         section: &[u8],
         item: &[u8],
     ) -> Result<Option<&str>, HgError> {
         self.get_parse(
             section,
             item,
             "ASCII or UTF-8 string",
             |value| str::from_utf8(value).ok(),
             false,
         )
     }
 
     /// Returns an `Err` if the first value found is not a valid unsigned
     /// integer. Otherwise, returns an `Ok(value)` if found, or `None`.
     pub fn get_u32(
         &self,
         section: &[u8],
         item: &[u8],
     ) -> Result<Option<u32>, HgError> {
         self.get_parse(
             section,
             item,
             "valid integer",
             |value| str::from_utf8(value).ok()?.parse().ok(),
             true,
         )
     }
 
     /// Returns an `Err` if the first value found is not a valid file size
     /// value such as `30` (default unit is bytes), `7 MB`, or `42.5 kb`.
     /// Otherwise, returns an `Ok(value_in_bytes)` if found, or `None`.
     pub fn get_byte_size(
         &self,
         section: &[u8],
         item: &[u8],
     ) -> Result<Option<u64>, HgError> {
         self.get_parse(
             section,
             item,
             "byte quantity",
             values::parse_byte_size,
             true,
         )
     }
 
     /// Returns an `Err` if the first value found is not a valid boolean.
     /// Otherwise, returns an `Ok(option)`, where `option` is the boolean if
     /// found, or `None`.
     pub fn get_option(
         &self,
         section: &[u8],
         item: &[u8],
     ) -> Result<Option<bool>, HgError> {
         self.get_parse(section, item, "boolean", values::parse_bool, true)
     }
 
     /// Same as `get_option`, but doesn't fall back to the default `configitem`
     /// if not defined in the user config.
     pub fn get_option_no_default(
         &self,
         section: &[u8],
         item: &[u8],
     ) -> Result<Option<bool>, HgError> {
         self.get_parse(section, item, "boolean", values::parse_bool, false)
     }
 
     /// Returns the corresponding boolean in the config. Returns `Ok(false)`
     /// if the value is not found, an `Err` if it's not a valid boolean.
     pub fn get_bool(
         &self,
         section: &[u8],
         item: &[u8],
     ) -> Result<bool, HgError> {
         Ok(self.get_option(section, item)?.unwrap_or(false))
     }
 
     /// Same as `get_bool`, but doesn't fall back to the default `configitem`
     /// if not defined in the user config.
     pub fn get_bool_no_default(
         &self,
         section: &[u8],
         item: &[u8],
     ) -> Result<bool, HgError> {
         Ok(self.get_option_no_default(section, item)?.unwrap_or(false))
     }
 
     /// Returns `true` if the extension is enabled, `false` otherwise
     pub fn is_extension_enabled(&self, extension: &[u8]) -> bool {
         let value = self.get(b"extensions", extension);
         match value {
             Some(c) => !c.starts_with(b"!"),
             None => false,
         }
     }
 
     /// If there is an `item` value in `section`, parse and return a list of
     /// byte strings.
     pub fn get_list(
         &self,
         section: &[u8],
         item: &[u8],
     ) -> Option<Vec<Vec<u8>>> {
         self.get(section, item).map(values::parse_list)
     }
 
     /// Returns the raw value bytes of the first one found, or `None`.
     pub fn get(&self, section: &[u8], item: &[u8]) -> Option<&[u8]> {
         self.get_inner(section, item)
             .map(|(_, value)| value.bytes.as_ref())
     }
 
     /// Returns the raw value bytes of the first one found, or `None`.
     pub fn get_with_origin(
         &self,
         section: &[u8],
         item: &[u8],
     ) -> Option<(&[u8], &ConfigOrigin)> {
         self.get_inner(section, item)
             .map(|(layer, value)| (value.bytes.as_ref(), &layer.origin))
     }
 
     /// Returns the layer and the value of the first one found, or `None`.
     fn get_inner(
         &self,
         section: &[u8],
         item: &[u8],
     ) -> Option<(&ConfigLayer, &ConfigValue)> {
         // Filter out the config items that are hidden by [PLAIN].
         // This differs from python hg where we delete them from the config.
         let should_ignore = should_ignore(&self.plain, section, item);
         for layer in self.layers.iter().rev() {
             if !layer.trusted {
                 continue;
             }
             //The [PLAIN] config should not affect the defaults.
             //
             // However, PLAIN should also affect the "tweaked" defaults (unless
             // "tweakdefault" is part of "HGPLAINEXCEPT").
             //
             // In practice the tweak-default layer is only added when it is
             // relevant, so we can safely always take it into
             // account here.
             if should_ignore && !(layer.origin == ConfigOrigin::Tweakdefaults)
             {
                 continue;
             }
             if let Some(v) = layer.get(section, item) {
                 return Some((layer, v));
             }
         }
         None
     }
 
     /// Return all keys defined for the given section
     pub fn get_section_keys(&self, section: &[u8]) -> HashSet<&[u8]> {
         self.layers
             .iter()
             .flat_map(|layer| layer.iter_keys(section))
             .collect()
     }
 
     /// Returns whether any key is defined in the given section
     pub fn has_non_empty_section(&self, section: &[u8]) -> bool {
         self.layers
             .iter()
             .any(|layer| layer.has_non_empty_section(section))
     }
 
     /// Yields (key, value) pairs for everything in the given section
     pub fn iter_section<'a>(
         &'a self,
         section: &'a [u8],
     ) -> impl Iterator<Item = (&[u8], &[u8])> + 'a {
         // Deduplicate keys redefined in multiple layers
         let mut keys_already_seen = HashSet::new();
         let mut key_is_new =
             move |&(key, _value): &(&'a [u8], &'a [u8])| -> bool {
                 keys_already_seen.insert(key)
             };
         // This is similar to `flat_map` + `filter_map`, except with a single
         // closure that owns `key_is_new` (and therefore the
         // `keys_already_seen` set):
         let mut layer_iters = self
             .layers
             .iter()
             .rev()
             .map(move |layer| layer.iter_section(section))
             .peekable();
         std::iter::from_fn(move || loop {
             if let Some(pair) = layer_iters.peek_mut()?.find(&mut key_is_new) {
                 return Some(pair);
             } else {
                 layer_iters.next();
             }
         })
     }
 
     /// Get raw values bytes from all layers (even untrusted ones) in order
     /// of precedence.
     #[cfg(test)]
     fn get_all(&self, section: &[u8], item: &[u8]) -> Vec<&[u8]> {
         let mut res = vec![];
         for layer in self.layers.iter().rev() {
             if let Some(v) = layer.get(section, item) {
                 res.push(v.bytes.as_ref());
             }
         }
         res
     }
 
     // a config layer that's introduced by ui.tweakdefaults
     fn tweakdefaults_layer() -> ConfigLayer {
         let mut layer = ConfigLayer::new(ConfigOrigin::Tweakdefaults);
 
         let mut add = |section: &[u8], item: &[u8], value: &[u8]| {
             layer.add(
                 section[..].into(),
                 item[..].into(),
                 value[..].into(),
                 None,
             );
         };
         // duplication of [tweakrc] from [ui.py]
         add(b"ui", b"rollback", b"False");
         add(b"ui", b"statuscopies", b"yes");
         add(b"ui", b"interface", b"curses");
         add(b"ui", b"relative-paths", b"yes");
         add(b"commands", b"grep.all-files", b"True");
         add(b"commands", b"update.check", b"noconflict");
         add(b"commands", b"status.verbose", b"True");
         add(b"commands", b"resolve.explicit-re-merge", b"True");
         add(b"git", b"git", b"1");
         add(b"git", b"showfunc", b"1");
         add(b"git", b"word-diff", b"1");
         layer
     }
 
     // introduce the tweaked defaults as implied by ui.tweakdefaults
     pub fn tweakdefaults(&mut self) {
         self.layers.insert(0, Config::tweakdefaults_layer());
     }
 }
 
 #[cfg(test)]
 mod tests {
     use super::*;
     use pretty_assertions::assert_eq;
     use std::fs::File;
     use std::io::Write;
 
     #[test]
     fn test_include_layer_ordering() {
         let tmpdir = tempfile::tempdir().unwrap();
         let tmpdir_path = tmpdir.path();
         let mut included_file =
-            File::create(&tmpdir_path.join("included.rc")).unwrap();
+            File::create(tmpdir_path.join("included.rc")).unwrap();
 
         included_file.write_all(b"[section]\nitem=value1").unwrap();
         let base_config_path = tmpdir_path.join("base.rc");
         let mut config_file = File::create(&base_config_path).unwrap();
         let data =
             b"[section]\nitem=value0\n%include included.rc\nitem=value2\n\
               [section2]\ncount = 4\nsize = 1.5 KB\nnot-count = 1.5\nnot-size = 1 ub";
         config_file.write_all(data).unwrap();
 
         let sources = vec![ConfigSource::AbsPath(base_config_path)];
         let config = Config::load_from_explicit_sources(sources)
             .expect("expected valid config");
 
         let (_, value) = config.get_inner(b"section", b"item").unwrap();
         assert_eq!(
             value,
             &ConfigValue {
                 bytes: b"value2".to_vec(),
                 line: Some(4)
             }
         );
 
         let value = config.get(b"section", b"item").unwrap();
         assert_eq!(value, b"value2",);
         assert_eq!(
             config.get_all(b"section", b"item"),
             [b"value2", b"value1", b"value0"]
         );
 
         assert_eq!(config.get_u32(b"section2", b"count").unwrap(), Some(4));
         assert_eq!(
             config.get_byte_size(b"section2", b"size").unwrap(),
             Some(1024 + 512)
         );
         assert!(config.get_u32(b"section2", b"not-count").is_err());
         assert!(config.get_byte_size(b"section2", b"not-size").is_err());
     }
 
     #[test]
     fn test_default_parse() {
         let config = Config::load_from_explicit_sources(vec![])
             .expect("expected valid config");
         let ret = config.get_byte_size(b"cmdserver", b"max-log-size");
         assert!(ret.is_ok(), "{:?}", ret);
 
         let ret = config.get_byte_size(b"ui", b"formatted");
         assert!(ret.unwrap().is_none());
     }
 }

rust/hg-core/src/dagops.rs

@@ -1,286 +1,285 b''
 // dagops.rs
 //
 // Copyright 2019 Georges Racinet <[e-mail]>
 //
 // This software may be used and distributed according to the terms of the
 // GNU General Public License version 2 or any later version.
 
 //! Miscellaneous DAG operations
 //!
 //! # Terminology
 //! - By *relative heads* of a collection of revision numbers (`Revision`), we
 //!   mean those revisions that have no children among the collection.
 //! - Similarly *relative roots* of a collection of `Revision`, we mean those
 //!   whose parents, if any, don't belong to the collection.
 use super::{Graph, GraphError, Revision, NULL_REVISION};
 use crate::ancestors::AncestorsIterator;
 use std::collections::{BTreeSet, HashSet};
 
 fn remove_parents<S: std::hash::BuildHasher>(
     graph: &impl Graph,
     rev: Revision,
     set: &mut HashSet<Revision, S>,
 ) -> Result<(), GraphError> {
     for parent in graph.parents(rev)?.iter() {
         if *parent != NULL_REVISION {
             set.remove(parent);
         }
     }
     Ok(())
 }
 
 /// Relative heads out of some revisions, passed as an iterator.
 ///
 /// These heads are defined as those revisions that have no children
 /// among those emitted by the iterator.
 ///
 /// # Performance notes
 /// Internally, this clones the iterator, and builds a `HashSet` out of it.
 ///
 /// This function takes an `Iterator` instead of `impl IntoIterator` to
 /// guarantee that cloning the iterator doesn't result in cloning the full
 /// construct it comes from.
 pub fn heads<'a>(
     graph: &impl Graph,
     iter_revs: impl Clone + Iterator<Item = &'a Revision>,
 ) -> Result<HashSet<Revision>, GraphError> {
     let mut heads: HashSet<Revision> = iter_revs.clone().cloned().collect();
     heads.remove(&NULL_REVISION);
     for rev in iter_revs {
         if *rev != NULL_REVISION {
             remove_parents(graph, *rev, &mut heads)?;
         }
     }
     Ok(heads)
 }
 
 /// Retain in `revs` only its relative heads.
 ///
 /// This is an in-place operation, so that control of the incoming
 /// set is left to the caller.
 /// - a direct Python binding would probably need to build its own `HashSet`
 ///   from an incoming iterable, even if its sole purpose is to extract the
 ///   heads.
 /// - a Rust caller can decide whether cloning beforehand is appropriate
 ///
 /// # Performance notes
 /// Internally, this function will store a full copy of `revs` in a `Vec`.
 pub fn retain_heads<S: std::hash::BuildHasher>(
     graph: &impl Graph,
     revs: &mut HashSet<Revision, S>,
 ) -> Result<(), GraphError> {
     revs.remove(&NULL_REVISION);
     // we need to construct an iterable copy of revs to avoid itering while
     // mutating
     let as_vec: Vec<Revision> = revs.iter().cloned().collect();
     for rev in as_vec {
         if rev != NULL_REVISION {
             remove_parents(graph, rev, revs)?;
         }
     }
     Ok(())
 }
 
 /// Roots of `revs`, passed as a `HashSet`
 ///
 /// They are returned in arbitrary order
 pub fn roots<G: Graph, S: std::hash::BuildHasher>(
     graph: &G,
     revs: &HashSet<Revision, S>,
 ) -> Result<Vec<Revision>, GraphError> {
     let mut roots: Vec<Revision> = Vec::new();
     for rev in revs {
         if graph
             .parents(*rev)?
             .iter()
             .filter(|p| **p != NULL_REVISION)
             .all(|p| !revs.contains(p))
         {
             roots.push(*rev);
         }
     }
     Ok(roots)
 }
 
 /// Compute the topological range between two collections of revisions
 ///
 /// This is equivalent to the revset `<roots>::<heads>`.
 ///
 /// Currently, the given `Graph` has to implement `Clone`, which means
 /// actually cloning just a reference-counted Python pointer if
 /// it's passed over through `rust-cpython`. This is due to the internal
 /// use of `AncestorsIterator`
 ///
 /// # Algorithmic details
 ///
 /// This is a two-pass swipe inspired from what `reachableroots2` from
 /// `mercurial.cext.parsers` does to obtain the same results.
 ///
 /// - first, we climb up the DAG from `heads` in topological order, keeping
 ///   them in the vector `heads_ancestors` vector, and adding any element of
 ///   `roots` we find among them to the resulting range.
 /// - Then, we iterate on that recorded vector so that a revision is always
 ///   emitted after its parents and add all revisions whose parents are already
 ///   in the range to the results.
 ///
 /// # Performance notes
 ///
 /// The main difference with the C implementation is that
 /// the latter uses a flat array with bit flags, instead of complex structures
 /// like `HashSet`, making it faster in most scenarios. In theory, it's
 /// possible that the present implementation could be more memory efficient
 /// for very large repositories with many branches.
 pub fn range(
     graph: &(impl Graph + Clone),
     roots: impl IntoIterator<Item = Revision>,
     heads: impl IntoIterator<Item = Revision>,
 ) -> Result<BTreeSet<Revision>, GraphError> {
     let mut range = BTreeSet::new();
     let roots: HashSet<Revision> = roots.into_iter().collect();
     let min_root: Revision = match roots.iter().cloned().min() {
         None => {
             return Ok(range);
         }
         Some(r) => r,
     };
 
     // Internally, AncestorsIterator currently maintains a `HashSet`
     // of all seen revision, which is also what we record, albeit in an ordered
     // way. There's room for improvement on this duplication.
     let ait = AncestorsIterator::new(graph.clone(), heads, min_root, true)?;
     let mut heads_ancestors: Vec<Revision> = Vec::new();
     for revres in ait {
         let rev = revres?;
         if roots.contains(&rev) {
             range.insert(rev);
         }
         heads_ancestors.push(rev);
     }
 
     for rev in heads_ancestors.into_iter().rev() {
         for parent in graph.parents(rev)?.iter() {
             if *parent != NULL_REVISION && range.contains(parent) {
                 range.insert(rev);
             }
         }
     }
     Ok(range)
 }
 
 #[cfg(test)]
 mod tests {
 
     use super::*;
     use crate::{testing::SampleGraph, BaseRevision};
 
     /// Apply `retain_heads()` to the given slice and return as a sorted `Vec`
     fn retain_heads_sorted(
         graph: &impl Graph,
         revs: &[BaseRevision],
     ) -> Result<Vec<Revision>, GraphError> {
         let mut revs: HashSet<Revision> =
             revs.iter().cloned().map(Revision).collect();
         retain_heads(graph, &mut revs)?;
         let mut as_vec: Vec<Revision> = revs.iter().cloned().collect();
         as_vec.sort_unstable();
         Ok(as_vec)
     }
 
     #[test]
     fn test_retain_heads() -> Result<(), GraphError> {
         assert_eq!(retain_heads_sorted(&SampleGraph, &[4, 5, 6])?, vec![5, 6]);
         assert_eq!(
             retain_heads_sorted(&SampleGraph, &[4, 1, 6, 12, 0])?,
             vec![1, 6, 12]
         );
         assert_eq!(
             retain_heads_sorted(&SampleGraph, &[1, 2, 3, 4, 5, 6, 7, 8, 9])?,
             vec![3, 5, 8, 9]
         );
         Ok(())
     }
 
     /// Apply `heads()` to the given slice and return as a sorted `Vec`
     fn heads_sorted(
         graph: &impl Graph,
         revs: &[BaseRevision],
     ) -> Result<Vec<Revision>, GraphError> {
-        let iter_revs: Vec<_> =
-            revs.into_iter().cloned().map(Revision).collect();
+        let iter_revs: Vec<_> = revs.iter().cloned().map(Revision).collect();
         let heads = heads(graph, iter_revs.iter())?;
         let mut as_vec: Vec<Revision> = heads.iter().cloned().collect();
         as_vec.sort_unstable();
         Ok(as_vec)
     }
 
     #[test]
     fn test_heads() -> Result<(), GraphError> {
         assert_eq!(heads_sorted(&SampleGraph, &[4, 5, 6])?, vec![5, 6]);
         assert_eq!(
             heads_sorted(&SampleGraph, &[4, 1, 6, 12, 0])?,
             vec![1, 6, 12]
         );
         assert_eq!(
             heads_sorted(&SampleGraph, &[1, 2, 3, 4, 5, 6, 7, 8, 9])?,
             vec![3, 5, 8, 9]
         );
         Ok(())
     }
 
     /// Apply `roots()` and sort the result for easier comparison
     fn roots_sorted(
         graph: &impl Graph,
         revs: &[BaseRevision],
     ) -> Result<Vec<Revision>, GraphError> {
         let set: HashSet<_> = revs.iter().cloned().map(Revision).collect();
         let mut as_vec = roots(graph, &set)?;
         as_vec.sort_unstable();
         Ok(as_vec)
     }
 
     #[test]
     fn test_roots() -> Result<(), GraphError> {
         assert_eq!(roots_sorted(&SampleGraph, &[4, 5, 6])?, vec![4]);
         assert_eq!(
             roots_sorted(&SampleGraph, &[4, 1, 6, 12, 0])?,
             vec![0, 4, 12]
         );
         assert_eq!(
             roots_sorted(&SampleGraph, &[1, 2, 3, 4, 5, 6, 7, 8, 9])?,
             vec![1, 8]
         );
         Ok(())
     }
 
     /// Apply `range()` and convert the result into a Vec for easier comparison
     fn range_vec(
         graph: impl Graph + Clone,
         roots: &[BaseRevision],
         heads: &[BaseRevision],
     ) -> Result<Vec<Revision>, GraphError> {
         range(
             &graph,
             roots.iter().cloned().map(Revision),
             heads.iter().cloned().map(Revision),
         )
         .map(|bs| bs.into_iter().collect())
     }
 
     #[test]
     fn test_range() -> Result<(), GraphError> {
         assert_eq!(range_vec(SampleGraph, &[0], &[4])?, vec![0, 1, 2, 4]);
         assert_eq!(
             range_vec(SampleGraph, &[0], &[8])?,
             Vec::<Revision>::new()
         );
         assert_eq!(
             range_vec(SampleGraph, &[5, 6], &[10, 11, 13])?,
             vec![5, 10]
         );
         assert_eq!(
             range_vec(SampleGraph, &[5, 6], &[10, 12])?,
             vec![5, 6, 9, 10, 12]
         );
         Ok(())
     }
 }

rust/hg-core/src/dirstate_tree/dirstate_map.rs

@@ -1,1937 +1,1937 b''
 use bytes_cast::BytesCast;
 use std::borrow::Cow;
 use std::path::PathBuf;
 
 use super::on_disk;
 use super::on_disk::DirstateV2ParseError;
 use super::owning::OwningDirstateMap;
 use super::path_with_basename::WithBasename;
 use crate::dirstate::parsers::pack_entry;
 use crate::dirstate::parsers::packed_entry_size;
 use crate::dirstate::parsers::parse_dirstate_entries;
 use crate::dirstate::CopyMapIter;
 use crate::dirstate::DirstateV2Data;
 use crate::dirstate::ParentFileData;
 use crate::dirstate::StateMapIter;
 use crate::dirstate::TruncatedTimestamp;
 use crate::matchers::Matcher;
 use crate::utils::filter_map_results;
 use crate::utils::hg_path::{HgPath, HgPathBuf};
 use crate::DirstateEntry;
 use crate::DirstateError;
 use crate::DirstateMapError;
 use crate::DirstateParents;
 use crate::DirstateStatus;
 use crate::FastHashbrownMap as FastHashMap;
 use crate::PatternFileWarning;
 use crate::StatusError;
 use crate::StatusOptions;
 
 /// Append to an existing data file if the amount of unreachable data (not used
 /// anymore) is less than this fraction of the total amount of existing data.
 const ACCEPTABLE_UNREACHABLE_BYTES_RATIO: f32 = 0.5;
 
 #[derive(Debug, PartialEq, Eq)]
 /// Version of the on-disk format
 pub enum DirstateVersion {
     V1,
     V2,
 }
 
 #[derive(Debug, PartialEq, Eq)]
 pub enum DirstateMapWriteMode {
     Auto,
     ForceNewDataFile,
     ForceAppend,
 }
 
 #[derive(Debug)]
 pub struct DirstateMap<'on_disk> {
     /// Contents of the `.hg/dirstate` file
     pub(super) on_disk: &'on_disk [u8],
 
     pub(super) root: ChildNodes<'on_disk>,
 
     /// Number of nodes anywhere in the tree that have `.entry.is_some()`.
     pub(super) nodes_with_entry_count: u32,
 
     /// Number of nodes anywhere in the tree that have
     /// `.copy_source.is_some()`.
     pub(super) nodes_with_copy_source_count: u32,
 
     /// See on_disk::Header
     pub(super) ignore_patterns_hash: on_disk::IgnorePatternsHash,
 
     /// How many bytes of `on_disk` are not used anymore
     pub(super) unreachable_bytes: u32,
 
     /// Size of the data used to first load this `DirstateMap`. Used in case
     /// we need to write some new metadata, but no new data on disk,
     /// as well as to detect writes that have happened in another process
     /// since first read.
     pub(super) old_data_size: usize,
 
     /// UUID used when first loading this `DirstateMap`. Used to check if
     /// the UUID has been changed by another process since first read.
     /// Can be `None` if using dirstate v1 or if it's a brand new dirstate.
     pub(super) old_uuid: Option<Vec<u8>>,
 
     /// Identity of the dirstate file (for dirstate-v1) or the docket file
     /// (v2). Used to detect if the file has changed from another process.
     /// Since it's always written atomically, we can compare the inode to
     /// check the file identity.
     ///
     /// TODO On non-Unix systems, something like hashing is a possibility?
     pub(super) identity: Option<u64>,
 
     pub(super) dirstate_version: DirstateVersion,
 
     /// Controlled by config option `devel.dirstate.v2.data_update_mode`
     pub(super) write_mode: DirstateMapWriteMode,
 }
 
 /// Using a plain `HgPathBuf` of the full path from the repository root as a
 /// map key would also work: all paths in a given map have the same parent
 /// path, so comparing full paths gives the same result as comparing base
 /// names. However `HashMap` would waste time always re-hashing the same
 /// string prefix.
 pub(super) type NodeKey<'on_disk> = WithBasename<Cow<'on_disk, HgPath>>;
 
 /// Similar to `&'tree Cow<'on_disk, HgPath>`, but can also be returned
 /// for on-disk nodes that don’t actually have a `Cow` to borrow.
 #[derive(Debug)]
 pub(super) enum BorrowedPath<'tree, 'on_disk> {
     InMemory(&'tree HgPathBuf),
     OnDisk(&'on_disk HgPath),
 }
 
 #[derive(Debug)]
 pub(super) enum ChildNodes<'on_disk> {
     InMemory(FastHashMap<NodeKey<'on_disk>, Node<'on_disk>>),
     OnDisk(&'on_disk [on_disk::Node]),
 }
 
 #[derive(Debug)]
 pub(super) enum ChildNodesRef<'tree, 'on_disk> {
     InMemory(&'tree FastHashMap<NodeKey<'on_disk>, Node<'on_disk>>),
     OnDisk(&'on_disk [on_disk::Node]),
 }
 
 #[derive(Debug)]
 pub(super) enum NodeRef<'tree, 'on_disk> {
     InMemory(&'tree NodeKey<'on_disk>, &'tree Node<'on_disk>),
     OnDisk(&'on_disk on_disk::Node),
 }
 
 impl<'tree, 'on_disk> BorrowedPath<'tree, 'on_disk> {
     pub fn detach_from_tree(&self) -> Cow<'on_disk, HgPath> {
         match *self {
             BorrowedPath::InMemory(in_memory) => Cow::Owned(in_memory.clone()),
             BorrowedPath::OnDisk(on_disk) => Cow::Borrowed(on_disk),
         }
     }
 }
 
 impl<'tree, 'on_disk> std::ops::Deref for BorrowedPath<'tree, 'on_disk> {
     type Target = HgPath;
 
     fn deref(&self) -> &HgPath {
         match *self {
             BorrowedPath::InMemory(in_memory) => in_memory,
             BorrowedPath::OnDisk(on_disk) => on_disk,
         }
     }
 }
 
 impl Default for ChildNodes<'_> {
     fn default() -> Self {
         ChildNodes::InMemory(Default::default())
     }
 }
 
 impl<'on_disk> ChildNodes<'on_disk> {
     pub(super) fn as_ref<'tree>(
         &'tree self,
     ) -> ChildNodesRef<'tree, 'on_disk> {
         match self {
             ChildNodes::InMemory(nodes) => ChildNodesRef::InMemory(nodes),
             ChildNodes::OnDisk(nodes) => ChildNodesRef::OnDisk(nodes),
         }
     }
 
     pub(super) fn is_empty(&self) -> bool {
         match self {
             ChildNodes::InMemory(nodes) => nodes.is_empty(),
             ChildNodes::OnDisk(nodes) => nodes.is_empty(),
         }
     }
 
     fn make_mut(
         &mut self,
         on_disk: &'on_disk [u8],
         unreachable_bytes: &mut u32,
     ) -> Result<
         &mut FastHashMap<NodeKey<'on_disk>, Node<'on_disk>>,
         DirstateV2ParseError,
     > {
         match self {
             ChildNodes::InMemory(nodes) => Ok(nodes),
             ChildNodes::OnDisk(nodes) => {
                 *unreachable_bytes +=
-                    std::mem::size_of_val::<[on_disk::Node]>(nodes) as u32;
+                    std::mem::size_of_val::<[on_disk::Node]>(*nodes) as u32;
                 let nodes = nodes
                     .iter()
                     .map(|node| {
                         Ok((
                             node.path(on_disk)?,
                             node.to_in_memory_node(on_disk)?,
                         ))
                     })
                     .collect::<Result<_, _>>()?;
                 *self = ChildNodes::InMemory(nodes);
                 match self {
                     ChildNodes::InMemory(nodes) => Ok(nodes),
                     ChildNodes::OnDisk(_) => unreachable!(),
                 }
             }
         }
     }
 }
 
 impl<'tree, 'on_disk> ChildNodesRef<'tree, 'on_disk> {
     pub(super) fn get(
         &self,
         base_name: &HgPath,
         on_disk: &'on_disk [u8],
     ) -> Result<Option<NodeRef<'tree, 'on_disk>>, DirstateV2ParseError> {
         match self {
             ChildNodesRef::InMemory(nodes) => Ok(nodes
                 .get_key_value(base_name)
                 .map(|(k, v)| NodeRef::InMemory(k, v))),
             ChildNodesRef::OnDisk(nodes) => {
                 let mut parse_result = Ok(());
                 let search_result = nodes.binary_search_by(|node| {
                     match node.base_name(on_disk) {
                         Ok(node_base_name) => node_base_name.cmp(base_name),
                         Err(e) => {
                             parse_result = Err(e);
                             // Dummy comparison result, `search_result` won’t
                             // be used since `parse_result` is an error
                             std::cmp::Ordering::Equal
                         }
                     }
                 });
                 parse_result.map(|()| {
                     search_result.ok().map(|i| NodeRef::OnDisk(&nodes[i]))
                 })
             }
         }
     }
 
     /// Iterate in undefined order
     pub(super) fn iter(
         &self,
     ) -> impl Iterator<Item = NodeRef<'tree, 'on_disk>> {
         match self {
             ChildNodesRef::InMemory(nodes) => itertools::Either::Left(
                 nodes.iter().map(|(k, v)| NodeRef::InMemory(k, v)),
             ),
             ChildNodesRef::OnDisk(nodes) => {
                 itertools::Either::Right(nodes.iter().map(NodeRef::OnDisk))
             }
         }
     }
 
     /// Iterate in parallel in undefined order
     pub(super) fn par_iter(
         &self,
     ) -> impl rayon::iter::ParallelIterator<Item = NodeRef<'tree, 'on_disk>>
     {
         use rayon::prelude::*;
         match self {
             ChildNodesRef::InMemory(nodes) => rayon::iter::Either::Left(
                 nodes.par_iter().map(|(k, v)| NodeRef::InMemory(k, v)),
             ),
             ChildNodesRef::OnDisk(nodes) => rayon::iter::Either::Right(
                 nodes.par_iter().map(NodeRef::OnDisk),
             ),
         }
     }
 
     pub(super) fn sorted(&self) -> Vec<NodeRef<'tree, 'on_disk>> {
         match self {
             ChildNodesRef::InMemory(nodes) => {
                 let mut vec: Vec<_> = nodes
                     .iter()
                     .map(|(k, v)| NodeRef::InMemory(k, v))
                     .collect();
                 fn sort_key<'a>(node: &'a NodeRef) -> &'a HgPath {
                     match node {
                         NodeRef::InMemory(path, _node) => path.base_name(),
                         NodeRef::OnDisk(_) => unreachable!(),
                     }
                 }
                 // `sort_unstable_by_key` doesn’t allow keys borrowing from the
                 // value: https://github.com/rust-lang/rust/issues/34162
                 vec.sort_unstable_by(|a, b| sort_key(a).cmp(sort_key(b)));
                 vec
             }
             ChildNodesRef::OnDisk(nodes) => {
                 // Nodes on disk are already sorted
                 nodes.iter().map(NodeRef::OnDisk).collect()
             }
         }
     }
 }
 
 impl<'tree, 'on_disk> NodeRef<'tree, 'on_disk> {
     pub(super) fn full_path(
         &self,
         on_disk: &'on_disk [u8],
     ) -> Result<&'tree HgPath, DirstateV2ParseError> {
         match self {
             NodeRef::InMemory(path, _node) => Ok(path.full_path()),
             NodeRef::OnDisk(node) => node.full_path(on_disk),
         }
     }
 
     /// Returns a `BorrowedPath`, which can be turned into a `Cow<'on_disk,
     /// HgPath>` detached from `'tree`
     pub(super) fn full_path_borrowed(
         &self,
         on_disk: &'on_disk [u8],
     ) -> Result<BorrowedPath<'tree, 'on_disk>, DirstateV2ParseError> {
         match self {
             NodeRef::InMemory(path, _node) => match path.full_path() {
                 Cow::Borrowed(on_disk) => Ok(BorrowedPath::OnDisk(on_disk)),
                 Cow::Owned(in_memory) => Ok(BorrowedPath::InMemory(in_memory)),
             },
             NodeRef::OnDisk(node) => {
                 Ok(BorrowedPath::OnDisk(node.full_path(on_disk)?))
             }
         }
     }
 
     pub(super) fn base_name(
         &self,
         on_disk: &'on_disk [u8],
     ) -> Result<&'tree HgPath, DirstateV2ParseError> {
         match self {
             NodeRef::InMemory(path, _node) => Ok(path.base_name()),
             NodeRef::OnDisk(node) => node.base_name(on_disk),
         }
     }
 
     pub(super) fn children(
         &self,
         on_disk: &'on_disk [u8],
     ) -> Result<ChildNodesRef<'tree, 'on_disk>, DirstateV2ParseError> {
         match self {
             NodeRef::InMemory(_path, node) => Ok(node.children.as_ref()),
             NodeRef::OnDisk(node) => {
                 Ok(ChildNodesRef::OnDisk(node.children(on_disk)?))
             }
         }
     }
 
     pub(super) fn has_copy_source(&self) -> bool {
         match self {
             NodeRef::InMemory(_path, node) => node.copy_source.is_some(),
             NodeRef::OnDisk(node) => node.has_copy_source(),
         }
     }
 
     pub(super) fn copy_source(
         &self,
         on_disk: &'on_disk [u8],
     ) -> Result<Option<&'tree HgPath>, DirstateV2ParseError> {
         match self {
             NodeRef::InMemory(_path, node) => Ok(node.copy_source.as_deref()),
             NodeRef::OnDisk(node) => node.copy_source(on_disk),
         }
     }
     /// Returns a `BorrowedPath`, which can be turned into a `Cow<'on_disk,
     /// HgPath>` detached from `'tree`
     pub(super) fn copy_source_borrowed(
         &self,
         on_disk: &'on_disk [u8],
     ) -> Result<Option<BorrowedPath<'tree, 'on_disk>>, DirstateV2ParseError>
     {
         Ok(match self {
             NodeRef::InMemory(_path, node) => {
                 node.copy_source.as_ref().map(|source| match source {
                     Cow::Borrowed(on_disk) => BorrowedPath::OnDisk(on_disk),
                     Cow::Owned(in_memory) => BorrowedPath::InMemory(in_memory),
                 })
             }
             NodeRef::OnDisk(node) => {
                 node.copy_source(on_disk)?.map(BorrowedPath::OnDisk)
             }
         })
     }
 
     pub(super) fn entry(
         &self,
     ) -> Result<Option<DirstateEntry>, DirstateV2ParseError> {
         match self {
             NodeRef::InMemory(_path, node) => {
                 Ok(node.data.as_entry().copied())
             }
             NodeRef::OnDisk(node) => node.entry(),
         }
     }
 
     pub(super) fn cached_directory_mtime(
         &self,
     ) -> Result<Option<TruncatedTimestamp>, DirstateV2ParseError> {
         match self {
             NodeRef::InMemory(_path, node) => Ok(match node.data {
                 NodeData::CachedDirectory { mtime } => Some(mtime),
                 _ => None,
             }),
             NodeRef::OnDisk(node) => node.cached_directory_mtime(),
         }
     }
 
     pub(super) fn descendants_with_entry_count(&self) -> u32 {
         match self {
             NodeRef::InMemory(_path, node) => {
                 node.descendants_with_entry_count
             }
             NodeRef::OnDisk(node) => node.descendants_with_entry_count.get(),
         }
     }
 
     pub(super) fn tracked_descendants_count(&self) -> u32 {
         match self {
             NodeRef::InMemory(_path, node) => node.tracked_descendants_count,
             NodeRef::OnDisk(node) => node.tracked_descendants_count.get(),
         }
     }
 }
 
 /// Represents a file or a directory
 #[derive(Default, Debug)]
 pub(super) struct Node<'on_disk> {
     pub(super) data: NodeData,
 
     pub(super) copy_source: Option<Cow<'on_disk, HgPath>>,
 
     pub(super) children: ChildNodes<'on_disk>,
 
     /// How many (non-inclusive) descendants of this node have an entry.
     pub(super) descendants_with_entry_count: u32,
 
     /// How many (non-inclusive) descendants of this node have an entry whose
     /// state is "tracked".
     pub(super) tracked_descendants_count: u32,
 }
 
 #[derive(Debug)]
 pub(super) enum NodeData {
     Entry(DirstateEntry),
     CachedDirectory { mtime: TruncatedTimestamp },
     None,
 }
 
 impl Default for NodeData {
     fn default() -> Self {
         NodeData::None
     }
 }
 
 impl NodeData {
     fn has_entry(&self) -> bool {
         matches!(self, NodeData::Entry(_))
     }
 
     fn as_entry(&self) -> Option<&DirstateEntry> {
         match self {
             NodeData::Entry(entry) => Some(entry),
             _ => None,
         }
     }
 
     fn as_entry_mut(&mut self) -> Option<&mut DirstateEntry> {
         match self {
             NodeData::Entry(entry) => Some(entry),
             _ => None,
         }
     }
 }
 
 impl<'on_disk> DirstateMap<'on_disk> {
     pub(super) fn empty(on_disk: &'on_disk [u8]) -> Self {
         Self {
             on_disk,
             root: ChildNodes::default(),
             nodes_with_entry_count: 0,
             nodes_with_copy_source_count: 0,
             ignore_patterns_hash: [0; on_disk::IGNORE_PATTERNS_HASH_LEN],
             unreachable_bytes: 0,
             old_data_size: 0,
             old_uuid: None,
             identity: None,
             dirstate_version: DirstateVersion::V1,
             write_mode: DirstateMapWriteMode::Auto,
         }
     }
 
     #[logging_timer::time("trace")]
     pub fn new_v2(
         on_disk: &'on_disk [u8],
         data_size: usize,
         metadata: &[u8],
         uuid: Vec<u8>,
         identity: Option<u64>,
     ) -> Result<Self, DirstateError> {
         if let Some(data) = on_disk.get(..data_size) {
             Ok(on_disk::read(data, metadata, uuid, identity)?)
         } else {
             Err(DirstateV2ParseError::new("not enough bytes on disk").into())
         }
     }
 
     #[logging_timer::time("trace")]
     pub fn new_v1(
         on_disk: &'on_disk [u8],
         identity: Option<u64>,
     ) -> Result<(Self, Option<DirstateParents>), DirstateError> {
         let mut map = Self::empty(on_disk);
         if map.on_disk.is_empty() {
             return Ok((map, None));
         }
 
         let parents = parse_dirstate_entries(
             map.on_disk,
             |path, entry, copy_source| {
                 let tracked = entry.tracked();
                 let node = Self::get_or_insert_node_inner(
                     map.on_disk,
                     &mut map.unreachable_bytes,
                     &mut map.root,
                     path,
                     WithBasename::to_cow_borrowed,
                     |ancestor| {
                         if tracked {
                             ancestor.tracked_descendants_count += 1
                         }
                         ancestor.descendants_with_entry_count += 1
                     },
                 )?;
                 assert!(
                     !node.data.has_entry(),
                     "duplicate dirstate entry in read"
                 );
                 assert!(
                     node.copy_source.is_none(),
                     "duplicate dirstate entry in read"
                 );
                 node.data = NodeData::Entry(*entry);
                 node.copy_source = copy_source.map(Cow::Borrowed);
                 map.nodes_with_entry_count += 1;
                 if copy_source.is_some() {
                     map.nodes_with_copy_source_count += 1
                 }
                 Ok(())
             },
         )?;
         let parents = Some(*parents);
         map.identity = identity;
 
         Ok((map, parents))
     }
 
     /// Assuming dirstate-v2 format, returns whether the next write should
     /// append to the existing data file that contains `self.on_disk` (true),
     /// or create a new data file from scratch (false).
     pub(super) fn write_should_append(&self) -> bool {
         match self.write_mode {
             DirstateMapWriteMode::ForceAppend => true,
             DirstateMapWriteMode::ForceNewDataFile => false,
             DirstateMapWriteMode::Auto => {
                 let ratio =
                     self.unreachable_bytes as f32 / self.on_disk.len() as f32;
                 ratio < ACCEPTABLE_UNREACHABLE_BYTES_RATIO
             }
         }
     }
 
     fn get_node<'tree>(
         &'tree self,
         path: &HgPath,
     ) -> Result<Option<NodeRef<'tree, 'on_disk>>, DirstateV2ParseError> {
         let mut children = self.root.as_ref();
         let mut components = path.components();
         let mut component =
             components.next().expect("expected at least one components");
         loop {
             if let Some(child) = children.get(component, self.on_disk)? {
                 if let Some(next_component) = components.next() {
                     component = next_component;
                     children = child.children(self.on_disk)?;
                 } else {
                     return Ok(Some(child));
                 }
             } else {
                 return Ok(None);
             }
         }
     }
 
     pub fn has_node(
         &self,
         path: &HgPath,
     ) -> Result<bool, DirstateV2ParseError> {
         let node = self.get_node(path)?;
         Ok(node.is_some())
     }
 
     /// Returns a mutable reference to the node at `path` if it exists
     ///
     /// `each_ancestor` is a callback that is called for each ancestor node
     /// when descending the tree. It is used to keep the different counters
     /// of the `DirstateMap` up-to-date.
     fn get_node_mut<'tree>(
         &'tree mut self,
         path: &HgPath,
         each_ancestor: impl FnMut(&mut Node),
     ) -> Result<Option<&'tree mut Node<'on_disk>>, DirstateV2ParseError> {
         Self::get_node_mut_inner(
             self.on_disk,
             &mut self.unreachable_bytes,
             &mut self.root,
             path,
             each_ancestor,
         )
     }
 
     /// Lower-level version of `get_node_mut`.
     ///
     /// This takes `root` instead of `&mut self` so that callers can mutate
     /// other fields while the returned borrow is still valid.
     ///
     /// `each_ancestor` is a callback that is called for each ancestor node
     /// when descending the tree. It is used to keep the different counters
     /// of the `DirstateMap` up-to-date.
     fn get_node_mut_inner<'tree>(
         on_disk: &'on_disk [u8],
         unreachable_bytes: &mut u32,
         root: &'tree mut ChildNodes<'on_disk>,
         path: &HgPath,
         mut each_ancestor: impl FnMut(&mut Node),
     ) -> Result<Option<&'tree mut Node<'on_disk>>, DirstateV2ParseError> {
         let mut children = root;
         let mut components = path.components();
         let mut component =
             components.next().expect("expected at least one components");
         loop {
             if let Some(child) = children
                 .make_mut(on_disk, unreachable_bytes)?
                 .get_mut(component)
             {
                 if let Some(next_component) = components.next() {
                     each_ancestor(child);
                     component = next_component;
                     children = &mut child.children;
                 } else {
                     return Ok(Some(child));
                 }
             } else {
                 return Ok(None);
             }
         }
     }
 
     /// Get a mutable reference to the node at `path`, creating it if it does
     /// not exist.
     ///
     /// `each_ancestor` is a callback that is called for each ancestor node
     /// when descending the tree. It is used to keep the different counters
     /// of the `DirstateMap` up-to-date.
     fn get_or_insert_node<'tree, 'path>(
         &'tree mut self,
         path: &'path HgPath,
         each_ancestor: impl FnMut(&mut Node),
     ) -> Result<&'tree mut Node<'on_disk>, DirstateV2ParseError> {
         Self::get_or_insert_node_inner(
             self.on_disk,
             &mut self.unreachable_bytes,
             &mut self.root,
             path,
             WithBasename::to_cow_owned,
             each_ancestor,
         )
     }
 
     /// Lower-level version of `get_or_insert_node_inner`, which is used when
     /// parsing disk data to remove allocations for new nodes.
     fn get_or_insert_node_inner<'tree, 'path>(
         on_disk: &'on_disk [u8],
         unreachable_bytes: &mut u32,
         root: &'tree mut ChildNodes<'on_disk>,
         path: &'path HgPath,
         to_cow: impl Fn(
             WithBasename<&'path HgPath>,
         ) -> WithBasename<Cow<'on_disk, HgPath>>,
         mut each_ancestor: impl FnMut(&mut Node),
     ) -> Result<&'tree mut Node<'on_disk>, DirstateV2ParseError> {
         let mut child_nodes = root;
         let mut inclusive_ancestor_paths =
             WithBasename::inclusive_ancestors_of(path);
         let mut ancestor_path = inclusive_ancestor_paths
             .next()
             .expect("expected at least one inclusive ancestor");
         loop {
             let (_, child_node) = child_nodes
                 .make_mut(on_disk, unreachable_bytes)?
                 .raw_entry_mut()
                 .from_key(ancestor_path.base_name())
                 .or_insert_with(|| (to_cow(ancestor_path), Node::default()));
             if let Some(next) = inclusive_ancestor_paths.next() {
                 each_ancestor(child_node);
                 ancestor_path = next;
                 child_nodes = &mut child_node.children;
             } else {
                 return Ok(child_node);
             }
         }
     }
 
     #[allow(clippy::too_many_arguments)]
     fn reset_state(
         &mut self,
         filename: &HgPath,
         old_entry_opt: Option<DirstateEntry>,
         wc_tracked: bool,
         p1_tracked: bool,
         p2_info: bool,
         has_meaningful_mtime: bool,
         parent_file_data_opt: Option<ParentFileData>,
     ) -> Result<(), DirstateError> {
         let (had_entry, was_tracked) = match old_entry_opt {
             Some(old_entry) => (true, old_entry.tracked()),
             None => (false, false),
         };
         let node = self.get_or_insert_node(filename, |ancestor| {
             if !had_entry {
                 ancestor.descendants_with_entry_count += 1;
             }
             if was_tracked {
                 if !wc_tracked {
                     ancestor.tracked_descendants_count = ancestor
                         .tracked_descendants_count
                         .checked_sub(1)
                         .expect("tracked count to be >= 0");
                 }
             } else if wc_tracked {
                 ancestor.tracked_descendants_count += 1;
             }
         })?;
 
         let v2_data = if let Some(parent_file_data) = parent_file_data_opt {
             DirstateV2Data {
                 wc_tracked,
                 p1_tracked,
                 p2_info,
                 mode_size: parent_file_data.mode_size,
                 mtime: if has_meaningful_mtime {
                     parent_file_data.mtime
                 } else {
                     None
                 },
                 ..Default::default()
             }
         } else {
             DirstateV2Data {
                 wc_tracked,
                 p1_tracked,
                 p2_info,
                 ..Default::default()
             }
         };
         node.data = NodeData::Entry(DirstateEntry::from_v2_data(v2_data));
         if !had_entry {
             self.nodes_with_entry_count += 1;
         }
         Ok(())
     }
 
     fn set_tracked(
         &mut self,
         filename: &HgPath,
         old_entry_opt: Option<DirstateEntry>,
     ) -> Result<bool, DirstateV2ParseError> {
         let was_tracked = old_entry_opt.map_or(false, |e| e.tracked());
         let had_entry = old_entry_opt.is_some();
-        let tracked_count_increment = if was_tracked { 0 } else { 1 };
+        let tracked_count_increment = u32::from(!was_tracked);
         let mut new = false;
 
         let node = self.get_or_insert_node(filename, |ancestor| {
             if !had_entry {
                 ancestor.descendants_with_entry_count += 1;
             }
 
             ancestor.tracked_descendants_count += tracked_count_increment;
         })?;
         if let Some(old_entry) = old_entry_opt {
             let mut e = old_entry;
             if e.tracked() {
                 // XXX
                 // This is probably overkill for more case, but we need this to
                 // fully replace the `normallookup` call with `set_tracked`
                 // one. Consider smoothing this in the future.
                 e.set_possibly_dirty();
             } else {
                 new = true;
                 e.set_tracked();
             }
             node.data = NodeData::Entry(e)
         } else {
             node.data = NodeData::Entry(DirstateEntry::new_tracked());
             self.nodes_with_entry_count += 1;
             new = true;
         };
         Ok(new)
     }
 
     /// Set a node as untracked in the dirstate.
     ///
     /// It is the responsibility of the caller to remove the copy source and/or
     /// the entry itself if appropriate.
     ///
     /// # Panics
     ///
     /// Panics if the node does not exist.
     fn set_untracked(
         &mut self,
         filename: &HgPath,
         old_entry: DirstateEntry,
     ) -> Result<(), DirstateV2ParseError> {
         let node = self
             .get_node_mut(filename, |ancestor| {
                 ancestor.tracked_descendants_count = ancestor
                     .tracked_descendants_count
                     .checked_sub(1)
                     .expect("tracked_descendants_count should be >= 0");
             })?
             .expect("node should exist");
         let mut new_entry = old_entry;
         new_entry.set_untracked();
         node.data = NodeData::Entry(new_entry);
         Ok(())
     }
 
     /// Set a node as clean in the dirstate.
     ///
     /// It is the responsibility of the caller to remove the copy source.
     ///
     /// # Panics
     ///
     /// Panics if the node does not exist.
     fn set_clean(
         &mut self,
         filename: &HgPath,
         old_entry: DirstateEntry,
         mode: u32,
         size: u32,
         mtime: TruncatedTimestamp,
     ) -> Result<(), DirstateError> {
         let node = self
             .get_node_mut(filename, |ancestor| {
                 if !old_entry.tracked() {
                     ancestor.tracked_descendants_count += 1;
                 }
             })?
             .expect("node should exist");
         let mut new_entry = old_entry;
         new_entry.set_clean(mode, size, mtime);
         node.data = NodeData::Entry(new_entry);
         Ok(())
     }
 
     /// Set a node as possibly dirty in the dirstate.
     ///
     /// # Panics
     ///
     /// Panics if the node does not exist.
     fn set_possibly_dirty(
         &mut self,
         filename: &HgPath,
     ) -> Result<(), DirstateError> {
         let node = self
             .get_node_mut(filename, |_ancestor| {})?
             .expect("node should exist");
         let entry = node.data.as_entry_mut().expect("entry should exist");
         entry.set_possibly_dirty();
         node.data = NodeData::Entry(*entry);
         Ok(())
     }
 
     /// Clears the cached mtime for the (potential) folder at `path`.
     pub(super) fn clear_cached_mtime(
         &mut self,
         path: &HgPath,
     ) -> Result<(), DirstateV2ParseError> {
         let node = match self.get_node_mut(path, |_ancestor| {})? {
             Some(node) => node,
             None => return Ok(()),
         };
         if let NodeData::CachedDirectory { .. } = &node.data {
             node.data = NodeData::None
         }
         Ok(())
     }
 
     /// Sets the cached mtime for the (potential) folder at `path`.
     pub(super) fn set_cached_mtime(
         &mut self,
         path: &HgPath,
         mtime: TruncatedTimestamp,
     ) -> Result<(), DirstateV2ParseError> {
         let node = match self.get_node_mut(path, |_ancestor| {})? {
             Some(node) => node,
             None => return Ok(()),
         };
         match &node.data {
             NodeData::Entry(_) => {} // Don’t overwrite an entry
             NodeData::CachedDirectory { .. } | NodeData::None => {
                 node.data = NodeData::CachedDirectory { mtime }
             }
         }
         Ok(())
     }
 
     fn iter_nodes<'tree>(
         &'tree self,
     ) -> impl Iterator<
         Item = Result<NodeRef<'tree, 'on_disk>, DirstateV2ParseError>,
     > + 'tree {
         // Depth first tree traversal.
         //
         // If we could afford internal iteration and recursion,
         // this would look like:
         //
         // ```
         // fn traverse_children(
         //     children: &ChildNodes,
         //     each: &mut impl FnMut(&Node),
         // ) {
         //     for child in children.values() {
         //         traverse_children(&child.children, each);
         //         each(child);
         //     }
         // }
         // ```
         //
         // However we want an external iterator and therefore can’t use the
         // call stack. Use an explicit stack instead:
         let mut stack = Vec::new();
         let mut iter = self.root.as_ref().iter();
         std::iter::from_fn(move || {
             while let Some(child_node) = iter.next() {
                 let children = match child_node.children(self.on_disk) {
                     Ok(children) => children,
                     Err(error) => return Some(Err(error)),
                 };
                 // Pseudo-recursion
                 let new_iter = children.iter();
                 let old_iter = std::mem::replace(&mut iter, new_iter);
                 stack.push((child_node, old_iter));
             }
             // Found the end of a `children.iter()` iterator.
             if let Some((child_node, next_iter)) = stack.pop() {
                 // "Return" from pseudo-recursion by restoring state from the
                 // explicit stack
                 iter = next_iter;
 
                 Some(Ok(child_node))
             } else {
                 // Reached the bottom of the stack, we’re done
                 None
             }
         })
     }
 
     fn count_dropped_path(unreachable_bytes: &mut u32, path: Cow<HgPath>) {
         if let Cow::Borrowed(path) = path {
             *unreachable_bytes += path.len() as u32
         }
     }
 
     pub(crate) fn set_write_mode(&mut self, write_mode: DirstateMapWriteMode) {
         self.write_mode = write_mode;
     }
 }
 
 type DebugDirstateTuple<'a> = (&'a HgPath, (u8, i32, i32, i32));
 
 impl OwningDirstateMap {
     pub fn clear(&mut self) {
         self.with_dmap_mut(|map| {
             map.root = Default::default();
             map.nodes_with_entry_count = 0;
             map.nodes_with_copy_source_count = 0;
             map.unreachable_bytes = map.on_disk.len() as u32;
         });
     }
 
     pub fn set_tracked(
         &mut self,
         filename: &HgPath,
     ) -> Result<bool, DirstateV2ParseError> {
         let old_entry_opt = self.get(filename)?;
         self.with_dmap_mut(|map| map.set_tracked(filename, old_entry_opt))
     }
 
     pub fn set_untracked(
         &mut self,
         filename: &HgPath,
     ) -> Result<bool, DirstateError> {
         let old_entry_opt = self.get(filename)?;
         match old_entry_opt {
             None => Ok(false),
             Some(old_entry) => {
                 if !old_entry.tracked() {
                     // `DirstateMap::set_untracked` is not a noop if
                     // already not tracked as it will decrement the
                     // tracked counters while going down.
                     return Ok(true);
                 }
                 if old_entry.added() {
                     // Untracking an "added" entry will just result in a
                     // worthless entry (and other parts of the code will
                     // complain about it), just drop it entirely.
                     self.drop_entry_and_copy_source(filename)?;
                     return Ok(true);
                 }
                 if !old_entry.p2_info() {
                     self.copy_map_remove(filename)?;
                 }
 
                 self.with_dmap_mut(|map| {
                     map.set_untracked(filename, old_entry)?;
                     Ok(true)
                 })
             }
         }
     }
 
     pub fn set_clean(
         &mut self,
         filename: &HgPath,
         mode: u32,
         size: u32,
         mtime: TruncatedTimestamp,
     ) -> Result<(), DirstateError> {
         let old_entry = match self.get(filename)? {
             None => {
                 return Err(
                     DirstateMapError::PathNotFound(filename.into()).into()
                 )
             }
             Some(e) => e,
         };
         self.copy_map_remove(filename)?;
         self.with_dmap_mut(|map| {
             map.set_clean(filename, old_entry, mode, size, mtime)
         })
     }
 
     pub fn set_possibly_dirty(
         &mut self,
         filename: &HgPath,
     ) -> Result<(), DirstateError> {
         if self.get(filename)?.is_none() {
             return Err(DirstateMapError::PathNotFound(filename.into()).into());
         }
         self.with_dmap_mut(|map| map.set_possibly_dirty(filename))
     }
 
     pub fn reset_state(
         &mut self,
         filename: &HgPath,
         wc_tracked: bool,
         p1_tracked: bool,
         p2_info: bool,
         has_meaningful_mtime: bool,
         parent_file_data_opt: Option<ParentFileData>,
     ) -> Result<(), DirstateError> {
         if !(p1_tracked || p2_info || wc_tracked) {
             self.drop_entry_and_copy_source(filename)?;
             return Ok(());
         }
         self.copy_map_remove(filename)?;
         let old_entry_opt = self.get(filename)?;
         self.with_dmap_mut(|map| {
             map.reset_state(
                 filename,
                 old_entry_opt,
                 wc_tracked,
                 p1_tracked,
                 p2_info,
                 has_meaningful_mtime,
                 parent_file_data_opt,
             )
         })
     }
 
     pub fn drop_entry_and_copy_source(
         &mut self,
         filename: &HgPath,
     ) -> Result<(), DirstateError> {
         let was_tracked = self.get(filename)?.map_or(false, |e| e.tracked());
         struct Dropped {
             was_tracked: bool,
             had_entry: bool,
             had_copy_source: bool,
         }
 
         /// If this returns `Ok(Some((dropped, removed)))`, then
         ///
         /// * `dropped` is about the leaf node that was at `filename`
         /// * `removed` is whether this particular level of recursion just
         ///   removed a node in `nodes`.
         fn recur<'on_disk>(
             on_disk: &'on_disk [u8],
             unreachable_bytes: &mut u32,
             nodes: &mut ChildNodes<'on_disk>,
             path: &HgPath,
         ) -> Result<Option<(Dropped, bool)>, DirstateV2ParseError> {
             let (first_path_component, rest_of_path) =
                 path.split_first_component();
             let nodes = nodes.make_mut(on_disk, unreachable_bytes)?;
             let node = if let Some(node) = nodes.get_mut(first_path_component)
             {
                 node
             } else {
                 return Ok(None);
             };
             let dropped;
             if let Some(rest) = rest_of_path {
                 if let Some((d, removed)) = recur(
                     on_disk,
                     unreachable_bytes,
                     &mut node.children,
                     rest,
                 )? {
                     dropped = d;
                     if dropped.had_entry {
                         node.descendants_with_entry_count = node
                             .descendants_with_entry_count
                             .checked_sub(1)
                             .expect(
                                 "descendants_with_entry_count should be >= 0",
                             );
                     }
                     if dropped.was_tracked {
                         node.tracked_descendants_count = node
                             .tracked_descendants_count
                             .checked_sub(1)
                             .expect(
                                 "tracked_descendants_count should be >= 0",
                             );
                     }
 
                     // Directory caches must be invalidated when removing a
                     // child node
                     if removed {
                         if let NodeData::CachedDirectory { .. } = &node.data {
                             node.data = NodeData::None
                         }
                     }
                 } else {
                     return Ok(None);
                 }
             } else {
                 let entry = node.data.as_entry();
                 let was_tracked = entry.map_or(false, |entry| entry.tracked());
                 let had_entry = entry.is_some();
                 if had_entry {
                     node.data = NodeData::None
                 }
                 let mut had_copy_source = false;
                 if let Some(source) = &node.copy_source {
                     DirstateMap::count_dropped_path(
                         unreachable_bytes,
                         Cow::Borrowed(source),
                     );
                     had_copy_source = true;
                     node.copy_source = None
                 }
                 dropped = Dropped {
                     was_tracked,
                     had_entry,
                     had_copy_source,
                 };
             }
             // After recursion, for both leaf (rest_of_path is None) nodes and
             // parent nodes, remove a node if it just became empty.
             let remove = !node.data.has_entry()
                 && node.copy_source.is_none()
                 && node.children.is_empty();
             if remove {
                 let (key, _) =
                     nodes.remove_entry(first_path_component).unwrap();
                 DirstateMap::count_dropped_path(
                     unreachable_bytes,
                     Cow::Borrowed(key.full_path()),
                 )
             }
             Ok(Some((dropped, remove)))
         }
 
         self.with_dmap_mut(|map| {
             if let Some((dropped, _removed)) = recur(
                 map.on_disk,
                 &mut map.unreachable_bytes,
                 &mut map.root,
                 filename,
             )? {
                 if dropped.had_entry {
                     map.nodes_with_entry_count = map
                         .nodes_with_entry_count
                         .checked_sub(1)
                         .expect("nodes_with_entry_count should be >= 0");
                 }
                 if dropped.had_copy_source {
                     map.nodes_with_copy_source_count = map
                         .nodes_with_copy_source_count
                         .checked_sub(1)
                         .expect("nodes_with_copy_source_count should be >= 0");
                 }
             } else {
                 debug_assert!(!was_tracked);
             }
             Ok(())
         })
     }
 
     pub fn has_tracked_dir(
         &mut self,
         directory: &HgPath,
     ) -> Result<bool, DirstateError> {
         self.with_dmap_mut(|map| {
             if let Some(node) = map.get_node(directory)? {
                 // A node without a `DirstateEntry` was created to hold child
                 // nodes, and is therefore a directory.
                 let is_dir = node.entry()?.is_none();
                 Ok(is_dir && node.tracked_descendants_count() > 0)
             } else {
                 Ok(false)
             }
         })
     }
 
     pub fn has_dir(
         &mut self,
         directory: &HgPath,
     ) -> Result<bool, DirstateError> {
         self.with_dmap_mut(|map| {
             if let Some(node) = map.get_node(directory)? {
                 // A node without a `DirstateEntry` was created to hold child
                 // nodes, and is therefore a directory.
                 let is_dir = node.entry()?.is_none();
                 Ok(is_dir && node.descendants_with_entry_count() > 0)
             } else {
                 Ok(false)
             }
         })
     }
 
     #[logging_timer::time("trace")]
     pub fn pack_v1(
         &self,
         parents: DirstateParents,
     ) -> Result<Vec<u8>, DirstateError> {
         let map = self.get_map();
         // Optizimation (to be measured?): pre-compute size to avoid `Vec`
         // reallocations
         let mut size = parents.as_bytes().len();
         for node in map.iter_nodes() {
             let node = node?;
             if node.entry()?.is_some() {
                 size += packed_entry_size(
                     node.full_path(map.on_disk)?,
                     node.copy_source(map.on_disk)?,
                 );
             }
         }
 
         let mut packed = Vec::with_capacity(size);
         packed.extend(parents.as_bytes());
 
         for node in map.iter_nodes() {
             let node = node?;
             if let Some(entry) = node.entry()? {
                 pack_entry(
                     node.full_path(map.on_disk)?,
                     &entry,
                     node.copy_source(map.on_disk)?,
                     &mut packed,
                 );
             }
         }
         Ok(packed)
     }
 
     /// Returns new data and metadata together with whether that data should be
     /// appended to the existing data file whose content is at
     /// `map.on_disk` (true), instead of written to a new data file
     /// (false), and the previous size of data on disk.
     #[logging_timer::time("trace")]
     pub fn pack_v2(
         &self,
         write_mode: DirstateMapWriteMode,
     ) -> Result<(Vec<u8>, on_disk::TreeMetadata, bool, usize), DirstateError>
     {
         let map = self.get_map();
         on_disk::write(map, write_mode)
     }
 
     /// `callback` allows the caller to process and do something with the
     /// results of the status. This is needed to do so efficiently (i.e.
     /// without cloning the `DirstateStatus` object with its paths) because
     /// we need to borrow from `Self`.
     pub fn with_status<R>(
         &mut self,
         matcher: &(dyn Matcher + Sync),
         root_dir: PathBuf,
         ignore_files: Vec<PathBuf>,
         options: StatusOptions,
         callback: impl for<'r> FnOnce(
             Result<(DirstateStatus<'r>, Vec<PatternFileWarning>), StatusError>,
         ) -> R,
     ) -> R {
         self.with_dmap_mut(|map| {
             callback(super::status::status(
                 map,
                 matcher,
                 root_dir,
                 ignore_files,
                 options,
             ))
         })
     }
 
     pub fn copy_map_len(&self) -> usize {
         let map = self.get_map();
         map.nodes_with_copy_source_count as usize
     }
 
     pub fn copy_map_iter(&self) -> CopyMapIter<'_> {
         let map = self.get_map();
         Box::new(filter_map_results(map.iter_nodes(), move |node| {
             Ok(if let Some(source) = node.copy_source(map.on_disk)? {
                 Some((node.full_path(map.on_disk)?, source))
             } else {
                 None
             })
         }))
     }
 
     pub fn copy_map_contains_key(
         &self,
         key: &HgPath,
     ) -> Result<bool, DirstateV2ParseError> {
         let map = self.get_map();
         Ok(if let Some(node) = map.get_node(key)? {
             node.has_copy_source()
         } else {
             false
         })
     }
 
     pub fn copy_map_get(
         &self,
         key: &HgPath,
     ) -> Result<Option<&HgPath>, DirstateV2ParseError> {
         let map = self.get_map();
         if let Some(node) = map.get_node(key)? {
             if let Some(source) = node.copy_source(map.on_disk)? {
                 return Ok(Some(source));
             }
         }
         Ok(None)
     }
 
     pub fn copy_map_remove(
         &mut self,
         key: &HgPath,
     ) -> Result<Option<HgPathBuf>, DirstateV2ParseError> {
         self.with_dmap_mut(|map| {
             let count = &mut map.nodes_with_copy_source_count;
             let unreachable_bytes = &mut map.unreachable_bytes;
             Ok(DirstateMap::get_node_mut_inner(
                 map.on_disk,
                 unreachable_bytes,
                 &mut map.root,
                 key,
                 |_ancestor| {},
             )?
             .and_then(|node| {
                 if let Some(source) = &node.copy_source {
                     *count = count
                         .checked_sub(1)
                         .expect("nodes_with_copy_source_count should be >= 0");
                     DirstateMap::count_dropped_path(
                         unreachable_bytes,
                         Cow::Borrowed(source),
                     );
                 }
                 node.copy_source.take().map(Cow::into_owned)
             }))
         })
     }
 
     pub fn copy_map_insert(
         &mut self,
         key: &HgPath,
         value: &HgPath,
     ) -> Result<Option<HgPathBuf>, DirstateV2ParseError> {
         self.with_dmap_mut(|map| {
             let node = map.get_or_insert_node(key, |_ancestor| {})?;
             let had_copy_source = node.copy_source.is_none();
             let old = node
                 .copy_source
                 .replace(value.to_owned().into())
                 .map(Cow::into_owned);
             if had_copy_source {
                 map.nodes_with_copy_source_count += 1
             }
             Ok(old)
         })
     }
 
     pub fn len(&self) -> usize {
         let map = self.get_map();
         map.nodes_with_entry_count as usize
     }
 
     pub fn is_empty(&self) -> bool {
         self.len() == 0
     }
 
     pub fn contains_key(
         &self,
         key: &HgPath,
     ) -> Result<bool, DirstateV2ParseError> {
         Ok(self.get(key)?.is_some())
     }
 
     pub fn get(
         &self,
         key: &HgPath,
     ) -> Result<Option<DirstateEntry>, DirstateV2ParseError> {
         let map = self.get_map();
         Ok(if let Some(node) = map.get_node(key)? {
             node.entry()?
         } else {
             None
         })
     }
 
     pub fn iter(&self) -> StateMapIter<'_> {
         let map = self.get_map();
         Box::new(filter_map_results(map.iter_nodes(), move |node| {
             Ok(if let Some(entry) = node.entry()? {
                 Some((node.full_path(map.on_disk)?, entry))
             } else {
                 None
             })
         }))
     }
 
     pub fn iter_tracked_dirs(
         &mut self,
     ) -> Result<
         Box<
             dyn Iterator<Item = Result<&HgPath, DirstateV2ParseError>>
                 + Send
                 + '_,
         >,
         DirstateError,
     > {
         let map = self.get_map();
         let on_disk = map.on_disk;
         Ok(Box::new(filter_map_results(
             map.iter_nodes(),
             move |node| {
                 Ok(if node.tracked_descendants_count() > 0 {
                     Some(node.full_path(on_disk)?)
                 } else {
                     None
                 })
             },
         )))
     }
 
     /// Only public because it needs to be exposed to the Python layer.
     /// It is not the full `setparents` logic, only the parts that mutate the
     /// entries.
     pub fn setparents_fixup(
         &mut self,
     ) -> Result<Vec<(HgPathBuf, HgPathBuf)>, DirstateV2ParseError> {
         // XXX
         // All the copying and re-querying is quite inefficient, but this is
         // still a lot better than doing it from Python.
         //
         // The better solution is to develop a mechanism for `iter_mut`,
         // which will be a lot more involved: we're dealing with a lazy,
         // append-mostly, tree-like data structure. This will do for now.
         let mut copies = vec![];
         let mut files_with_p2_info = vec![];
         for res in self.iter() {
             let (path, entry) = res?;
             if entry.p2_info() {
                 files_with_p2_info.push(path.to_owned())
             }
         }
         self.with_dmap_mut(|map| {
             for path in files_with_p2_info.iter() {
                 let node = map.get_or_insert_node(path, |_| {})?;
                 let entry =
                     node.data.as_entry_mut().expect("entry should exist");
                 entry.drop_merge_data();
                 if let Some(source) = node.copy_source.take().as_deref() {
                     copies.push((path.to_owned(), source.to_owned()));
                 }
             }
             Ok(copies)
         })
     }
 
     pub fn debug_iter(
         &self,
         all: bool,
     ) -> Box<
         dyn Iterator<Item = Result<DebugDirstateTuple, DirstateV2ParseError>>
             + Send
             + '_,
     > {
         let map = self.get_map();
         Box::new(filter_map_results(map.iter_nodes(), move |node| {
             let debug_tuple = if let Some(entry) = node.entry()? {
                 entry.debug_tuple()
             } else if !all {
                 return Ok(None);
             } else if let Some(mtime) = node.cached_directory_mtime()? {
                 (b' ', 0, -1, mtime.truncated_seconds() as i32)
             } else {
                 (b' ', 0, -1, -1)
             };
             Ok(Some((node.full_path(map.on_disk)?, debug_tuple)))
         }))
     }
 }
 #[cfg(test)]
 mod tests {
     use super::*;
 
     /// Shortcut to return tracked descendants of a path.
     /// Panics if the path does not exist.
     fn tracked_descendants(map: &OwningDirstateMap, path: &[u8]) -> u32 {
         let path = dbg!(HgPath::new(path));
         let node = map.get_map().get_node(path);
         node.unwrap().unwrap().tracked_descendants_count()
     }
 
     /// Shortcut to return descendants with an entry.
     /// Panics if the path does not exist.
     fn descendants_with_an_entry(map: &OwningDirstateMap, path: &[u8]) -> u32 {
         let path = dbg!(HgPath::new(path));
         let node = map.get_map().get_node(path);
         node.unwrap().unwrap().descendants_with_entry_count()
     }
 
     fn assert_does_not_exist(map: &OwningDirstateMap, path: &[u8]) {
         let path = dbg!(HgPath::new(path));
         let node = map.get_map().get_node(path);
         assert!(node.unwrap().is_none());
     }
 
     /// Shortcut for path creation in tests
     fn p(b: &[u8]) -> &HgPath {
         HgPath::new(b)
     }
 
     /// Test the very simple case a single tracked file
     #[test]
     fn test_tracked_descendants_simple() -> Result<(), DirstateError> {
         let mut map = OwningDirstateMap::new_empty(vec![]);
         assert_eq!(map.len(), 0);
 
         map.set_tracked(p(b"some/nested/path"))?;
 
         assert_eq!(map.len(), 1);
         assert_eq!(tracked_descendants(&map, b"some"), 1);
         assert_eq!(tracked_descendants(&map, b"some/nested"), 1);
         assert_eq!(tracked_descendants(&map, b"some/nested/path"), 0);
 
         map.set_untracked(p(b"some/nested/path"))?;
         assert_eq!(map.len(), 0);
         assert!(map.get_map().get_node(p(b"some"))?.is_none());
 
         Ok(())
     }
 
     /// Test the simple case of all tracked, but multiple files
     #[test]
     fn test_tracked_descendants_multiple() -> Result<(), DirstateError> {
         let mut map = OwningDirstateMap::new_empty(vec![]);
 
         map.set_tracked(p(b"some/nested/path"))?;
         map.set_tracked(p(b"some/nested/file"))?;
         // one layer without any files to test deletion cascade
         map.set_tracked(p(b"some/other/nested/path"))?;
         map.set_tracked(p(b"root_file"))?;
         map.set_tracked(p(b"some/file"))?;
         map.set_tracked(p(b"some/file2"))?;
         map.set_tracked(p(b"some/file3"))?;
 
         assert_eq!(map.len(), 7);
         assert_eq!(tracked_descendants(&map, b"some"), 6);
         assert_eq!(tracked_descendants(&map, b"some/nested"), 2);
         assert_eq!(tracked_descendants(&map, b"some/other"), 1);
         assert_eq!(tracked_descendants(&map, b"some/other/nested"), 1);
         assert_eq!(tracked_descendants(&map, b"some/nested/path"), 0);
 
         map.set_untracked(p(b"some/nested/path"))?;
         assert_eq!(map.len(), 6);
         assert_eq!(tracked_descendants(&map, b"some"), 5);
         assert_eq!(tracked_descendants(&map, b"some/nested"), 1);
         assert_eq!(tracked_descendants(&map, b"some/other"), 1);
         assert_eq!(tracked_descendants(&map, b"some/other/nested"), 1);
 
         map.set_untracked(p(b"some/nested/file"))?;
         assert_eq!(map.len(), 5);
         assert_eq!(tracked_descendants(&map, b"some"), 4);
         assert_eq!(tracked_descendants(&map, b"some/other"), 1);
         assert_eq!(tracked_descendants(&map, b"some/other/nested"), 1);
         assert_does_not_exist(&map, b"some_nested");
 
         map.set_untracked(p(b"some/other/nested/path"))?;
         assert_eq!(map.len(), 4);
         assert_eq!(tracked_descendants(&map, b"some"), 3);
         assert_does_not_exist(&map, b"some/other");
 
         map.set_untracked(p(b"root_file"))?;
         assert_eq!(map.len(), 3);
         assert_eq!(tracked_descendants(&map, b"some"), 3);
         assert_does_not_exist(&map, b"root_file");
 
         map.set_untracked(p(b"some/file"))?;
         assert_eq!(map.len(), 2);
         assert_eq!(tracked_descendants(&map, b"some"), 2);
         assert_does_not_exist(&map, b"some/file");
 
         map.set_untracked(p(b"some/file2"))?;
         assert_eq!(map.len(), 1);
         assert_eq!(tracked_descendants(&map, b"some"), 1);
         assert_does_not_exist(&map, b"some/file2");
 
         map.set_untracked(p(b"some/file3"))?;
         assert_eq!(map.len(), 0);
         assert_does_not_exist(&map, b"some/file3");
 
         Ok(())
     }
 
     /// Check with a mix of tracked and non-tracked items
     #[test]
     fn test_tracked_descendants_different() -> Result<(), DirstateError> {
         let mut map = OwningDirstateMap::new_empty(vec![]);
 
         // A file that was just added
         map.set_tracked(p(b"some/nested/path"))?;
         // This has no information, the dirstate should ignore it
         map.reset_state(p(b"some/file"), false, false, false, false, None)?;
         assert_does_not_exist(&map, b"some/file");
 
         // A file that was removed
         map.reset_state(
             p(b"some/nested/file"),
             false,
             true,
             false,
             false,
             None,
         )?;
         assert!(!map.get(p(b"some/nested/file"))?.unwrap().tracked());
         // Only present in p2
         map.reset_state(p(b"some/file3"), false, false, true, false, None)?;
         assert!(!map.get(p(b"some/file3"))?.unwrap().tracked());
         // A file that was merged
         map.reset_state(p(b"root_file"), true, true, true, false, None)?;
         assert!(map.get(p(b"root_file"))?.unwrap().tracked());
         // A file that is added, with info from p2
         // XXX is that actually possible?
         map.reset_state(p(b"some/file2"), true, false, true, false, None)?;
         assert!(map.get(p(b"some/file2"))?.unwrap().tracked());
         // A clean file
         // One layer without any files to test deletion cascade
         map.reset_state(
             p(b"some/other/nested/path"),
             true,
             true,
             false,
             false,
             None,
         )?;
         assert!(map.get(p(b"some/other/nested/path"))?.unwrap().tracked());
 
         assert_eq!(map.len(), 6);
         assert_eq!(tracked_descendants(&map, b"some"), 3);
         assert_eq!(descendants_with_an_entry(&map, b"some"), 5);
         assert_eq!(tracked_descendants(&map, b"some/other/nested"), 1);
         assert_eq!(descendants_with_an_entry(&map, b"some/other/nested"), 1);
         assert_eq!(tracked_descendants(&map, b"some/other/nested/path"), 0);
         assert_eq!(
             descendants_with_an_entry(&map, b"some/other/nested/path"),
             0
         );
         assert_eq!(tracked_descendants(&map, b"some/nested"), 1);
         assert_eq!(descendants_with_an_entry(&map, b"some/nested"), 2);
 
         // might as well check this
         map.set_untracked(p(b"path/does/not/exist"))?;
         assert_eq!(map.len(), 6);
 
         map.set_untracked(p(b"some/other/nested/path"))?;
         // It is set untracked but not deleted since it held other information
         assert_eq!(map.len(), 6);
         assert_eq!(tracked_descendants(&map, b"some"), 2);
         assert_eq!(descendants_with_an_entry(&map, b"some"), 5);
         assert_eq!(descendants_with_an_entry(&map, b"some/other"), 1);
         assert_eq!(descendants_with_an_entry(&map, b"some/other/nested"), 1);
         assert_eq!(tracked_descendants(&map, b"some/nested"), 1);
         assert_eq!(descendants_with_an_entry(&map, b"some/nested"), 2);
 
         map.set_untracked(p(b"some/nested/path"))?;
         // It is set untracked *and* deleted since it was only added
         assert_eq!(map.len(), 5);
         assert_eq!(tracked_descendants(&map, b"some"), 1);
         assert_eq!(descendants_with_an_entry(&map, b"some"), 4);
         assert_eq!(tracked_descendants(&map, b"some/nested"), 0);
         assert_eq!(descendants_with_an_entry(&map, b"some/nested"), 1);
         assert_does_not_exist(&map, b"some/nested/path");
 
         map.set_untracked(p(b"root_file"))?;
         // Untracked but not deleted
         assert_eq!(map.len(), 5);
         assert!(map.get(p(b"root_file"))?.is_some());
 
         map.set_untracked(p(b"some/file2"))?;
         assert_eq!(map.len(), 5);
         assert_eq!(tracked_descendants(&map, b"some"), 0);
         assert!(map.get(p(b"some/file2"))?.is_some());
 
         map.set_untracked(p(b"some/file3"))?;
         assert_eq!(map.len(), 5);
         assert_eq!(tracked_descendants(&map, b"some"), 0);
         assert!(map.get(p(b"some/file3"))?.is_some());
 
         Ok(())
     }
 
     /// Check that copies counter is correctly updated
     #[test]
     fn test_copy_source() -> Result<(), DirstateError> {
         let mut map = OwningDirstateMap::new_empty(vec![]);
 
         // Clean file
         map.reset_state(p(b"files/clean"), true, true, false, false, None)?;
         // Merged file
         map.reset_state(p(b"files/from_p2"), true, true, true, false, None)?;
         // Removed file
         map.reset_state(p(b"removed"), false, true, false, false, None)?;
         // Added file
         map.reset_state(p(b"files/added"), true, false, false, false, None)?;
         // Add copy
         map.copy_map_insert(p(b"files/clean"), p(b"clean_copy_source"))?;
         assert_eq!(map.copy_map_len(), 1);
 
         // Copy override
         map.copy_map_insert(p(b"files/clean"), p(b"other_clean_copy_source"))?;
         assert_eq!(map.copy_map_len(), 1);
 
         // Multiple copies
         map.copy_map_insert(p(b"removed"), p(b"removed_copy_source"))?;
         assert_eq!(map.copy_map_len(), 2);
 
         map.copy_map_insert(p(b"files/added"), p(b"added_copy_source"))?;
         assert_eq!(map.copy_map_len(), 3);
 
         // Added, so the entry is completely removed
         map.set_untracked(p(b"files/added"))?;
         assert_does_not_exist(&map, b"files/added");
         assert_eq!(map.copy_map_len(), 2);
 
         // Removed, so the entry is kept around, so is its copy
         map.set_untracked(p(b"removed"))?;
         assert!(map.get(p(b"removed"))?.is_some());
         assert_eq!(map.copy_map_len(), 2);
 
         // Clean, so the entry is kept around, but not its copy
         map.set_untracked(p(b"files/clean"))?;
         assert!(map.get(p(b"files/clean"))?.is_some());
         assert_eq!(map.copy_map_len(), 1);
 
         map.copy_map_insert(p(b"files/from_p2"), p(b"from_p2_copy_source"))?;
         assert_eq!(map.copy_map_len(), 2);
 
         // Info from p2, so its copy source info is kept around
         map.set_untracked(p(b"files/from_p2"))?;
         assert!(map.get(p(b"files/from_p2"))?.is_some());
         assert_eq!(map.copy_map_len(), 2);
 
         Ok(())
     }
 
     /// Test with "on disk" data. For the sake of this test, the "on disk" data
     /// does not actually come from the disk, but it's opaque to the code being
     /// tested.
     #[test]
     fn test_on_disk() -> Result<(), DirstateError> {
         // First let's create some data to put "on disk"
         let mut map = OwningDirstateMap::new_empty(vec![]);
 
         // A file that was just added
         map.set_tracked(p(b"some/nested/added"))?;
         map.copy_map_insert(p(b"some/nested/added"), p(b"added_copy_source"))?;
 
         // A file that was removed
         map.reset_state(
             p(b"some/nested/removed"),
             false,
             true,
             false,
             false,
             None,
         )?;
         // Only present in p2
         map.reset_state(
             p(b"other/p2_info_only"),
             false,
             false,
             true,
             false,
             None,
         )?;
         map.copy_map_insert(
             p(b"other/p2_info_only"),
             p(b"other/p2_info_copy_source"),
         )?;
         // A file that was merged
         map.reset_state(p(b"merged"), true, true, true, false, None)?;
         // A file that is added, with info from p2
         // XXX is that actually possible?
         map.reset_state(
             p(b"other/added_with_p2"),
             true,
             false,
             true,
             false,
             None,
         )?;
         // One layer without any files to test deletion cascade
         // A clean file
         map.reset_state(
             p(b"some/other/nested/clean"),
             true,
             true,
             false,
             false,
             None,
         )?;
 
         let (packed, metadata, _should_append, _old_data_size) =
             map.pack_v2(DirstateMapWriteMode::ForceNewDataFile)?;
         let packed_len = packed.len();
         assert!(packed_len > 0);
 
         // Recreate "from disk"
         let mut map = OwningDirstateMap::new_v2(
             packed,
             packed_len,
             metadata.as_bytes(),
             vec![],
             None,
         )?;
 
         // Check that everything is accounted for
         assert!(map.contains_key(p(b"some/nested/added"))?);
         assert!(map.contains_key(p(b"some/nested/removed"))?);
         assert!(map.contains_key(p(b"merged"))?);
         assert!(map.contains_key(p(b"other/p2_info_only"))?);
         assert!(map.contains_key(p(b"other/added_with_p2"))?);
         assert!(map.contains_key(p(b"some/other/nested/clean"))?);
         assert_eq!(
             map.copy_map_get(p(b"some/nested/added"))?,
             Some(p(b"added_copy_source"))
         );
         assert_eq!(
             map.copy_map_get(p(b"other/p2_info_only"))?,
             Some(p(b"other/p2_info_copy_source"))
         );
         assert_eq!(tracked_descendants(&map, b"some"), 2);
         assert_eq!(descendants_with_an_entry(&map, b"some"), 3);
         assert_eq!(tracked_descendants(&map, b"other"), 1);
         assert_eq!(descendants_with_an_entry(&map, b"other"), 2);
         assert_eq!(tracked_descendants(&map, b"some/other"), 1);
         assert_eq!(descendants_with_an_entry(&map, b"some/other"), 1);
         assert_eq!(tracked_descendants(&map, b"some/other/nested"), 1);
         assert_eq!(descendants_with_an_entry(&map, b"some/other/nested"), 1);
         assert_eq!(tracked_descendants(&map, b"some/nested"), 1);
         assert_eq!(descendants_with_an_entry(&map, b"some/nested"), 2);
         assert_eq!(map.len(), 6);
         assert_eq!(map.get_map().unreachable_bytes, 0);
         assert_eq!(map.copy_map_len(), 2);
 
         // Shouldn't change anything since it's already not tracked
         map.set_untracked(p(b"some/nested/removed"))?;
         assert_eq!(map.get_map().unreachable_bytes, 0);
 
         if let ChildNodes::InMemory(_) = map.get_map().root {
             panic!("root should not have been mutated")
         }
         // We haven't mutated enough (nothing, actually), we should still be in
         // the append strategy
         assert!(map.get_map().write_should_append());
 
         // But this mutates the structure, so there should be unreachable_bytes
         assert!(map.set_untracked(p(b"some/nested/added"))?);
         let unreachable_bytes = map.get_map().unreachable_bytes;
         assert!(unreachable_bytes > 0);
 
         if let ChildNodes::OnDisk(_) = map.get_map().root {
             panic!("root should have been mutated")
         }
 
         // This should not mutate the structure either, since `root` has
         // already been mutated along with its direct children.
         map.set_untracked(p(b"merged"))?;
         assert_eq!(map.get_map().unreachable_bytes, unreachable_bytes);
 
         if let NodeRef::InMemory(_, _) =
             map.get_map().get_node(p(b"other/added_with_p2"))?.unwrap()
         {
             panic!("'other/added_with_p2' should not have been mutated")
         }
         // But this should, since it's in a different path
         // than `<root>some/nested/add`
         map.set_untracked(p(b"other/added_with_p2"))?;
         assert!(map.get_map().unreachable_bytes > unreachable_bytes);
 
         if let NodeRef::OnDisk(_) =
             map.get_map().get_node(p(b"other/added_with_p2"))?.unwrap()
         {
             panic!("'other/added_with_p2' should have been mutated")
         }
 
         // We have rewritten most of the tree, we should create a new file
         assert!(!map.get_map().write_should_append());
 
         Ok(())
     }
 }

rust/hg-core/src/dirstate_tree/status.rs

@@ -1,1055 +1,1055 b''
 use crate::dirstate::entry::TruncatedTimestamp;
 use crate::dirstate::status::IgnoreFnType;
 use crate::dirstate::status::StatusPath;
 use crate::dirstate_tree::dirstate_map::BorrowedPath;
 use crate::dirstate_tree::dirstate_map::ChildNodesRef;
 use crate::dirstate_tree::dirstate_map::DirstateMap;
 use crate::dirstate_tree::dirstate_map::DirstateVersion;
 use crate::dirstate_tree::dirstate_map::NodeRef;
 use crate::dirstate_tree::on_disk::DirstateV2ParseError;
 use crate::matchers::get_ignore_function;
 use crate::matchers::{Matcher, VisitChildrenSet};
 use crate::utils::files::get_bytes_from_os_string;
 use crate::utils::files::get_bytes_from_path;
 use crate::utils::files::get_path_from_bytes;
 use crate::utils::hg_path::hg_path_to_path_buf;
 use crate::utils::hg_path::HgPath;
 use crate::BadMatch;
 use crate::BadType;
 use crate::DirstateStatus;
 use crate::HgPathCow;
 use crate::PatternFileWarning;
 use crate::StatusError;
 use crate::StatusOptions;
 use once_cell::sync::OnceCell;
 use rayon::prelude::*;
 use sha1::{Digest, Sha1};
 use std::borrow::Cow;
 use std::io;
 use std::os::unix::prelude::FileTypeExt;
 use std::path::Path;
 use std::path::PathBuf;
 use std::sync::Mutex;
 use std::time::SystemTime;
 
 /// Returns the status of the working directory compared to its parent
 /// changeset.
 ///
 /// This algorithm is based on traversing the filesystem tree (`fs` in function
 /// and variable names) and dirstate tree at the same time. The core of this
 /// traversal is the recursive `traverse_fs_directory_and_dirstate` function
 /// and its use of `itertools::merge_join_by`. When reaching a path that only
 /// exists in one of the two trees, depending on information requested by
 /// `options` we may need to traverse the remaining subtree.
 #[logging_timer::time("trace")]
 pub fn status<'dirstate>(
     dmap: &'dirstate mut DirstateMap,
     matcher: &(dyn Matcher + Sync),
     root_dir: PathBuf,
     ignore_files: Vec<PathBuf>,
     options: StatusOptions,
 ) -> Result<(DirstateStatus<'dirstate>, Vec<PatternFileWarning>), StatusError>
 {
     // Also cap for a Python caller of this function, but don't complain if
     // the global threadpool has already been set since this code path is also
     // being used by `rhg`, which calls this early.
     let _ = crate::utils::cap_default_rayon_threads();
 
     let (ignore_fn, warnings, patterns_changed): (IgnoreFnType, _, _) =
         if options.list_ignored || options.list_unknown {
             let (ignore_fn, warnings, changed) = match dmap.dirstate_version {
                 DirstateVersion::V1 => {
                     let (ignore_fn, warnings) = get_ignore_function(
                         ignore_files,
                         &root_dir,
                         &mut |_source, _pattern_bytes| {},
                     )?;
                     (ignore_fn, warnings, None)
                 }
                 DirstateVersion::V2 => {
                     let mut hasher = Sha1::new();
                     let (ignore_fn, warnings) = get_ignore_function(
                         ignore_files,
                         &root_dir,
                         &mut |source, pattern_bytes| {
                             // If inside the repo, use the relative version to
                             // make it deterministic inside tests.
                             // The performance hit should be negligible.
                             let source = source
                                 .strip_prefix(&root_dir)
                                 .unwrap_or(source);
                             let source = get_bytes_from_path(source);
 
                             let mut subhasher = Sha1::new();
                             subhasher.update(pattern_bytes);
                             let patterns_hash = subhasher.finalize();
 
                             hasher.update(source);
                             hasher.update(b" ");
                             hasher.update(patterns_hash);
                             hasher.update(b"\n");
                         },
                     )?;
                     let new_hash = *hasher.finalize().as_ref();
                     let changed = new_hash != dmap.ignore_patterns_hash;
                     dmap.ignore_patterns_hash = new_hash;
                     (ignore_fn, warnings, Some(changed))
                 }
             };
             (ignore_fn, warnings, changed)
         } else {
             (Box::new(|&_| true), vec![], None)
         };
 
     let filesystem_time_at_status_start =
         filesystem_now(&root_dir).ok().map(TruncatedTimestamp::from);
 
     // If the repository is under the current directory, prefer using a
     // relative path, so the kernel needs to traverse fewer directory in every
     // call to `read_dir` or `symlink_metadata`.
     // This is effective in the common case where the current directory is the
     // repository root.
 
     // TODO: Better yet would be to use libc functions like `openat` and
     // `fstatat` to remove such repeated traversals entirely, but the standard
     // library does not provide APIs based on those.
     // Maybe with a crate like https://crates.io/crates/openat instead?
     let root_dir = if let Some(relative) = std::env::current_dir()
         .ok()
         .and_then(|cwd| root_dir.strip_prefix(cwd).ok())
     {
         relative
     } else {
         &root_dir
     };
 
     let outcome = DirstateStatus {
         filesystem_time_at_status_start,
         ..Default::default()
     };
     let common = StatusCommon {
         dmap,
         options,
         matcher,
         ignore_fn,
         outcome: Mutex::new(outcome),
         ignore_patterns_have_changed: patterns_changed,
         new_cacheable_directories: Default::default(),
         outdated_cached_directories: Default::default(),
         filesystem_time_at_status_start,
     };
     let is_at_repo_root = true;
     let hg_path = &BorrowedPath::OnDisk(HgPath::new(""));
     let has_ignored_ancestor = HasIgnoredAncestor::create(None, hg_path);
     let root_cached_mtime = None;
     // If the path we have for the repository root is a symlink, do follow it.
     // (As opposed to symlinks within the working directory which are not
     // followed, using `std::fs::symlink_metadata`.)
     common.traverse_fs_directory_and_dirstate(
         &has_ignored_ancestor,
         dmap.root.as_ref(),
         hg_path,
         &DirEntry {
             hg_path: Cow::Borrowed(HgPath::new(b"")),
             fs_path: Cow::Borrowed(root_dir),
             symlink_metadata: None,
             file_type: FakeFileType::Directory,
         },
         root_cached_mtime,
         is_at_repo_root,
     )?;
     if let Some(file_set) = common.matcher.file_set() {
         for file in file_set {
             if !file.is_empty() && !dmap.has_node(file)? {
                 let path = hg_path_to_path_buf(file)?;
                 if let io::Result::Err(error) =
                     root_dir.join(path).symlink_metadata()
                 {
                     common.io_error(error, file)
                 }
             }
         }
     }
     let mut outcome = common.outcome.into_inner().unwrap();
     let new_cacheable = common.new_cacheable_directories.into_inner().unwrap();
     let outdated = common.outdated_cached_directories.into_inner().unwrap();
 
     outcome.dirty = common.ignore_patterns_have_changed == Some(true)
         || !outdated.is_empty()
         || (!new_cacheable.is_empty()
             && dmap.dirstate_version == DirstateVersion::V2);
 
     // Remove outdated mtimes before adding new mtimes, in case a given
     // directory is both
     for path in &outdated {
         dmap.clear_cached_mtime(path)?;
     }
     for (path, mtime) in &new_cacheable {
         dmap.set_cached_mtime(path, *mtime)?;
     }
 
     Ok((outcome, warnings))
 }
 
 /// Bag of random things needed by various parts of the algorithm. Reduces the
 /// number of parameters passed to functions.
 struct StatusCommon<'a, 'tree, 'on_disk: 'tree> {
     dmap: &'tree DirstateMap<'on_disk>,
     options: StatusOptions,
     matcher: &'a (dyn Matcher + Sync),
     ignore_fn: IgnoreFnType<'a>,
     outcome: Mutex<DirstateStatus<'on_disk>>,
     /// New timestamps of directories to be used for caching their readdirs
     new_cacheable_directories:
         Mutex<Vec<(Cow<'on_disk, HgPath>, TruncatedTimestamp)>>,
     /// Used to invalidate the readdir cache of directories
     outdated_cached_directories: Mutex<Vec<Cow<'on_disk, HgPath>>>,
 
     /// Whether ignore files like `.hgignore` have changed since the previous
     /// time a `status()` call wrote their hash to the dirstate. `None` means
     /// we don’t know as this run doesn’t list either ignored or uknown files
     /// and therefore isn’t reading `.hgignore`.
     ignore_patterns_have_changed: Option<bool>,
 
     /// The current time at the start of the `status()` algorithm, as measured
     /// and possibly truncated by the filesystem.
     filesystem_time_at_status_start: Option<TruncatedTimestamp>,
 }
 
 enum Outcome {
     Modified,
     Added,
     Removed,
     Deleted,
     Clean,
     Ignored,
     Unknown,
     Unsure,
 }
 
 /// Lazy computation of whether a given path has a hgignored
 /// ancestor.
 struct HasIgnoredAncestor<'a> {
     /// `path` and `parent` constitute the inputs to the computation,
     /// `cache` stores the outcome.
     path: &'a HgPath,
     parent: Option<&'a HasIgnoredAncestor<'a>>,
     cache: OnceCell<bool>,
 }
 
 impl<'a> HasIgnoredAncestor<'a> {
     fn create(
         parent: Option<&'a HasIgnoredAncestor<'a>>,
         path: &'a HgPath,
     ) -> HasIgnoredAncestor<'a> {
         Self {
             path,
             parent,
             cache: OnceCell::new(),
         }
     }
 
-    fn force<'b>(&self, ignore_fn: &IgnoreFnType<'b>) -> bool {
+    fn force(&self, ignore_fn: &IgnoreFnType<'_>) -> bool {
         match self.parent {
             None => false,
             Some(parent) => {
                 *(self.cache.get_or_init(|| {
                     parent.force(ignore_fn) || ignore_fn(self.path)
                 }))
             }
         }
     }
 }
 
 impl<'a, 'tree, 'on_disk> StatusCommon<'a, 'tree, 'on_disk> {
     fn push_outcome(
         &self,
         which: Outcome,
         dirstate_node: &NodeRef<'tree, 'on_disk>,
     ) -> Result<(), DirstateV2ParseError> {
         let path = dirstate_node
             .full_path_borrowed(self.dmap.on_disk)?
             .detach_from_tree();
         let copy_source = if self.options.list_copies {
             dirstate_node
                 .copy_source_borrowed(self.dmap.on_disk)?
                 .map(|source| source.detach_from_tree())
         } else {
             None
         };
         self.push_outcome_common(which, path, copy_source);
         Ok(())
     }
 
     fn push_outcome_without_copy_source(
         &self,
         which: Outcome,
         path: &BorrowedPath<'_, 'on_disk>,
     ) {
         self.push_outcome_common(which, path.detach_from_tree(), None)
     }
 
     fn push_outcome_common(
         &self,
         which: Outcome,
         path: HgPathCow<'on_disk>,
         copy_source: Option<HgPathCow<'on_disk>>,
     ) {
         let mut outcome = self.outcome.lock().unwrap();
         let vec = match which {
             Outcome::Modified => &mut outcome.modified,
             Outcome::Added => &mut outcome.added,
             Outcome::Removed => &mut outcome.removed,
             Outcome::Deleted => &mut outcome.deleted,
             Outcome::Clean => &mut outcome.clean,
             Outcome::Ignored => &mut outcome.ignored,
             Outcome::Unknown => &mut outcome.unknown,
             Outcome::Unsure => &mut outcome.unsure,
         };
         vec.push(StatusPath { path, copy_source });
     }
 
     fn read_dir(
         &self,
         hg_path: &HgPath,
         fs_path: &Path,
         is_at_repo_root: bool,
     ) -> Result<Vec<DirEntry>, ()> {
         DirEntry::read_dir(fs_path, is_at_repo_root)
             .map_err(|error| self.io_error(error, hg_path))
     }
 
     fn io_error(&self, error: std::io::Error, hg_path: &HgPath) {
         let errno = error.raw_os_error().expect("expected real OS error");
         self.outcome
             .lock()
             .unwrap()
             .bad
             .push((hg_path.to_owned().into(), BadMatch::OsError(errno)))
     }
 
     fn check_for_outdated_directory_cache(
         &self,
         dirstate_node: &NodeRef<'tree, 'on_disk>,
     ) -> Result<bool, DirstateV2ParseError> {
         if self.ignore_patterns_have_changed == Some(true)
             && dirstate_node.cached_directory_mtime()?.is_some()
         {
             self.outdated_cached_directories.lock().unwrap().push(
                 dirstate_node
                     .full_path_borrowed(self.dmap.on_disk)?
                     .detach_from_tree(),
             );
             return Ok(true);
         }
         Ok(false)
     }
 
     /// If this returns true, we can get accurate results by only using
     /// `symlink_metadata` for child nodes that exist in the dirstate and don’t
     /// need to call `read_dir`.
     fn can_skip_fs_readdir(
         &self,
         directory_entry: &DirEntry,
         cached_directory_mtime: Option<TruncatedTimestamp>,
     ) -> bool {
         if !self.options.list_unknown && !self.options.list_ignored {
             // All states that we care about listing have corresponding
             // dirstate entries.
             // This happens for example with `hg status -mard`.
             return true;
         }
         if !self.options.list_ignored
             && self.ignore_patterns_have_changed == Some(false)
         {
             if let Some(cached_mtime) = cached_directory_mtime {
                 // The dirstate contains a cached mtime for this directory, set
                 // by a previous run of the `status` algorithm which found this
                 // directory eligible for `read_dir` caching.
                 if let Ok(meta) = directory_entry.symlink_metadata() {
                     if cached_mtime
                         .likely_equal_to_mtime_of(&meta)
                         .unwrap_or(false)
                     {
                         // The mtime of that directory has not changed
                         // since then, which means that the results of
                         // `read_dir` should also be unchanged.
                         return true;
                     }
                 }
             }
         }
         false
     }
 
     fn should_visit(set: &VisitChildrenSet, basename: &HgPath) -> bool {
         match set {
             VisitChildrenSet::This | VisitChildrenSet::Recursive => true,
             VisitChildrenSet::Empty => false,
             VisitChildrenSet::Set(children_to_visit) => {
                 children_to_visit.contains(basename)
             }
         }
     }
 
     /// Returns whether all child entries of the filesystem directory have a
     /// corresponding dirstate node or are ignored.
     fn traverse_fs_directory_and_dirstate<'ancestor>(
         &self,
         has_ignored_ancestor: &'ancestor HasIgnoredAncestor<'ancestor>,
         dirstate_nodes: ChildNodesRef<'tree, 'on_disk>,
         directory_hg_path: &BorrowedPath<'tree, 'on_disk>,
         directory_entry: &DirEntry,
         cached_directory_mtime: Option<TruncatedTimestamp>,
         is_at_repo_root: bool,
     ) -> Result<bool, DirstateV2ParseError> {
         let children_set = self.matcher.visit_children_set(directory_hg_path);
         if let VisitChildrenSet::Empty = children_set {
             return Ok(false);
         }
         if self.can_skip_fs_readdir(directory_entry, cached_directory_mtime) {
             dirstate_nodes
                 .par_iter()
                 .map(|dirstate_node| {
                     let fs_path = &directory_entry.fs_path;
                     let basename =
                         dirstate_node.base_name(self.dmap.on_disk)?.as_bytes();
                     let fs_path = fs_path.join(get_path_from_bytes(basename));
                     if !Self::should_visit(
                         &children_set,
                         HgPath::new(basename),
                     ) {
                         return Ok(());
                     }
                     match std::fs::symlink_metadata(&fs_path) {
                         Ok(fs_metadata) => {
                             let file_type = fs_metadata.file_type().into();
                             let entry = DirEntry {
                                 hg_path: Cow::Borrowed(
                                     dirstate_node
                                         .full_path(self.dmap.on_disk)?,
                                 ),
                                 fs_path: Cow::Borrowed(&fs_path),
                                 symlink_metadata: Some(fs_metadata),
                                 file_type,
                             };
                             self.traverse_fs_and_dirstate(
                                 &entry,
                                 dirstate_node,
                                 has_ignored_ancestor,
                             )
                         }
                         Err(e) if e.kind() == std::io::ErrorKind::NotFound => {
                             self.traverse_dirstate_only(dirstate_node)
                         }
                         Err(error) => {
                             let hg_path =
                                 dirstate_node.full_path(self.dmap.on_disk)?;
                             self.io_error(error, hg_path);
                             Ok(())
                         }
                     }
                 })
                 .collect::<Result<_, _>>()?;
 
             // We don’t know, so conservatively say this isn’t the case
             let children_all_have_dirstate_node_or_are_ignored = false;
 
             return Ok(children_all_have_dirstate_node_or_are_ignored);
         }
 
         let readdir_succeeded;
         let mut fs_entries = if let Ok(entries) = self.read_dir(
             directory_hg_path,
             &directory_entry.fs_path,
             is_at_repo_root,
         ) {
             readdir_succeeded = true;
             entries
         } else {
             // Treat an unreadable directory (typically because of insufficient
             // permissions) like an empty directory. `self.read_dir` has
             // already called `self.io_error` so a warning will be emitted.
             // We still need to remember that there was an error so that we
             // know not to cache this result.
             readdir_succeeded = false;
             Vec::new()
         };
 
         // `merge_join_by` requires both its input iterators to be sorted:
 
         let dirstate_nodes = dirstate_nodes.sorted();
         // `sort_unstable_by_key` doesn’t allow keys borrowing from the value:
         // https://github.com/rust-lang/rust/issues/34162
         fs_entries.sort_unstable_by(|e1, e2| e1.hg_path.cmp(&e2.hg_path));
 
         // Propagate here any error that would happen inside the comparison
         // callback below
         for dirstate_node in &dirstate_nodes {
             dirstate_node.base_name(self.dmap.on_disk)?;
         }
         itertools::merge_join_by(
             dirstate_nodes,
             &fs_entries,
             |dirstate_node, fs_entry| {
                 // This `unwrap` never panics because we already propagated
                 // those errors above
                 dirstate_node
                     .base_name(self.dmap.on_disk)
                     .unwrap()
                     .cmp(&fs_entry.hg_path)
             },
         )
         .par_bridge()
         .map(|pair| {
             use itertools::EitherOrBoth::*;
             let basename = match &pair {
                 Left(dirstate_node) | Both(dirstate_node, _) => HgPath::new(
                     dirstate_node.base_name(self.dmap.on_disk)?.as_bytes(),
                 ),
                 Right(fs_entry) => &fs_entry.hg_path,
             };
             if !Self::should_visit(&children_set, basename) {
                 return Ok(false);
             }
             let has_dirstate_node_or_is_ignored = match pair {
                 Both(dirstate_node, fs_entry) => {
                     self.traverse_fs_and_dirstate(
                         fs_entry,
                         dirstate_node,
                         has_ignored_ancestor,
                     )?;
                     true
                 }
                 Left(dirstate_node) => {
                     self.traverse_dirstate_only(dirstate_node)?;
                     true
                 }
                 Right(fs_entry) => self.traverse_fs_only(
                     has_ignored_ancestor.force(&self.ignore_fn),
                     directory_hg_path,
                     fs_entry,
                 ),
             };
             Ok(has_dirstate_node_or_is_ignored)
         })
         .try_reduce(|| true, |a, b| Ok(a && b))
         .map(|res| res && readdir_succeeded)
     }
 
     fn traverse_fs_and_dirstate<'ancestor>(
         &self,
         fs_entry: &DirEntry,
         dirstate_node: NodeRef<'tree, 'on_disk>,
         has_ignored_ancestor: &'ancestor HasIgnoredAncestor<'ancestor>,
     ) -> Result<(), DirstateV2ParseError> {
         let outdated_dircache =
             self.check_for_outdated_directory_cache(&dirstate_node)?;
         let hg_path = &dirstate_node.full_path_borrowed(self.dmap.on_disk)?;
         let file_or_symlink = fs_entry.is_file() || fs_entry.is_symlink();
         if !file_or_symlink {
             // If we previously had a file here, it was removed (with
             // `hg rm` or similar) or deleted before it could be
             // replaced by a directory or something else.
             self.mark_removed_or_deleted_if_file(&dirstate_node)?;
         }
         if let Some(bad_type) = fs_entry.is_bad() {
             if self.matcher.exact_match(hg_path) {
                 let path = dirstate_node.full_path(self.dmap.on_disk)?;
                 self.outcome.lock().unwrap().bad.push((
                     path.to_owned().into(),
                     BadMatch::BadType(bad_type),
                 ))
             }
         }
         if fs_entry.is_dir() {
             if self.options.collect_traversed_dirs {
                 self.outcome
                     .lock()
                     .unwrap()
                     .traversed
                     .push(hg_path.detach_from_tree())
             }
             let is_ignored = HasIgnoredAncestor::create(
                 Some(has_ignored_ancestor),
                 hg_path,
             );
             let is_at_repo_root = false;
             let children_all_have_dirstate_node_or_are_ignored = self
                 .traverse_fs_directory_and_dirstate(
                     &is_ignored,
                     dirstate_node.children(self.dmap.on_disk)?,
                     hg_path,
                     fs_entry,
                     dirstate_node.cached_directory_mtime()?,
                     is_at_repo_root,
                 )?;
             self.maybe_save_directory_mtime(
                 children_all_have_dirstate_node_or_are_ignored,
                 fs_entry,
                 dirstate_node,
                 outdated_dircache,
             )?
         } else {
             if file_or_symlink && self.matcher.matches(hg_path) {
                 if let Some(entry) = dirstate_node.entry()? {
                     if !entry.any_tracked() {
                         // Forward-compat if we start tracking unknown/ignored
                         // files for caching reasons
                         self.mark_unknown_or_ignored(
                             has_ignored_ancestor.force(&self.ignore_fn),
                             hg_path,
                         );
                     }
                     if entry.added() {
                         self.push_outcome(Outcome::Added, &dirstate_node)?;
                     } else if entry.removed() {
                         self.push_outcome(Outcome::Removed, &dirstate_node)?;
                     } else if entry.modified() {
                         self.push_outcome(Outcome::Modified, &dirstate_node)?;
                     } else {
                         self.handle_normal_file(&dirstate_node, fs_entry)?;
                     }
                 } else {
                     // `node.entry.is_none()` indicates a "directory"
                     // node, but the filesystem has a file
                     self.mark_unknown_or_ignored(
                         has_ignored_ancestor.force(&self.ignore_fn),
                         hg_path,
                     );
                 }
             }
 
             for child_node in dirstate_node.children(self.dmap.on_disk)?.iter()
             {
                 self.traverse_dirstate_only(child_node)?
             }
         }
         Ok(())
     }
 
     /// Save directory mtime if applicable.
     ///
     /// `outdated_directory_cache` is `true` if we've just invalidated the
     /// cache for this directory in `check_for_outdated_directory_cache`,
     /// which forces the update.
     fn maybe_save_directory_mtime(
         &self,
         children_all_have_dirstate_node_or_are_ignored: bool,
         directory_entry: &DirEntry,
         dirstate_node: NodeRef<'tree, 'on_disk>,
         outdated_directory_cache: bool,
     ) -> Result<(), DirstateV2ParseError> {
         if !children_all_have_dirstate_node_or_are_ignored {
             return Ok(());
         }
         // All filesystem directory entries from `read_dir` have a
         // corresponding node in the dirstate, so we can reconstitute the
         // names of those entries without calling `read_dir` again.
 
         // TODO: use let-else here and below when available:
         // https://github.com/rust-lang/rust/issues/87335
         let status_start = if let Some(status_start) =
             &self.filesystem_time_at_status_start
         {
             status_start
         } else {
             return Ok(());
         };
 
         // Although the Rust standard library’s `SystemTime` type
         // has nanosecond precision, the times reported for a
         // directory’s (or file’s) modified time may have lower
         // resolution based on the filesystem (for example ext3
         // only stores integer seconds), kernel (see
         // https://stackoverflow.com/a/14393315/1162888), etc.
         let metadata = match directory_entry.symlink_metadata() {
             Ok(meta) => meta,
             Err(_) => return Ok(()),
         };
 
         let directory_mtime = match TruncatedTimestamp::for_reliable_mtime_of(
             &metadata,
             status_start,
         ) {
             Ok(Some(directory_mtime)) => directory_mtime,
             Ok(None) => {
                 // The directory was modified too recently,
                 // don’t cache its `read_dir` results.
                 //
                 // 1. A change to this directory (direct child was
                 //    added or removed) cause its mtime to be set
                 //    (possibly truncated) to `directory_mtime`
                 // 2. This `status` algorithm calls `read_dir`
                 // 3. An other change is made to the same directory is
                 //    made so that calling `read_dir` agin would give
                 //    different results, but soon enough after 1. that
                 //    the mtime stays the same
                 //
                 // On a system where the time resolution poor, this
                 // scenario is not unlikely if all three steps are caused
                 // by the same script.
                 return Ok(());
             }
             Err(_) => {
                 // OS/libc does not support mtime?
                 return Ok(());
             }
         };
         // We’ve observed (through `status_start`) that time has
         // “progressed” since `directory_mtime`, so any further
         // change to this directory is extremely likely to cause a
         // different mtime.
         //
         // Having the same mtime again is not entirely impossible
         // since the system clock is not monotonous. It could jump
         // backward to some point before `directory_mtime`, then a
         // directory change could potentially happen during exactly
         // the wrong tick.
         //
         // We deem this scenario (unlike the previous one) to be
         // unlikely enough in practice.
 
         let is_up_to_date = if let Some(cached) =
             dirstate_node.cached_directory_mtime()?
         {
             !outdated_directory_cache && cached.likely_equal(directory_mtime)
         } else {
             false
         };
         if !is_up_to_date {
             let hg_path = dirstate_node
                 .full_path_borrowed(self.dmap.on_disk)?
                 .detach_from_tree();
             self.new_cacheable_directories
                 .lock()
                 .unwrap()
                 .push((hg_path, directory_mtime))
         }
         Ok(())
     }
 
     /// A file that is clean in the dirstate was found in the filesystem
     fn handle_normal_file(
         &self,
         dirstate_node: &NodeRef<'tree, 'on_disk>,
         fs_entry: &DirEntry,
     ) -> Result<(), DirstateV2ParseError> {
         // Keep the low 31 bits
         fn truncate_u64(value: u64) -> i32 {
             (value & 0x7FFF_FFFF) as i32
         }
 
         let fs_metadata = match fs_entry.symlink_metadata() {
             Ok(meta) => meta,
             Err(_) => return Ok(()),
         };
 
         let entry = dirstate_node
             .entry()?
             .expect("handle_normal_file called with entry-less node");
         let mode_changed =
             || self.options.check_exec && entry.mode_changed(&fs_metadata);
         let size = entry.size();
         let size_changed = size != truncate_u64(fs_metadata.len());
         if size >= 0 && size_changed && fs_metadata.file_type().is_symlink() {
             // issue6456: Size returned may be longer due to encryption
             // on EXT-4 fscrypt. TODO maybe only do it on EXT4?
             self.push_outcome(Outcome::Unsure, dirstate_node)?
         } else if dirstate_node.has_copy_source()
             || entry.is_from_other_parent()
             || (size >= 0 && (size_changed || mode_changed()))
         {
             self.push_outcome(Outcome::Modified, dirstate_node)?
         } else {
             let mtime_looks_clean = if let Some(dirstate_mtime) =
                 entry.truncated_mtime()
             {
                 let fs_mtime = TruncatedTimestamp::for_mtime_of(&fs_metadata)
                     .expect("OS/libc does not support mtime?");
                 // There might be a change in the future if for example the
                 // internal clock become off while process run, but this is a
                 // case where the issues the user would face
                 // would be a lot worse and there is nothing we
                 // can really do.
                 fs_mtime.likely_equal(dirstate_mtime)
             } else {
                 // No mtime in the dirstate entry
                 false
             };
             if !mtime_looks_clean {
                 self.push_outcome(Outcome::Unsure, dirstate_node)?
             } else if self.options.list_clean {
                 self.push_outcome(Outcome::Clean, dirstate_node)?
             }
         }
         Ok(())
     }
 
     /// A node in the dirstate tree has no corresponding filesystem entry
     fn traverse_dirstate_only(
         &self,
         dirstate_node: NodeRef<'tree, 'on_disk>,
     ) -> Result<(), DirstateV2ParseError> {
         self.check_for_outdated_directory_cache(&dirstate_node)?;
         self.mark_removed_or_deleted_if_file(&dirstate_node)?;
         dirstate_node
             .children(self.dmap.on_disk)?
             .par_iter()
             .map(|child_node| self.traverse_dirstate_only(child_node))
             .collect()
     }
 
     /// A node in the dirstate tree has no corresponding *file* on the
     /// filesystem
     ///
     /// Does nothing on a "directory" node
     fn mark_removed_or_deleted_if_file(
         &self,
         dirstate_node: &NodeRef<'tree, 'on_disk>,
     ) -> Result<(), DirstateV2ParseError> {
         if let Some(entry) = dirstate_node.entry()? {
             if !entry.any_tracked() {
                 // Future-compat for when we start storing ignored and unknown
                 // files for caching reasons
                 return Ok(());
             }
             let path = dirstate_node.full_path(self.dmap.on_disk)?;
             if self.matcher.matches(path) {
                 if entry.removed() {
                     self.push_outcome(Outcome::Removed, dirstate_node)?
                 } else {
                     self.push_outcome(Outcome::Deleted, dirstate_node)?
                 }
             }
         }
         Ok(())
     }
 
     /// Something in the filesystem has no corresponding dirstate node
     ///
     /// Returns whether that path is ignored
     fn traverse_fs_only(
         &self,
         has_ignored_ancestor: bool,
         directory_hg_path: &HgPath,
         fs_entry: &DirEntry,
     ) -> bool {
         let hg_path = directory_hg_path.join(&fs_entry.hg_path);
         let file_or_symlink = fs_entry.is_file() || fs_entry.is_symlink();
         if fs_entry.is_dir() {
             let is_ignored =
                 has_ignored_ancestor || (self.ignore_fn)(&hg_path);
             let traverse_children = if is_ignored {
                 // Descendants of an ignored directory are all ignored
                 self.options.list_ignored
             } else {
                 // Descendants of an unknown directory may be either unknown or
                 // ignored
                 self.options.list_unknown || self.options.list_ignored
             };
             if traverse_children {
                 let is_at_repo_root = false;
                 if let Ok(children_fs_entries) =
                     self.read_dir(&hg_path, &fs_entry.fs_path, is_at_repo_root)
                 {
                     children_fs_entries.par_iter().for_each(|child_fs_entry| {
                         self.traverse_fs_only(
                             is_ignored,
                             &hg_path,
                             child_fs_entry,
                         );
                     })
                 }
                 if self.options.collect_traversed_dirs {
                     self.outcome.lock().unwrap().traversed.push(hg_path.into())
                 }
             }
             is_ignored
         } else if file_or_symlink {
             if self.matcher.matches(&hg_path) {
                 self.mark_unknown_or_ignored(
                     has_ignored_ancestor,
                     &BorrowedPath::InMemory(&hg_path),
                 )
             } else {
                 // We haven’t computed whether this path is ignored. It
                 // might not be, and a future run of status might have a
                 // different matcher that matches it. So treat it as not
                 // ignored. That is, inhibit readdir caching of the parent
                 // directory.
                 false
             }
         } else {
             // This is neither a directory, a plain file, or a symlink.
             // Treat it like an ignored file.
             true
         }
     }
 
     /// Returns whether that path is ignored
     fn mark_unknown_or_ignored(
         &self,
         has_ignored_ancestor: bool,
         hg_path: &BorrowedPath<'_, 'on_disk>,
     ) -> bool {
         let is_ignored = has_ignored_ancestor || (self.ignore_fn)(hg_path);
         if is_ignored {
             if self.options.list_ignored {
                 self.push_outcome_without_copy_source(
                     Outcome::Ignored,
                     hg_path,
                 )
             }
         } else if self.options.list_unknown {
             self.push_outcome_without_copy_source(Outcome::Unknown, hg_path)
         }
         is_ignored
     }
 }
 
 /// Since [`std::fs::FileType`] cannot be built directly, we emulate what we
 /// care about.
 #[derive(Copy, Clone, Debug, PartialEq, Eq)]
 enum FakeFileType {
     File,
     Directory,
     Symlink,
     BadType(BadType),
 }
 
 impl From<std::fs::FileType> for FakeFileType {
     fn from(f: std::fs::FileType) -> Self {
         if f.is_dir() {
             Self::Directory
         } else if f.is_file() {
             Self::File
         } else if f.is_symlink() {
             Self::Symlink
         } else if f.is_fifo() {
             Self::BadType(BadType::FIFO)
         } else if f.is_block_device() {
             Self::BadType(BadType::BlockDevice)
         } else if f.is_char_device() {
             Self::BadType(BadType::CharacterDevice)
         } else if f.is_socket() {
             Self::BadType(BadType::Socket)
         } else {
             Self::BadType(BadType::Unknown)
         }
     }
 }
 
 struct DirEntry<'a> {
     /// Path as stored in the dirstate, or just the filename for optimization.
     hg_path: HgPathCow<'a>,
     /// Filesystem path
     fs_path: Cow<'a, Path>,
     /// Lazily computed
     symlink_metadata: Option<std::fs::Metadata>,
     /// Already computed for ergonomics.
     file_type: FakeFileType,
 }
 
 impl<'a> DirEntry<'a> {
     /// Returns **unsorted** entries in the given directory, with name,
     /// metadata and file type.
     ///
     /// If a `.hg` sub-directory is encountered:
     ///
     /// * At the repository root, ignore that sub-directory
     /// * Elsewhere, we’re listing the content of a sub-repo. Return an empty
     ///   list instead.
     fn read_dir(path: &Path, is_at_repo_root: bool) -> io::Result<Vec<Self>> {
         // `read_dir` returns a "not found" error for the empty path
         let at_cwd = path == Path::new("");
         let read_dir_path = if at_cwd { Path::new(".") } else { path };
         let mut results = Vec::new();
         for entry in read_dir_path.read_dir()? {
             let entry = entry?;
             let file_type = match entry.file_type() {
                 Ok(v) => v,
                 Err(e) => {
                     // race with file deletion?
                     if e.kind() == std::io::ErrorKind::NotFound {
                         continue;
                     } else {
                         return Err(e);
                     }
                 }
             };
             let file_name = entry.file_name();
             // FIXME don't do this when cached
             if file_name == ".hg" {
                 if is_at_repo_root {
                     // Skip the repo’s own .hg (might be a symlink)
                     continue;
                 } else if file_type.is_dir() {
                     // A .hg sub-directory at another location means a subrepo,
                     // skip it entirely.
                     return Ok(Vec::new());
                 }
             }
             let full_path = if at_cwd {
                 file_name.clone().into()
             } else {
                 entry.path()
             };
             let filename =
                 Cow::Owned(get_bytes_from_os_string(file_name).into());
             let file_type = FakeFileType::from(file_type);
             results.push(DirEntry {
                 hg_path: filename,
                 fs_path: Cow::Owned(full_path.to_path_buf()),
                 symlink_metadata: None,
                 file_type,
             })
         }
         Ok(results)
     }
 
     fn symlink_metadata(&self) -> Result<std::fs::Metadata, std::io::Error> {
         match &self.symlink_metadata {
             Some(meta) => Ok(meta.clone()),
             None => std::fs::symlink_metadata(&self.fs_path),
         }
     }
 
     fn is_dir(&self) -> bool {
         self.file_type == FakeFileType::Directory
     }
 
     fn is_file(&self) -> bool {
         self.file_type == FakeFileType::File
     }
 
     fn is_symlink(&self) -> bool {
         self.file_type == FakeFileType::Symlink
     }
 
     fn is_bad(&self) -> Option<BadType> {
         match self.file_type {
             FakeFileType::BadType(ty) => Some(ty),
             _ => None,
         }
     }
 }
 
 /// Return the `mtime` of a temporary file newly-created in the `.hg` directory
 /// of the give repository.
 ///
 /// This is similar to `SystemTime::now()`, with the result truncated to the
 /// same time resolution as other files’ modification times. Using `.hg`
 /// instead of the system’s default temporary directory (such as `/tmp`) makes
 /// it more likely the temporary file is in the same disk partition as contents
 /// of the working directory, which can matter since different filesystems may
 /// store timestamps with different resolutions.
 ///
 /// This may fail, typically if we lack write permissions. In that case we
 /// should continue the `status()` algoritm anyway and consider the current
 /// date/time to be unknown.
 fn filesystem_now(repo_root: &Path) -> Result<SystemTime, io::Error> {
     tempfile::tempfile_in(repo_root.join(".hg"))?
         .metadata()?
         .modified()
 }

rust/hg-core/src/discovery.rs

@@ -1,711 +1,711 b''
 // discovery.rs
 //
 // Copyright 2019 Georges Racinet <[e-mail]>
 //
 // This software may be used and distributed according to the terms of the
 // GNU General Public License version 2 or any later version.
 
 //! Discovery operations
 //!
 //! This is a Rust counterpart to the `partialdiscovery` class of
 //! `mercurial.setdiscovery`
 
 use super::{Graph, GraphError, Revision, NULL_REVISION};
 use crate::{ancestors::MissingAncestors, dagops, FastHashMap};
 use rand::seq::SliceRandom;
 use rand::{thread_rng, RngCore, SeedableRng};
 use std::cmp::{max, min};
 use std::collections::{HashSet, VecDeque};
 
 type Rng = rand_pcg::Pcg32;
 type Seed = [u8; 16];
 
 pub struct PartialDiscovery<G: Graph + Clone> {
     target_heads: Option<Vec<Revision>>,
     graph: G, // plays the role of self._repo
     common: MissingAncestors<G>,
     undecided: Option<HashSet<Revision>>,
     children_cache: Option<FastHashMap<Revision, Vec<Revision>>>,
     missing: HashSet<Revision>,
     rng: Rng,
     respect_size: bool,
     randomize: bool,
 }
 
 pub struct DiscoveryStats {
     pub undecided: Option<usize>,
 }
 
 /// Update an existing sample to match the expected size
 ///
 /// The sample is updated with revisions exponentially distant from each
 /// element of `heads`.
 ///
 /// If a target size is specified, the sampling will stop once this size is
 /// reached. Otherwise sampling will happen until roots of the <revs> set are
 /// reached.
 ///
 /// - `revs`: set of revs we want to discover (if None, `assume` the whole dag
 ///   represented by `parentfn`
 /// - `heads`: set of DAG head revs
 /// - `sample`: a sample to update
 /// - `parentfn`: a callable to resolve parents for a revision
 /// - `quicksamplesize`: optional target size of the sample
 fn update_sample<I>(
     revs: Option<&HashSet<Revision>>,
     heads: impl IntoIterator<Item = Revision>,
     sample: &mut HashSet<Revision>,
     parentsfn: impl Fn(Revision) -> Result<I, GraphError>,
     quicksamplesize: Option<usize>,
 ) -> Result<(), GraphError>
 where
     I: Iterator<Item = Revision>,
 {
     let mut distances: FastHashMap<Revision, u32> = FastHashMap::default();
     let mut visit: VecDeque<Revision> = heads.into_iter().collect();
     let mut factor: u32 = 1;
     let mut seen: HashSet<Revision> = HashSet::new();
     while let Some(current) = visit.pop_front() {
         if !seen.insert(current) {
             continue;
         }
 
         let d = *distances.entry(current).or_insert(1);
         if d > factor {
             factor *= 2;
         }
         if d == factor {
             sample.insert(current);
             if let Some(sz) = quicksamplesize {
                 if sample.len() >= sz {
                     return Ok(());
                 }
             }
         }
         for p in parentsfn(current)? {
             if let Some(revs) = revs {
                 if !revs.contains(&p) {
                     continue;
                 }
             }
             distances.entry(p).or_insert(d + 1);
             visit.push_back(p);
         }
     }
     Ok(())
 }
 
 struct ParentsIterator {
     parents: [Revision; 2],
     cur: usize,
 }
 
 impl ParentsIterator {
     fn graph_parents(
         graph: &impl Graph,
         r: Revision,
     ) -> Result<ParentsIterator, GraphError> {
         Ok(ParentsIterator {
             parents: graph.parents(r)?,
             cur: 0,
         })
     }
 }
 
 impl Iterator for ParentsIterator {
     type Item = Revision;
 
     fn next(&mut self) -> Option<Revision> {
         if self.cur > 1 {
             return None;
         }
         let rev = self.parents[self.cur];
         self.cur += 1;
         if rev == NULL_REVISION {
             return self.next();
         }
         Some(rev)
     }
 }
 
 impl<G: Graph + Clone> PartialDiscovery<G> {
     /// Create a PartialDiscovery object, with the intent
     /// of comparing our `::<target_heads>` revset to the contents of another
     /// repo.
     ///
     /// For now `target_heads` is passed as a vector, and will be used
     /// at the first call to `ensure_undecided()`.
     ///
     /// If we want to make the signature more flexible,
     /// we'll have to make it a type argument of `PartialDiscovery` or a trait
     /// object since we'll keep it in the meanwhile
     ///
     /// The `respect_size` boolean controls how the sampling methods
     /// will interpret the size argument requested by the caller. If it's
     /// `false`, they are allowed to produce a sample whose size is more
     /// appropriate to the situation (typically bigger).
     ///
     /// The `randomize` boolean affects sampling, and specifically how
     /// limiting or last-minute expanding is been done:
     ///
     /// If `true`, both will perform random picking from `self.undecided`.
     /// This is currently the best for actual discoveries.
     ///
     /// If `false`, a reproductible picking strategy is performed. This is
     /// useful for integration tests.
     pub fn new(
         graph: G,
         target_heads: Vec<Revision>,
         respect_size: bool,
         randomize: bool,
     ) -> Self {
         let mut seed = [0; 16];
         if randomize {
             thread_rng().fill_bytes(&mut seed);
         }
         Self::new_with_seed(graph, target_heads, seed, respect_size, randomize)
     }
 
     pub fn new_with_seed(
         graph: G,
         target_heads: Vec<Revision>,
         seed: Seed,
         respect_size: bool,
         randomize: bool,
     ) -> Self {
         PartialDiscovery {
             undecided: None,
             children_cache: None,
             target_heads: Some(target_heads),
             graph: graph.clone(),
             common: MissingAncestors::new(graph, vec![]),
             missing: HashSet::new(),
             rng: Rng::from_seed(seed),
             respect_size,
             randomize,
         }
     }
 
     /// Extract at most `size` random elements from sample and return them
     /// as a vector
     fn limit_sample(
         &mut self,
         mut sample: Vec<Revision>,
         size: usize,
     ) -> Vec<Revision> {
         if !self.randomize {
             sample.sort_unstable();
             sample.truncate(size);
             return sample;
         }
         let sample_len = sample.len();
         if sample_len <= size {
             return sample;
         }
         let rng = &mut self.rng;
         let dropped_size = sample_len - size;
         let limited_slice = if size < dropped_size {
             sample.partial_shuffle(rng, size).0
         } else {
             sample.partial_shuffle(rng, dropped_size).1
         };
         limited_slice.to_owned()
     }
 
     /// Register revisions known as being common
     pub fn add_common_revisions(
         &mut self,
         common: impl IntoIterator<Item = Revision>,
     ) -> Result<(), GraphError> {
         let before_len = self.common.get_bases().len();
         self.common.add_bases(common);
         if self.common.get_bases().len() == before_len {
             return Ok(());
         }
         if let Some(ref mut undecided) = self.undecided {
             self.common.remove_ancestors_from(undecided)?;
         }
         Ok(())
     }
 
     /// Register revisions known as being missing
     ///
     /// # Performance note
     ///
     /// Except in the most trivial case, the first call of this method has
     /// the side effect of computing `self.undecided` set for the first time,
     /// and the related caches it might need for efficiency of its internal
     /// computation. This is typically faster if more information is
     /// available in `self.common`. Therefore, for good performance, the
     /// caller should avoid calling this too early.
     pub fn add_missing_revisions(
         &mut self,
         missing: impl IntoIterator<Item = Revision>,
     ) -> Result<(), GraphError> {
         let mut tovisit: VecDeque<Revision> = missing.into_iter().collect();
         if tovisit.is_empty() {
             return Ok(());
         }
         self.ensure_children_cache()?;
         self.ensure_undecided()?; // for safety of possible future refactors
         let children = self.children_cache.as_ref().unwrap();
         let mut seen: HashSet<Revision> = HashSet::new();
         let undecided_mut = self.undecided.as_mut().unwrap();
         while let Some(rev) = tovisit.pop_front() {
             if !self.missing.insert(rev) {
                 // either it's known to be missing from a previous
                 // invocation, and there's no need to iterate on its
                 // children (we now they are all missing)
                 // or it's from a previous iteration of this loop
                 // and its children have already been queued
                 continue;
             }
             undecided_mut.remove(&rev);
             match children.get(&rev) {
                 None => {
                     continue;
                 }
                 Some(this_children) => {
                     for child in this_children.iter().cloned() {
                         if seen.insert(child) {
                             tovisit.push_back(child);
                         }
                     }
                 }
             }
         }
         Ok(())
     }
 
     /// Do we have any information about the peer?
     pub fn has_info(&self) -> bool {
         self.common.has_bases()
     }
 
     /// Did we acquire full knowledge of our Revisions that the peer has?
     pub fn is_complete(&self) -> bool {
         self.undecided.as_ref().map_or(false, HashSet::is_empty)
     }
 
     /// Return the heads of the currently known common set of revisions.
     ///
     /// If the discovery process is not complete (see `is_complete()`), the
     /// caller must be aware that this is an intermediate state.
     ///
     /// On the other hand, if it is complete, then this is currently
     /// the only way to retrieve the end results of the discovery process.
     ///
     /// We may introduce in the future an `into_common_heads` call that
     /// would be more appropriate for normal Rust callers, dropping `self`
     /// if it is complete.
     pub fn common_heads(&self) -> Result<HashSet<Revision>, GraphError> {
         self.common.bases_heads()
     }
 
     /// Force first computation of `self.undecided`
     ///
     /// After this, `self.undecided.as_ref()` and `.as_mut()` can be
     /// unwrapped to get workable immutable or mutable references without
     /// any panic.
     ///
     /// This is an imperative call instead of an access with added lazyness
     /// to reduce easily the scope of mutable borrow for the caller,
     /// compared to undecided(&'a mut self) -> &'a… that would keep it
     /// as long as the resulting immutable one.
     fn ensure_undecided(&mut self) -> Result<(), GraphError> {
         if self.undecided.is_some() {
             return Ok(());
         }
         let tgt = self.target_heads.take().unwrap();
         self.undecided =
             Some(self.common.missing_ancestors(tgt)?.into_iter().collect());
         Ok(())
     }
 
     fn ensure_children_cache(&mut self) -> Result<(), GraphError> {
         if self.children_cache.is_some() {
             return Ok(());
         }
         self.ensure_undecided()?;
 
         let mut children: FastHashMap<Revision, Vec<Revision>> =
             FastHashMap::default();
         for &rev in self.undecided.as_ref().unwrap() {
             for p in ParentsIterator::graph_parents(&self.graph, rev)? {
-                children.entry(p).or_insert_with(Vec::new).push(rev);
+                children.entry(p).or_default().push(rev);
             }
         }
         self.children_cache = Some(children);
         Ok(())
     }
 
     /// Provide statistics about the current state of the discovery process
     pub fn stats(&self) -> DiscoveryStats {
         DiscoveryStats {
             undecided: self.undecided.as_ref().map(HashSet::len),
         }
     }
 
     pub fn take_quick_sample(
         &mut self,
         headrevs: impl IntoIterator<Item = Revision>,
         size: usize,
     ) -> Result<Vec<Revision>, GraphError> {
         self.ensure_undecided()?;
         let mut sample = {
             let undecided = self.undecided.as_ref().unwrap();
             if undecided.len() <= size {
                 return Ok(undecided.iter().cloned().collect());
             }
             dagops::heads(&self.graph, undecided.iter())?
         };
         if sample.len() >= size {
             return Ok(self.limit_sample(sample.into_iter().collect(), size));
         }
         update_sample(
             None,
             headrevs,
             &mut sample,
             |r| ParentsIterator::graph_parents(&self.graph, r),
             Some(size),
         )?;
         Ok(sample.into_iter().collect())
     }
 
     /// Extract a sample from `self.undecided`, going from its heads and roots.
     ///
     /// The `size` parameter is used to avoid useless computations if
     /// it turns out to be bigger than the whole set of undecided Revisions.
     ///
     /// The sample is taken by using `update_sample` from the heads, then
     /// from the roots, working on the reverse DAG,
     /// expressed by `self.children_cache`.
     ///
     /// No effort is being made to complete or limit the sample to `size`
     /// but this method returns another interesting size that it derives
     /// from its knowledge of the structure of the various sets, leaving
     /// to the caller the decision to use it or not.
     fn bidirectional_sample(
         &mut self,
         size: usize,
     ) -> Result<(HashSet<Revision>, usize), GraphError> {
         self.ensure_undecided()?;
         {
             // we don't want to compute children_cache before this
             // but doing it after extracting self.undecided takes a mutable
             // ref to self while a shareable one is still active.
             let undecided = self.undecided.as_ref().unwrap();
             if undecided.len() <= size {
                 return Ok((undecided.clone(), size));
             }
         }
 
         self.ensure_children_cache()?;
         let revs = self.undecided.as_ref().unwrap();
         let mut sample: HashSet<Revision> = revs.clone();
 
         // it's possible that leveraging the children cache would be more
         // efficient here
         dagops::retain_heads(&self.graph, &mut sample)?;
         let revsheads = sample.clone(); // was again heads(revs) in python
 
         // update from heads
         update_sample(
             Some(revs),
             revsheads.iter().cloned(),
             &mut sample,
             |r| ParentsIterator::graph_parents(&self.graph, r),
             None,
         )?;
 
         // update from roots
         let revroots: HashSet<Revision> =
             dagops::roots(&self.graph, revs)?.into_iter().collect();
         let prescribed_size = max(size, min(revroots.len(), revsheads.len()));
 
         let children = self.children_cache.as_ref().unwrap();
         let empty_vec: Vec<Revision> = Vec::new();
         update_sample(
             Some(revs),
             revroots,
             &mut sample,
             |r| Ok(children.get(&r).unwrap_or(&empty_vec).iter().cloned()),
             None,
         )?;
         Ok((sample, prescribed_size))
     }
 
     /// Fill up sample up to the wished size with random undecided Revisions.
     ///
     /// This is intended to be used as a last resort completion if the
     /// regular sampling algorithm returns too few elements.
     fn random_complete_sample(
         &mut self,
         sample: &mut Vec<Revision>,
         size: usize,
     ) {
         let sample_len = sample.len();
         if size <= sample_len {
             return;
         }
         let take_from: Vec<Revision> = self
             .undecided
             .as_ref()
             .unwrap()
             .iter()
             .filter(|&r| !sample.contains(r))
             .cloned()
             .collect();
         sample.extend(self.limit_sample(take_from, size - sample_len));
     }
 
     pub fn take_full_sample(
         &mut self,
         size: usize,
     ) -> Result<Vec<Revision>, GraphError> {
         let (sample_set, prescribed_size) = self.bidirectional_sample(size)?;
         let size = if self.respect_size {
             size
         } else {
             prescribed_size
         };
         let mut sample =
             self.limit_sample(sample_set.into_iter().collect(), size);
         self.random_complete_sample(&mut sample, size);
         Ok(sample)
     }
 }
 
 #[cfg(test)]
 mod tests {
     use super::*;
     use crate::testing::SampleGraph;
 
     /// Shorthand to reduce boilerplate when creating [`Revision`] for testing
     macro_rules! R {
         ($revision:literal) => {
             Revision($revision)
         };
     }
 
     /// A PartialDiscovery as for pushing all the heads of `SampleGraph`
     ///
     /// To avoid actual randomness in these tests, we give it a fixed
     /// random seed, but by default we'll test the random version.
     fn full_disco() -> PartialDiscovery<SampleGraph> {
         PartialDiscovery::new_with_seed(
             SampleGraph,
             vec![R!(10), R!(11), R!(12), R!(13)],
             [0; 16],
             true,
             true,
         )
     }
 
     /// A PartialDiscovery as for pushing the 12 head of `SampleGraph`
     ///
     /// To avoid actual randomness in tests, we give it a fixed random seed.
     fn disco12() -> PartialDiscovery<SampleGraph> {
         PartialDiscovery::new_with_seed(
             SampleGraph,
             vec![R!(12)],
             [0; 16],
             true,
             true,
         )
     }
 
     fn sorted_undecided(
         disco: &PartialDiscovery<SampleGraph>,
     ) -> Vec<Revision> {
         let mut as_vec: Vec<Revision> =
             disco.undecided.as_ref().unwrap().iter().cloned().collect();
         as_vec.sort_unstable();
         as_vec
     }
 
     fn sorted_missing(disco: &PartialDiscovery<SampleGraph>) -> Vec<Revision> {
         let mut as_vec: Vec<Revision> =
             disco.missing.iter().cloned().collect();
         as_vec.sort_unstable();
         as_vec
     }
 
     fn sorted_common_heads(
         disco: &PartialDiscovery<SampleGraph>,
     ) -> Result<Vec<Revision>, GraphError> {
         let mut as_vec: Vec<Revision> =
             disco.common_heads()?.iter().cloned().collect();
         as_vec.sort_unstable();
         Ok(as_vec)
     }
 
     #[test]
     fn test_add_common_get_undecided() -> Result<(), GraphError> {
         let mut disco = full_disco();
         assert_eq!(disco.undecided, None);
         assert!(!disco.has_info());
         assert_eq!(disco.stats().undecided, None);
 
         disco.add_common_revisions(vec![R!(11), R!(12)])?;
         assert!(disco.has_info());
         assert!(!disco.is_complete());
         assert!(disco.missing.is_empty());
 
         // add_common_revisions did not trigger a premature computation
         // of `undecided`, let's check that and ask for them
         assert_eq!(disco.undecided, None);
         disco.ensure_undecided()?;
         assert_eq!(sorted_undecided(&disco), vec![5, 8, 10, 13]);
         assert_eq!(disco.stats().undecided, Some(4));
         Ok(())
     }
 
     /// in this test, we pretend that our peer misses exactly (8+10)::
     /// and we're comparing all our repo to it (as in a bare push)
     #[test]
     fn test_discovery() -> Result<(), GraphError> {
         let mut disco = full_disco();
         disco.add_common_revisions(vec![R!(11), R!(12)])?;
         disco.add_missing_revisions(vec![R!(8), R!(10)])?;
         assert_eq!(sorted_undecided(&disco), vec![5]);
         assert_eq!(sorted_missing(&disco), vec![8, 10, 13]);
         assert!(!disco.is_complete());
 
         disco.add_common_revisions(vec![R!(5)])?;
         assert_eq!(sorted_undecided(&disco), Vec::<Revision>::new());
         assert_eq!(sorted_missing(&disco), vec![8, 10, 13]);
         assert!(disco.is_complete());
         assert_eq!(sorted_common_heads(&disco)?, vec![5, 11, 12]);
         Ok(())
     }
 
     #[test]
     fn test_add_missing_early_continue() -> Result<(), GraphError> {
         eprintln!("test_add_missing_early_stop");
         let mut disco = full_disco();
         disco.add_common_revisions(vec![R!(13), R!(3), R!(4)])?;
         disco.ensure_children_cache()?;
         // 12 is grand-child of 6 through 9
         // passing them in this order maximizes the chances of the
         // early continue to do the wrong thing
         disco.add_missing_revisions(vec![R!(6), R!(9), R!(12)])?;
         assert_eq!(sorted_undecided(&disco), vec![5, 7, 10, 11]);
         assert_eq!(sorted_missing(&disco), vec![6, 9, 12]);
         assert!(!disco.is_complete());
         Ok(())
     }
 
     #[test]
     fn test_limit_sample_no_need_to() {
         let sample = vec![R!(1), R!(2), R!(3), R!(4)];
         assert_eq!(full_disco().limit_sample(sample, 10), vec![1, 2, 3, 4]);
     }
 
     #[test]
     fn test_limit_sample_less_than_half() {
         assert_eq!(
             full_disco().limit_sample((1..6).map(Revision).collect(), 2),
             vec![2, 5]
         );
     }
 
     #[test]
     fn test_limit_sample_more_than_half() {
         assert_eq!(
             full_disco().limit_sample((1..4).map(Revision).collect(), 2),
             vec![1, 2]
         );
     }
 
     #[test]
     fn test_limit_sample_no_random() {
         let mut disco = full_disco();
         disco.randomize = false;
         assert_eq!(
             disco.limit_sample(
                 vec![R!(1), R!(8), R!(13), R!(5), R!(7), R!(3)],
                 4
             ),
             vec![1, 3, 5, 7]
         );
     }
 
     #[test]
     fn test_quick_sample_enough_undecided_heads() -> Result<(), GraphError> {
         let mut disco = full_disco();
         disco.undecided = Some((1..=13).map(Revision).collect());
 
         let mut sample_vec = disco.take_quick_sample(vec![], 4)?;
         sample_vec.sort_unstable();
         assert_eq!(sample_vec, vec![10, 11, 12, 13]);
         Ok(())
     }
 
     #[test]
     fn test_quick_sample_climbing_from_12() -> Result<(), GraphError> {
         let mut disco = disco12();
         disco.ensure_undecided()?;
 
         let mut sample_vec = disco.take_quick_sample(vec![R!(12)], 4)?;
         sample_vec.sort_unstable();
         // r12's only parent is r9, whose unique grand-parent through the
         // diamond shape is r4. This ends there because the distance from r4
         // to the root is only 3.
         assert_eq!(sample_vec, vec![4, 9, 12]);
         Ok(())
     }
 
     #[test]
     fn test_children_cache() -> Result<(), GraphError> {
         let mut disco = full_disco();
         disco.ensure_children_cache()?;
 
         let cache = disco.children_cache.unwrap();
         assert_eq!(cache.get(&R!(2)).cloned(), Some(vec![R!(4)]));
         assert_eq!(cache.get(&R!(10)).cloned(), None);
 
         let mut children_4 = cache.get(&R!(4)).cloned().unwrap();
         children_4.sort_unstable();
         assert_eq!(children_4, vec![R!(5), R!(6), R!(7)]);
 
         let mut children_7 = cache.get(&R!(7)).cloned().unwrap();
         children_7.sort_unstable();
         assert_eq!(children_7, vec![R!(9), R!(11)]);
 
         Ok(())
     }
 
     #[test]
     fn test_complete_sample() {
         let mut disco = full_disco();
         let undecided: HashSet<Revision> =
             [4, 7, 9, 2, 3].iter().cloned().map(Revision).collect();
         disco.undecided = Some(undecided);
 
         let mut sample = vec![R!(0)];
         disco.random_complete_sample(&mut sample, 3);
         assert_eq!(sample.len(), 3);
 
         let mut sample = vec![R!(2), R!(4), R!(7)];
         disco.random_complete_sample(&mut sample, 1);
         assert_eq!(sample.len(), 3);
     }
 
     #[test]
     fn test_bidirectional_sample() -> Result<(), GraphError> {
         let mut disco = full_disco();
-        disco.undecided = Some((0..=13).into_iter().map(Revision).collect());
+        disco.undecided = Some((0..=13).map(Revision).collect());
 
         let (sample_set, size) = disco.bidirectional_sample(7)?;
         assert_eq!(size, 7);
         let mut sample: Vec<Revision> = sample_set.into_iter().collect();
         sample.sort_unstable();
         // our DAG is a bit too small for the results to be really interesting
         // at least it shows that
         // - we went both ways
         // - we didn't take all Revisions (6 is not in the sample)
         assert_eq!(sample, vec![0, 1, 2, 3, 4, 5, 7, 8, 9, 10, 11, 12, 13]);
         Ok(())
     }
 }

rust/hg-core/src/filepatterns.rs

@@ -1,876 +1,874 b''
 // filepatterns.rs
 //
 // Copyright 2019 Raphaël Gomès <[e-mail]>
 //
 // This software may be used and distributed according to the terms of the
 // GNU General Public License version 2 or any later version.
 
 //! Handling of Mercurial-specific patterns.
 
 use crate::{
     utils::{
         files::{canonical_path, get_bytes_from_path, get_path_from_bytes},
         hg_path::{path_to_hg_path_buf, HgPathBuf, HgPathError},
         SliceExt,
     },
     FastHashMap, PatternError,
 };
 use lazy_static::lazy_static;
 use regex::bytes::{NoExpand, Regex};
 use std::ops::Deref;
 use std::path::{Path, PathBuf};
 use std::vec::Vec;
 
 lazy_static! {
     static ref RE_ESCAPE: Vec<Vec<u8>> = {
         let mut v: Vec<Vec<u8>> = (0..=255).map(|byte| vec![byte]).collect();
         let to_escape = b"()[]{}?*+-|^$\\.&~#\t\n\r\x0b\x0c";
         for byte in to_escape {
             v[*byte as usize].insert(0, b'\\');
         }
         v
     };
 }
 
 /// These are matched in order
 const GLOB_REPLACEMENTS: &[(&[u8], &[u8])] =
     &[(b"*/", b"(?:.*/)?"), (b"*", b".*"), (b"", b"[^/]*")];
 
 #[derive(Debug, Clone, PartialEq, Eq)]
 pub enum PatternSyntax {
     /// A regular expression
     Regexp,
     /// Glob that matches at the front of the path
     RootGlob,
     /// Glob that matches at any suffix of the path (still anchored at
     /// slashes)
     Glob,
     /// a path relative to repository root, which is matched recursively
     Path,
     /// a single exact path relative to repository root
     FilePath,
     /// A path relative to cwd
     RelPath,
     /// an unrooted glob (*.rs matches Rust files in all dirs)
     RelGlob,
     /// A regexp that needn't match the start of a name
     RelRegexp,
     /// A path relative to repository root, which is matched non-recursively
     /// (will not match subdirectories)
     RootFiles,
     /// A file of patterns to read and include
     Include,
     /// A file of patterns to match against files under the same directory
     SubInclude,
     /// SubInclude with the result of parsing the included file
     ///
     /// Note: there is no ExpandedInclude because that expansion can be done
     /// in place by replacing the Include pattern by the included patterns.
     /// SubInclude requires more handling.
     ///
     /// Note: `Box` is used to minimize size impact on other enum variants
     ExpandedSubInclude(Box<SubInclude>),
 }
 
 /// Transforms a glob pattern into a regex
 fn glob_to_re(pat: &[u8]) -> Vec<u8> {
     let mut input = pat;
     let mut res: Vec<u8> = vec![];
     let mut group_depth = 0;
 
     while let Some((c, rest)) = input.split_first() {
         input = rest;
 
         match c {
             b'*' => {
                 for (source, repl) in GLOB_REPLACEMENTS {
                     if let Some(rest) = input.drop_prefix(source) {
                         input = rest;
                         res.extend(*repl);
                         break;
                     }
                 }
             }
             b'?' => res.extend(b"."),
             b'[' => {
                 match input.iter().skip(1).position(|b| *b == b']') {
                     None => res.extend(b"\\["),
                     Some(end) => {
                         // Account for the one we skipped
                         let end = end + 1;
 
                         res.extend(b"[");
 
                         for (i, b) in input[..end].iter().enumerate() {
                             if *b == b'!' && i == 0 {
                                 res.extend(b"^")
                             } else if *b == b'^' && i == 0 {
                                 res.extend(b"\\^")
                             } else if *b == b'\\' {
                                 res.extend(b"\\\\")
                             } else {
                                 res.push(*b)
                             }
                         }
                         res.extend(b"]");
                         input = &input[end + 1..];
                     }
                 }
             }
             b'{' => {
                 group_depth += 1;
                 res.extend(b"(?:")
             }
             b'}' if group_depth > 0 => {
                 group_depth -= 1;
                 res.extend(b")");
             }
             b',' if group_depth > 0 => res.extend(b"|"),
             b'\\' => {
                 let c = {
                     if let Some((c, rest)) = input.split_first() {
                         input = rest;
                         c
                     } else {
                         c
                     }
                 };
                 res.extend(&RE_ESCAPE[*c as usize])
             }
             _ => res.extend(&RE_ESCAPE[*c as usize]),
         }
     }
     res
 }
 
 fn escape_pattern(pattern: &[u8]) -> Vec<u8> {
     pattern
         .iter()
         .flat_map(|c| RE_ESCAPE[*c as usize].clone())
         .collect()
 }
 
 pub fn parse_pattern_syntax(
     kind: &[u8],
 ) -> Result<PatternSyntax, PatternError> {
     match kind {
         b"re:" => Ok(PatternSyntax::Regexp),
         b"path:" => Ok(PatternSyntax::Path),
         b"filepath:" => Ok(PatternSyntax::FilePath),
         b"relpath:" => Ok(PatternSyntax::RelPath),
         b"rootfilesin:" => Ok(PatternSyntax::RootFiles),
         b"relglob:" => Ok(PatternSyntax::RelGlob),
         b"relre:" => Ok(PatternSyntax::RelRegexp),
         b"glob:" => Ok(PatternSyntax::Glob),
         b"rootglob:" => Ok(PatternSyntax::RootGlob),
         b"include:" => Ok(PatternSyntax::Include),
         b"subinclude:" => Ok(PatternSyntax::SubInclude),
         _ => Err(PatternError::UnsupportedSyntax(
             String::from_utf8_lossy(kind).to_string(),
         )),
     }
 }
 
 lazy_static! {
     static ref FLAG_RE: Regex = Regex::new(r"^\(\?[aiLmsux]+\)").unwrap();
 }
 
 /// Builds the regex that corresponds to the given pattern.
 /// If within a `syntax: regexp` context, returns the pattern,
 /// otherwise, returns the corresponding regex.
 fn _build_single_regex(entry: &IgnorePattern, glob_suffix: &[u8]) -> Vec<u8> {
     let IgnorePattern {
         syntax, pattern, ..
     } = entry;
     if pattern.is_empty() {
         return vec![];
     }
     match syntax {
         PatternSyntax::Regexp => pattern.to_owned(),
         PatternSyntax::RelRegexp => {
             // The `regex` crate accepts `**` while `re2` and Python's `re`
             // do not. Checking for `*` correctly triggers the same error all
             // engines.
             if pattern[0] == b'^'
                 || pattern[0] == b'*'
                 || pattern.starts_with(b".*")
             {
                 return pattern.to_owned();
             }
             match FLAG_RE.find(pattern) {
                 Some(mat) => {
                     let s = mat.start();
                     let e = mat.end();
                     [
                         &b"(?"[..],
                         &pattern[s + 2..e - 1],
                         &b":"[..],
                         if pattern[e] == b'^'
                             || pattern[e] == b'*'
                             || pattern[e..].starts_with(b".*")
                         {
                             &b""[..]
                         } else {
                             &b".*"[..]
                         },
                         &pattern[e..],
                         &b")"[..],
                     ]
                     .concat()
                 }
                 None => [&b".*"[..], pattern].concat(),
             }
         }
         PatternSyntax::Path | PatternSyntax::RelPath => {
             if pattern == b"." {
                 return vec![];
             }
             [escape_pattern(pattern).as_slice(), b"(?:/|$)"].concat()
         }
         PatternSyntax::RootFiles => {
             let mut res = if pattern == b"." {
                 vec![]
             } else {
                 // Pattern is a directory name.
                 [escape_pattern(pattern).as_slice(), b"/"].concat()
             };
 
             // Anything after the pattern must be a non-directory.
             res.extend(b"[^/]+$");
             res
         }
         PatternSyntax::RelGlob => {
             let glob_re = glob_to_re(pattern);
             if let Some(rest) = glob_re.drop_prefix(b"[^/]*") {
                 [b".*", rest, glob_suffix].concat()
             } else {
                 [b"(?:.*/)?", glob_re.as_slice(), glob_suffix].concat()
             }
         }
         PatternSyntax::Glob | PatternSyntax::RootGlob => {
             [glob_to_re(pattern).as_slice(), glob_suffix].concat()
         }
         PatternSyntax::Include
         | PatternSyntax::SubInclude
         | PatternSyntax::ExpandedSubInclude(_)
         | PatternSyntax::FilePath => unreachable!(),
     }
 }
 
 const GLOB_SPECIAL_CHARACTERS: [u8; 7] =
     [b'*', b'?', b'[', b']', b'{', b'}', b'\\'];
 
 /// TODO support other platforms
 #[cfg(unix)]
 pub fn normalize_path_bytes(bytes: &[u8]) -> Vec<u8> {
     if bytes.is_empty() {
         return b".".to_vec();
     }
     let sep = b'/';
 
     let mut initial_slashes = bytes.iter().take_while(|b| **b == sep).count();
     if initial_slashes > 2 {
         // POSIX allows one or two initial slashes, but treats three or more
         // as single slash.
         initial_slashes = 1;
     }
     let components = bytes
         .split(|b| *b == sep)
         .filter(|c| !(c.is_empty() || c == b"."))
         .fold(vec![], |mut acc, component| {
             if component != b".."
                 || (initial_slashes == 0 && acc.is_empty())
                 || (!acc.is_empty() && acc[acc.len() - 1] == b"..")
             {
                 acc.push(component)
             } else if !acc.is_empty() {
                 acc.pop();
             }
             acc
         });
     let mut new_bytes = components.join(&sep);
 
     if initial_slashes > 0 {
         let mut buf: Vec<_> = (0..initial_slashes).map(|_| sep).collect();
         buf.extend(new_bytes);
         new_bytes = buf;
     }
     if new_bytes.is_empty() {
         b".".to_vec()
     } else {
         new_bytes
     }
 }
 
 /// Wrapper function to `_build_single_regex` that short-circuits 'exact' globs
 /// that don't need to be transformed into a regex.
 pub fn build_single_regex(
     entry: &IgnorePattern,
     glob_suffix: &[u8],
 ) -> Result<Option<Vec<u8>>, PatternError> {
     let IgnorePattern {
         pattern, syntax, ..
     } = entry;
     let pattern = match syntax {
         PatternSyntax::RootGlob
         | PatternSyntax::Path
         | PatternSyntax::RelGlob
         | PatternSyntax::RelPath
         | PatternSyntax::RootFiles => normalize_path_bytes(pattern),
         PatternSyntax::Include | PatternSyntax::SubInclude => {
             return Err(PatternError::NonRegexPattern(entry.clone()))
         }
         _ => pattern.to_owned(),
     };
     let is_simple_rootglob = *syntax == PatternSyntax::RootGlob
         && !pattern.iter().any(|b| GLOB_SPECIAL_CHARACTERS.contains(b));
     if is_simple_rootglob || syntax == &PatternSyntax::FilePath {
         Ok(None)
     } else {
         let mut entry = entry.clone();
         entry.pattern = pattern;
         Ok(Some(_build_single_regex(&entry, glob_suffix)))
     }
 }
 
 lazy_static! {
     static ref SYNTAXES: FastHashMap<&'static [u8], PatternSyntax> = {
         let mut m = FastHashMap::default();
 
         m.insert(b"re:".as_ref(), PatternSyntax::Regexp);
         m.insert(b"regexp:".as_ref(), PatternSyntax::Regexp);
         m.insert(b"path:".as_ref(), PatternSyntax::Path);
         m.insert(b"filepath:".as_ref(), PatternSyntax::FilePath);
         m.insert(b"relpath:".as_ref(), PatternSyntax::RelPath);
         m.insert(b"rootfilesin:".as_ref(), PatternSyntax::RootFiles);
         m.insert(b"relglob:".as_ref(), PatternSyntax::RelGlob);
         m.insert(b"relre:".as_ref(), PatternSyntax::RelRegexp);
         m.insert(b"glob:".as_ref(), PatternSyntax::Glob);
         m.insert(b"rootglob:".as_ref(), PatternSyntax::RootGlob);
         m.insert(b"include:".as_ref(), PatternSyntax::Include);
         m.insert(b"subinclude:".as_ref(), PatternSyntax::SubInclude);
 
         m
     };
 }
 
 #[derive(Debug)]
 pub enum PatternFileWarning {
     /// (file path, syntax bytes)
     InvalidSyntax(PathBuf, Vec<u8>),
     /// File path
     NoSuchFile(PathBuf),
 }
 
 pub fn parse_one_pattern(
     pattern: &[u8],
     source: &Path,
     default: PatternSyntax,
     normalize: bool,
 ) -> IgnorePattern {
     let mut pattern_bytes: &[u8] = pattern;
     let mut syntax = default;
 
     for (s, val) in SYNTAXES.iter() {
         if let Some(rest) = pattern_bytes.drop_prefix(s) {
             syntax = val.clone();
             pattern_bytes = rest;
             break;
         }
     }
 
     let pattern = match syntax {
         PatternSyntax::RootGlob
         | PatternSyntax::Path
         | PatternSyntax::Glob
         | PatternSyntax::RelGlob
         | PatternSyntax::RelPath
         | PatternSyntax::RootFiles
             if normalize =>
         {
             normalize_path_bytes(pattern_bytes)
         }
         _ => pattern_bytes.to_vec(),
     };
 
     IgnorePattern {
         syntax,
         pattern,
         source: source.to_owned(),
     }
 }
 
 pub fn parse_pattern_file_contents(
     lines: &[u8],
     file_path: &Path,
     default_syntax_override: Option<PatternSyntax>,
     warn: bool,
     relativize: bool,
 ) -> Result<(Vec<IgnorePattern>, Vec<PatternFileWarning>), PatternError> {
     let comment_regex = Regex::new(r"((?:^|[^\\])(?:\\\\)*)#.*").unwrap();
 
     #[allow(clippy::trivial_regex)]
     let comment_escape_regex = Regex::new(r"\\#").unwrap();
     let mut inputs: Vec<IgnorePattern> = vec![];
     let mut warnings: Vec<PatternFileWarning> = vec![];
 
     let mut current_syntax =
         default_syntax_override.unwrap_or(PatternSyntax::RelRegexp);
 
     for mut line in lines.split(|c| *c == b'\n') {
         let line_buf;
         if line.contains(&b'#') {
             if let Some(cap) = comment_regex.captures(line) {
                 line = &line[..cap.get(1).unwrap().end()]
             }
             line_buf = comment_escape_regex.replace_all(line, NoExpand(b"#"));
             line = &line_buf;
         }
 
         let line = line.trim_end();
 
         if line.is_empty() {
             continue;
         }
 
         if let Some(syntax) = line.drop_prefix(b"syntax:") {
             let syntax = syntax.trim();
 
             if let Some(parsed) =
                 SYNTAXES.get([syntax, &b":"[..]].concat().as_slice())
             {
                 current_syntax = parsed.clone();
             } else if warn {
                 warnings.push(PatternFileWarning::InvalidSyntax(
                     file_path.to_owned(),
                     syntax.to_owned(),
                 ));
             }
         } else {
             let pattern = parse_one_pattern(
                 line,
                 file_path,
                 current_syntax.clone(),
                 false,
             );
             inputs.push(if relativize {
                 pattern.to_relative()
             } else {
                 pattern
             })
         }
     }
     Ok((inputs, warnings))
 }
 
 pub fn parse_pattern_args(
     patterns: Vec<Vec<u8>>,
     cwd: &Path,
     root: &Path,
 ) -> Result<Vec<IgnorePattern>, HgPathError> {
     let mut ignore_patterns: Vec<IgnorePattern> = Vec::new();
     for pattern in patterns {
         let pattern = parse_one_pattern(
             &pattern,
             Path::new("<args>"),
             PatternSyntax::RelPath,
             true,
         );
         match pattern.syntax {
             PatternSyntax::RelGlob | PatternSyntax::RelPath => {
                 let name = get_path_from_bytes(&pattern.pattern);
                 let canon = canonical_path(root, cwd, name)?;
                 ignore_patterns.push(IgnorePattern {
                     syntax: pattern.syntax,
                     pattern: get_bytes_from_path(canon),
                     source: pattern.source,
                 })
             }
             _ => ignore_patterns.push(pattern.to_owned()),
         };
     }
     Ok(ignore_patterns)
 }
 
 pub fn read_pattern_file(
     file_path: &Path,
     warn: bool,
     inspect_pattern_bytes: &mut impl FnMut(&Path, &[u8]),
 ) -> Result<(Vec<IgnorePattern>, Vec<PatternFileWarning>), PatternError> {
     match std::fs::read(file_path) {
         Ok(contents) => {
             inspect_pattern_bytes(file_path, &contents);
             parse_pattern_file_contents(&contents, file_path, None, warn, true)
         }
         Err(e) if e.kind() == std::io::ErrorKind::NotFound => Ok((
             vec![],
             vec![PatternFileWarning::NoSuchFile(file_path.to_owned())],
         )),
         Err(e) => Err(e.into()),
     }
 }
 
 /// Represents an entry in an "ignore" file.
 #[derive(Debug, Eq, PartialEq, Clone)]
 pub struct IgnorePattern {
     pub syntax: PatternSyntax,
     pub pattern: Vec<u8>,
     pub source: PathBuf,
 }
 
 impl IgnorePattern {
     pub fn new(syntax: PatternSyntax, pattern: &[u8], source: &Path) -> Self {
         Self {
             syntax,
             pattern: pattern.to_owned(),
             source: source.to_owned(),
         }
     }
 
     pub fn to_relative(self) -> Self {
         let Self {
             syntax,
             pattern,
             source,
         } = self;
         Self {
             syntax: match syntax {
                 PatternSyntax::Regexp => PatternSyntax::RelRegexp,
                 PatternSyntax::Glob => PatternSyntax::RelGlob,
                 x => x,
             },
             pattern,
             source,
         }
     }
 }
 
 pub type PatternResult<T> = Result<T, PatternError>;
 
 /// Wrapper for `read_pattern_file` that also recursively expands `include:`
 /// and `subinclude:` patterns.
 ///
 /// The former are expanded in place, while `PatternSyntax::ExpandedSubInclude`
 /// is used for the latter to form a tree of patterns.
 pub fn get_patterns_from_file(
     pattern_file: &Path,
     root_dir: &Path,
     inspect_pattern_bytes: &mut impl FnMut(&Path, &[u8]),
 ) -> PatternResult<(Vec<IgnorePattern>, Vec<PatternFileWarning>)> {
     let (patterns, mut warnings) =
         read_pattern_file(pattern_file, true, inspect_pattern_bytes)?;
     let patterns = patterns
         .into_iter()
         .flat_map(|entry| -> PatternResult<_> {
             Ok(match &entry.syntax {
                 PatternSyntax::Include => {
                     let inner_include =
                         root_dir.join(get_path_from_bytes(&entry.pattern));
                     let (inner_pats, inner_warnings) = get_patterns_from_file(
                         &inner_include,
                         root_dir,
                         inspect_pattern_bytes,
                     )?;
                     warnings.extend(inner_warnings);
                     inner_pats
                 }
                 PatternSyntax::SubInclude => {
                     let mut sub_include = SubInclude::new(
                         root_dir,
                         &entry.pattern,
                         &entry.source,
                     )?;
                     let (inner_patterns, inner_warnings) =
                         get_patterns_from_file(
                             &sub_include.path,
                             &sub_include.root,
                             inspect_pattern_bytes,
                         )?;
                     sub_include.included_patterns = inner_patterns;
                     warnings.extend(inner_warnings);
                     vec![IgnorePattern {
                         syntax: PatternSyntax::ExpandedSubInclude(Box::new(
                             sub_include,
                         )),
                         ..entry
                     }]
                 }
                 _ => vec![entry],
             })
         })
         .flatten()
         .collect();
 
     Ok((patterns, warnings))
 }
 
 /// Holds all the information needed to handle a `subinclude:` pattern.
 #[derive(Debug, PartialEq, Eq, Clone)]
 pub struct SubInclude {
     /// Will be used for repository (hg) paths that start with this prefix.
     /// It is relative to the current working directory, so comparing against
     /// repository paths is painless.
     pub prefix: HgPathBuf,
     /// The file itself, containing the patterns
     pub path: PathBuf,
     /// Folder in the filesystem where this it applies
     pub root: PathBuf,
 
     pub included_patterns: Vec<IgnorePattern>,
 }
 
 impl SubInclude {
     pub fn new(
         root_dir: &Path,
         pattern: &[u8],
         source: &Path,
     ) -> Result<SubInclude, HgPathError> {
         let normalized_source =
             normalize_path_bytes(&get_bytes_from_path(source));
 
         let source_root = get_path_from_bytes(&normalized_source);
-        let source_root =
-            source_root.parent().unwrap_or_else(|| source_root.deref());
+        let source_root = source_root.parent().unwrap_or(source_root);
 
         let path = source_root.join(get_path_from_bytes(pattern));
         let new_root = path.parent().unwrap_or_else(|| path.deref());
 
         let prefix = canonical_path(root_dir, root_dir, new_root)?;
 
         Ok(Self {
             prefix: path_to_hg_path_buf(prefix).map(|mut p| {
                 if !p.is_empty() {
                     p.push_byte(b'/');
                 }
                 p
             })?,
             path: path.to_owned(),
             root: new_root.to_owned(),
             included_patterns: Vec::new(),
         })
     }
 }
 
 /// Separate and pre-process subincludes from other patterns for the "ignore"
 /// phase.
 pub fn filter_subincludes(
     ignore_patterns: Vec<IgnorePattern>,
 ) -> Result<(Vec<SubInclude>, Vec<IgnorePattern>), HgPathError> {
     let mut subincludes = vec![];
     let mut others = vec![];
 
     for pattern in ignore_patterns {
         if let PatternSyntax::ExpandedSubInclude(sub_include) = pattern.syntax
         {
             subincludes.push(*sub_include);
         } else {
             others.push(pattern)
         }
     }
     Ok((subincludes, others))
 }
 
 #[cfg(test)]
 mod tests {
     use super::*;
     use pretty_assertions::assert_eq;
 
     #[test]
     fn escape_pattern_test() {
         let untouched =
             br#"!"%',/0123456789:;<=>@ABCDEFGHIJKLMNOPQRSTUVWXYZ_`abcdefghijklmnopqrstuvwxyz"#;
         assert_eq!(escape_pattern(untouched), untouched.to_vec());
         // All escape codes
         assert_eq!(
-            escape_pattern(br#"()[]{}?*+-|^$\\.&~#\t\n\r\v\f"#),
-            br#"\(\)\[\]\{\}\?\*\+\-\|\^\$\\\\\.\&\~\#\\t\\n\\r\\v\\f"#
-                .to_vec()
+            escape_pattern(br"()[]{}?*+-|^$\\.&~#\t\n\r\v\f"),
+            br"\(\)\[\]\{\}\?\*\+\-\|\^\$\\\\\.\&\~\#\\t\\n\\r\\v\\f".to_vec()
         );
     }
 
     #[test]
     fn glob_test() {
-        assert_eq!(glob_to_re(br#"?"#), br#"."#);
-        assert_eq!(glob_to_re(br#"*"#), br#"[^/]*"#);
-        assert_eq!(glob_to_re(br#"**"#), br#".*"#);
-        assert_eq!(glob_to_re(br#"**/a"#), br#"(?:.*/)?a"#);
-        assert_eq!(glob_to_re(br#"a/**/b"#), br#"a/(?:.*/)?b"#);
-        assert_eq!(glob_to_re(br#"[a*?!^][^b][!c]"#), br#"[a*?!^][\^b][^c]"#);
-        assert_eq!(glob_to_re(br#"{a,b}"#), br#"(?:a|b)"#);
-        assert_eq!(glob_to_re(br#".\*\?"#), br#"\.\*\?"#);
+        assert_eq!(glob_to_re(br"?"), br".");
+        assert_eq!(glob_to_re(br"*"), br"[^/]*");
+        assert_eq!(glob_to_re(br"**"), br".*");
+        assert_eq!(glob_to_re(br"**/a"), br"(?:.*/)?a");
+        assert_eq!(glob_to_re(br"a/**/b"), br"a/(?:.*/)?b");
+        assert_eq!(glob_to_re(br"[a*?!^][^b][!c]"), br"[a*?!^][\^b][^c]");
+        assert_eq!(glob_to_re(br"{a,b}"), br"(?:a|b)");
+        assert_eq!(glob_to_re(br".\*\?"), br"\.\*\?");
     }
 
     #[test]
     fn test_parse_pattern_file_contents() {
         let lines = b"syntax: glob\n*.elc";
 
         assert_eq!(
             parse_pattern_file_contents(
                 lines,
                 Path::new("file_path"),
                 None,
                 false,
                 true,
             )
             .unwrap()
             .0,
             vec![IgnorePattern::new(
                 PatternSyntax::RelGlob,
                 b"*.elc",
                 Path::new("file_path")
             )],
         );
 
         let lines = b"syntax: include\nsyntax: glob";
 
         assert_eq!(
             parse_pattern_file_contents(
                 lines,
                 Path::new("file_path"),
                 None,
                 false,
                 true,
             )
             .unwrap()
             .0,
             vec![]
         );
         let lines = b"glob:**.o";
         assert_eq!(
             parse_pattern_file_contents(
                 lines,
                 Path::new("file_path"),
                 None,
                 false,
                 true,
             )
             .unwrap()
             .0,
             vec![IgnorePattern::new(
                 PatternSyntax::RelGlob,
                 b"**.o",
                 Path::new("file_path")
             )]
         );
     }
 
     #[test]
     fn test_build_single_regex() {
         assert_eq!(
             build_single_regex(
                 &IgnorePattern::new(
                     PatternSyntax::RelGlob,
                     b"rust/target/",
                     Path::new("")
                 ),
                 b"(?:/|$)"
             )
             .unwrap(),
             Some(br"(?:.*/)?rust/target(?:/|$)".to_vec()),
         );
         assert_eq!(
             build_single_regex(
                 &IgnorePattern::new(
                     PatternSyntax::Regexp,
                     br"rust/target/\d+",
                     Path::new("")
                 ),
                 b"(?:/|$)"
             )
             .unwrap(),
             Some(br"rust/target/\d+".to_vec()),
         );
     }
 
     #[test]
     fn test_build_single_regex_shortcut() {
         assert_eq!(
             build_single_regex(
                 &IgnorePattern::new(
                     PatternSyntax::RootGlob,
                     b"",
                     Path::new("")
                 ),
                 b"(?:/|$)"
             )
             .unwrap(),
             None,
         );
         assert_eq!(
             build_single_regex(
                 &IgnorePattern::new(
                     PatternSyntax::RootGlob,
                     b"whatever",
                     Path::new("")
                 ),
                 b"(?:/|$)"
             )
             .unwrap(),
             None,
         );
         assert_eq!(
             build_single_regex(
                 &IgnorePattern::new(
                     PatternSyntax::RootGlob,
                     b"*.o",
                     Path::new("")
                 ),
                 b"(?:/|$)"
             )
             .unwrap(),
             Some(br"[^/]*\.o(?:/|$)".to_vec()),
         );
     }
 
     #[test]
     fn test_build_single_relregex() {
         assert_eq!(
             build_single_regex(
                 &IgnorePattern::new(
                     PatternSyntax::RelRegexp,
                     b"^ba{2}r",
                     Path::new("")
                 ),
                 b"(?:/|$)"
             )
             .unwrap(),
             Some(b"^ba{2}r".to_vec()),
         );
         assert_eq!(
             build_single_regex(
                 &IgnorePattern::new(
                     PatternSyntax::RelRegexp,
                     b"ba{2}r",
                     Path::new("")
                 ),
                 b"(?:/|$)"
             )
             .unwrap(),
             Some(b".*ba{2}r".to_vec()),
         );
         assert_eq!(
             build_single_regex(
                 &IgnorePattern::new(
                     PatternSyntax::RelRegexp,
                     b"(?ia)ba{2}r",
                     Path::new("")
                 ),
                 b"(?:/|$)"
             )
             .unwrap(),
             Some(b"(?ia:.*ba{2}r)".to_vec()),
         );
         assert_eq!(
             build_single_regex(
                 &IgnorePattern::new(
                     PatternSyntax::RelRegexp,
                     b"(?ia)^ba{2}r",
                     Path::new("")
                 ),
                 b"(?:/|$)"
             )
             .unwrap(),
             Some(b"(?ia:^ba{2}r)".to_vec()),
         );
     }
 }

rust/hg-core/src/matchers.rs

@@ -1,2110 +1,2109 b''
 // matchers.rs
 //
 // Copyright 2019 Raphaël Gomès <[e-mail]>
 //
 // This software may be used and distributed according to the terms of the
 // GNU General Public License version 2 or any later version.
 
 //! Structs and types for matching files and directories.
 
 use format_bytes::format_bytes;
 use once_cell::sync::OnceCell;
 
 use crate::{
     dirstate::dirs_multiset::DirsChildrenMultiset,
     filepatterns::{
         build_single_regex, filter_subincludes, get_patterns_from_file,
         PatternFileWarning, PatternResult,
     },
     utils::{
         files::find_dirs,
         hg_path::{HgPath, HgPathBuf, HgPathError},
         Escaped,
     },
     DirsMultiset, FastHashMap, IgnorePattern, PatternError, PatternSyntax,
 };
 
 use crate::dirstate::status::IgnoreFnType;
 use crate::filepatterns::normalize_path_bytes;
 use std::collections::HashSet;
 use std::fmt::{Display, Error, Formatter};
-use std::ops::Deref;
 use std::path::{Path, PathBuf};
 use std::{borrow::ToOwned, collections::BTreeSet};
 
 #[derive(Debug, PartialEq)]
 pub enum VisitChildrenSet {
     /// Don't visit anything
     Empty,
     /// Only visit this directory
     This,
     /// Visit this directory and these subdirectories
     /// TODO Should we implement a `NonEmptyHashSet`?
     Set(HashSet<HgPathBuf>),
     /// Visit this directory and all subdirectories
     Recursive,
 }
 
 pub trait Matcher: core::fmt::Debug {
     /// Explicitly listed files
     fn file_set(&self) -> Option<&HashSet<HgPathBuf>>;
     /// Returns whether `filename` is in `file_set`
     fn exact_match(&self, filename: &HgPath) -> bool;
     /// Returns whether `filename` is matched by this matcher
     fn matches(&self, filename: &HgPath) -> bool;
     /// Decides whether a directory should be visited based on whether it
     /// has potential matches in it or one of its subdirectories, and
     /// potentially lists which subdirectories of that directory should be
     /// visited. This is based on the match's primary, included, and excluded
     /// patterns.
     ///
     /// # Example
     ///
     /// Assume matchers `['path:foo/bar', 'rootfilesin:qux']`, we would
     /// return the following values (assuming the implementation of
     /// visit_children_set is capable of recognizing this; some implementations
     /// are not).
     ///
     /// ```text
     /// ```ignore
     /// '' -> {'foo', 'qux'}
     /// 'baz' -> set()
     /// 'foo' -> {'bar'}
     /// // Ideally this would be `Recursive`, but since the prefix nature of
     /// // matchers is applied to the entire matcher, we have to downgrade this
     /// // to `This` due to the (yet to be implemented in Rust) non-prefix
     /// // `RootFilesIn'-kind matcher being mixed in.
     /// 'foo/bar' -> 'this'
     /// 'qux' -> 'this'
     /// ```
     /// # Important
     ///
     /// Most matchers do not know if they're representing files or
     /// directories. They see `['path:dir/f']` and don't know whether `f` is a
     /// file or a directory, so `visit_children_set('dir')` for most matchers
     /// will return `HashSet{ HgPath { "f" } }`, but if the matcher knows it's
     /// a file (like the yet to be implemented in Rust `ExactMatcher` does),
     /// it may return `VisitChildrenSet::This`.
     /// Do not rely on the return being a `HashSet` indicating that there are
     /// no files in this dir to investigate (or equivalently that if there are
     /// files to investigate in 'dir' that it will always return
     /// `VisitChildrenSet::This`).
     fn visit_children_set(&self, directory: &HgPath) -> VisitChildrenSet;
     /// Matcher will match everything and `files_set()` will be empty:
     /// optimization might be possible.
     fn matches_everything(&self) -> bool;
     /// Matcher will match exactly the files in `files_set()`: optimization
     /// might be possible.
     fn is_exact(&self) -> bool;
 }
 
 /// Matches everything.
 ///```
 /// use hg::{ matchers::{Matcher, AlwaysMatcher}, utils::hg_path::HgPath };
 ///
 /// let matcher = AlwaysMatcher;
 ///
 /// assert_eq!(matcher.matches(HgPath::new(b"whatever")), true);
 /// assert_eq!(matcher.matches(HgPath::new(b"b.txt")), true);
 /// assert_eq!(matcher.matches(HgPath::new(b"main.c")), true);
 /// assert_eq!(matcher.matches(HgPath::new(br"re:.*\.c$")), true);
 /// ```
 #[derive(Debug)]
 pub struct AlwaysMatcher;
 
 impl Matcher for AlwaysMatcher {
     fn file_set(&self) -> Option<&HashSet<HgPathBuf>> {
         None
     }
     fn exact_match(&self, _filename: &HgPath) -> bool {
         false
     }
     fn matches(&self, _filename: &HgPath) -> bool {
         true
     }
     fn visit_children_set(&self, _directory: &HgPath) -> VisitChildrenSet {
         VisitChildrenSet::Recursive
     }
     fn matches_everything(&self) -> bool {
         true
     }
     fn is_exact(&self) -> bool {
         false
     }
 }
 
 /// Matches nothing.
 #[derive(Debug)]
 pub struct NeverMatcher;
 
 impl Matcher for NeverMatcher {
     fn file_set(&self) -> Option<&HashSet<HgPathBuf>> {
         None
     }
     fn exact_match(&self, _filename: &HgPath) -> bool {
         false
     }
     fn matches(&self, _filename: &HgPath) -> bool {
         false
     }
     fn visit_children_set(&self, _directory: &HgPath) -> VisitChildrenSet {
         VisitChildrenSet::Empty
     }
     fn matches_everything(&self) -> bool {
         false
     }
     fn is_exact(&self) -> bool {
         true
     }
 }
 
 /// Matches the input files exactly. They are interpreted as paths, not
 /// patterns.
 ///
 ///```
 /// use hg::{ matchers::{Matcher, FileMatcher}, utils::hg_path::{HgPath, HgPathBuf} };
 ///
 /// let files = vec![HgPathBuf::from_bytes(b"a.txt"), HgPathBuf::from_bytes(br"re:.*\.c$")];
 /// let matcher = FileMatcher::new(files).unwrap();
 ///
 /// assert_eq!(matcher.matches(HgPath::new(b"a.txt")), true);
 /// assert_eq!(matcher.matches(HgPath::new(b"b.txt")), false);
 /// assert_eq!(matcher.matches(HgPath::new(b"main.c")), false);
 /// assert_eq!(matcher.matches(HgPath::new(br"re:.*\.c$")), true);
 /// ```
 #[derive(Debug)]
 pub struct FileMatcher {
     files: HashSet<HgPathBuf>,
     dirs: DirsMultiset,
     sorted_visitchildrenset_candidates: OnceCell<BTreeSet<HgPathBuf>>,
 }
 
 impl FileMatcher {
     pub fn new(files: Vec<HgPathBuf>) -> Result<Self, HgPathError> {
         let dirs = DirsMultiset::from_manifest(&files)?;
         Ok(Self {
-            files: HashSet::from_iter(files.into_iter()),
+            files: HashSet::from_iter(files),
             dirs,
             sorted_visitchildrenset_candidates: OnceCell::new(),
         })
     }
     fn inner_matches(&self, filename: &HgPath) -> bool {
         self.files.contains(filename.as_ref())
     }
 }
 
 impl Matcher for FileMatcher {
     fn file_set(&self) -> Option<&HashSet<HgPathBuf>> {
         Some(&self.files)
     }
     fn exact_match(&self, filename: &HgPath) -> bool {
         self.inner_matches(filename)
     }
     fn matches(&self, filename: &HgPath) -> bool {
         self.inner_matches(filename)
     }
     fn visit_children_set(&self, directory: &HgPath) -> VisitChildrenSet {
         if self.files.is_empty() || !self.dirs.contains(directory) {
             return VisitChildrenSet::Empty;
         }
 
         let compute_candidates = || -> BTreeSet<HgPathBuf> {
             let mut candidates: BTreeSet<HgPathBuf> =
                 self.dirs.iter().cloned().collect();
             candidates.extend(self.files.iter().cloned());
             candidates.remove(HgPath::new(b""));
             candidates
         };
         let candidates =
             if directory.as_ref().is_empty() {
                 compute_candidates()
             } else {
                 let sorted_candidates = self
                     .sorted_visitchildrenset_candidates
                     .get_or_init(compute_candidates);
                 let directory_bytes = directory.as_ref().as_bytes();
                 let start: HgPathBuf =
                     format_bytes!(b"{}/", directory_bytes).into();
                 let start_len = start.len();
                 // `0` sorts after `/`
                 let end = format_bytes!(b"{}0", directory_bytes).into();
                 BTreeSet::from_iter(sorted_candidates.range(start..end).map(
                     |c| HgPathBuf::from_bytes(&c.as_bytes()[start_len..]),
                 ))
             };
 
         // `self.dirs` includes all of the directories, recursively, so if
         // we're attempting to match 'foo/bar/baz.txt', it'll have '', 'foo',
         // 'foo/bar' in it. Thus we can safely ignore a candidate that has a
         // '/' in it, indicating it's for a subdir-of-a-subdir; the immediate
         // subdir will be in there without a slash.
         VisitChildrenSet::Set(
             candidates
                 .into_iter()
                 .filter_map(|c| {
                     if c.bytes().all(|b| *b != b'/') {
                         Some(c)
                     } else {
                         None
                     }
                 })
                 .collect(),
         )
     }
     fn matches_everything(&self) -> bool {
         false
     }
     fn is_exact(&self) -> bool {
         true
     }
 }
 
 /// Matches a set of (kind, pat, source) against a 'root' directory.
 /// (Currently the 'root' directory is effectively always empty)
 /// ```
 /// use hg::{
 ///     matchers::{PatternMatcher, Matcher},
 ///     IgnorePattern,
 ///     PatternSyntax,
 ///     utils::hg_path::{HgPath, HgPathBuf}
 /// };
 /// use std::collections::HashSet;
 /// use std::path::Path;
 /// ///
 /// let ignore_patterns : Vec<IgnorePattern> =
 ///     vec![IgnorePattern::new(PatternSyntax::Regexp, br".*\.c$", Path::new("")),
 ///          IgnorePattern::new(PatternSyntax::Path, b"foo/a", Path::new("")),
 ///          IgnorePattern::new(PatternSyntax::RelPath, b"b", Path::new("")),
 ///          IgnorePattern::new(PatternSyntax::Glob, b"*.h", Path::new("")),
 ///     ];
 /// let matcher = PatternMatcher::new(ignore_patterns).unwrap();
 /// ///
 /// assert_eq!(matcher.matches(HgPath::new(b"main.c")), true); // matches re:.*\.c$
 /// assert_eq!(matcher.matches(HgPath::new(b"b.txt")), false);
 /// assert_eq!(matcher.matches(HgPath::new(b"foo/a")), true); // matches path:foo/a
 /// assert_eq!(matcher.matches(HgPath::new(b"a")), false); // does not match path:b, since 'root' is 'foo'
 /// assert_eq!(matcher.matches(HgPath::new(b"b")), true); // matches relpath:b, since 'root' is 'foo'
 /// assert_eq!(matcher.matches(HgPath::new(b"lib.h")), true); // matches glob:*.h
 /// assert_eq!(matcher.file_set().unwrap(),
 ///            &HashSet::from([HgPathBuf::from_bytes(b""), HgPathBuf::from_bytes(b"foo/a"),
 ///                            HgPathBuf::from_bytes(b""), HgPathBuf::from_bytes(b"b")]));
 /// assert_eq!(matcher.exact_match(HgPath::new(b"foo/a")), true);
 /// assert_eq!(matcher.exact_match(HgPath::new(b"b")), true);
 /// assert_eq!(matcher.exact_match(HgPath::new(b"lib.h")), false); // exact matches are for (rel)path kinds
 /// ```
 pub struct PatternMatcher<'a> {
     patterns: Vec<u8>,
     match_fn: IgnoreFnType<'a>,
     /// Whether all the patterns match a prefix (i.e. recursively)
     prefix: bool,
     files: HashSet<HgPathBuf>,
     dirs: DirsMultiset,
 }
 
 impl core::fmt::Debug for PatternMatcher<'_> {
     fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result {
         f.debug_struct("PatternMatcher")
             .field("patterns", &String::from_utf8_lossy(&self.patterns))
             .field("prefix", &self.prefix)
             .field("files", &self.files)
             .field("dirs", &self.dirs)
             .finish()
     }
 }
 
 impl<'a> PatternMatcher<'a> {
     pub fn new(ignore_patterns: Vec<IgnorePattern>) -> PatternResult<Self> {
         let (files, _) = roots_and_dirs(&ignore_patterns);
         let dirs = DirsMultiset::from_manifest(&files)?;
-        let files: HashSet<HgPathBuf> = HashSet::from_iter(files.into_iter());
+        let files: HashSet<HgPathBuf> = HashSet::from_iter(files);
 
         let prefix = ignore_patterns.iter().all(|k| {
             matches!(k.syntax, PatternSyntax::Path | PatternSyntax::RelPath)
         });
         let (patterns, match_fn) = build_match(ignore_patterns, b"$")?;
 
         Ok(Self {
             patterns,
             match_fn,
             prefix,
             files,
             dirs,
         })
     }
 }
 
 impl<'a> Matcher for PatternMatcher<'a> {
     fn file_set(&self) -> Option<&HashSet<HgPathBuf>> {
         Some(&self.files)
     }
 
     fn exact_match(&self, filename: &HgPath) -> bool {
         self.files.contains(filename)
     }
 
     fn matches(&self, filename: &HgPath) -> bool {
         if self.files.contains(filename) {
             return true;
         }
         (self.match_fn)(filename)
     }
 
     fn visit_children_set(&self, directory: &HgPath) -> VisitChildrenSet {
         if self.prefix && self.files.contains(directory) {
             return VisitChildrenSet::Recursive;
         }
         let path_or_parents_in_set = find_dirs(directory)
             .any(|parent_dir| self.files.contains(parent_dir));
         if self.dirs.contains(directory) || path_or_parents_in_set {
             VisitChildrenSet::This
         } else {
             VisitChildrenSet::Empty
         }
     }
 
     fn matches_everything(&self) -> bool {
         false
     }
 
     fn is_exact(&self) -> bool {
         false
     }
 }
 
 /// Matches files that are included in the ignore rules.
 /// ```
 /// use hg::{
 ///     matchers::{IncludeMatcher, Matcher},
 ///     IgnorePattern,
 ///     PatternSyntax,
 ///     utils::hg_path::HgPath
 /// };
 /// use std::path::Path;
 /// ///
 /// let ignore_patterns =
 /// vec![IgnorePattern::new(PatternSyntax::RootGlob, b"this*", Path::new(""))];
 /// let matcher = IncludeMatcher::new(ignore_patterns).unwrap();
 /// ///
 /// assert_eq!(matcher.matches(HgPath::new(b"testing")), false);
 /// assert_eq!(matcher.matches(HgPath::new(b"this should work")), true);
 /// assert_eq!(matcher.matches(HgPath::new(b"this also")), true);
 /// assert_eq!(matcher.matches(HgPath::new(b"but not this")), false);
 /// ```
 pub struct IncludeMatcher<'a> {
     patterns: Vec<u8>,
     match_fn: IgnoreFnType<'a>,
     /// Whether all the patterns match a prefix (i.e. recursively)
     prefix: bool,
     roots: HashSet<HgPathBuf>,
     dirs: HashSet<HgPathBuf>,
     parents: HashSet<HgPathBuf>,
 }
 
 impl core::fmt::Debug for IncludeMatcher<'_> {
     fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result {
         f.debug_struct("IncludeMatcher")
             .field("patterns", &String::from_utf8_lossy(&self.patterns))
             .field("prefix", &self.prefix)
             .field("roots", &self.roots)
             .field("dirs", &self.dirs)
             .field("parents", &self.parents)
             .finish()
     }
 }
 
 impl<'a> Matcher for IncludeMatcher<'a> {
     fn file_set(&self) -> Option<&HashSet<HgPathBuf>> {
         None
     }
 
     fn exact_match(&self, _filename: &HgPath) -> bool {
         false
     }
 
     fn matches(&self, filename: &HgPath) -> bool {
         (self.match_fn)(filename)
     }
 
     fn visit_children_set(&self, directory: &HgPath) -> VisitChildrenSet {
         let dir = directory;
         if self.prefix && self.roots.contains(dir) {
             return VisitChildrenSet::Recursive;
         }
         if self.roots.contains(HgPath::new(b""))
             || self.roots.contains(dir)
             || self.dirs.contains(dir)
             || find_dirs(dir).any(|parent_dir| self.roots.contains(parent_dir))
         {
             return VisitChildrenSet::This;
         }
 
         if self.parents.contains(dir.as_ref()) {
             let multiset = self.get_all_parents_children();
             if let Some(children) = multiset.get(dir) {
                 return VisitChildrenSet::Set(
                     children.iter().map(HgPathBuf::from).collect(),
                 );
             }
         }
         VisitChildrenSet::Empty
     }
 
     fn matches_everything(&self) -> bool {
         false
     }
 
     fn is_exact(&self) -> bool {
         false
     }
 }
 
 /// The union of multiple matchers. Will match if any of the matchers match.
 #[derive(Debug)]
 pub struct UnionMatcher {
     matchers: Vec<Box<dyn Matcher + Sync>>,
 }
 
 impl Matcher for UnionMatcher {
     fn file_set(&self) -> Option<&HashSet<HgPathBuf>> {
         None
     }
 
     fn exact_match(&self, _filename: &HgPath) -> bool {
         false
     }
 
     fn matches(&self, filename: &HgPath) -> bool {
         self.matchers.iter().any(|m| m.matches(filename))
     }
 
     fn visit_children_set(&self, directory: &HgPath) -> VisitChildrenSet {
         let mut result = HashSet::new();
         let mut this = false;
         for matcher in self.matchers.iter() {
             let visit = matcher.visit_children_set(directory);
             match visit {
                 VisitChildrenSet::Empty => continue,
                 VisitChildrenSet::This => {
                     this = true;
                     // Don't break, we might have an 'all' in here.
                     continue;
                 }
                 VisitChildrenSet::Set(set) => {
                     result.extend(set);
                 }
                 VisitChildrenSet::Recursive => {
                     return visit;
                 }
             }
         }
         if this {
             return VisitChildrenSet::This;
         }
         if result.is_empty() {
             VisitChildrenSet::Empty
         } else {
             VisitChildrenSet::Set(result)
         }
     }
 
     fn matches_everything(&self) -> bool {
         // TODO Maybe if all are AlwaysMatcher?
         false
     }
 
     fn is_exact(&self) -> bool {
         false
     }
 }
 
 impl UnionMatcher {
     pub fn new(matchers: Vec<Box<dyn Matcher + Sync>>) -> Self {
         Self { matchers }
     }
 }
 
 #[derive(Debug)]
 pub struct IntersectionMatcher {
     m1: Box<dyn Matcher + Sync>,
     m2: Box<dyn Matcher + Sync>,
     files: Option<HashSet<HgPathBuf>>,
 }
 
 impl Matcher for IntersectionMatcher {
     fn file_set(&self) -> Option<&HashSet<HgPathBuf>> {
         self.files.as_ref()
     }
 
     fn exact_match(&self, filename: &HgPath) -> bool {
         self.files.as_ref().map_or(false, |f| f.contains(filename))
     }
 
     fn matches(&self, filename: &HgPath) -> bool {
         self.m1.matches(filename) && self.m2.matches(filename)
     }
 
     fn visit_children_set(&self, directory: &HgPath) -> VisitChildrenSet {
         let m1_set = self.m1.visit_children_set(directory);
         if m1_set == VisitChildrenSet::Empty {
             return VisitChildrenSet::Empty;
         }
         let m2_set = self.m2.visit_children_set(directory);
         if m2_set == VisitChildrenSet::Empty {
             return VisitChildrenSet::Empty;
         }
 
         if m1_set == VisitChildrenSet::Recursive {
             return m2_set;
         } else if m2_set == VisitChildrenSet::Recursive {
             return m1_set;
         }
 
         match (&m1_set, &m2_set) {
             (VisitChildrenSet::Recursive, _) => m2_set,
             (_, VisitChildrenSet::Recursive) => m1_set,
             (VisitChildrenSet::This, _) | (_, VisitChildrenSet::This) => {
                 VisitChildrenSet::This
             }
             (VisitChildrenSet::Set(m1), VisitChildrenSet::Set(m2)) => {
                 let set: HashSet<_> = m1.intersection(m2).cloned().collect();
                 if set.is_empty() {
                     VisitChildrenSet::Empty
                 } else {
                     VisitChildrenSet::Set(set)
                 }
             }
             _ => unreachable!(),
         }
     }
 
     fn matches_everything(&self) -> bool {
         self.m1.matches_everything() && self.m2.matches_everything()
     }
 
     fn is_exact(&self) -> bool {
         self.m1.is_exact() || self.m2.is_exact()
     }
 }
 
 impl IntersectionMatcher {
     pub fn new(
         mut m1: Box<dyn Matcher + Sync>,
         mut m2: Box<dyn Matcher + Sync>,
     ) -> Self {
         let files = if m1.is_exact() || m2.is_exact() {
             if !m1.is_exact() {
                 std::mem::swap(&mut m1, &mut m2);
             }
             m1.file_set().map(|m1_files| {
                 m1_files.iter().cloned().filter(|f| m2.matches(f)).collect()
             })
         } else {
             // without exact input file sets, we can't do an exact
             // intersection, so we must over-approximate by
             // unioning instead
             m1.file_set().map(|m1_files| match m2.file_set() {
                 Some(m2_files) => m1_files.union(m2_files).cloned().collect(),
                 None => m1_files.iter().cloned().collect(),
             })
         };
         Self { m1, m2, files }
     }
 }
 
 #[derive(Debug)]
 pub struct DifferenceMatcher {
     base: Box<dyn Matcher + Sync>,
     excluded: Box<dyn Matcher + Sync>,
     files: Option<HashSet<HgPathBuf>>,
 }
 
 impl Matcher for DifferenceMatcher {
     fn file_set(&self) -> Option<&HashSet<HgPathBuf>> {
         self.files.as_ref()
     }
 
     fn exact_match(&self, filename: &HgPath) -> bool {
         self.files.as_ref().map_or(false, |f| f.contains(filename))
     }
 
     fn matches(&self, filename: &HgPath) -> bool {
         self.base.matches(filename) && !self.excluded.matches(filename)
     }
 
     fn visit_children_set(&self, directory: &HgPath) -> VisitChildrenSet {
         let excluded_set = self.excluded.visit_children_set(directory);
         if excluded_set == VisitChildrenSet::Recursive {
             return VisitChildrenSet::Empty;
         }
         let base_set = self.base.visit_children_set(directory);
         // Possible values for base: 'recursive', 'this', set(...), set()
         // Possible values for excluded:          'this', set(...), set()
         // If excluded has nothing under here that we care about, return base,
         // even if it's 'recursive'.
         if excluded_set == VisitChildrenSet::Empty {
             return base_set;
         }
         match base_set {
             VisitChildrenSet::This | VisitChildrenSet::Recursive => {
                 // Never return 'recursive' here if excluded_set is any kind of
                 // non-empty (either 'this' or set(foo)), since excluded might
                 // return set() for a subdirectory.
                 VisitChildrenSet::This
             }
             set => {
                 // Possible values for base:         set(...), set()
                 // Possible values for excluded: 'this', set(...)
                 // We ignore excluded set results. They're possibly incorrect:
                 //  base = path:dir/subdir
                 //  excluded=rootfilesin:dir,
                 //  visit_children_set(''):
                 //   base returns {'dir'}, excluded returns {'dir'}, if we
                 //   subtracted we'd return set(), which is *not* correct, we
                 //   still need to visit 'dir'!
                 set
             }
         }
     }
 
     fn matches_everything(&self) -> bool {
         false
     }
 
     fn is_exact(&self) -> bool {
         self.base.is_exact()
     }
 }
 
 impl DifferenceMatcher {
     pub fn new(
         base: Box<dyn Matcher + Sync>,
         excluded: Box<dyn Matcher + Sync>,
     ) -> Self {
         let base_is_exact = base.is_exact();
         let base_files = base.file_set().map(ToOwned::to_owned);
         let mut new = Self {
             base,
             excluded,
             files: None,
         };
         if base_is_exact {
             new.files = base_files.map(|files| {
                 files.iter().cloned().filter(|f| new.matches(f)).collect()
             });
         }
         new
     }
 }
 
 /// Wraps [`regex::bytes::Regex`] to improve performance in multithreaded
 /// contexts.
 ///
 /// The `status` algorithm makes heavy use of threads, and calling `is_match`
 /// from many threads at once is prone to contention, probably within the
 /// scratch space needed as the regex DFA is built lazily.
 ///
 /// We are in the process of raising the issue upstream, but for now
 /// the workaround used here is to store the `Regex` in a lazily populated
 /// thread-local variable, sharing the initial read-only compilation, but
 /// not the lazy dfa scratch space mentioned above.
 ///
 /// This reduces the contention observed with 16+ threads, but does not
 /// completely remove it. Hopefully this can be addressed upstream.
 struct RegexMatcher {
     /// Compiled at the start of the status algorithm, used as a base for
     /// cloning in each thread-local `self.local`, thus sharing the expensive
     /// first compilation.
     base: regex::bytes::Regex,
     /// Thread-local variable that holds the `Regex` that is actually queried
     /// from each thread.
     local: thread_local::ThreadLocal<regex::bytes::Regex>,
 }
 
 impl RegexMatcher {
     /// Returns whether the path matches the stored `Regex`.
     pub fn is_match(&self, path: &HgPath) -> bool {
         self.local
             .get_or(|| self.base.clone())
             .is_match(path.as_bytes())
     }
 }
 
 /// Returns a function that matches an `HgPath` against the given regex
 /// pattern.
 ///
 /// This can fail when the pattern is invalid or not supported by the
 /// underlying engine (the `regex` crate), for instance anything with
 /// back-references.
 #[logging_timer::time("trace")]
 fn re_matcher(pattern: &[u8]) -> PatternResult<RegexMatcher> {
     use std::io::Write;
 
     // The `regex` crate adds `.*` to the start and end of expressions if there
     // are no anchors, so add the start anchor.
     let mut escaped_bytes = vec![b'^', b'(', b'?', b':'];
     for byte in pattern {
         if *byte > 127 {
             write!(escaped_bytes, "\\x{:x}", *byte).unwrap();
         } else {
             escaped_bytes.push(*byte);
         }
     }
     escaped_bytes.push(b')');
 
     // Avoid the cost of UTF8 checking
     //
     // # Safety
     // This is safe because we escaped all non-ASCII bytes.
     let pattern_string = unsafe { String::from_utf8_unchecked(escaped_bytes) };
     let re = regex::bytes::RegexBuilder::new(&pattern_string)
         .unicode(false)
         // Big repos with big `.hgignore` will hit the default limit and
         // incur a significant performance hit. One repo's `hg status` hit
         // multiple *minutes*.
         .dfa_size_limit(50 * (1 << 20))
         .build()
         .map_err(|e| PatternError::UnsupportedSyntax(e.to_string()))?;
 
     Ok(RegexMatcher {
         base: re,
         local: Default::default(),
     })
 }
 
 /// Returns the regex pattern and a function that matches an `HgPath` against
 /// said regex formed by the given ignore patterns.
-fn build_regex_match<'a, 'b>(
-    ignore_patterns: &'a [IgnorePattern],
+fn build_regex_match<'a>(
+    ignore_patterns: &[IgnorePattern],
     glob_suffix: &[u8],
-) -> PatternResult<(Vec<u8>, IgnoreFnType<'b>)> {
+) -> PatternResult<(Vec<u8>, IgnoreFnType<'a>)> {
     let mut regexps = vec![];
     let mut exact_set = HashSet::new();
 
     for pattern in ignore_patterns {
         if let Some(re) = build_single_regex(pattern, glob_suffix)? {
             regexps.push(re);
         } else {
             let exact = normalize_path_bytes(&pattern.pattern);
             exact_set.insert(HgPathBuf::from_bytes(&exact));
         }
     }
 
     let full_regex = regexps.join(&b'|');
 
     // An empty pattern would cause the regex engine to incorrectly match the
     // (empty) root directory
     let func = if !(regexps.is_empty()) {
         let matcher = re_matcher(&full_regex)?;
         let func = move |filename: &HgPath| {
             exact_set.contains(filename) || matcher.is_match(filename)
         };
         Box::new(func) as IgnoreFnType
     } else {
         let func = move |filename: &HgPath| exact_set.contains(filename);
         Box::new(func) as IgnoreFnType
     };
 
     Ok((full_regex, func))
 }
 
 /// Returns roots and directories corresponding to each pattern.
 ///
 /// This calculates the roots and directories exactly matching the patterns and
 /// returns a tuple of (roots, dirs). It does not return other directories
 /// which may also need to be considered, like the parent directories.
 fn roots_and_dirs(
     ignore_patterns: &[IgnorePattern],
 ) -> (Vec<HgPathBuf>, Vec<HgPathBuf>) {
     let mut roots = Vec::new();
     let mut dirs = Vec::new();
 
     for ignore_pattern in ignore_patterns {
         let IgnorePattern {
             syntax, pattern, ..
         } = ignore_pattern;
         match syntax {
             PatternSyntax::RootGlob | PatternSyntax::Glob => {
                 let mut root = HgPathBuf::new();
                 for p in pattern.split(|c| *c == b'/') {
                     if p.iter()
                         .any(|c| matches!(*c, b'[' | b'{' | b'*' | b'?'))
                     {
                         break;
                     }
                     root.push(HgPathBuf::from_bytes(p).as_ref());
                 }
                 roots.push(root);
             }
             PatternSyntax::Path
             | PatternSyntax::RelPath
             | PatternSyntax::FilePath => {
                 let pat = HgPath::new(if pattern == b"." {
                     &[] as &[u8]
                 } else {
                     pattern
                 });
                 roots.push(pat.to_owned());
             }
             PatternSyntax::RootFiles => {
                 let pat = if pattern == b"." {
                     &[] as &[u8]
                 } else {
                     pattern
                 };
                 dirs.push(HgPathBuf::from_bytes(pat));
             }
             _ => {
                 roots.push(HgPathBuf::new());
             }
         }
     }
     (roots, dirs)
 }
 
 /// Paths extracted from patterns
 #[derive(Debug, PartialEq)]
 struct RootsDirsAndParents {
     /// Directories to match recursively
     pub roots: HashSet<HgPathBuf>,
     /// Directories to match non-recursively
     pub dirs: HashSet<HgPathBuf>,
     /// Implicitly required directories to go to items in either roots or dirs
     pub parents: HashSet<HgPathBuf>,
 }
 
 /// Extract roots, dirs and parents from patterns.
 fn roots_dirs_and_parents(
     ignore_patterns: &[IgnorePattern],
 ) -> PatternResult<RootsDirsAndParents> {
     let (roots, dirs) = roots_and_dirs(ignore_patterns);
 
     let mut parents = HashSet::new();
 
     parents.extend(
         DirsMultiset::from_manifest(&dirs)?
             .iter()
             .map(ToOwned::to_owned),
     );
     parents.extend(
         DirsMultiset::from_manifest(&roots)?
             .iter()
             .map(ToOwned::to_owned),
     );
 
     Ok(RootsDirsAndParents {
         roots: HashSet::from_iter(roots),
         dirs: HashSet::from_iter(dirs),
         parents,
     })
 }
 
 /// Returns a function that checks whether a given file (in the general sense)
 /// should be matched.
 fn build_match<'a>(
     ignore_patterns: Vec<IgnorePattern>,
     glob_suffix: &[u8],
 ) -> PatternResult<(Vec<u8>, IgnoreFnType<'a>)> {
     let mut match_funcs: Vec<IgnoreFnType<'a>> = vec![];
     // For debugging and printing
     let mut patterns = vec![];
 
     let (subincludes, ignore_patterns) = filter_subincludes(ignore_patterns)?;
 
     if !subincludes.is_empty() {
         // Build prefix-based matcher functions for subincludes
         let mut submatchers = FastHashMap::default();
         let mut prefixes = vec![];
 
         for sub_include in subincludes {
             let matcher = IncludeMatcher::new(sub_include.included_patterns)?;
             let match_fn =
                 Box::new(move |path: &HgPath| matcher.matches(path));
             prefixes.push(sub_include.prefix.clone());
             submatchers.insert(sub_include.prefix.clone(), match_fn);
         }
 
         let match_subinclude = move |filename: &HgPath| {
             for prefix in prefixes.iter() {
                 if let Some(rel) = filename.relative_to(prefix) {
                     if (submatchers[prefix])(rel) {
                         return true;
                     }
                 }
             }
             false
         };
 
         match_funcs.push(Box::new(match_subinclude));
     }
 
     if !ignore_patterns.is_empty() {
         // Either do dumb matching if all patterns are rootfiles, or match
         // with a regex.
         if ignore_patterns
             .iter()
             .all(|k| k.syntax == PatternSyntax::RootFiles)
         {
             let dirs: HashSet<_> = ignore_patterns
                 .iter()
                 .map(|k| k.pattern.to_owned())
                 .collect();
             let mut dirs_vec: Vec<_> = dirs.iter().cloned().collect();
 
             let match_func = move |path: &HgPath| -> bool {
                 let path = path.as_bytes();
                 let i = path.iter().rfind(|a| **a == b'/');
                 let dir = if let Some(i) = i {
                     &path[..*i as usize]
                 } else {
                     b"."
                 };
-                dirs.contains(dir.deref())
+                dirs.contains(dir)
             };
             match_funcs.push(Box::new(match_func));
 
             patterns.extend(b"rootfilesin: ");
             dirs_vec.sort();
             patterns.extend(dirs_vec.escaped_bytes());
         } else {
             let (new_re, match_func) =
                 build_regex_match(&ignore_patterns, glob_suffix)?;
             patterns = new_re;
             match_funcs.push(match_func)
         }
     }
 
     Ok(if match_funcs.len() == 1 {
         (patterns, match_funcs.remove(0))
     } else {
         (
             patterns,
             Box::new(move |f: &HgPath| -> bool {
                 match_funcs.iter().any(|match_func| match_func(f))
             }),
         )
     })
 }
 
 /// Parses all "ignore" files with their recursive includes and returns a
 /// function that checks whether a given file (in the general sense) should be
 /// ignored.
 pub fn get_ignore_matcher<'a>(
     mut all_pattern_files: Vec<PathBuf>,
     root_dir: &Path,
     inspect_pattern_bytes: &mut impl FnMut(&Path, &[u8]),
 ) -> PatternResult<(IncludeMatcher<'a>, Vec<PatternFileWarning>)> {
     let mut all_patterns = vec![];
     let mut all_warnings = vec![];
 
     // Sort to make the ordering of calls to `inspect_pattern_bytes`
     // deterministic even if the ordering of `all_pattern_files` is not (such
     // as when a iteration order of a Python dict or Rust HashMap is involved).
     // Sort by "string" representation instead of the default by component
     // (with a Rust-specific definition of a component)
     all_pattern_files
         .sort_unstable_by(|a, b| a.as_os_str().cmp(b.as_os_str()));
 
     for pattern_file in &all_pattern_files {
         let (patterns, warnings) = get_patterns_from_file(
             pattern_file,
             root_dir,
             inspect_pattern_bytes,
         )?;
 
         all_patterns.extend(patterns.to_owned());
         all_warnings.extend(warnings);
     }
     let matcher = IncludeMatcher::new(all_patterns)?;
     Ok((matcher, all_warnings))
 }
 
 /// Parses all "ignore" files with their recursive includes and returns a
 /// function that checks whether a given file (in the general sense) should be
 /// ignored.
 pub fn get_ignore_function<'a>(
     all_pattern_files: Vec<PathBuf>,
     root_dir: &Path,
     inspect_pattern_bytes: &mut impl FnMut(&Path, &[u8]),
 ) -> PatternResult<(IgnoreFnType<'a>, Vec<PatternFileWarning>)> {
     let res =
         get_ignore_matcher(all_pattern_files, root_dir, inspect_pattern_bytes);
     res.map(|(matcher, all_warnings)| {
         let res: IgnoreFnType<'a> =
             Box::new(move |path: &HgPath| matcher.matches(path));
 
         (res, all_warnings)
     })
 }
 
 impl<'a> IncludeMatcher<'a> {
     pub fn new(ignore_patterns: Vec<IgnorePattern>) -> PatternResult<Self> {
         let RootsDirsAndParents {
             roots,
             dirs,
             parents,
         } = roots_dirs_and_parents(&ignore_patterns)?;
         let prefix = ignore_patterns.iter().all(|k| {
             matches!(k.syntax, PatternSyntax::Path | PatternSyntax::RelPath)
         });
         let (patterns, match_fn) = build_match(ignore_patterns, b"(?:/|$)")?;
 
         Ok(Self {
             patterns,
             match_fn,
             prefix,
             roots,
             dirs,
             parents,
         })
     }
 
     fn get_all_parents_children(&self) -> DirsChildrenMultiset {
         // TODO cache
         let thing = self
             .dirs
             .iter()
             .chain(self.roots.iter())
             .chain(self.parents.iter());
         DirsChildrenMultiset::new(thing, Some(&self.parents))
     }
 
     pub fn debug_get_patterns(&self) -> &[u8] {
         self.patterns.as_ref()
     }
 }
 
 impl<'a> Display for IncludeMatcher<'a> {
     fn fmt(&self, f: &mut Formatter<'_>) -> Result<(), Error> {
         // XXX What about exact matches?
         // I'm not sure it's worth it to clone the HashSet and keep it
         // around just in case someone wants to display the matcher, plus
         // it's going to be unreadable after a few entries, but we need to
         // inform in this display that exact matches are being used and are
         // (on purpose) missing from the `includes`.
         write!(
             f,
             "IncludeMatcher(includes='{}')",
             String::from_utf8_lossy(&self.patterns.escaped_bytes())
         )
     }
 }
 
 #[cfg(test)]
 mod tests {
     use super::*;
     use pretty_assertions::assert_eq;
     use std::path::Path;
 
     #[test]
     fn test_roots_and_dirs() {
         let pats = vec![
             IgnorePattern::new(PatternSyntax::Glob, b"g/h/*", Path::new("")),
             IgnorePattern::new(PatternSyntax::Glob, b"g/h", Path::new("")),
             IgnorePattern::new(PatternSyntax::Glob, b"g*", Path::new("")),
         ];
         let (roots, dirs) = roots_and_dirs(&pats);
 
         assert_eq!(
             roots,
             vec!(
                 HgPathBuf::from_bytes(b"g/h"),
                 HgPathBuf::from_bytes(b"g/h"),
                 HgPathBuf::new()
             ),
         );
         assert_eq!(dirs, vec!());
     }
 
     #[test]
     fn test_roots_dirs_and_parents() {
         let pats = vec![
             IgnorePattern::new(PatternSyntax::Glob, b"g/h/*", Path::new("")),
             IgnorePattern::new(PatternSyntax::Glob, b"g/h", Path::new("")),
             IgnorePattern::new(PatternSyntax::Glob, b"g*", Path::new("")),
         ];
 
         let mut roots = HashSet::new();
         roots.insert(HgPathBuf::from_bytes(b"g/h"));
         roots.insert(HgPathBuf::new());
 
         let dirs = HashSet::new();
 
         let mut parents = HashSet::new();
         parents.insert(HgPathBuf::new());
         parents.insert(HgPathBuf::from_bytes(b"g"));
 
         assert_eq!(
             roots_dirs_and_parents(&pats).unwrap(),
             RootsDirsAndParents {
                 roots,
                 dirs,
                 parents
             }
         );
     }
 
     #[test]
     fn test_filematcher_visit_children_set() {
         // Visitchildrenset
         let files = vec![HgPathBuf::from_bytes(b"dir/subdir/foo.txt")];
         let matcher = FileMatcher::new(files).unwrap();
 
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"dir"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::Set(set)
         );
 
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"subdir"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir")),
             VisitChildrenSet::Set(set)
         );
 
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"foo.txt"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir")),
             VisitChildrenSet::Set(set)
         );
 
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/x")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/foo.txt")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
     }
 
     #[test]
     fn test_filematcher_visit_children_set_files_and_dirs() {
         let files = vec![
             HgPathBuf::from_bytes(b"rootfile.txt"),
             HgPathBuf::from_bytes(b"a/file1.txt"),
             HgPathBuf::from_bytes(b"a/b/file2.txt"),
             // No file in a/b/c
             HgPathBuf::from_bytes(b"a/b/c/d/file4.txt"),
         ];
         let matcher = FileMatcher::new(files).unwrap();
 
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"a"));
         set.insert(HgPathBuf::from_bytes(b"rootfile.txt"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::Set(set)
         );
 
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"b"));
         set.insert(HgPathBuf::from_bytes(b"file1.txt"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"a")),
             VisitChildrenSet::Set(set)
         );
 
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"c"));
         set.insert(HgPathBuf::from_bytes(b"file2.txt"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"a/b")),
             VisitChildrenSet::Set(set)
         );
 
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"d"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"a/b/c")),
             VisitChildrenSet::Set(set)
         );
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"file4.txt"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"a/b/c/d")),
             VisitChildrenSet::Set(set)
         );
 
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"a/b/c/d/e")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
     }
 
     #[test]
     fn test_patternmatcher() {
         // VisitdirPrefix
         let m = PatternMatcher::new(vec![IgnorePattern::new(
             PatternSyntax::Path,
             b"dir/subdir",
             Path::new(""),
         )])
         .unwrap();
         assert_eq!(
             m.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::This
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir")),
             VisitChildrenSet::This
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir/subdir")),
             VisitChildrenSet::Recursive
         );
         // OPT: This should probably be Recursive if its parent is?
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir/subdir/x")),
             VisitChildrenSet::This
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
 
         // VisitchildrensetPrefix
         let m = PatternMatcher::new(vec![IgnorePattern::new(
             PatternSyntax::Path,
             b"dir/subdir",
             Path::new(""),
         )])
         .unwrap();
         assert_eq!(
             m.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::This
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir")),
             VisitChildrenSet::This
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir/subdir")),
             VisitChildrenSet::Recursive
         );
         // OPT: This should probably be Recursive if its parent is?
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir/subdir/x")),
             VisitChildrenSet::This
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
 
         // VisitdirRootfilesin
         let m = PatternMatcher::new(vec![IgnorePattern::new(
             PatternSyntax::RootFiles,
             b"dir/subdir",
             Path::new(""),
         )])
         .unwrap();
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir/subdir/x")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
         // FIXME: These should probably be This.
         assert_eq!(
             m.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir/subdir")),
             VisitChildrenSet::Empty
         );
 
         // VisitchildrensetRootfilesin
         let m = PatternMatcher::new(vec![IgnorePattern::new(
             PatternSyntax::RootFiles,
             b"dir/subdir",
             Path::new(""),
         )])
         .unwrap();
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir/subdir/x")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
         // FIXME: These should probably be {'dir'}, {'subdir'} and This,
         // respectively, or at least This for all three.
         assert_eq!(
             m.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir/subdir")),
             VisitChildrenSet::Empty
         );
 
         // VisitdirGlob
         let m = PatternMatcher::new(vec![IgnorePattern::new(
             PatternSyntax::Glob,
             b"dir/z*",
             Path::new(""),
         )])
         .unwrap();
         assert_eq!(
             m.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::This
         );
         // FIXME: This probably should be This
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
         // OPT: these should probably be False.
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir/subdir")),
             VisitChildrenSet::This
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir/subdir/x")),
             VisitChildrenSet::This
         );
 
         // VisitchildrensetGlob
         let m = PatternMatcher::new(vec![IgnorePattern::new(
             PatternSyntax::Glob,
             b"dir/z*",
             Path::new(""),
         )])
         .unwrap();
         assert_eq!(
             m.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::This
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
         // FIXME: This probably should be This
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir")),
             VisitChildrenSet::Empty
         );
         // OPT: these should probably be Empty
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir/subdir")),
             VisitChildrenSet::This
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir/subdir/x")),
             VisitChildrenSet::This
         );
 
         // VisitdirFilepath
         let m = PatternMatcher::new(vec![IgnorePattern::new(
             PatternSyntax::FilePath,
             b"dir/z",
             Path::new(""),
         )])
         .unwrap();
         assert_eq!(
             m.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::This
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir")),
             VisitChildrenSet::This
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir/subdir")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir/subdir/x")),
             VisitChildrenSet::Empty
         );
 
         // VisitchildrensetFilepath
         let m = PatternMatcher::new(vec![IgnorePattern::new(
             PatternSyntax::FilePath,
             b"dir/z",
             Path::new(""),
         )])
         .unwrap();
         assert_eq!(
             m.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::This
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir")),
             VisitChildrenSet::This
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir/subdir")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             m.visit_children_set(HgPath::new(b"dir/subdir/x")),
             VisitChildrenSet::Empty
         );
     }
 
     #[test]
     fn test_includematcher() {
         // VisitchildrensetPrefix
         let matcher = IncludeMatcher::new(vec![IgnorePattern::new(
             PatternSyntax::RelPath,
             b"dir/subdir",
             Path::new(""),
         )])
         .unwrap();
 
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"dir"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::Set(set)
         );
 
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"subdir"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir")),
             VisitChildrenSet::Set(set)
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir")),
             VisitChildrenSet::Recursive
         );
         // OPT: This should probably be 'all' if its parent is?
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/x")),
             VisitChildrenSet::This
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
 
         // VisitchildrensetRootfilesin
         let matcher = IncludeMatcher::new(vec![IgnorePattern::new(
             PatternSyntax::RootFiles,
             b"dir/subdir",
             Path::new(""),
         )])
         .unwrap();
 
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"dir"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::Set(set)
         );
 
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"subdir"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir")),
             VisitChildrenSet::Set(set)
         );
 
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir")),
             VisitChildrenSet::This
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/x")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
 
         // VisitchildrensetGlob
         let matcher = IncludeMatcher::new(vec![IgnorePattern::new(
             PatternSyntax::Glob,
             b"dir/z*",
             Path::new(""),
         )])
         .unwrap();
 
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"dir"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::Set(set)
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir")),
             VisitChildrenSet::This
         );
         // OPT: these should probably be set().
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir")),
             VisitChildrenSet::This
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/x")),
             VisitChildrenSet::This
         );
 
         // VisitchildrensetFilePath
         let matcher = IncludeMatcher::new(vec![IgnorePattern::new(
             PatternSyntax::FilePath,
             b"dir/z",
             Path::new(""),
         )])
         .unwrap();
 
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"dir"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::Set(set)
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"z"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir")),
             VisitChildrenSet::Set(set)
         );
         // OPT: these should probably be set().
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/x")),
             VisitChildrenSet::Empty
         );
 
         // Test multiple patterns
         let matcher = IncludeMatcher::new(vec![
             IgnorePattern::new(PatternSyntax::RelPath, b"foo", Path::new("")),
             IgnorePattern::new(PatternSyntax::Glob, b"g*", Path::new("")),
         ])
         .unwrap();
 
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::This
         );
 
         // Test multiple patterns
         let matcher = IncludeMatcher::new(vec![IgnorePattern::new(
             PatternSyntax::Glob,
             b"**/*.exe",
             Path::new(""),
         )])
         .unwrap();
 
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::This
         );
     }
 
     #[test]
     fn test_unionmatcher() {
         // Path + Rootfiles
         let m1 = IncludeMatcher::new(vec![IgnorePattern::new(
             PatternSyntax::RelPath,
             b"dir/subdir",
             Path::new(""),
         )])
         .unwrap();
         let m2 = IncludeMatcher::new(vec![IgnorePattern::new(
             PatternSyntax::RootFiles,
             b"dir",
             Path::new(""),
         )])
         .unwrap();
         let matcher = UnionMatcher::new(vec![Box::new(m1), Box::new(m2)]);
 
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"dir"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::Set(set)
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir")),
             VisitChildrenSet::This
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir")),
             VisitChildrenSet::Recursive
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/foo")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
 
         // OPT: These next two could be 'all' instead of 'this'.
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/z")),
             VisitChildrenSet::This
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/x")),
             VisitChildrenSet::This
         );
 
         // Path + unrelated Path
         let m1 = IncludeMatcher::new(vec![IgnorePattern::new(
             PatternSyntax::RelPath,
             b"dir/subdir",
             Path::new(""),
         )])
         .unwrap();
         let m2 = IncludeMatcher::new(vec![IgnorePattern::new(
             PatternSyntax::RelPath,
             b"folder",
             Path::new(""),
         )])
         .unwrap();
         let matcher = UnionMatcher::new(vec![Box::new(m1), Box::new(m2)]);
 
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"folder"));
         set.insert(HgPathBuf::from_bytes(b"dir"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::Set(set)
         );
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"subdir"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir")),
             VisitChildrenSet::Set(set)
         );
 
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir")),
             VisitChildrenSet::Recursive
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/foo")),
             VisitChildrenSet::Empty
         );
 
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Recursive
         );
         // OPT: These next two could be 'all' instead of 'this'.
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/z")),
             VisitChildrenSet::This
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/x")),
             VisitChildrenSet::This
         );
 
         // Path + subpath
         let m1 = IncludeMatcher::new(vec![IgnorePattern::new(
             PatternSyntax::RelPath,
             b"dir/subdir/x",
             Path::new(""),
         )])
         .unwrap();
         let m2 = IncludeMatcher::new(vec![IgnorePattern::new(
             PatternSyntax::RelPath,
             b"dir/subdir",
             Path::new(""),
         )])
         .unwrap();
         let matcher = UnionMatcher::new(vec![Box::new(m1), Box::new(m2)]);
 
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"dir"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::Set(set)
         );
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"subdir"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir")),
             VisitChildrenSet::Set(set)
         );
 
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir")),
             VisitChildrenSet::Recursive
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/foo")),
             VisitChildrenSet::Empty
         );
 
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/x")),
             VisitChildrenSet::Recursive
         );
         // OPT: this should probably be 'all' not 'this'.
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/z")),
             VisitChildrenSet::This
         );
     }
 
     #[test]
     fn test_intersectionmatcher() {
         // Include path + Include rootfiles
         let m1 = Box::new(
             IncludeMatcher::new(vec![IgnorePattern::new(
                 PatternSyntax::RelPath,
                 b"dir/subdir",
                 Path::new(""),
             )])
             .unwrap(),
         );
         let m2 = Box::new(
             IncludeMatcher::new(vec![IgnorePattern::new(
                 PatternSyntax::RootFiles,
                 b"dir",
                 Path::new(""),
             )])
             .unwrap(),
         );
         let matcher = IntersectionMatcher::new(m1, m2);
 
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"dir"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::Set(set)
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir")),
             VisitChildrenSet::This
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/foo")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/z")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/x")),
             VisitChildrenSet::Empty
         );
 
         // Non intersecting paths
         let m1 = Box::new(
             IncludeMatcher::new(vec![IgnorePattern::new(
                 PatternSyntax::RelPath,
                 b"dir/subdir",
                 Path::new(""),
             )])
             .unwrap(),
         );
         let m2 = Box::new(
             IncludeMatcher::new(vec![IgnorePattern::new(
                 PatternSyntax::RelPath,
                 b"folder",
                 Path::new(""),
             )])
             .unwrap(),
         );
         let matcher = IntersectionMatcher::new(m1, m2);
 
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/foo")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/z")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/x")),
             VisitChildrenSet::Empty
         );
 
         // Nested paths
         let m1 = Box::new(
             IncludeMatcher::new(vec![IgnorePattern::new(
                 PatternSyntax::RelPath,
                 b"dir/subdir/x",
                 Path::new(""),
             )])
             .unwrap(),
         );
         let m2 = Box::new(
             IncludeMatcher::new(vec![IgnorePattern::new(
                 PatternSyntax::RelPath,
                 b"dir/subdir",
                 Path::new(""),
             )])
             .unwrap(),
         );
         let matcher = IntersectionMatcher::new(m1, m2);
 
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"dir"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::Set(set)
         );
 
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"subdir"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir")),
             VisitChildrenSet::Set(set)
         );
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"x"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir")),
             VisitChildrenSet::Set(set)
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/foo")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/z")),
             VisitChildrenSet::Empty
         );
         // OPT: this should probably be 'all' not 'this'.
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/x")),
             VisitChildrenSet::This
         );
 
         // Diverging paths
         let m1 = Box::new(
             IncludeMatcher::new(vec![IgnorePattern::new(
                 PatternSyntax::RelPath,
                 b"dir/subdir/x",
                 Path::new(""),
             )])
             .unwrap(),
         );
         let m2 = Box::new(
             IncludeMatcher::new(vec![IgnorePattern::new(
                 PatternSyntax::RelPath,
                 b"dir/subdir/z",
                 Path::new(""),
             )])
             .unwrap(),
         );
         let matcher = IntersectionMatcher::new(m1, m2);
 
         // OPT: these next two could probably be Empty as well.
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"dir"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::Set(set)
         );
         // OPT: these next two could probably be Empty as well.
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"subdir"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir")),
             VisitChildrenSet::Set(set)
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/foo")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/z")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/x")),
             VisitChildrenSet::Empty
         );
     }
 
     #[test]
     fn test_differencematcher() {
         // Two alwaysmatchers should function like a nevermatcher
         let m1 = AlwaysMatcher;
         let m2 = AlwaysMatcher;
         let matcher = DifferenceMatcher::new(Box::new(m1), Box::new(m2));
 
         for case in &[
             &b""[..],
             b"dir",
             b"dir/subdir",
             b"dir/subdir/z",
             b"dir/foo",
             b"dir/subdir/x",
             b"folder",
         ] {
             assert_eq!(
                 matcher.visit_children_set(HgPath::new(case)),
                 VisitChildrenSet::Empty
             );
         }
 
         // One always and one never should behave the same as an always
         let m1 = AlwaysMatcher;
         let m2 = NeverMatcher;
         let matcher = DifferenceMatcher::new(Box::new(m1), Box::new(m2));
 
         for case in &[
             &b""[..],
             b"dir",
             b"dir/subdir",
             b"dir/subdir/z",
             b"dir/foo",
             b"dir/subdir/x",
             b"folder",
         ] {
             assert_eq!(
                 matcher.visit_children_set(HgPath::new(case)),
                 VisitChildrenSet::Recursive
             );
         }
 
         // Two include matchers
         let m1 = Box::new(
             IncludeMatcher::new(vec![IgnorePattern::new(
                 PatternSyntax::RelPath,
                 b"dir/subdir",
                 Path::new("/repo"),
             )])
             .unwrap(),
         );
         let m2 = Box::new(
             IncludeMatcher::new(vec![IgnorePattern::new(
                 PatternSyntax::RootFiles,
                 b"dir",
                 Path::new("/repo"),
             )])
             .unwrap(),
         );
 
         let matcher = DifferenceMatcher::new(m1, m2);
 
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"dir"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"")),
             VisitChildrenSet::Set(set)
         );
 
         let mut set = HashSet::new();
         set.insert(HgPathBuf::from_bytes(b"subdir"));
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir")),
             VisitChildrenSet::Set(set)
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir")),
             VisitChildrenSet::Recursive
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/foo")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"folder")),
             VisitChildrenSet::Empty
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/z")),
             VisitChildrenSet::This
         );
         assert_eq!(
             matcher.visit_children_set(HgPath::new(b"dir/subdir/x")),
             VisitChildrenSet::This
         );
     }
 }

rust/hg-core/src/repo.rs

@@ -1,782 +1,782 b''
 use crate::changelog::Changelog;
 use crate::config::{Config, ConfigError, ConfigParseError};
 use crate::dirstate::DirstateParents;
 use crate::dirstate_tree::dirstate_map::DirstateMapWriteMode;
 use crate::dirstate_tree::on_disk::Docket as DirstateDocket;
 use crate::dirstate_tree::owning::OwningDirstateMap;
 use crate::errors::HgResultExt;
 use crate::errors::{HgError, IoResultExt};
 use crate::lock::{try_with_lock_no_wait, LockError};
 use crate::manifest::{Manifest, Manifestlog};
 use crate::revlog::filelog::Filelog;
 use crate::revlog::RevlogError;
 use crate::utils::debug::debug_wait_for_file_or_print;
 use crate::utils::files::get_path_from_bytes;
 use crate::utils::hg_path::HgPath;
 use crate::utils::SliceExt;
 use crate::vfs::{is_dir, is_file, Vfs};
 use crate::DirstateError;
 use crate::{requirements, NodePrefix, UncheckedRevision};
 use std::cell::{Ref, RefCell, RefMut};
 use std::collections::HashSet;
 use std::io::Seek;
 use std::io::SeekFrom;
 use std::io::Write as IoWrite;
 use std::path::{Path, PathBuf};
 
 const V2_MAX_READ_ATTEMPTS: usize = 5;
 
 type DirstateMapIdentity = (Option<u64>, Option<Vec<u8>>, usize);
 
 /// A repository on disk
 pub struct Repo {
     working_directory: PathBuf,
     dot_hg: PathBuf,
     store: PathBuf,
     requirements: HashSet<String>,
     config: Config,
     dirstate_parents: LazyCell<DirstateParents>,
     dirstate_map: LazyCell<OwningDirstateMap>,
     changelog: LazyCell<Changelog>,
     manifestlog: LazyCell<Manifestlog>,
 }
 
 #[derive(Debug, derive_more::From)]
 pub enum RepoError {
     NotFound {
         at: PathBuf,
     },
     #[from]
     ConfigParseError(ConfigParseError),
     #[from]
     Other(HgError),
 }
 
 impl From<ConfigError> for RepoError {
     fn from(error: ConfigError) -> Self {
         match error {
             ConfigError::Parse(error) => error.into(),
             ConfigError::Other(error) => error.into(),
         }
     }
 }
 
 impl Repo {
     /// tries to find nearest repository root in current working directory or
     /// its ancestors
     pub fn find_repo_root() -> Result<PathBuf, RepoError> {
         let current_directory = crate::utils::current_dir()?;
         // ancestors() is inclusive: it first yields `current_directory`
         // as-is.
         for ancestor in current_directory.ancestors() {
             if is_dir(ancestor.join(".hg"))? {
                 return Ok(ancestor.to_path_buf());
             }
         }
         Err(RepoError::NotFound {
             at: current_directory,
         })
     }
 
     /// Find a repository, either at the given path (which must contain a `.hg`
     /// sub-directory) or by searching the current directory and its
     /// ancestors.
     ///
     /// A method with two very different "modes" like this usually a code smell
     /// to make two methods instead, but in this case an `Option` is what rhg
     /// sub-commands get from Clap for the `-R` / `--repository` CLI argument.
     /// Having two methods would just move that `if` to almost all callers.
     pub fn find(
         config: &Config,
         explicit_path: Option<PathBuf>,
     ) -> Result<Self, RepoError> {
         if let Some(root) = explicit_path {
             if is_dir(root.join(".hg"))? {
                 Self::new_at_path(root, config)
             } else if is_file(&root)? {
                 Err(HgError::unsupported("bundle repository").into())
             } else {
                 Err(RepoError::NotFound { at: root })
             }
         } else {
             let root = Self::find_repo_root()?;
             Self::new_at_path(root, config)
         }
     }
 
     /// To be called after checking that `.hg` is a sub-directory
     fn new_at_path(
         working_directory: PathBuf,
         config: &Config,
     ) -> Result<Self, RepoError> {
         let dot_hg = working_directory.join(".hg");
 
         let mut repo_config_files =
             vec![dot_hg.join("hgrc"), dot_hg.join("hgrc-not-shared")];
 
         let hg_vfs = Vfs { base: &dot_hg };
         let mut reqs = requirements::load_if_exists(hg_vfs)?;
         let relative =
             reqs.contains(requirements::RELATIVE_SHARED_REQUIREMENT);
         let shared =
             reqs.contains(requirements::SHARED_REQUIREMENT) || relative;
 
         // From `mercurial/localrepo.py`:
         //
         // if .hg/requires contains the sharesafe requirement, it means
         // there exists a `.hg/store/requires` too and we should read it
         // NOTE: presence of SHARESAFE_REQUIREMENT imply that store requirement
         // is present. We never write SHARESAFE_REQUIREMENT for a repo if store
         // is not present, refer checkrequirementscompat() for that
         //
         // However, if SHARESAFE_REQUIREMENT is not present, it means that the
         // repository was shared the old way. We check the share source
         // .hg/requires for SHARESAFE_REQUIREMENT to detect whether the
         // current repository needs to be reshared
         let share_safe = reqs.contains(requirements::SHARESAFE_REQUIREMENT);
 
         let store_path;
         if !shared {
             store_path = dot_hg.join("store");
         } else {
             let bytes = hg_vfs.read("sharedpath")?;
             let mut shared_path =
                 get_path_from_bytes(bytes.trim_end_matches(|b| b == b'\n'))
                     .to_owned();
             if relative {
                 shared_path = dot_hg.join(shared_path)
             }
             if !is_dir(&shared_path)? {
                 return Err(HgError::corrupted(format!(
                     ".hg/sharedpath points to nonexistent directory {}",
                     shared_path.display()
                 ))
                 .into());
             }
 
             store_path = shared_path.join("store");
 
             let source_is_share_safe =
                 requirements::load(Vfs { base: &shared_path })?
                     .contains(requirements::SHARESAFE_REQUIREMENT);
 
             if share_safe != source_is_share_safe {
                 return Err(HgError::unsupported("share-safe mismatch").into());
             }
 
             if share_safe {
                 repo_config_files.insert(0, shared_path.join("hgrc"))
             }
         }
         if share_safe {
             reqs.extend(requirements::load(Vfs { base: &store_path })?);
         }
 
         let repo_config = if std::env::var_os("HGRCSKIPREPO").is_none() {
             config.combine_with_repo(&repo_config_files)?
         } else {
             config.clone()
         };
 
         let repo = Self {
             requirements: reqs,
             working_directory,
             store: store_path,
             dot_hg,
             config: repo_config,
             dirstate_parents: LazyCell::new(),
             dirstate_map: LazyCell::new(),
             changelog: LazyCell::new(),
             manifestlog: LazyCell::new(),
         };
 
         requirements::check(&repo)?;
 
         Ok(repo)
     }
 
     pub fn working_directory_path(&self) -> &Path {
         &self.working_directory
     }
 
     pub fn requirements(&self) -> &HashSet<String> {
         &self.requirements
     }
 
     pub fn config(&self) -> &Config {
         &self.config
     }
 
     /// For accessing repository files (in `.hg`), except for the store
     /// (`.hg/store`).
     pub fn hg_vfs(&self) -> Vfs<'_> {
         Vfs { base: &self.dot_hg }
     }
 
     /// For accessing repository store files (in `.hg/store`)
     pub fn store_vfs(&self) -> Vfs<'_> {
         Vfs { base: &self.store }
     }
 
     /// For accessing the working copy
     pub fn working_directory_vfs(&self) -> Vfs<'_> {
         Vfs {
             base: &self.working_directory,
         }
     }
 
     pub fn try_with_wlock_no_wait<R>(
         &self,
         f: impl FnOnce() -> R,
     ) -> Result<R, LockError> {
         try_with_lock_no_wait(self.hg_vfs(), "wlock", f)
     }
 
     /// Whether this repo should use dirstate-v2.
     /// The presence of `dirstate-v2` in the requirements does not mean that
     /// the on-disk dirstate is necessarily in version 2. In most cases,
     /// a dirstate-v2 file will indeed be found, but in rare cases (like the
     /// upgrade mechanism being cut short), the on-disk version will be a
     /// v1 file.
     /// Semantically, having a requirement only means that a client cannot
     /// properly understand or properly update the repo if it lacks the support
     /// for the required feature, but not that that feature is actually used
     /// in all occasions.
     pub fn use_dirstate_v2(&self) -> bool {
         self.requirements
             .contains(requirements::DIRSTATE_V2_REQUIREMENT)
     }
 
     pub fn has_sparse(&self) -> bool {
         self.requirements.contains(requirements::SPARSE_REQUIREMENT)
     }
 
     pub fn has_narrow(&self) -> bool {
         self.requirements.contains(requirements::NARROW_REQUIREMENT)
     }
 
     pub fn has_nodemap(&self) -> bool {
         self.requirements
             .contains(requirements::NODEMAP_REQUIREMENT)
     }
 
     fn dirstate_file_contents(&self) -> Result<Vec<u8>, HgError> {
         Ok(self
             .hg_vfs()
             .read("dirstate")
             .io_not_found_as_none()?
             .unwrap_or_default())
     }
 
     fn dirstate_identity(&self) -> Result<Option<u64>, HgError> {
         use std::os::unix::fs::MetadataExt;
         Ok(self
             .hg_vfs()
             .symlink_metadata("dirstate")
             .io_not_found_as_none()?
             .map(|meta| meta.ino()))
     }
 
     pub fn dirstate_parents(&self) -> Result<DirstateParents, HgError> {
         Ok(*self
             .dirstate_parents
             .get_or_init(|| self.read_dirstate_parents())?)
     }
 
     fn read_dirstate_parents(&self) -> Result<DirstateParents, HgError> {
         let dirstate = self.dirstate_file_contents()?;
         let parents = if dirstate.is_empty() {
             DirstateParents::NULL
         } else if self.use_dirstate_v2() {
             let docket_res =
                 crate::dirstate_tree::on_disk::read_docket(&dirstate);
             match docket_res {
                 Ok(docket) => docket.parents(),
                 Err(_) => {
                     log::info!(
                         "Parsing dirstate docket failed, \
                         falling back to dirstate-v1"
                     );
                     *crate::dirstate::parsers::parse_dirstate_parents(
                         &dirstate,
                     )?
                 }
             }
         } else {
             *crate::dirstate::parsers::parse_dirstate_parents(&dirstate)?
         };
         self.dirstate_parents.set(parents);
         Ok(parents)
     }
 
     /// Returns the information read from the dirstate docket necessary to
     /// check if the data file has been updated/deleted by another process
     /// since we last read the dirstate.
     /// Namely, the inode, data file uuid and the data size.
     fn get_dirstate_data_file_integrity(
         &self,
     ) -> Result<DirstateMapIdentity, HgError> {
         assert!(
             self.use_dirstate_v2(),
             "accessing dirstate data file ID without dirstate-v2"
         );
         // Get the identity before the contents since we could have a race
         // between the two. Having an identity that is too old is fine, but
         // one that is younger than the content change is bad.
         let identity = self.dirstate_identity()?;
         let dirstate = self.dirstate_file_contents()?;
         if dirstate.is_empty() {
             self.dirstate_parents.set(DirstateParents::NULL);
             Ok((identity, None, 0))
         } else {
             let docket_res =
                 crate::dirstate_tree::on_disk::read_docket(&dirstate);
             match docket_res {
                 Ok(docket) => {
                     self.dirstate_parents.set(docket.parents());
                     Ok((
                         identity,
                         Some(docket.uuid.to_owned()),
                         docket.data_size(),
                     ))
                 }
                 Err(_) => {
                     log::info!(
                         "Parsing dirstate docket failed, \
                         falling back to dirstate-v1"
                     );
                     let parents =
                         *crate::dirstate::parsers::parse_dirstate_parents(
                             &dirstate,
                         )?;
                     self.dirstate_parents.set(parents);
                     Ok((identity, None, 0))
                 }
             }
         }
     }
 
     fn new_dirstate_map(&self) -> Result<OwningDirstateMap, DirstateError> {
         if self.use_dirstate_v2() {
             // The v2 dirstate is split into a docket and a data file.
             // Since we don't always take the `wlock` to read it
             // (like in `hg status`), it is susceptible to races.
             // A simple retry method should be enough since full rewrites
             // only happen when too much garbage data is present and
             // this race is unlikely.
             let mut tries = 0;
 
             while tries < V2_MAX_READ_ATTEMPTS {
                 tries += 1;
                 match self.read_docket_and_data_file() {
                     Ok(m) => {
                         return Ok(m);
                     }
                     Err(e) => match e {
                         DirstateError::Common(HgError::RaceDetected(
                             context,
                         )) => {
                             log::info!(
                                 "dirstate read race detected {} (retry {}/{})",
                                 context,
                                 tries,
                                 V2_MAX_READ_ATTEMPTS,
                             );
                             continue;
                         }
                         _ => {
                             log::info!(
                                 "Reading dirstate v2 failed, \
                                 falling back to v1"
                             );
                             return self.new_dirstate_map_v1();
                         }
                     },
                 }
             }
             let error = HgError::abort(
                 format!("dirstate read race happened {tries} times in a row"),
                 255,
                 None,
             );
             Err(DirstateError::Common(error))
         } else {
             self.new_dirstate_map_v1()
         }
     }
 
     fn new_dirstate_map_v1(&self) -> Result<OwningDirstateMap, DirstateError> {
         debug_wait_for_file_or_print(self.config(), "dirstate.pre-read-file");
         let identity = self.dirstate_identity()?;
         let dirstate_file_contents = self.dirstate_file_contents()?;
         if dirstate_file_contents.is_empty() {
             self.dirstate_parents.set(DirstateParents::NULL);
             Ok(OwningDirstateMap::new_empty(Vec::new()))
         } else {
             let (map, parents) =
                 OwningDirstateMap::new_v1(dirstate_file_contents, identity)?;
             self.dirstate_parents.set(parents);
             Ok(map)
         }
     }
 
     fn read_docket_and_data_file(
         &self,
     ) -> Result<OwningDirstateMap, DirstateError> {
         debug_wait_for_file_or_print(self.config(), "dirstate.pre-read-file");
         let dirstate_file_contents = self.dirstate_file_contents()?;
         let identity = self.dirstate_identity()?;
         if dirstate_file_contents.is_empty() {
             self.dirstate_parents.set(DirstateParents::NULL);
             return Ok(OwningDirstateMap::new_empty(Vec::new()));
         }
         let docket = crate::dirstate_tree::on_disk::read_docket(
             &dirstate_file_contents,
         )?;
         debug_wait_for_file_or_print(
             self.config(),
             "dirstate.post-docket-read-file",
         );
         self.dirstate_parents.set(docket.parents());
         let uuid = docket.uuid.to_owned();
         let data_size = docket.data_size();
 
         let context = "between reading dirstate docket and data file";
         let race_error = HgError::RaceDetected(context.into());
         let metadata = docket.tree_metadata();
 
         let mut map = if crate::vfs::is_on_nfs_mount(docket.data_filename()) {
             // Don't mmap on NFS to prevent `SIGBUS` error on deletion
             let contents = self.hg_vfs().read(docket.data_filename());
             let contents = match contents {
                 Ok(c) => c,
                 Err(HgError::IoError { error, context }) => {
                     match error.raw_os_error().expect("real os error") {
                         // 2 = ENOENT, No such file or directory
                         // 116 = ESTALE, Stale NFS file handle
                         //
                         // TODO match on `error.kind()` when
                         // `ErrorKind::StaleNetworkFileHandle` is stable.
                         2 | 116 => {
                             // Race where the data file was deleted right after
                             // we read the docket, try again
                             return Err(race_error.into());
                         }
                         _ => {
                             return Err(
                                 HgError::IoError { error, context }.into()
                             )
                         }
                     }
                 }
                 Err(e) => return Err(e.into()),
             };
             OwningDirstateMap::new_v2(
                 contents, data_size, metadata, uuid, identity,
             )
         } else {
             match self
                 .hg_vfs()
                 .mmap_open(docket.data_filename())
                 .io_not_found_as_none()
             {
                 Ok(Some(data_mmap)) => OwningDirstateMap::new_v2(
                     data_mmap, data_size, metadata, uuid, identity,
                 ),
                 Ok(None) => {
                     // Race where the data file was deleted right after we
                     // read the docket, try again
                     return Err(race_error.into());
                 }
                 Err(e) => return Err(e.into()),
             }
         }?;
 
         let write_mode_config = self
             .config()
             .get_str(b"devel", b"dirstate.v2.data_update_mode")
             .unwrap_or(Some("auto"))
             .unwrap_or("auto"); // don't bother for devel options
         let write_mode = match write_mode_config {
             "auto" => DirstateMapWriteMode::Auto,
             "force-new" => DirstateMapWriteMode::ForceNewDataFile,
             "force-append" => DirstateMapWriteMode::ForceAppend,
             _ => DirstateMapWriteMode::Auto,
         };
 
         map.with_dmap_mut(|m| m.set_write_mode(write_mode));
 
         Ok(map)
     }
 
     pub fn dirstate_map(
         &self,
     ) -> Result<Ref<OwningDirstateMap>, DirstateError> {
         self.dirstate_map.get_or_init(|| self.new_dirstate_map())
     }
 
     pub fn dirstate_map_mut(
         &self,
     ) -> Result<RefMut<OwningDirstateMap>, DirstateError> {
         self.dirstate_map
             .get_mut_or_init(|| self.new_dirstate_map())
     }
 
     fn new_changelog(&self) -> Result<Changelog, HgError> {
         Changelog::open(&self.store_vfs(), self.has_nodemap())
     }
 
     pub fn changelog(&self) -> Result<Ref<Changelog>, HgError> {
         self.changelog.get_or_init(|| self.new_changelog())
     }
 
     pub fn changelog_mut(&self) -> Result<RefMut<Changelog>, HgError> {
         self.changelog.get_mut_or_init(|| self.new_changelog())
     }
 
     fn new_manifestlog(&self) -> Result<Manifestlog, HgError> {
         Manifestlog::open(&self.store_vfs(), self.has_nodemap())
     }
 
     pub fn manifestlog(&self) -> Result<Ref<Manifestlog>, HgError> {
         self.manifestlog.get_or_init(|| self.new_manifestlog())
     }
 
     pub fn manifestlog_mut(&self) -> Result<RefMut<Manifestlog>, HgError> {
         self.manifestlog.get_mut_or_init(|| self.new_manifestlog())
     }
 
     /// Returns the manifest of the *changeset* with the given node ID
     pub fn manifest_for_node(
         &self,
         node: impl Into<NodePrefix>,
     ) -> Result<Manifest, RevlogError> {
         self.manifestlog()?.data_for_node(
             self.changelog()?
                 .data_for_node(node.into())?
                 .manifest_node()?
                 .into(),
         )
     }
 
     /// Returns the manifest of the *changeset* with the given revision number
     pub fn manifest_for_rev(
         &self,
         revision: UncheckedRevision,
     ) -> Result<Manifest, RevlogError> {
         self.manifestlog()?.data_for_node(
             self.changelog()?
                 .data_for_rev(revision)?
                 .manifest_node()?
                 .into(),
         )
     }
 
     pub fn has_subrepos(&self) -> Result<bool, DirstateError> {
         if let Some(entry) = self.dirstate_map()?.get(HgPath::new(".hgsub"))? {
             Ok(entry.tracked())
         } else {
             Ok(false)
         }
     }
 
     pub fn filelog(&self, path: &HgPath) -> Result<Filelog, HgError> {
         Filelog::open(self, path)
     }
 
     /// Write to disk any updates that were made through `dirstate_map_mut`.
     ///
     /// The "wlock" must be held while calling this.
     /// See for example `try_with_wlock_no_wait`.
     ///
     /// TODO: have a `WritableRepo` type only accessible while holding the
     /// lock?
     pub fn write_dirstate(&self) -> Result<(), DirstateError> {
         let map = self.dirstate_map()?;
         // TODO: Maintain a `DirstateMap::dirty` flag, and return early here if
         // it’s unset
         let parents = self.dirstate_parents()?;
         let (packed_dirstate, old_uuid_to_remove) = if self.use_dirstate_v2() {
             let (identity, uuid, data_size) =
                 self.get_dirstate_data_file_integrity()?;
             let identity_changed = identity != map.old_identity();
             let uuid_changed = uuid.as_deref() != map.old_uuid();
             let data_length_changed = data_size != map.old_data_size();
 
             if identity_changed || uuid_changed || data_length_changed {
                 // If any of identity, uuid or length have changed since
                 // last disk read, don't write.
                 // This is fine because either we're in a command that doesn't
                 // write anything too important (like `hg status`), or we're in
                 // `hg add` and we're supposed to have taken the lock before
                 // reading anyway.
                 //
                 // TODO complain loudly if we've changed anything important
                 // without taking the lock.
                 // (see `hg help config.format.use-dirstate-tracked-hint`)
                 log::debug!(
                     "dirstate has changed since last read, not updating."
                 );
                 return Ok(());
             }
 
             let uuid_opt = map.old_uuid();
             let write_mode = if uuid_opt.is_some() {
                 DirstateMapWriteMode::Auto
             } else {
                 DirstateMapWriteMode::ForceNewDataFile
             };
             let (data, tree_metadata, append, old_data_size) =
                 map.pack_v2(write_mode)?;
 
             // Reuse the uuid, or generate a new one, keeping the old for
             // deletion.
             let (uuid, old_uuid) = match uuid_opt {
                 Some(uuid) => {
                     let as_str = std::str::from_utf8(uuid)
                         .map_err(|_| {
                             HgError::corrupted(
                                 "non-UTF-8 dirstate data file ID",
                             )
                         })?
                         .to_owned();
                     if append {
                         (as_str, None)
                     } else {
                         (DirstateDocket::new_uid(), Some(as_str))
                     }
                 }
                 None => (DirstateDocket::new_uid(), None),
             };
 
             let data_filename = format!("dirstate.{}", uuid);
             let data_filename = self.hg_vfs().join(data_filename);
             let mut options = std::fs::OpenOptions::new();
             options.write(true);
 
             // Why are we not using the O_APPEND flag when appending?
             //
             // - O_APPEND makes it trickier to deal with garbage at the end of
             //   the file, left by a previous uncommitted transaction. By
             //   starting the write at [old_data_size] we make sure we erase
             //   all such garbage.
             //
             // - O_APPEND requires to special-case 0-byte writes, whereas we
             //   don't need that.
             //
             // - Some OSes have bugs in implementation O_APPEND:
             //   revlog.py talks about a Solaris bug, but we also saw some ZFS
             //   bug: https://github.com/openzfs/zfs/pull/3124,
             //   https://github.com/openzfs/zfs/issues/13370
             //
             if !append {
                 log::trace!("creating a new dirstate data file");
                 options.create_new(true);
             } else {
                 log::trace!("appending to the dirstate data file");
             }
 
             let data_size = (|| {
                 // TODO: loop and try another random ID if !append and this
                 // returns `ErrorKind::AlreadyExists`? Collision chance of two
                 // random IDs is one in 2**32
                 let mut file = options.open(&data_filename)?;
                 if append {
                     file.seek(SeekFrom::Start(old_data_size as u64))?;
                 }
                 file.write_all(&data)?;
                 file.flush()?;
-                file.seek(SeekFrom::Current(0))
+                file.stream_position()
             })()
             .when_writing_file(&data_filename)?;
 
             let packed_dirstate = DirstateDocket::serialize(
                 parents,
                 tree_metadata,
                 data_size,
                 uuid.as_bytes(),
             )
             .map_err(|_: std::num::TryFromIntError| {
                 HgError::corrupted("overflow in dirstate docket serialization")
             })?;
 
             (packed_dirstate, old_uuid)
         } else {
             let identity = self.dirstate_identity()?;
             if identity != map.old_identity() {
                 // If identity changed since last disk read, don't write.
                 // This is fine because either we're in a command that doesn't
                 // write anything too important (like `hg status`), or we're in
                 // `hg add` and we're supposed to have taken the lock before
                 // reading anyway.
                 //
                 // TODO complain loudly if we've changed anything important
                 // without taking the lock.
                 // (see `hg help config.format.use-dirstate-tracked-hint`)
                 log::debug!(
                     "dirstate has changed since last read, not updating."
                 );
                 return Ok(());
             }
             (map.pack_v1(parents)?, None)
         };
 
         let vfs = self.hg_vfs();
         vfs.atomic_write("dirstate", &packed_dirstate)?;
         if let Some(uuid) = old_uuid_to_remove {
             // Remove the old data file after the new docket pointing to the
             // new data file was written.
             vfs.remove_file(format!("dirstate.{}", uuid))?;
         }
         Ok(())
     }
 }
 
 /// Lazily-initialized component of `Repo` with interior mutability
 ///
 /// This differs from `OnceCell` in that the value can still be "deinitialized"
 /// later by setting its inner `Option` to `None`. It also takes the
 /// initialization function as an argument when the value is requested, not
 /// when the instance is created.
 struct LazyCell<T> {
     value: RefCell<Option<T>>,
 }
 
 impl<T> LazyCell<T> {
     fn new() -> Self {
         Self {
             value: RefCell::new(None),
         }
     }
 
     fn set(&self, value: T) {
         *self.value.borrow_mut() = Some(value)
     }
 
     fn get_or_init<E>(
         &self,
         init: impl Fn() -> Result<T, E>,
     ) -> Result<Ref<T>, E> {
         let mut borrowed = self.value.borrow();
         if borrowed.is_none() {
             drop(borrowed);
             // Only use `borrow_mut` if it is really needed to avoid panic in
             // case there is another outstanding borrow but mutation is not
             // needed.
             *self.value.borrow_mut() = Some(init()?);
             borrowed = self.value.borrow()
         }
         Ok(Ref::map(borrowed, |option| option.as_ref().unwrap()))
     }
 
     fn get_mut_or_init<E>(
         &self,
         init: impl Fn() -> Result<T, E>,
     ) -> Result<RefMut<T>, E> {
         let mut borrowed = self.value.borrow_mut();
         if borrowed.is_none() {
             *borrowed = Some(init()?);
         }
         Ok(RefMut::map(borrowed, |option| option.as_mut().unwrap()))
     }
 }

rust/hg-core/src/revlog/mod.rs

@@ -1,965 +1,965 b''
 // Copyright 2018-2023 Georges Racinet <[e-mail]>
 //           and Mercurial contributors
 //
 // This software may be used and distributed according to the terms of the
 // GNU General Public License version 2 or any later version.
 //! Mercurial concepts for handling revision history
 
 pub mod node;
 pub mod nodemap;
 mod nodemap_docket;
 pub mod path_encode;
 pub use node::{FromHexError, Node, NodePrefix};
 pub mod changelog;
 pub mod filelog;
 pub mod index;
 pub mod manifest;
 pub mod patch;
 
 use std::borrow::Cow;
 use std::io::Read;
 use std::ops::Deref;
 use std::path::Path;
 
 use flate2::read::ZlibDecoder;
 use sha1::{Digest, Sha1};
 use std::cell::RefCell;
 use zstd;
 
 use self::node::{NODE_BYTES_LENGTH, NULL_NODE};
 use self::nodemap_docket::NodeMapDocket;
 use super::index::Index;
 use super::nodemap::{NodeMap, NodeMapError};
 use crate::errors::HgError;
 use crate::vfs::Vfs;
 
 /// As noted in revlog.c, revision numbers are actually encoded in
 /// 4 bytes, and are liberally converted to ints, whence the i32
 pub type BaseRevision = i32;
 
 /// Mercurial revision numbers
 /// In contrast to the more general [`UncheckedRevision`], these are "checked"
 /// in the sense that they should only be used for revisions that are
 /// valid for a given index (i.e. in bounds).
 #[derive(
     Debug,
     derive_more::Display,
     Clone,
     Copy,
     Hash,
     PartialEq,
     Eq,
     PartialOrd,
     Ord,
 )]
 pub struct Revision(pub BaseRevision);
 
 impl format_bytes::DisplayBytes for Revision {
     fn display_bytes(
         &self,
         output: &mut dyn std::io::Write,
     ) -> std::io::Result<()> {
         self.0.display_bytes(output)
     }
 }
 
 /// Unchecked Mercurial revision numbers.
 ///
 /// Values of this type have no guarantee of being a valid revision number
 /// in any context. Use method `check_revision` to get a valid revision within
 /// the appropriate index object.
 #[derive(
     Debug,
     derive_more::Display,
     Clone,
     Copy,
     Hash,
     PartialEq,
     Eq,
     PartialOrd,
     Ord,
 )]
 pub struct UncheckedRevision(pub BaseRevision);
 
 impl format_bytes::DisplayBytes for UncheckedRevision {
     fn display_bytes(
         &self,
         output: &mut dyn std::io::Write,
     ) -> std::io::Result<()> {
         self.0.display_bytes(output)
     }
 }
 
 impl From<Revision> for UncheckedRevision {
     fn from(value: Revision) -> Self {
         Self(value.0)
     }
 }
 
 impl From<BaseRevision> for UncheckedRevision {
     fn from(value: BaseRevision) -> Self {
         Self(value)
     }
 }
 
 /// Marker expressing the absence of a parent
 ///
 /// Independently of the actual representation, `NULL_REVISION` is guaranteed
 /// to be smaller than all existing revisions.
 pub const NULL_REVISION: Revision = Revision(-1);
 
 /// Same as `mercurial.node.wdirrev`
 ///
 /// This is also equal to `i32::max_value()`, but it's better to spell
 /// it out explicitely, same as in `mercurial.node`
 #[allow(clippy::unreadable_literal)]
 pub const WORKING_DIRECTORY_REVISION: UncheckedRevision =
     UncheckedRevision(0x7fffffff);
 
 pub const WORKING_DIRECTORY_HEX: &str =
     "ffffffffffffffffffffffffffffffffffffffff";
 
 /// The simplest expression of what we need of Mercurial DAGs.
 pub trait Graph {
     /// Return the two parents of the given `Revision`.
     ///
     /// Each of the parents can be independently `NULL_REVISION`
     fn parents(&self, rev: Revision) -> Result<[Revision; 2], GraphError>;
 }
 
 #[derive(Clone, Debug, PartialEq)]
 pub enum GraphError {
     ParentOutOfRange(Revision),
 }
 
 /// The Mercurial Revlog Index
 ///
 /// This is currently limited to the minimal interface that is needed for
 /// the [`nodemap`](nodemap/index.html) module
 pub trait RevlogIndex {
     /// Total number of Revisions referenced in this index
     fn len(&self) -> usize;
 
     fn is_empty(&self) -> bool {
         self.len() == 0
     }
 
     /// Return a reference to the Node or `None` for `NULL_REVISION`
     fn node(&self, rev: Revision) -> Option<&Node>;
 
     /// Return a [`Revision`] if `rev` is a valid revision number for this
     /// index
     fn check_revision(&self, rev: UncheckedRevision) -> Option<Revision> {
         let rev = rev.0;
 
         if rev == NULL_REVISION.0 || (rev >= 0 && (rev as usize) < self.len())
         {
             Some(Revision(rev))
         } else {
             None
         }
     }
 }
 
 const REVISION_FLAG_CENSORED: u16 = 1 << 15;
 const REVISION_FLAG_ELLIPSIS: u16 = 1 << 14;
 const REVISION_FLAG_EXTSTORED: u16 = 1 << 13;
 const REVISION_FLAG_HASCOPIESINFO: u16 = 1 << 12;
 
 // Keep this in sync with REVIDX_KNOWN_FLAGS in
 // mercurial/revlogutils/flagutil.py
 const REVIDX_KNOWN_FLAGS: u16 = REVISION_FLAG_CENSORED
     | REVISION_FLAG_ELLIPSIS
     | REVISION_FLAG_EXTSTORED
     | REVISION_FLAG_HASCOPIESINFO;
 
 const NULL_REVLOG_ENTRY_FLAGS: u16 = 0;
 
 #[derive(Debug, derive_more::From, derive_more::Display)]
 pub enum RevlogError {
     InvalidRevision,
     /// Working directory is not supported
     WDirUnsupported,
     /// Found more than one entry whose ID match the requested prefix
     AmbiguousPrefix,
     #[from]
     Other(HgError),
 }
 
 impl From<NodeMapError> for RevlogError {
     fn from(error: NodeMapError) -> Self {
         match error {
             NodeMapError::MultipleResults => RevlogError::AmbiguousPrefix,
             NodeMapError::RevisionNotInIndex(rev) => RevlogError::corrupted(
                 format!("nodemap point to revision {} not in index", rev),
             ),
         }
     }
 }
 
 fn corrupted<S: AsRef<str>>(context: S) -> HgError {
     HgError::corrupted(format!("corrupted revlog, {}", context.as_ref()))
 }
 
 impl RevlogError {
     fn corrupted<S: AsRef<str>>(context: S) -> Self {
         RevlogError::Other(corrupted(context))
     }
 }
 
 /// Read only implementation of revlog.
 pub struct Revlog {
     /// When index and data are not interleaved: bytes of the revlog index.
     /// When index and data are interleaved: bytes of the revlog index and
     /// data.
     index: Index,
     /// When index and data are not interleaved: bytes of the revlog data
     data_bytes: Option<Box<dyn Deref<Target = [u8]> + Send>>,
     /// When present on disk: the persistent nodemap for this revlog
     nodemap: Option<nodemap::NodeTree>,
 }
 
 impl Graph for Revlog {
     fn parents(&self, rev: Revision) -> Result<[Revision; 2], GraphError> {
         self.index.parents(rev)
     }
 }
 
 impl Revlog {
     /// Open a revlog index file.
     ///
     /// It will also open the associated data file if index and data are not
     /// interleaved.
     pub fn open(
         store_vfs: &Vfs,
         index_path: impl AsRef<Path>,
         data_path: Option<&Path>,
         use_nodemap: bool,
     ) -> Result<Self, HgError> {
         Self::open_gen(store_vfs, index_path, data_path, use_nodemap, None)
     }
 
     fn open_gen(
         store_vfs: &Vfs,
         index_path: impl AsRef<Path>,
         data_path: Option<&Path>,
         use_nodemap: bool,
         nodemap_for_test: Option<nodemap::NodeTree>,
     ) -> Result<Self, HgError> {
         let index_path = index_path.as_ref();
         let index = {
-            match store_vfs.mmap_open_opt(&index_path)? {
-                None => Index::new(Box::new(vec![])),
+            match store_vfs.mmap_open_opt(index_path)? {
+                None => Index::new(Box::<Vec<_>>::default()),
                 Some(index_mmap) => {
                     let index = Index::new(Box::new(index_mmap))?;
                     Ok(index)
                 }
             }
         }?;
 
         let default_data_path = index_path.with_extension("d");
 
         // type annotation required
         // won't recognize Mmap as Deref<Target = [u8]>
         let data_bytes: Option<Box<dyn Deref<Target = [u8]> + Send>> =
             if index.is_inline() {
                 None
             } else {
                 let data_path = data_path.unwrap_or(&default_data_path);
                 let data_mmap = store_vfs.mmap_open(data_path)?;
                 Some(Box::new(data_mmap))
             };
 
         let nodemap = if index.is_inline() || !use_nodemap {
             None
         } else {
             NodeMapDocket::read_from_file(store_vfs, index_path)?.map(
                 |(docket, data)| {
                     nodemap::NodeTree::load_bytes(
                         Box::new(data),
                         docket.data_length,
                     )
                 },
             )
         };
 
         let nodemap = nodemap_for_test.or(nodemap);
 
         Ok(Revlog {
             index,
             data_bytes,
             nodemap,
         })
     }
 
     /// Return number of entries of the `Revlog`.
     pub fn len(&self) -> usize {
         self.index.len()
     }
 
     /// Returns `true` if the `Revlog` has zero `entries`.
     pub fn is_empty(&self) -> bool {
         self.index.is_empty()
     }
 
     /// Returns the node ID for the given revision number, if it exists in this
     /// revlog
     pub fn node_from_rev(&self, rev: UncheckedRevision) -> Option<&Node> {
         if rev == NULL_REVISION.into() {
             return Some(&NULL_NODE);
         }
         let rev = self.index.check_revision(rev)?;
         Some(self.index.get_entry(rev)?.hash())
     }
 
     /// Return the revision number for the given node ID, if it exists in this
     /// revlog
     pub fn rev_from_node(
         &self,
         node: NodePrefix,
     ) -> Result<Revision, RevlogError> {
         if let Some(nodemap) = &self.nodemap {
             nodemap
                 .find_bin(&self.index, node)?
                 .ok_or(RevlogError::InvalidRevision)
         } else {
             self.rev_from_node_no_persistent_nodemap(node)
         }
     }
 
     /// Same as `rev_from_node`, without using a persistent nodemap
     ///
     /// This is used as fallback when a persistent nodemap is not present.
     /// This happens when the persistent-nodemap experimental feature is not
     /// enabled, or for small revlogs.
     fn rev_from_node_no_persistent_nodemap(
         &self,
         node: NodePrefix,
     ) -> Result<Revision, RevlogError> {
         // Linear scan of the revlog
         // TODO: consider building a non-persistent nodemap in memory to
         // optimize these cases.
         let mut found_by_prefix = None;
         for rev in (-1..self.len() as BaseRevision).rev() {
             let rev = Revision(rev as BaseRevision);
             let candidate_node = if rev == Revision(-1) {
                 NULL_NODE
             } else {
                 let index_entry =
                     self.index.get_entry(rev).ok_or_else(|| {
                         HgError::corrupted(
                             "revlog references a revision not in the index",
                         )
                     })?;
                 *index_entry.hash()
             };
             if node == candidate_node {
                 return Ok(rev);
             }
             if node.is_prefix_of(&candidate_node) {
                 if found_by_prefix.is_some() {
                     return Err(RevlogError::AmbiguousPrefix);
                 }
                 found_by_prefix = Some(rev)
             }
         }
         found_by_prefix.ok_or(RevlogError::InvalidRevision)
     }
 
     /// Returns whether the given revision exists in this revlog.
     pub fn has_rev(&self, rev: UncheckedRevision) -> bool {
         self.index.check_revision(rev).is_some()
     }
 
     /// Return the full data associated to a revision.
     ///
     /// All entries required to build the final data out of deltas will be
     /// retrieved as needed, and the deltas will be applied to the inital
     /// snapshot to rebuild the final data.
     pub fn get_rev_data(
         &self,
         rev: UncheckedRevision,
     ) -> Result<Cow<[u8]>, RevlogError> {
         if rev == NULL_REVISION.into() {
             return Ok(Cow::Borrowed(&[]));
         };
         self.get_entry(rev)?.data()
     }
 
     /// [`Self::get_rev_data`] for checked revisions.
     pub fn get_rev_data_for_checked_rev(
         &self,
         rev: Revision,
     ) -> Result<Cow<[u8]>, RevlogError> {
         if rev == NULL_REVISION {
             return Ok(Cow::Borrowed(&[]));
         };
         self.get_entry_for_checked_rev(rev)?.data()
     }
 
     /// Check the hash of some given data against the recorded hash.
     pub fn check_hash(
         &self,
         p1: Revision,
         p2: Revision,
         expected: &[u8],
         data: &[u8],
     ) -> bool {
         let e1 = self.index.get_entry(p1);
         let h1 = match e1 {
             Some(ref entry) => entry.hash(),
             None => &NULL_NODE,
         };
         let e2 = self.index.get_entry(p2);
         let h2 = match e2 {
             Some(ref entry) => entry.hash(),
             None => &NULL_NODE,
         };
 
         hash(data, h1.as_bytes(), h2.as_bytes()) == expected
     }
 
     /// Build the full data of a revision out its snapshot
     /// and its deltas.
     fn build_data_from_deltas(
         snapshot: RevlogEntry,
         deltas: &[RevlogEntry],
     ) -> Result<Vec<u8>, HgError> {
         let snapshot = snapshot.data_chunk()?;
         let deltas = deltas
             .iter()
             .rev()
             .map(RevlogEntry::data_chunk)
             .collect::<Result<Vec<_>, _>>()?;
         let patches: Vec<_> =
             deltas.iter().map(|d| patch::PatchList::new(d)).collect();
         let patch = patch::fold_patch_lists(&patches);
         Ok(patch.apply(&snapshot))
     }
 
     /// Return the revlog data.
     fn data(&self) -> &[u8] {
         match &self.data_bytes {
             Some(data_bytes) => data_bytes,
             None => panic!(
                 "forgot to load the data or trying to access inline data"
             ),
         }
     }
 
     pub fn make_null_entry(&self) -> RevlogEntry {
         RevlogEntry {
             revlog: self,
             rev: NULL_REVISION,
             bytes: b"",
             compressed_len: 0,
             uncompressed_len: 0,
             base_rev_or_base_of_delta_chain: None,
             p1: NULL_REVISION,
             p2: NULL_REVISION,
             flags: NULL_REVLOG_ENTRY_FLAGS,
             hash: NULL_NODE,
         }
     }
 
     fn get_entry_for_checked_rev(
         &self,
         rev: Revision,
     ) -> Result<RevlogEntry, RevlogError> {
         if rev == NULL_REVISION {
             return Ok(self.make_null_entry());
         }
         let index_entry = self
             .index
             .get_entry(rev)
             .ok_or(RevlogError::InvalidRevision)?;
         let start = index_entry.offset();
         let end = start + index_entry.compressed_len() as usize;
         let data = if self.index.is_inline() {
             self.index.data(start, end)
         } else {
             &self.data()[start..end]
         };
         let base_rev = self
             .index
             .check_revision(index_entry.base_revision_or_base_of_delta_chain())
             .ok_or_else(|| {
                 RevlogError::corrupted(format!(
                     "base revision for rev {} is invalid",
                     rev
                 ))
             })?;
         let p1 =
             self.index.check_revision(index_entry.p1()).ok_or_else(|| {
                 RevlogError::corrupted(format!(
                     "p1 for rev {} is invalid",
                     rev
                 ))
             })?;
         let p2 =
             self.index.check_revision(index_entry.p2()).ok_or_else(|| {
                 RevlogError::corrupted(format!(
                     "p2 for rev {} is invalid",
                     rev
                 ))
             })?;
         let entry = RevlogEntry {
             revlog: self,
             rev,
             bytes: data,
             compressed_len: index_entry.compressed_len(),
             uncompressed_len: index_entry.uncompressed_len(),
             base_rev_or_base_of_delta_chain: if base_rev == rev {
                 None
             } else {
                 Some(base_rev)
             },
             p1,
             p2,
             flags: index_entry.flags(),
             hash: *index_entry.hash(),
         };
         Ok(entry)
     }
 
     /// Get an entry of the revlog.
     pub fn get_entry(
         &self,
         rev: UncheckedRevision,
     ) -> Result<RevlogEntry, RevlogError> {
         if rev == NULL_REVISION.into() {
             return Ok(self.make_null_entry());
         }
         let rev = self.index.check_revision(rev).ok_or_else(|| {
             RevlogError::corrupted(format!("rev {} is invalid", rev))
         })?;
         self.get_entry_for_checked_rev(rev)
     }
 }
 
 /// The revlog entry's bytes and the necessary informations to extract
 /// the entry's data.
 #[derive(Clone)]
 pub struct RevlogEntry<'revlog> {
     revlog: &'revlog Revlog,
     rev: Revision,
     bytes: &'revlog [u8],
     compressed_len: u32,
     uncompressed_len: i32,
     base_rev_or_base_of_delta_chain: Option<Revision>,
     p1: Revision,
     p2: Revision,
     flags: u16,
     hash: Node,
 }
 
 thread_local! {
   // seems fine to [unwrap] here: this can only fail due to memory allocation
   // failing, and it's normal for that to cause panic.
   static ZSTD_DECODER : RefCell<zstd::bulk::Decompressor<'static>> =
       RefCell::new(zstd::bulk::Decompressor::new().ok().unwrap());
 }
 
 fn zstd_decompress_to_buffer(
     bytes: &[u8],
     buf: &mut Vec<u8>,
 ) -> Result<usize, std::io::Error> {
     ZSTD_DECODER
         .with(|decoder| decoder.borrow_mut().decompress_to_buffer(bytes, buf))
 }
 
 impl<'revlog> RevlogEntry<'revlog> {
     pub fn revision(&self) -> Revision {
         self.rev
     }
 
     pub fn node(&self) -> &Node {
         &self.hash
     }
 
     pub fn uncompressed_len(&self) -> Option<u32> {
         u32::try_from(self.uncompressed_len).ok()
     }
 
     pub fn has_p1(&self) -> bool {
         self.p1 != NULL_REVISION
     }
 
     pub fn p1_entry(
         &self,
     ) -> Result<Option<RevlogEntry<'revlog>>, RevlogError> {
         if self.p1 == NULL_REVISION {
             Ok(None)
         } else {
             Ok(Some(self.revlog.get_entry_for_checked_rev(self.p1)?))
         }
     }
 
     pub fn p2_entry(
         &self,
     ) -> Result<Option<RevlogEntry<'revlog>>, RevlogError> {
         if self.p2 == NULL_REVISION {
             Ok(None)
         } else {
             Ok(Some(self.revlog.get_entry_for_checked_rev(self.p2)?))
         }
     }
 
     pub fn p1(&self) -> Option<Revision> {
         if self.p1 == NULL_REVISION {
             None
         } else {
             Some(self.p1)
         }
     }
 
     pub fn p2(&self) -> Option<Revision> {
         if self.p2 == NULL_REVISION {
             None
         } else {
             Some(self.p2)
         }
     }
 
     pub fn is_censored(&self) -> bool {
         (self.flags & REVISION_FLAG_CENSORED) != 0
     }
 
     pub fn has_length_affecting_flag_processor(&self) -> bool {
         // Relevant Python code: revlog.size()
         // note: ELLIPSIS is known to not change the content
         (self.flags & (REVIDX_KNOWN_FLAGS ^ REVISION_FLAG_ELLIPSIS)) != 0
     }
 
     /// The data for this entry, after resolving deltas if any.
     pub fn rawdata(&self) -> Result<Cow<'revlog, [u8]>, RevlogError> {
         let mut entry = self.clone();
         let mut delta_chain = vec![];
 
         // The meaning of `base_rev_or_base_of_delta_chain` depends on
         // generaldelta. See the doc on `ENTRY_DELTA_BASE` in
         // `mercurial/revlogutils/constants.py` and the code in
         // [_chaininfo] and in [index_deltachain].
         let uses_generaldelta = self.revlog.index.uses_generaldelta();
         while let Some(base_rev) = entry.base_rev_or_base_of_delta_chain {
             entry = if uses_generaldelta {
                 delta_chain.push(entry);
                 self.revlog.get_entry_for_checked_rev(base_rev)?
             } else {
                 let base_rev = UncheckedRevision(entry.rev.0 - 1);
                 delta_chain.push(entry);
                 self.revlog.get_entry(base_rev)?
             };
         }
 
         let data = if delta_chain.is_empty() {
             entry.data_chunk()?
         } else {
             Revlog::build_data_from_deltas(entry, &delta_chain)?.into()
         };
 
         Ok(data)
     }
 
     fn check_data(
         &self,
         data: Cow<'revlog, [u8]>,
     ) -> Result<Cow<'revlog, [u8]>, RevlogError> {
         if self.revlog.check_hash(
             self.p1,
             self.p2,
             self.hash.as_bytes(),
             &data,
         ) {
             Ok(data)
         } else {
             if (self.flags & REVISION_FLAG_ELLIPSIS) != 0 {
                 return Err(HgError::unsupported(
                     "ellipsis revisions are not supported by rhg",
                 )
                 .into());
             }
             Err(corrupted(format!(
                 "hash check failed for revision {}",
                 self.rev
             ))
             .into())
         }
     }
 
     pub fn data(&self) -> Result<Cow<'revlog, [u8]>, RevlogError> {
         let data = self.rawdata()?;
         if self.rev == NULL_REVISION {
             return Ok(data);
         }
         if self.is_censored() {
             return Err(HgError::CensoredNodeError.into());
         }
         self.check_data(data)
     }
 
     /// Extract the data contained in the entry.
     /// This may be a delta. (See `is_delta`.)
     fn data_chunk(&self) -> Result<Cow<'revlog, [u8]>, HgError> {
         if self.bytes.is_empty() {
             return Ok(Cow::Borrowed(&[]));
         }
         match self.bytes[0] {
             // Revision data is the entirety of the entry, including this
             // header.
             b'\0' => Ok(Cow::Borrowed(self.bytes)),
             // Raw revision data follows.
             b'u' => Ok(Cow::Borrowed(&self.bytes[1..])),
             // zlib (RFC 1950) data.
             b'x' => Ok(Cow::Owned(self.uncompressed_zlib_data()?)),
             // zstd data.
             b'\x28' => Ok(Cow::Owned(self.uncompressed_zstd_data()?)),
             // A proper new format should have had a repo/store requirement.
             format_type => Err(corrupted(format!(
                 "unknown compression header '{}'",
                 format_type
             ))),
         }
     }
 
     fn uncompressed_zlib_data(&self) -> Result<Vec<u8>, HgError> {
         let mut decoder = ZlibDecoder::new(self.bytes);
         if self.is_delta() {
             let mut buf = Vec::with_capacity(self.compressed_len as usize);
             decoder
                 .read_to_end(&mut buf)
                 .map_err(|e| corrupted(e.to_string()))?;
             Ok(buf)
         } else {
             let cap = self.uncompressed_len.max(0) as usize;
             let mut buf = vec![0; cap];
             decoder
                 .read_exact(&mut buf)
                 .map_err(|e| corrupted(e.to_string()))?;
             Ok(buf)
         }
     }
 
     fn uncompressed_zstd_data(&self) -> Result<Vec<u8>, HgError> {
         let cap = self.uncompressed_len.max(0) as usize;
         if self.is_delta() {
             // [cap] is usually an over-estimate of the space needed because
             // it's the length of delta-decoded data, but we're interested
             // in the size of the delta.
             // This means we have to [shrink_to_fit] to avoid holding on
             // to a large chunk of memory, but it also means we must have a
             // fallback branch, for the case when the delta is longer than
             // the original data (surprisingly, this does happen in practice)
             let mut buf = Vec::with_capacity(cap);
             match zstd_decompress_to_buffer(self.bytes, &mut buf) {
                 Ok(_) => buf.shrink_to_fit(),
                 Err(_) => {
                     buf.clear();
                     zstd::stream::copy_decode(self.bytes, &mut buf)
                         .map_err(|e| corrupted(e.to_string()))?;
                 }
             };
             Ok(buf)
         } else {
             let mut buf = Vec::with_capacity(cap);
             let len = zstd_decompress_to_buffer(self.bytes, &mut buf)
                 .map_err(|e| corrupted(e.to_string()))?;
             if len != self.uncompressed_len as usize {
                 Err(corrupted("uncompressed length does not match"))
             } else {
                 Ok(buf)
             }
         }
     }
 
     /// Tell if the entry is a snapshot or a delta
     /// (influences on decompression).
     fn is_delta(&self) -> bool {
         self.base_rev_or_base_of_delta_chain.is_some()
     }
 }
 
 /// Calculate the hash of a revision given its data and its parents.
 fn hash(
     data: &[u8],
     p1_hash: &[u8],
     p2_hash: &[u8],
 ) -> [u8; NODE_BYTES_LENGTH] {
     let mut hasher = Sha1::new();
     let (a, b) = (p1_hash, p2_hash);
     if a > b {
         hasher.update(b);
         hasher.update(a);
     } else {
         hasher.update(a);
         hasher.update(b);
     }
     hasher.update(data);
     *hasher.finalize().as_ref()
 }
 
 #[cfg(test)]
 mod tests {
     use super::*;
     use crate::index::{IndexEntryBuilder, INDEX_ENTRY_SIZE};
     use itertools::Itertools;
 
     #[test]
     fn test_empty() {
         let temp = tempfile::tempdir().unwrap();
         let vfs = Vfs { base: temp.path() };
         std::fs::write(temp.path().join("foo.i"), b"").unwrap();
         let revlog = Revlog::open(&vfs, "foo.i", None, false).unwrap();
         assert!(revlog.is_empty());
         assert_eq!(revlog.len(), 0);
         assert!(revlog.get_entry(0.into()).is_err());
         assert!(!revlog.has_rev(0.into()));
         assert_eq!(
             revlog.rev_from_node(NULL_NODE.into()).unwrap(),
             NULL_REVISION
         );
         let null_entry = revlog.get_entry(NULL_REVISION.into()).ok().unwrap();
         assert_eq!(null_entry.revision(), NULL_REVISION);
         assert!(null_entry.data().unwrap().is_empty());
     }
 
     #[test]
     fn test_inline() {
         let temp = tempfile::tempdir().unwrap();
         let vfs = Vfs { base: temp.path() };
         let node0 = Node::from_hex("2ed2a3912a0b24502043eae84ee4b279c18b90dd")
             .unwrap();
         let node1 = Node::from_hex("b004912a8510032a0350a74daa2803dadfb00e12")
             .unwrap();
         let node2 = Node::from_hex("dd6ad206e907be60927b5a3117b97dffb2590582")
             .unwrap();
         let entry0_bytes = IndexEntryBuilder::new()
             .is_first(true)
             .with_version(1)
             .with_inline(true)
             .with_offset(INDEX_ENTRY_SIZE)
             .with_node(node0)
             .build();
         let entry1_bytes = IndexEntryBuilder::new()
             .with_offset(INDEX_ENTRY_SIZE)
             .with_node(node1)
             .build();
         let entry2_bytes = IndexEntryBuilder::new()
             .with_offset(INDEX_ENTRY_SIZE)
             .with_p1(Revision(0))
             .with_p2(Revision(1))
             .with_node(node2)
             .build();
         let contents = vec![entry0_bytes, entry1_bytes, entry2_bytes]
             .into_iter()
             .flatten()
             .collect_vec();
         std::fs::write(temp.path().join("foo.i"), contents).unwrap();
         let revlog = Revlog::open(&vfs, "foo.i", None, false).unwrap();
 
         let entry0 = revlog.get_entry(0.into()).ok().unwrap();
         assert_eq!(entry0.revision(), Revision(0));
         assert_eq!(*entry0.node(), node0);
         assert!(!entry0.has_p1());
         assert_eq!(entry0.p1(), None);
         assert_eq!(entry0.p2(), None);
         let p1_entry = entry0.p1_entry().unwrap();
         assert!(p1_entry.is_none());
         let p2_entry = entry0.p2_entry().unwrap();
         assert!(p2_entry.is_none());
 
         let entry1 = revlog.get_entry(1.into()).ok().unwrap();
         assert_eq!(entry1.revision(), Revision(1));
         assert_eq!(*entry1.node(), node1);
         assert!(!entry1.has_p1());
         assert_eq!(entry1.p1(), None);
         assert_eq!(entry1.p2(), None);
         let p1_entry = entry1.p1_entry().unwrap();
         assert!(p1_entry.is_none());
         let p2_entry = entry1.p2_entry().unwrap();
         assert!(p2_entry.is_none());
 
         let entry2 = revlog.get_entry(2.into()).ok().unwrap();
         assert_eq!(entry2.revision(), Revision(2));
         assert_eq!(*entry2.node(), node2);
         assert!(entry2.has_p1());
         assert_eq!(entry2.p1(), Some(Revision(0)));
         assert_eq!(entry2.p2(), Some(Revision(1)));
         let p1_entry = entry2.p1_entry().unwrap();
         assert!(p1_entry.is_some());
         assert_eq!(p1_entry.unwrap().revision(), Revision(0));
         let p2_entry = entry2.p2_entry().unwrap();
         assert!(p2_entry.is_some());
         assert_eq!(p2_entry.unwrap().revision(), Revision(1));
     }
 
     #[test]
     fn test_nodemap() {
         let temp = tempfile::tempdir().unwrap();
         let vfs = Vfs { base: temp.path() };
 
         // building a revlog with a forced Node starting with zeros
         // This is a corruption, but it does not preclude using the nodemap
         // if we don't try and access the data
         let node0 = Node::from_hex("00d2a3912a0b24502043eae84ee4b279c18b90dd")
             .unwrap();
         let node1 = Node::from_hex("b004912a8510032a0350a74daa2803dadfb00e12")
             .unwrap();
         let entry0_bytes = IndexEntryBuilder::new()
             .is_first(true)
             .with_version(1)
             .with_inline(true)
             .with_offset(INDEX_ENTRY_SIZE)
             .with_node(node0)
             .build();
         let entry1_bytes = IndexEntryBuilder::new()
             .with_offset(INDEX_ENTRY_SIZE)
             .with_node(node1)
             .build();
         let contents = vec![entry0_bytes, entry1_bytes]
             .into_iter()
             .flatten()
             .collect_vec();
         std::fs::write(temp.path().join("foo.i"), contents).unwrap();
 
         let mut idx = nodemap::tests::TestNtIndex::new();
         idx.insert_node(Revision(0), node0).unwrap();
         idx.insert_node(Revision(1), node1).unwrap();
 
         let revlog =
             Revlog::open_gen(&vfs, "foo.i", None, true, Some(idx.nt)).unwrap();
 
         // accessing the data shows the corruption
         revlog.get_entry(0.into()).unwrap().data().unwrap_err();
 
         assert_eq!(
             revlog.rev_from_node(NULL_NODE.into()).unwrap(),
             Revision(-1)
         );
         assert_eq!(revlog.rev_from_node(node0.into()).unwrap(), Revision(0));
         assert_eq!(revlog.rev_from_node(node1.into()).unwrap(), Revision(1));
         assert_eq!(
             revlog
                 .rev_from_node(NodePrefix::from_hex("000").unwrap())
                 .unwrap(),
             Revision(-1)
         );
         assert_eq!(
             revlog
                 .rev_from_node(NodePrefix::from_hex("b00").unwrap())
                 .unwrap(),
             Revision(1)
         );
         // RevlogError does not implement PartialEq
         // (ultimately because io::Error does not)
         match revlog
             .rev_from_node(NodePrefix::from_hex("00").unwrap())
             .expect_err("Expected to give AmbiguousPrefix error")
         {
             RevlogError::AmbiguousPrefix => (),
             e => {
                 panic!("Got another error than AmbiguousPrefix: {:?}", e);
             }
         };
     }
 }

rust/hg-core/src/revlog/node.rs

@@ -1,429 +1,429 b''
 // Copyright 2019-2020 Georges Racinet <[e-mail]>
 //
 // This software may be used and distributed according to the terms of the
 // GNU General Public License version 2 or any later version.
 
 //! Definitions and utilities for Revision nodes
 //!
 //! In Mercurial code base, it is customary to call "a node" the binary SHA
 //! of a revision.
 
 use crate::errors::HgError;
 use bytes_cast::BytesCast;
 use std::fmt;
 
 /// The length in bytes of a `Node`
 ///
 /// This constant is meant to ease refactors of this module, and
 /// are private so that calling code does not expect all nodes have
 /// the same size, should we support several formats concurrently in
 /// the future.
 pub const NODE_BYTES_LENGTH: usize = 20;
 
 /// Id of the null node.
 ///
 /// Used to indicate the absence of node.
 pub const NULL_NODE_ID: [u8; NODE_BYTES_LENGTH] = [0u8; NODE_BYTES_LENGTH];
 
 /// The length in bytes of a `Node`
 ///
 /// see also `NODES_BYTES_LENGTH` about it being private.
 const NODE_NYBBLES_LENGTH: usize = 2 * NODE_BYTES_LENGTH;
 
 /// Default for UI presentation
 const SHORT_PREFIX_DEFAULT_NYBBLES_LENGTH: u8 = 12;
 
 /// Private alias for readability and to ease future change
 type NodeData = [u8; NODE_BYTES_LENGTH];
 
 /// Binary revision SHA
 ///
 /// ## Future changes of hash size
 ///
 /// To accomodate future changes of hash size, Rust callers
 /// should use the conversion methods at the boundaries (FFI, actual
 /// computation of hashes and I/O) only, and only if required.
 ///
 /// All other callers outside of unit tests should just handle `Node` values
 /// and never make any assumption on the actual length, using [`nybbles_len`]
 /// if they need a loop boundary.
 ///
 /// All methods that create a `Node` either take a type that enforces
 /// the size or return an error at runtime.
 ///
 /// [`nybbles_len`]: #method.nybbles_len
 #[derive(Copy, Clone, PartialEq, BytesCast, derive_more::From)]
 #[repr(transparent)]
 pub struct Node {
     data: NodeData,
 }
 
 impl fmt::Debug for Node {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
         let n = format!("{:x?}", self.data);
         // We're using debug_tuple because it makes the output a little
         // more compact without losing data.
         f.debug_tuple("Node").field(&n).finish()
     }
 }
 
 /// The node value for NULL_REVISION
 pub const NULL_NODE: Node = Node {
     data: [0; NODE_BYTES_LENGTH],
 };
 
 /// Return an error if the slice has an unexpected length
 impl<'a> TryFrom<&'a [u8]> for &'a Node {
     type Error = ();
 
     #[inline]
     fn try_from(bytes: &'a [u8]) -> Result<Self, Self::Error> {
         match Node::from_bytes(bytes) {
             Ok((node, rest)) if rest.is_empty() => Ok(node),
             _ => Err(()),
         }
     }
 }
 
 /// Return an error if the slice has an unexpected length
 impl TryFrom<&'_ [u8]> for Node {
     type Error = std::array::TryFromSliceError;
 
     #[inline]
     fn try_from(bytes: &'_ [u8]) -> Result<Self, Self::Error> {
         let data = bytes.try_into()?;
         Ok(Self { data })
     }
 }
 
 impl From<&'_ NodeData> for Node {
     #[inline]
     fn from(data: &'_ NodeData) -> Self {
         Self { data: *data }
     }
 }
 
 impl fmt::LowerHex for Node {
     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
         for &byte in &self.data {
             write!(f, "{:02x}", byte)?
         }
         Ok(())
     }
 }
 
 #[derive(Debug)]
 pub struct FromHexError;
 
 /// Low level utility function, also for prefixes
 fn get_nybble(s: &[u8], i: usize) -> u8 {
     if i % 2 == 0 {
         s[i / 2] >> 4
     } else {
         s[i / 2] & 0x0f
     }
 }
 
 impl Node {
     /// Retrieve the `i`th half-byte of the binary data.
     ///
     /// This is also the `i`th hexadecimal digit in numeric form,
     /// also called a [nybble](https://en.wikipedia.org/wiki/Nibble).
     pub fn get_nybble(&self, i: usize) -> u8 {
         get_nybble(&self.data, i)
     }
 
     /// Length of the data, in nybbles
     pub fn nybbles_len(&self) -> usize {
         // public exposure as an instance method only, so that we can
         // easily support several sizes of hashes if needed in the future.
         NODE_NYBBLES_LENGTH
     }
 
     /// Convert from hexadecimal string representation
     ///
     /// Exact length is required.
     ///
     /// To be used in FFI and I/O only, in order to facilitate future
     /// changes of hash format.
     pub fn from_hex(hex: impl AsRef<[u8]>) -> Result<Node, FromHexError> {
         let prefix = NodePrefix::from_hex(hex)?;
         if prefix.nybbles_len() == NODE_NYBBLES_LENGTH {
             Ok(Self { data: prefix.data })
         } else {
             Err(FromHexError)
         }
     }
 
     /// `from_hex`, but for input from an internal file of the repository such
     /// as a changelog or manifest entry.
     ///
     /// An error is treated as repository corruption.
     pub fn from_hex_for_repo(hex: impl AsRef<[u8]>) -> Result<Node, HgError> {
         Self::from_hex(hex.as_ref()).map_err(|FromHexError| {
             HgError::CorruptedRepository(format!(
                 "Expected a full hexadecimal node ID, found {}",
                 String::from_utf8_lossy(hex.as_ref())
             ))
         })
     }
 
     /// Provide access to binary data
     ///
     /// This is needed by FFI layers, for instance to return expected
     /// binary values to Python.
     pub fn as_bytes(&self) -> &[u8] {
         &self.data
     }
 
     pub fn short(&self) -> NodePrefix {
         NodePrefix {
             nybbles_len: SHORT_PREFIX_DEFAULT_NYBBLES_LENGTH,
             data: self.data,
         }
     }
 
     pub fn pad_to_256_bits(&self) -> [u8; 32] {
         let mut bits = [0; 32];
         bits[..NODE_BYTES_LENGTH].copy_from_slice(&self.data);
         bits
     }
 }
 
 /// The beginning of a binary revision SHA.
 ///
 /// Since it can potentially come from an hexadecimal representation with
 /// odd length, it needs to carry around whether the last 4 bits are relevant
 /// or not.
 #[derive(Debug, PartialEq, Copy, Clone)]
 pub struct NodePrefix {
     /// In `1..=NODE_NYBBLES_LENGTH`
     nybbles_len: u8,
     /// The first `4 * length_in_nybbles` bits are used (considering bits
     /// within a bytes in big-endian: most significant first), the rest
     /// are zero.
     data: NodeData,
 }
 
 impl NodePrefix {
     /// Convert from hexadecimal string representation
     ///
     /// Similarly to `hex::decode`, can be used with Unicode string types
     /// (`String`, `&str`) as well as bytes.
     ///
     /// To be used in FFI and I/O only, in order to facilitate future
     /// changes of hash format.
     pub fn from_hex(hex: impl AsRef<[u8]>) -> Result<Self, FromHexError> {
         let hex = hex.as_ref();
         let len = hex.len();
         if len > NODE_NYBBLES_LENGTH || len == 0 {
             return Err(FromHexError);
         }
 
         let mut data = [0; NODE_BYTES_LENGTH];
         let mut nybbles_len = 0;
         for &ascii_byte in hex {
             let nybble = match char::from(ascii_byte).to_digit(16) {
                 Some(digit) => digit as u8,
                 None => return Err(FromHexError),
             };
             // Fill in the upper half of a byte first, then the lower half.
             let shift = if nybbles_len % 2 == 0 { 4 } else { 0 };
             data[nybbles_len as usize / 2] |= nybble << shift;
             nybbles_len += 1;
         }
         Ok(Self { data, nybbles_len })
     }
 
     pub fn nybbles_len(&self) -> usize {
         self.nybbles_len as _
     }
 
     pub fn is_prefix_of(&self, node: &Node) -> bool {
         let full_bytes = self.nybbles_len() / 2;
         if self.data[..full_bytes] != node.data[..full_bytes] {
             return false;
         }
         if self.nybbles_len() % 2 == 0 {
             return true;
         }
         let last = self.nybbles_len() - 1;
         self.get_nybble(last) == node.get_nybble(last)
     }
 
     /// Retrieve the `i`th half-byte from the prefix.
     ///
     /// This is also the `i`th hexadecimal digit in numeric form,
     /// also called a [nybble](https://en.wikipedia.org/wiki/Nibble).
     pub fn get_nybble(&self, i: usize) -> u8 {
         assert!(i < self.nybbles_len());
         get_nybble(&self.data, i)
     }
 
     fn iter_nybbles(&self) -> impl Iterator<Item = u8> + '_ {
         (0..self.nybbles_len()).map(move |i| get_nybble(&self.data, i))
     }
 
     /// Return the index first nybble that's different from `node`
     ///
     /// If the return value is `None` that means that `self` is
     /// a prefix of `node`, but the current method is a bit slower
     /// than `is_prefix_of`.
     ///
     /// Returned index is as in `get_nybble`, i.e., starting at 0.
     pub fn first_different_nybble(&self, node: &Node) -> Option<usize> {
         self.iter_nybbles()
             .zip(NodePrefix::from(*node).iter_nybbles())
             .position(|(a, b)| a != b)
     }
 }
 
 impl fmt::LowerHex for NodePrefix {
     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
         let full_bytes = self.nybbles_len() / 2;
         for &byte in &self.data[..full_bytes] {
             write!(f, "{:02x}", byte)?
         }
         if self.nybbles_len() % 2 == 1 {
             let last = self.nybbles_len() - 1;
             write!(f, "{:x}", self.get_nybble(last))?
         }
         Ok(())
     }
 }
 
 /// A shortcut for full `Node` references
 impl From<&'_ Node> for NodePrefix {
     fn from(node: &'_ Node) -> Self {
         NodePrefix {
             nybbles_len: node.nybbles_len() as _,
             data: node.data,
         }
     }
 }
 
 /// A shortcut for full `Node` references
 impl From<Node> for NodePrefix {
     fn from(node: Node) -> Self {
         NodePrefix {
             nybbles_len: node.nybbles_len() as _,
             data: node.data,
         }
     }
 }
 
 impl PartialEq<Node> for NodePrefix {
     fn eq(&self, other: &Node) -> bool {
         self.data == other.data && self.nybbles_len() == other.nybbles_len()
     }
 }
 
 #[cfg(test)]
 mod tests {
     use super::*;
 
     const SAMPLE_NODE_HEX: &str = "0123456789abcdeffedcba9876543210deadbeef";
     const SAMPLE_NODE: Node = Node {
         data: [
             0x01, 0x23, 0x45, 0x67, 0x89, 0xab, 0xcd, 0xef, 0xfe, 0xdc, 0xba,
             0x98, 0x76, 0x54, 0x32, 0x10, 0xde, 0xad, 0xbe, 0xef,
         ],
     };
 
     /// Pad an hexadecimal string to reach `NODE_NYBBLES_LENGTH`
     /// The padding is made with zeros.
     pub fn hex_pad_right(hex: &str) -> String {
         let mut res = hex.to_string();
         while res.len() < NODE_NYBBLES_LENGTH {
             res.push('0');
         }
         res
     }
 
     #[test]
     fn test_node_from_hex() {
         let not_hex = "012... oops";
         let too_short = "0123";
         let too_long = format!("{}0", SAMPLE_NODE_HEX);
         assert_eq!(Node::from_hex(SAMPLE_NODE_HEX).unwrap(), SAMPLE_NODE);
         assert!(Node::from_hex(not_hex).is_err());
         assert!(Node::from_hex(too_short).is_err());
-        assert!(Node::from_hex(&too_long).is_err());
+        assert!(Node::from_hex(too_long).is_err());
     }
 
     #[test]
     fn test_node_encode_hex() {
         assert_eq!(format!("{:x}", SAMPLE_NODE), SAMPLE_NODE_HEX);
     }
 
     #[test]
     fn test_prefix_from_to_hex() -> Result<(), FromHexError> {
         assert_eq!(format!("{:x}", NodePrefix::from_hex("0e1")?), "0e1");
         assert_eq!(format!("{:x}", NodePrefix::from_hex("0e1a")?), "0e1a");
         assert_eq!(
             format!("{:x}", NodePrefix::from_hex(SAMPLE_NODE_HEX)?),
             SAMPLE_NODE_HEX
         );
         Ok(())
     }
 
     #[test]
     fn test_prefix_from_hex_errors() {
         assert!(NodePrefix::from_hex("testgr").is_err());
         let mut long = format!("{:x}", NULL_NODE);
         long.push('c');
         assert!(NodePrefix::from_hex(&long).is_err())
     }
 
     #[test]
     fn test_is_prefix_of() -> Result<(), FromHexError> {
         let mut node_data = [0; NODE_BYTES_LENGTH];
         node_data[0] = 0x12;
         node_data[1] = 0xca;
         let node = Node::from(node_data);
         assert!(NodePrefix::from_hex("12")?.is_prefix_of(&node));
         assert!(!NodePrefix::from_hex("1a")?.is_prefix_of(&node));
         assert!(NodePrefix::from_hex("12c")?.is_prefix_of(&node));
         assert!(!NodePrefix::from_hex("12d")?.is_prefix_of(&node));
         Ok(())
     }
 
     #[test]
     fn test_get_nybble() -> Result<(), FromHexError> {
         let prefix = NodePrefix::from_hex("dead6789cafe")?;
         assert_eq!(prefix.get_nybble(0), 13);
         assert_eq!(prefix.get_nybble(7), 9);
         Ok(())
     }
 
     #[test]
     fn test_first_different_nybble_even_prefix() {
         let prefix = NodePrefix::from_hex("12ca").unwrap();
         let mut node = Node::from([0; NODE_BYTES_LENGTH]);
         assert_eq!(prefix.first_different_nybble(&node), Some(0));
         node.data[0] = 0x13;
         assert_eq!(prefix.first_different_nybble(&node), Some(1));
         node.data[0] = 0x12;
         assert_eq!(prefix.first_different_nybble(&node), Some(2));
         node.data[1] = 0xca;
         // now it is a prefix
         assert_eq!(prefix.first_different_nybble(&node), None);
     }
 
     #[test]
     fn test_first_different_nybble_odd_prefix() {
         let prefix = NodePrefix::from_hex("12c").unwrap();
         let mut node = Node::from([0; NODE_BYTES_LENGTH]);
         assert_eq!(prefix.first_different_nybble(&node), Some(0));
         node.data[0] = 0x13;
         assert_eq!(prefix.first_different_nybble(&node), Some(1));
         node.data[0] = 0x12;
         assert_eq!(prefix.first_different_nybble(&node), Some(2));
         node.data[1] = 0xca;
         // now it is a prefix
         assert_eq!(prefix.first_different_nybble(&node), None);
     }
 }
 
 #[cfg(test)]
 pub use tests::hex_pad_right;

rust/hg-core/src/revlog/nodemap.rs

@@ -1,1102 +1,1108 b''
 // Copyright 2018-2020 Georges Racinet <[e-mail]>
 //           and Mercurial contributors
 //
 // This software may be used and distributed according to the terms of the
 // GNU General Public License version 2 or any later version.
 //! Indexing facilities for fast retrieval of `Revision` from `Node`
 //!
 //! This provides a variation on the 16-ary radix tree that is
 //! provided as "nodetree" in revlog.c, ready for append-only persistence
 //! on disk.
 //!
 //! Following existing implicit conventions, the "nodemap" terminology
 //! is used in a more abstract context.
 
 use crate::UncheckedRevision;
 
 use super::{
     node::NULL_NODE, Node, NodePrefix, Revision, RevlogIndex, NULL_REVISION,
 };
 
 use bytes_cast::{unaligned, BytesCast};
 use std::cmp::max;
 use std::fmt;
 use std::mem::{self, align_of, size_of};
 use std::ops::Deref;
 use std::ops::Index;
 
 #[derive(Debug, PartialEq)]
 pub enum NodeMapError {
     /// A `NodePrefix` matches several [`Revision`]s.
     ///
     /// This can be returned by methods meant for (at most) one match.
     MultipleResults,
     /// A `Revision` stored in the nodemap could not be found in the index
     RevisionNotInIndex(UncheckedRevision),
 }
 
 /// Mapping system from Mercurial nodes to revision numbers.
 ///
 /// ## `RevlogIndex` and `NodeMap`
 ///
 /// One way to think about their relationship is that
 /// the `NodeMap` is a prefix-oriented reverse index of the [`Node`]
 /// information carried by a [`RevlogIndex`].
 ///
 /// Many of the methods in this trait take a `RevlogIndex` argument
 /// which is used for validation of their results. This index must naturally
 /// be the one the `NodeMap` is about, and it must be consistent.
 ///
 /// Notably, the `NodeMap` must not store
 /// information about more `Revision` values than there are in the index.
 /// In these methods, an encountered `Revision` is not in the index, a
 /// [RevisionNotInIndex](NodeMapError) error is returned.
 ///
 /// In insert operations, the rule is thus that the `NodeMap` must always
 /// be updated after the `RevlogIndex` it is about.
 pub trait NodeMap {
     /// Find the unique `Revision` having the given `Node`
     ///
     /// If no Revision matches the given `Node`, `Ok(None)` is returned.
     fn find_node(
         &self,
         index: &impl RevlogIndex,
         node: &Node,
     ) -> Result<Option<Revision>, NodeMapError> {
         self.find_bin(index, node.into())
     }
 
     /// Find the unique Revision whose `Node` starts with a given binary prefix
     ///
     /// If no Revision matches the given prefix, `Ok(None)` is returned.
     ///
     /// If several Revisions match the given prefix, a
     /// [MultipleResults](NodeMapError)  error is returned.
     fn find_bin(
         &self,
         idx: &impl RevlogIndex,
         prefix: NodePrefix,
     ) -> Result<Option<Revision>, NodeMapError>;
 
     /// Give the size of the shortest node prefix that determines
     /// the revision uniquely.
     ///
     /// From a binary node prefix, if it is matched in the node map, this
     /// returns the number of hexadecimal digits that would had sufficed
     /// to find the revision uniquely.
     ///
     /// Returns `None` if no [`Revision`] could be found for the prefix.
     ///
     /// If several Revisions match the given prefix, a
     /// [MultipleResults](NodeMapError)  error is returned.
     fn unique_prefix_len_bin(
         &self,
         idx: &impl RevlogIndex,
         node_prefix: NodePrefix,
     ) -> Result<Option<usize>, NodeMapError>;
 
     /// Same as [unique_prefix_len_bin](Self::unique_prefix_len_bin), with
     /// a full [`Node`] as input
     fn unique_prefix_len_node(
         &self,
         idx: &impl RevlogIndex,
         node: &Node,
     ) -> Result<Option<usize>, NodeMapError> {
         self.unique_prefix_len_bin(idx, node.into())
     }
 }
 
 pub trait MutableNodeMap: NodeMap {
     fn insert<I: RevlogIndex>(
         &mut self,
         index: &I,
         node: &Node,
         rev: Revision,
     ) -> Result<(), NodeMapError>;
 }
 
 /// Low level NodeTree [`Block`] elements
 ///
 /// These are exactly as for instance on persistent storage.
 type RawElement = unaligned::I32Be;
 
 /// High level representation of values in NodeTree
 /// [`Blocks`](struct.Block.html)
 ///
 /// This is the high level representation that most algorithms should
 /// use.
 #[derive(Clone, Debug, Eq, PartialEq)]
 enum Element {
     // This is not a Mercurial revision. It's a `i32` because this is the
     // right type for this structure.
     Rev(i32),
     Block(usize),
     None,
 }
 
 impl From<RawElement> for Element {
     /// Conversion from low level representation, after endianness conversion.
     ///
     /// See [`Block`](struct.Block.html) for explanation about the encoding.
     fn from(raw: RawElement) -> Element {
         let int = raw.get();
         if int >= 0 {
             Element::Block(int as usize)
         } else if int == -1 {
             Element::None
         } else {
             Element::Rev(-int - 2)
         }
     }
 }
 
 impl From<Element> for RawElement {
     fn from(element: Element) -> RawElement {
         RawElement::from(match element {
             Element::None => 0,
             Element::Block(i) => i as i32,
             Element::Rev(rev) => -rev - 2,
         })
     }
 }
 
 const ELEMENTS_PER_BLOCK: usize = 16; // number of different values in a nybble
 
 /// A logical block of the [`NodeTree`], packed with a fixed size.
 ///
 /// These are always used in container types implementing `Index<Block>`,
 /// such as `&Block`
 ///
 /// As an array of integers, its ith element encodes that the
 /// ith potential edge from the block, representing the ith hexadecimal digit
 /// (nybble) `i` is either:
 ///
 /// - absent (value -1)
 /// - another `Block` in the same indexable container (value ≥ 0)
 ///  - a [`Revision`] leaf (value ≤ -2)
 ///
 /// Endianness has to be fixed for consistency on shared storage across
 /// different architectures.
 ///
 /// A key difference with the C `nodetree` is that we need to be
 /// able to represent the [`Block`] at index 0, hence -1 is the empty marker
 /// rather than 0 and the [`Revision`] range upper limit of -2 instead of -1.
 ///
 /// Another related difference is that `NULL_REVISION` (-1) is not
 /// represented at all, because we want an immutable empty nodetree
 /// to be valid.
 #[derive(Copy, Clone, BytesCast, PartialEq)]
 #[repr(transparent)]
 pub struct Block([RawElement; ELEMENTS_PER_BLOCK]);
 
 impl Block {
     fn new() -> Self {
         let absent_node = RawElement::from(-1);
         Block([absent_node; ELEMENTS_PER_BLOCK])
     }
 
     fn get(&self, nybble: u8) -> Element {
         self.0[nybble as usize].into()
     }
 
     fn set(&mut self, nybble: u8, element: Element) {
         self.0[nybble as usize] = element.into()
     }
 }
 
 impl fmt::Debug for Block {
     /// sparse representation for testing and debugging purposes
     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
         f.debug_map()
             .entries((0..16).filter_map(|i| match self.get(i) {
                 Element::None => None,
                 element => Some((i, element)),
             }))
             .finish()
     }
 }
 
 /// A mutable 16-radix tree with the root block logically at the end
 ///
 /// Because of the append only nature of our node trees, we need to
 /// keep the original untouched and store new blocks separately.
 ///
 /// The mutable root [`Block`] is kept apart so that we don't have to rebump
 /// it on each insertion.
 pub struct NodeTree {
     readonly: Box<dyn Deref<Target = [Block]> + Send>,
     growable: Vec<Block>,
     root: Block,
     masked_inner_blocks: usize,
 }
 
 impl Index<usize> for NodeTree {
     type Output = Block;
 
     fn index(&self, i: usize) -> &Block {
         let ro_len = self.readonly.len();
         if i < ro_len {
             &self.readonly[i]
         } else if i == ro_len + self.growable.len() {
             &self.root
         } else {
             &self.growable[i - ro_len]
         }
     }
 }
 
 /// Return `None` unless the [`Node`] for `rev` has given prefix in `idx`.
 fn has_prefix_or_none(
     idx: &impl RevlogIndex,
     prefix: NodePrefix,
     rev: UncheckedRevision,
 ) -> Result<Option<Revision>, NodeMapError> {
     match idx.check_revision(rev) {
         Some(checked) => idx
             .node(checked)
             .ok_or(NodeMapError::RevisionNotInIndex(rev))
             .map(|node| {
                 if prefix.is_prefix_of(node) {
                     Some(checked)
                 } else {
                     None
                 }
             }),
         None => Err(NodeMapError::RevisionNotInIndex(rev)),
     }
 }
 
 /// validate that the candidate's node starts indeed with given prefix,
 /// and treat ambiguities related to [`NULL_REVISION`].
 ///
 /// From the data in the NodeTree, one can only conclude that some
 /// revision is the only one for a *subprefix* of the one being looked up.
 fn validate_candidate(
     idx: &impl RevlogIndex,
     prefix: NodePrefix,
     candidate: (Option<UncheckedRevision>, usize),
 ) -> Result<(Option<Revision>, usize), NodeMapError> {
     let (rev, steps) = candidate;
     if let Some(nz_nybble) = prefix.first_different_nybble(&NULL_NODE) {
         rev.map_or(Ok((None, steps)), |r| {
             has_prefix_or_none(idx, prefix, r)
                 .map(|opt| (opt, max(steps, nz_nybble + 1)))
         })
     } else {
         // the prefix is only made of zeros; NULL_REVISION always matches it
         // and any other *valid* result is an ambiguity
         match rev {
             None => Ok((Some(NULL_REVISION), steps + 1)),
             Some(r) => match has_prefix_or_none(idx, prefix, r)? {
                 None => Ok((Some(NULL_REVISION), steps + 1)),
                 _ => Err(NodeMapError::MultipleResults),
             },
         }
     }
 }
 
 impl NodeTree {
     /// Initiate a NodeTree from an immutable slice-like of `Block`
     ///
     /// We keep `readonly` and clone its root block if it isn't empty.
     fn new(readonly: Box<dyn Deref<Target = [Block]> + Send>) -> Self {
         let root = readonly.last().cloned().unwrap_or_else(Block::new);
         NodeTree {
             readonly,
             growable: Vec::new(),
             root,
             masked_inner_blocks: 0,
         }
     }
 
     /// Create from an opaque bunch of bytes
     ///
     /// The created [`NodeTreeBytes`] from `bytes`,
     /// of which exactly `amount` bytes are used.
     ///
     /// - `buffer` could be derived from `PyBuffer` and `Mmap` objects.
     /// - `amount` is expressed in bytes, and is not automatically derived from
     ///   `bytes`, so that a caller that manages them atomically can perform
     ///   temporary disk serializations and still rollback easily if needed.
     ///   First use-case for this would be to support Mercurial shell hooks.
     ///
     /// panics if `buffer` is smaller than `amount`
     pub fn load_bytes(
         bytes: Box<dyn Deref<Target = [u8]> + Send>,
         amount: usize,
     ) -> Self {
         NodeTree::new(Box::new(NodeTreeBytes::new(bytes, amount)))
     }
 
     /// Retrieve added [`Block`]s and the original immutable data
     pub fn into_readonly_and_added(
         self,
     ) -> (Box<dyn Deref<Target = [Block]> + Send>, Vec<Block>) {
         let mut vec = self.growable;
         let readonly = self.readonly;
         if readonly.last() != Some(&self.root) {
             vec.push(self.root);
         }
         (readonly, vec)
     }
 
     /// Retrieve added [`Block]s as bytes, ready to be written to persistent
     /// storage
     pub fn into_readonly_and_added_bytes(
         self,
     ) -> (Box<dyn Deref<Target = [Block]> + Send>, Vec<u8>) {
         let (readonly, vec) = self.into_readonly_and_added();
         // Prevent running `v`'s destructor so we are in complete control
         // of the allocation.
         let vec = mem::ManuallyDrop::new(vec);
 
         // Transmute the `Vec<Block>` to a `Vec<u8>`. Blocks are contiguous
         // bytes, so this is perfectly safe.
         let bytes = unsafe {
             // Check for compatible allocation layout.
             // (Optimized away by constant-folding + dead code elimination.)
             assert_eq!(size_of::<Block>(), 64);
             assert_eq!(align_of::<Block>(), 1);
 
             // /!\ Any use of `vec` after this is use-after-free.
             // TODO: use `into_raw_parts` once stabilized
             Vec::from_raw_parts(
                 vec.as_ptr() as *mut u8,
                 vec.len() * size_of::<Block>(),
                 vec.capacity() * size_of::<Block>(),
             )
         };
         (readonly, bytes)
     }
 
     /// Total number of blocks
     fn len(&self) -> usize {
         self.readonly.len() + self.growable.len() + 1
     }
 
     /// Implemented for completeness
     ///
     /// A `NodeTree` always has at least the mutable root block.
     #[allow(dead_code)]
     fn is_empty(&self) -> bool {
         false
     }
 
     /// Main working method for `NodeTree` searches
     ///
     /// The first returned value is the result of analysing `NodeTree` data
     /// *alone*: whereas `None` guarantees that the given prefix is absent
     /// from the [`NodeTree`] data (but still could match [`NULL_NODE`]), with
     /// `Some(rev)`, it is to be understood that `rev` is the unique
     /// [`Revision`] that could match the prefix. Actually, all that can
     /// be inferred from
     /// the `NodeTree` data is that `rev` is the revision with the longest
     /// common node prefix with the given prefix.
     /// We return an [`UncheckedRevision`] because we have no guarantee that
     /// the revision we found is valid for the index.
     ///
     /// The second returned value is the size of the smallest subprefix
     /// of `prefix` that would give the same result, i.e. not the
     /// [MultipleResults](NodeMapError) error variant (again, using only the
     /// data of the [`NodeTree`]).
     fn lookup(
         &self,
         prefix: NodePrefix,
     ) -> Result<(Option<UncheckedRevision>, usize), NodeMapError> {
         for (i, visit_item) in self.visit(prefix).enumerate() {
             if let Some(opt) = visit_item.final_revision() {
                 return Ok((opt, i + 1));
             }
         }
         Err(NodeMapError::MultipleResults)
     }
 
     fn visit(&self, prefix: NodePrefix) -> NodeTreeVisitor {
         NodeTreeVisitor {
             nt: self,
             prefix,
             visit: self.len() - 1,
             nybble_idx: 0,
             done: false,
         }
     }
     /// Return a mutable reference for `Block` at index `idx`.
     ///
     /// If `idx` lies in the immutable area, then the reference is to
     /// a newly appended copy.
     ///
     /// Returns (new_idx, glen, mut_ref) where
     ///
     /// - `new_idx` is the index of the mutable `Block`
     /// - `mut_ref` is a mutable reference to the mutable Block.
     /// - `glen` is the new length of `self.growable`
     ///
     /// Note: the caller wouldn't be allowed to query `self.growable.len()`
     /// itself because of the mutable borrow taken with the returned `Block`
     fn mutable_block(&mut self, idx: usize) -> (usize, &mut Block, usize) {
         let ro_blocks = &self.readonly;
         let ro_len = ro_blocks.len();
         let glen = self.growable.len();
         if idx < ro_len {
             self.masked_inner_blocks += 1;
             self.growable.push(ro_blocks[idx]);
             (glen + ro_len, &mut self.growable[glen], glen + 1)
         } else if glen + ro_len == idx {
             (idx, &mut self.root, glen)
         } else {
             (idx, &mut self.growable[idx - ro_len], glen)
         }
     }
 
     /// Main insertion method
     ///
     /// This will dive in the node tree to find the deepest `Block` for
     /// `node`, split it as much as needed and record `node` in there.
     /// The method then backtracks, updating references in all the visited
     /// blocks from the root.
     ///
     /// All the mutated `Block` are copied first to the growable part if
     /// needed. That happens for those in the immutable part except the root.
     pub fn insert<I: RevlogIndex>(
         &mut self,
         index: &I,
         node: &Node,
         rev: Revision,
     ) -> Result<(), NodeMapError> {
         let ro_len = &self.readonly.len();
 
         let mut visit_steps: Vec<_> = self.visit(node.into()).collect();
         let read_nybbles = visit_steps.len();
         // visit_steps cannot be empty, since we always visit the root block
         let deepest = visit_steps.pop().unwrap();
 
         let (mut block_idx, mut block, mut glen) =
             self.mutable_block(deepest.block_idx);
 
         if let Element::Rev(old_rev) = deepest.element {
             let old_node = index
                 .check_revision(old_rev.into())
                 .and_then(|rev| index.node(rev))
                 .ok_or_else(|| {
                     NodeMapError::RevisionNotInIndex(old_rev.into())
                 })?;
             if old_node == node {
                 return Ok(()); // avoid creating lots of useless blocks
             }
 
             // Looping over the tail of nybbles in both nodes, creating
             // new blocks until we find the difference
             let mut new_block_idx = ro_len + glen;
             let mut nybble = deepest.nybble;
             for nybble_pos in read_nybbles..node.nybbles_len() {
                 block.set(nybble, Element::Block(new_block_idx));
 
                 let new_nybble = node.get_nybble(nybble_pos);
                 let old_nybble = old_node.get_nybble(nybble_pos);
 
                 if old_nybble == new_nybble {
                     self.growable.push(Block::new());
                     block = &mut self.growable[glen];
                     glen += 1;
                     new_block_idx += 1;
                     nybble = new_nybble;
                 } else {
                     let mut new_block = Block::new();
                     new_block.set(old_nybble, Element::Rev(old_rev));
                     new_block.set(new_nybble, Element::Rev(rev.0));
                     self.growable.push(new_block);
                     break;
                 }
             }
         } else {
             // Free slot in the deepest block: no splitting has to be done
             block.set(deepest.nybble, Element::Rev(rev.0));
         }
 
         // Backtrack over visit steps to update references
         while let Some(visited) = visit_steps.pop() {
             let to_write = Element::Block(block_idx);
             if visit_steps.is_empty() {
                 self.root.set(visited.nybble, to_write);
                 break;
             }
             let (new_idx, block, _) = self.mutable_block(visited.block_idx);
             if block.get(visited.nybble) == to_write {
                 break;
             }
             block.set(visited.nybble, to_write);
             block_idx = new_idx;
         }
         Ok(())
     }
 
     /// Make the whole `NodeTree` logically empty, without touching the
     /// immutable part.
     pub fn invalidate_all(&mut self) {
         self.root = Block::new();
         self.growable = Vec::new();
         self.masked_inner_blocks = self.readonly.len();
     }
 
     /// Return the number of blocks in the readonly part that are currently
     /// masked in the mutable part.
     ///
     /// The `NodeTree` structure has no efficient way to know how many blocks
     /// are already unreachable in the readonly part.
     ///
     /// After a call to `invalidate_all()`, the returned number can be actually
     /// bigger than the whole readonly part, a conventional way to mean that
     /// all the readonly blocks have been masked. This is what is really
     /// useful to the caller and does not require to know how many were
     /// actually unreachable to begin with.
     pub fn masked_readonly_blocks(&self) -> usize {
         if let Some(readonly_root) = self.readonly.last() {
             if readonly_root == &self.root {
                 return 0;
             }
         } else {
             return 0;
         }
         self.masked_inner_blocks + 1
     }
 }
 
 pub struct NodeTreeBytes {
     buffer: Box<dyn Deref<Target = [u8]> + Send>,
     len_in_blocks: usize,
 }
 
 impl NodeTreeBytes {
     fn new(
         buffer: Box<dyn Deref<Target = [u8]> + Send>,
         amount: usize,
     ) -> Self {
         assert!(buffer.len() >= amount);
         let len_in_blocks = amount / size_of::<Block>();
         NodeTreeBytes {
             buffer,
             len_in_blocks,
         }
     }
 }
 
 impl Deref for NodeTreeBytes {
     type Target = [Block];
 
     fn deref(&self) -> &[Block] {
         Block::slice_from_bytes(&self.buffer, self.len_in_blocks)
             // `NodeTreeBytes::new` already asserted that `self.buffer` is
             // large enough.
             .unwrap()
             .0
     }
 }
 
 struct NodeTreeVisitor<'n> {
     nt: &'n NodeTree,
     prefix: NodePrefix,
     visit: usize,
     nybble_idx: usize,
     done: bool,
 }
 
 #[derive(Debug, PartialEq, Clone)]
 struct NodeTreeVisitItem {
     block_idx: usize,
     nybble: u8,
     element: Element,
 }
 
 impl<'n> Iterator for NodeTreeVisitor<'n> {
     type Item = NodeTreeVisitItem;
 
     fn next(&mut self) -> Option<Self::Item> {
         if self.done || self.nybble_idx >= self.prefix.nybbles_len() {
             return None;
         }
 
         let nybble = self.prefix.get_nybble(self.nybble_idx);
         self.nybble_idx += 1;
 
         let visit = self.visit;
         let element = self.nt[visit].get(nybble);
         if let Element::Block(idx) = element {
             self.visit = idx;
         } else {
             self.done = true;
         }
 
         Some(NodeTreeVisitItem {
             block_idx: visit,
             nybble,
             element,
         })
     }
 }
 
 impl NodeTreeVisitItem {
     // Return `Some(opt)` if this item is final, with `opt` being the
     // `UncheckedRevision` that it may represent.
     //
     // If the item is not terminal, return `None`
     fn final_revision(&self) -> Option<Option<UncheckedRevision>> {
         match self.element {
             Element::Block(_) => None,
             Element::Rev(r) => Some(Some(r.into())),
             Element::None => Some(None),
         }
     }
 }
 
 impl From<Vec<Block>> for NodeTree {
     fn from(vec: Vec<Block>) -> Self {
         Self::new(Box::new(vec))
     }
 }
 
 impl fmt::Debug for NodeTree {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-        let readonly: &[Block] = &*self.readonly;
+        let readonly: &[Block] = &self.readonly;
         write!(
             f,
             "readonly: {:?}, growable: {:?}, root: {:?}",
             readonly, self.growable, self.root
         )
     }
 }
 
 impl Default for NodeTree {
     /// Create a fully mutable empty NodeTree
     fn default() -> Self {
-        NodeTree::new(Box::new(Vec::new()))
+        NodeTree::new(Box::<Vec<_>>::default())
     }
 }
 
 impl NodeMap for NodeTree {
     fn find_bin<'a>(
         &self,
         idx: &impl RevlogIndex,
         prefix: NodePrefix,
     ) -> Result<Option<Revision>, NodeMapError> {
         validate_candidate(idx, prefix, self.lookup(prefix)?)
             .map(|(opt, _shortest)| opt)
     }
 
     fn unique_prefix_len_bin<'a>(
         &self,
         idx: &impl RevlogIndex,
         prefix: NodePrefix,
     ) -> Result<Option<usize>, NodeMapError> {
         validate_candidate(idx, prefix, self.lookup(prefix)?)
             .map(|(opt, shortest)| opt.map(|_rev| shortest))
     }
 }
 
 #[cfg(test)]
 pub mod tests {
     use super::NodeMapError::*;
     use super::*;
     use crate::revlog::node::{hex_pad_right, Node};
     use std::collections::HashMap;
 
     /// Creates a `Block` using a syntax close to the `Debug` output
     macro_rules! block {
         {$($nybble:tt : $variant:ident($val:tt)),*} => (
             {
                 let mut block = Block::new();
                 $(block.set($nybble, Element::$variant($val)));*;
                 block
             }
         )
     }
 
     /// Shorthand to reduce boilerplate when creating [`Revision`] for testing
     macro_rules! R {
         ($revision:literal) => {
             Revision($revision)
         };
     }
 
     #[test]
     fn test_block_debug() {
         let mut block = Block::new();
         block.set(1, Element::Rev(3));
         block.set(10, Element::Block(0));
         assert_eq!(format!("{:?}", block), "{1: Rev(3), 10: Block(0)}");
     }
 
     #[test]
     fn test_block_macro() {
         let block = block! {5: Block(2)};
         assert_eq!(format!("{:?}", block), "{5: Block(2)}");
 
         let block = block! {13: Rev(15), 5: Block(2)};
         assert_eq!(format!("{:?}", block), "{5: Block(2), 13: Rev(15)}");
     }
 
     #[test]
     fn test_raw_block() {
         let mut raw = [255u8; 64];
 
         let mut counter = 0;
         for val in [0_i32, 15, -2, -1, -3].iter() {
             for byte in val.to_be_bytes().iter() {
                 raw[counter] = *byte;
                 counter += 1;
             }
         }
         let (block, _) = Block::from_bytes(&raw).unwrap();
         assert_eq!(block.get(0), Element::Block(0));
         assert_eq!(block.get(1), Element::Block(15));
         assert_eq!(block.get(3), Element::None);
         assert_eq!(block.get(2), Element::Rev(0));
         assert_eq!(block.get(4), Element::Rev(1));
     }
 
     type TestIndex = HashMap<UncheckedRevision, Node>;
 
     impl RevlogIndex for TestIndex {
         fn node(&self, rev: Revision) -> Option<&Node> {
             self.get(&rev.into())
         }
 
         fn len(&self) -> usize {
             self.len()
         }
 
         fn check_revision(&self, rev: UncheckedRevision) -> Option<Revision> {
             self.get(&rev).map(|_| Revision(rev.0))
         }
     }
 
     /// Pad hexadecimal Node prefix with zeros on the right
     ///
     /// This avoids having to repeatedly write very long hexadecimal
     /// strings for test data, and brings actual hash size independency.
     #[cfg(test)]
     fn pad_node(hex: &str) -> Node {
-        Node::from_hex(&hex_pad_right(hex)).unwrap()
+        Node::from_hex(hex_pad_right(hex)).unwrap()
     }
 
     /// Pad hexadecimal Node prefix with zeros on the right, then insert
     fn pad_insert(idx: &mut TestIndex, rev: Revision, hex: &str) {
         idx.insert(rev.into(), pad_node(hex));
     }
 
     fn sample_nodetree() -> NodeTree {
         NodeTree::from(vec![
             block![0: Rev(9)],
             block![0: Rev(0), 1: Rev(9)],
             block![0: Block(1), 1:Rev(1)],
         ])
     }
 
     fn hex(s: &str) -> NodePrefix {
         NodePrefix::from_hex(s).unwrap()
     }
 
     #[test]
     fn test_nt_debug() {
         let nt = sample_nodetree();
         assert_eq!(
             format!("{:?}", nt),
             "readonly: \
              [{0: Rev(9)}, {0: Rev(0), 1: Rev(9)}, {0: Block(1), 1: Rev(1)}], \
              growable: [], \
              root: {0: Block(1), 1: Rev(1)}",
         );
     }
 
     #[test]
     fn test_immutable_find_simplest() -> Result<(), NodeMapError> {
         let mut idx: TestIndex = HashMap::new();
         pad_insert(&mut idx, R!(1), "1234deadcafe");
 
         let nt = NodeTree::from(vec![block! {1: Rev(1)}]);
         assert_eq!(nt.find_bin(&idx, hex("1"))?, Some(R!(1)));
         assert_eq!(nt.find_bin(&idx, hex("12"))?, Some(R!(1)));
         assert_eq!(nt.find_bin(&idx, hex("1234de"))?, Some(R!(1)));
         assert_eq!(nt.find_bin(&idx, hex("1a"))?, None);
         assert_eq!(nt.find_bin(&idx, hex("ab"))?, None);
 
         // and with full binary Nodes
         assert_eq!(
             nt.find_node(&idx, idx.get(&1.into()).unwrap())?,
             Some(R!(1))
         );
-        let unknown = Node::from_hex(&hex_pad_right("3d")).unwrap();
+        let unknown = Node::from_hex(hex_pad_right("3d")).unwrap();
         assert_eq!(nt.find_node(&idx, &unknown)?, None);
         Ok(())
     }
 
     #[test]
     fn test_immutable_find_one_jump() {
         let mut idx = TestIndex::new();
         pad_insert(&mut idx, R!(9), "012");
         pad_insert(&mut idx, R!(0), "00a");
 
         let nt = sample_nodetree();
 
         assert_eq!(nt.find_bin(&idx, hex("0")), Err(MultipleResults));
         assert_eq!(nt.find_bin(&idx, hex("01")), Ok(Some(R!(9))));
         assert_eq!(nt.find_bin(&idx, hex("00")), Err(MultipleResults));
         assert_eq!(nt.find_bin(&idx, hex("00a")), Ok(Some(R!(0))));
         assert_eq!(nt.unique_prefix_len_bin(&idx, hex("00a")), Ok(Some(3)));
         assert_eq!(nt.find_bin(&idx, hex("000")), Ok(Some(NULL_REVISION)));
     }
 
     #[test]
     fn test_mutated_find() -> Result<(), NodeMapError> {
         let mut idx = TestIndex::new();
         pad_insert(&mut idx, R!(9), "012");
         pad_insert(&mut idx, R!(0), "00a");
         pad_insert(&mut idx, R!(2), "cafe");
         pad_insert(&mut idx, R!(3), "15");
         pad_insert(&mut idx, R!(1), "10");
 
         let nt = NodeTree {
             readonly: sample_nodetree().readonly,
             growable: vec![block![0: Rev(1), 5: Rev(3)]],
             root: block![0: Block(1), 1:Block(3), 12: Rev(2)],
             masked_inner_blocks: 1,
         };
         assert_eq!(nt.find_bin(&idx, hex("10"))?, Some(R!(1)));
         assert_eq!(nt.find_bin(&idx, hex("c"))?, Some(R!(2)));
         assert_eq!(nt.unique_prefix_len_bin(&idx, hex("c"))?, Some(1));
         assert_eq!(nt.find_bin(&idx, hex("00")), Err(MultipleResults));
         assert_eq!(nt.find_bin(&idx, hex("000"))?, Some(NULL_REVISION));
         assert_eq!(nt.unique_prefix_len_bin(&idx, hex("000"))?, Some(3));
         assert_eq!(nt.find_bin(&idx, hex("01"))?, Some(R!(9)));
         assert_eq!(nt.masked_readonly_blocks(), 2);
         Ok(())
     }
 
     pub struct TestNtIndex {
         pub index: TestIndex,
         pub nt: NodeTree,
     }
 
     impl TestNtIndex {
         pub fn new() -> Self {
             TestNtIndex {
                 index: HashMap::new(),
                 nt: NodeTree::default(),
             }
         }
 
         pub fn insert_node(
             &mut self,
             rev: Revision,
             node: Node,
         ) -> Result<(), NodeMapError> {
             self.index.insert(rev.into(), node);
             self.nt.insert(&self.index, &node, rev)?;
             Ok(())
         }
 
         pub fn insert(
             &mut self,
             rev: Revision,
             hex: &str,
         ) -> Result<(), NodeMapError> {
             let node = pad_node(hex);
-            return self.insert_node(rev, node);
+            self.insert_node(rev, node)
         }
 
         fn find_hex(
             &self,
             prefix: &str,
         ) -> Result<Option<Revision>, NodeMapError> {
             self.nt.find_bin(&self.index, hex(prefix))
         }
 
         fn unique_prefix_len_hex(
             &self,
             prefix: &str,
         ) -> Result<Option<usize>, NodeMapError> {
             self.nt.unique_prefix_len_bin(&self.index, hex(prefix))
         }
 
         /// Drain `added` and restart a new one
         fn commit(self) -> Self {
             let mut as_vec: Vec<Block> =
                 self.nt.readonly.iter().copied().collect();
             as_vec.extend(self.nt.growable);
             as_vec.push(self.nt.root);
 
             Self {
                 index: self.index,
                 nt: NodeTree::from(as_vec),
             }
         }
     }
 
+    impl Default for TestNtIndex {
+        fn default() -> Self {
+            Self::new()
+        }
+    }
+
     #[test]
     fn test_insert_full_mutable() -> Result<(), NodeMapError> {
         let mut idx = TestNtIndex::new();
         idx.insert(Revision(0), "1234")?;
         assert_eq!(idx.find_hex("1")?, Some(R!(0)));
         assert_eq!(idx.find_hex("12")?, Some(R!(0)));
 
         // let's trigger a simple split
         idx.insert(Revision(1), "1a34")?;
         assert_eq!(idx.nt.growable.len(), 1);
         assert_eq!(idx.find_hex("12")?, Some(R!(0)));
         assert_eq!(idx.find_hex("1a")?, Some(R!(1)));
 
         // reinserting is a no_op
         idx.insert(Revision(1), "1a34")?;
         assert_eq!(idx.nt.growable.len(), 1);
         assert_eq!(idx.find_hex("12")?, Some(R!(0)));
         assert_eq!(idx.find_hex("1a")?, Some(R!(1)));
 
         idx.insert(Revision(2), "1a01")?;
         assert_eq!(idx.nt.growable.len(), 2);
         assert_eq!(idx.find_hex("1a"), Err(NodeMapError::MultipleResults));
         assert_eq!(idx.find_hex("12")?, Some(R!(0)));
         assert_eq!(idx.find_hex("1a3")?, Some(R!(1)));
         assert_eq!(idx.find_hex("1a0")?, Some(R!(2)));
         assert_eq!(idx.find_hex("1a12")?, None);
 
         // now let's make it split and create more than one additional block
         idx.insert(Revision(3), "1a345")?;
         assert_eq!(idx.nt.growable.len(), 4);
         assert_eq!(idx.find_hex("1a340")?, Some(R!(1)));
         assert_eq!(idx.find_hex("1a345")?, Some(R!(3)));
         assert_eq!(idx.find_hex("1a341")?, None);
 
         // there's no readonly block to mask
         assert_eq!(idx.nt.masked_readonly_blocks(), 0);
         Ok(())
     }
 
     #[test]
     fn test_unique_prefix_len_zero_prefix() {
         let mut idx = TestNtIndex::new();
         idx.insert(Revision(0), "00000abcd").unwrap();
 
         assert_eq!(idx.find_hex("000"), Err(NodeMapError::MultipleResults));
         // in the nodetree proper, this will be found at the first nybble
         // yet the correct answer for unique_prefix_len is not 1, nor 1+1,
         // but the first difference with `NULL_NODE`
         assert_eq!(idx.unique_prefix_len_hex("00000a"), Ok(Some(6)));
         assert_eq!(idx.unique_prefix_len_hex("00000ab"), Ok(Some(6)));
 
         // same with odd result
         idx.insert(Revision(1), "00123").unwrap();
         assert_eq!(idx.unique_prefix_len_hex("001"), Ok(Some(3)));
         assert_eq!(idx.unique_prefix_len_hex("0012"), Ok(Some(3)));
 
         // these are unchanged of course
         assert_eq!(idx.unique_prefix_len_hex("00000a"), Ok(Some(6)));
         assert_eq!(idx.unique_prefix_len_hex("00000ab"), Ok(Some(6)));
     }
 
     #[test]
     fn test_insert_extreme_splitting() -> Result<(), NodeMapError> {
         // check that the splitting loop is long enough
         let mut nt_idx = TestNtIndex::new();
         let nt = &mut nt_idx.nt;
         let idx = &mut nt_idx.index;
 
         let node0_hex = hex_pad_right("444444");
         let mut node1_hex = hex_pad_right("444444");
         node1_hex.pop();
         node1_hex.push('5');
-        let node0 = Node::from_hex(&node0_hex).unwrap();
+        let node0 = Node::from_hex(node0_hex).unwrap();
         let node1 = Node::from_hex(&node1_hex).unwrap();
 
         idx.insert(0.into(), node0);
         nt.insert(idx, &node0, R!(0))?;
         idx.insert(1.into(), node1);
         nt.insert(idx, &node1, R!(1))?;
 
         assert_eq!(nt.find_bin(idx, (&node0).into())?, Some(R!(0)));
         assert_eq!(nt.find_bin(idx, (&node1).into())?, Some(R!(1)));
         Ok(())
     }
 
     #[test]
     fn test_insert_partly_immutable() -> Result<(), NodeMapError> {
         let mut idx = TestNtIndex::new();
         idx.insert(Revision(0), "1234")?;
         idx.insert(Revision(1), "1235")?;
         idx.insert(Revision(2), "131")?;
         idx.insert(Revision(3), "cafe")?;
         let mut idx = idx.commit();
         assert_eq!(idx.find_hex("1234")?, Some(R!(0)));
         assert_eq!(idx.find_hex("1235")?, Some(R!(1)));
         assert_eq!(idx.find_hex("131")?, Some(R!(2)));
         assert_eq!(idx.find_hex("cafe")?, Some(R!(3)));
         // we did not add anything since init from readonly
         assert_eq!(idx.nt.masked_readonly_blocks(), 0);
 
         idx.insert(Revision(4), "123A")?;
         assert_eq!(idx.find_hex("1234")?, Some(R!(0)));
         assert_eq!(idx.find_hex("1235")?, Some(R!(1)));
         assert_eq!(idx.find_hex("131")?, Some(R!(2)));
         assert_eq!(idx.find_hex("cafe")?, Some(R!(3)));
         assert_eq!(idx.find_hex("123A")?, Some(R!(4)));
         // we masked blocks for all prefixes of "123", including the root
         assert_eq!(idx.nt.masked_readonly_blocks(), 4);
 
         eprintln!("{:?}", idx.nt);
         idx.insert(Revision(5), "c0")?;
         assert_eq!(idx.find_hex("cafe")?, Some(R!(3)));
         assert_eq!(idx.find_hex("c0")?, Some(R!(5)));
         assert_eq!(idx.find_hex("c1")?, None);
         assert_eq!(idx.find_hex("1234")?, Some(R!(0)));
         // inserting "c0" is just splitting the 'c' slot of the mutable root,
         // it doesn't mask anything
         assert_eq!(idx.nt.masked_readonly_blocks(), 4);
 
         Ok(())
     }
 
     #[test]
     fn test_invalidate_all() -> Result<(), NodeMapError> {
         let mut idx = TestNtIndex::new();
         idx.insert(Revision(0), "1234")?;
         idx.insert(Revision(1), "1235")?;
         idx.insert(Revision(2), "131")?;
         idx.insert(Revision(3), "cafe")?;
         let mut idx = idx.commit();
 
         idx.nt.invalidate_all();
 
         assert_eq!(idx.find_hex("1234")?, None);
         assert_eq!(idx.find_hex("1235")?, None);
         assert_eq!(idx.find_hex("131")?, None);
         assert_eq!(idx.find_hex("cafe")?, None);
         // all the readonly blocks have been masked, this is the
         // conventional expected response
         assert_eq!(idx.nt.masked_readonly_blocks(), idx.nt.readonly.len() + 1);
         Ok(())
     }
 
     #[test]
     fn test_into_added_empty() {
         assert!(sample_nodetree().into_readonly_and_added().1.is_empty());
         assert!(sample_nodetree()
             .into_readonly_and_added_bytes()
             .1
             .is_empty());
     }
 
     #[test]
     fn test_into_added_bytes() -> Result<(), NodeMapError> {
         let mut idx = TestNtIndex::new();
         idx.insert(Revision(0), "1234")?;
         let mut idx = idx.commit();
         idx.insert(Revision(4), "cafe")?;
         let (_, bytes) = idx.nt.into_readonly_and_added_bytes();
 
         // only the root block has been changed
         assert_eq!(bytes.len(), size_of::<Block>());
         // big endian for -2
         assert_eq!(&bytes[4..2 * 4], [255, 255, 255, 254]);
         // big endian for -6
         assert_eq!(&bytes[12 * 4..13 * 4], [255, 255, 255, 250]);
         Ok(())
     }
 }

rust/hg-core/src/revlog/nodemap_docket.rs

@@ -1,108 +1,108 b''
 use crate::errors::{HgError, HgResultExt};
 use bytes_cast::{unaligned, BytesCast};
 use memmap2::Mmap;
 use std::path::{Path, PathBuf};
 
 use crate::vfs::Vfs;
 
 const ONDISK_VERSION: u8 = 1;
 
 pub(super) struct NodeMapDocket {
     pub data_length: usize,
     // TODO: keep here more of the data from `parse()` when we need it
 }
 
 #[derive(BytesCast)]
 #[repr(C)]
 struct DocketHeader {
     uid_size: u8,
     _tip_rev: unaligned::U64Be,
     data_length: unaligned::U64Be,
     _data_unused: unaligned::U64Be,
     tip_node_size: unaligned::U64Be,
 }
 
 impl NodeMapDocket {
     /// Return `Ok(None)` when the caller should proceed without a persistent
     /// nodemap:
     ///
     /// * This revlog does not have a `.n` docket file (it is not generated for
     ///   small revlogs), or
     /// * The docket has an unsupported version number (repositories created by
     ///   later hg, maybe that should be a requirement instead?), or
     /// * The docket file points to a missing (likely deleted) data file (this
     ///   can happen in a rare race condition).
     pub fn read_from_file(
         store_vfs: &Vfs,
         index_path: &Path,
     ) -> Result<Option<(Self, Mmap)>, HgError> {
         let docket_path = index_path.with_extension("n");
         let docket_bytes = if let Some(bytes) =
             store_vfs.read(&docket_path).io_not_found_as_none()?
         {
             bytes
         } else {
             return Ok(None);
         };
 
         let input = if let Some((&ONDISK_VERSION, rest)) =
             docket_bytes.split_first()
         {
             rest
         } else {
             return Ok(None);
         };
 
         /// Treat any error as a parse error
         fn parse<T, E>(result: Result<T, E>) -> Result<T, HgError> {
             result
                 .map_err(|_| HgError::corrupted("nodemap docket parse error"))
         }
 
         let (header, rest) = parse(DocketHeader::from_bytes(input))?;
         let uid_size = header.uid_size as usize;
         // TODO: do we care about overflow for 4 GB+ nodemap files on 32-bit
         // systems?
         let tip_node_size = header.tip_node_size.get() as usize;
         let data_length = header.data_length.get() as usize;
         let (uid, rest) = parse(u8::slice_from_bytes(rest, uid_size))?;
         let (_tip_node, _rest) =
             parse(u8::slice_from_bytes(rest, tip_node_size))?;
         let uid = parse(std::str::from_utf8(uid))?;
         let docket = NodeMapDocket { data_length };
 
         let data_path = rawdata_path(&docket_path, uid);
         // TODO: use `vfs.read()` here when the `persistent-nodemap.mmap`
         // config is false?
         if let Some(mmap) =
-            store_vfs.mmap_open(&data_path).io_not_found_as_none()?
+            store_vfs.mmap_open(data_path).io_not_found_as_none()?
         {
             if mmap.len() >= data_length {
                 Ok(Some((docket, mmap)))
             } else {
                 Err(HgError::corrupted("persistent nodemap too short"))
             }
         } else {
             // Even if .hg/requires opted in, some revlogs are deemed small
             // enough to not need a persistent nodemap.
             Ok(None)
         }
     }
 }
 
 fn rawdata_path(docket_path: &Path, uid: &str) -> PathBuf {
     let docket_name = docket_path
         .file_name()
         .expect("expected a base name")
         .to_str()
         .expect("expected an ASCII file name in the store");
     let prefix = docket_name
         .strip_suffix(".n.a")
         .or_else(|| docket_name.strip_suffix(".n"))
         .expect("expected docket path in .n or .n.a");
     let name = format!("{}-{}.nd", prefix, uid);
     docket_path
         .parent()
         .expect("expected a non-root path")
         .join(name)
 }

rust/hg-core/src/utils/files.rs

@@ -1,436 +1,436 b''
 // files.rs
 //
 // Copyright 2019
 // Raphaël Gomès <[e-mail]>,
 // Yuya Nishihara <[e-mail]>
 //
 // This software may be used and distributed according to the terms of the
 // GNU General Public License version 2 or any later version.
 
 //! Functions for fiddling with files.
 
 use crate::utils::{
     hg_path::{path_to_hg_path_buf, HgPath, HgPathBuf, HgPathError},
     path_auditor::PathAuditor,
     replace_slice,
 };
 use lazy_static::lazy_static;
 use same_file::is_same_file;
 use std::borrow::{Cow, ToOwned};
 use std::ffi::{OsStr, OsString};
 use std::iter::FusedIterator;
 use std::ops::Deref;
 use std::path::{Path, PathBuf};
 
 pub fn get_os_str_from_bytes(bytes: &[u8]) -> &OsStr {
     let os_str;
     #[cfg(unix)]
     {
         use std::os::unix::ffi::OsStrExt;
         os_str = std::ffi::OsStr::from_bytes(bytes);
     }
     // TODO Handle other platforms
     // TODO: convert from WTF8 to Windows MBCS (ANSI encoding).
     // Perhaps, the return type would have to be Result<PathBuf>.
     os_str
 }
 
 pub fn get_path_from_bytes(bytes: &[u8]) -> &Path {
     Path::new(get_os_str_from_bytes(bytes))
 }
 
 // TODO: need to convert from WTF8 to MBCS bytes on Windows.
 // that's why Vec<u8> is returned.
 #[cfg(unix)]
 pub fn get_bytes_from_path(path: impl AsRef<Path>) -> Vec<u8> {
     get_bytes_from_os_str(path.as_ref())
 }
 
 #[cfg(unix)]
 pub fn get_bytes_from_os_str(str: impl AsRef<OsStr>) -> Vec<u8> {
     use std::os::unix::ffi::OsStrExt;
     str.as_ref().as_bytes().to_vec()
 }
 
 #[cfg(unix)]
 pub fn get_bytes_from_os_string(str: OsString) -> Vec<u8> {
     use std::os::unix::ffi::OsStringExt;
     str.into_vec()
 }
 
 /// An iterator over repository path yielding itself and its ancestors.
 #[derive(Copy, Clone, Debug)]
 pub struct Ancestors<'a> {
     next: Option<&'a HgPath>,
 }
 
 impl<'a> Iterator for Ancestors<'a> {
     type Item = &'a HgPath;
 
     fn next(&mut self) -> Option<Self::Item> {
         let next = self.next;
         self.next = match self.next {
             Some(s) if s.is_empty() => None,
             Some(s) => {
                 let p = s.bytes().rposition(|c| *c == b'/').unwrap_or(0);
                 Some(HgPath::new(&s.as_bytes()[..p]))
             }
             None => None,
         };
         next
     }
 }
 
 impl<'a> FusedIterator for Ancestors<'a> {}
 
 /// An iterator over repository path yielding itself and its ancestors.
 #[derive(Copy, Clone, Debug)]
 pub(crate) struct AncestorsWithBase<'a> {
     next: Option<(&'a HgPath, &'a HgPath)>,
 }
 
 impl<'a> Iterator for AncestorsWithBase<'a> {
     type Item = (&'a HgPath, &'a HgPath);
 
     fn next(&mut self) -> Option<Self::Item> {
         let next = self.next;
         self.next = match self.next {
             Some((s, _)) if s.is_empty() => None,
             Some((s, _)) => Some(s.split_filename()),
             None => None,
         };
         next
     }
 }
 
 impl<'a> FusedIterator for AncestorsWithBase<'a> {}
 
 /// Returns an iterator yielding ancestor directories of the given repository
 /// path.
 ///
 /// The path is separated by '/', and must not start with '/'.
 ///
 /// The path itself isn't included unless it is b"" (meaning the root
 /// directory.)
 pub fn find_dirs(path: &HgPath) -> Ancestors {
     let mut dirs = Ancestors { next: Some(path) };
     if !path.is_empty() {
         dirs.next(); // skip itself
     }
     dirs
 }
 
 /// Returns an iterator yielding ancestor directories of the given repository
 /// path.
 ///
 /// The path is separated by '/', and must not start with '/'.
 ///
 /// The path itself isn't included unless it is b"" (meaning the root
 /// directory.)
 pub(crate) fn find_dirs_with_base(path: &HgPath) -> AncestorsWithBase {
     let mut dirs = AncestorsWithBase {
         next: Some((path, HgPath::new(b""))),
     };
     if !path.is_empty() {
         dirs.next(); // skip itself
     }
     dirs
 }
 
 /// TODO more than ASCII?
 pub fn normalize_case(path: &HgPath) -> HgPathBuf {
     #[cfg(windows)] // NTFS compares via upper()
     return path.to_ascii_uppercase();
     #[cfg(unix)]
     path.to_ascii_lowercase()
 }
 
 lazy_static! {
     static ref IGNORED_CHARS: Vec<Vec<u8>> = {
         [
             0x200c, 0x200d, 0x200e, 0x200f, 0x202a, 0x202b, 0x202c, 0x202d,
             0x202e, 0x206a, 0x206b, 0x206c, 0x206d, 0x206e, 0x206f, 0xfeff,
         ]
         .iter()
         .map(|code| {
             std::char::from_u32(*code)
                 .unwrap()
                 .encode_utf8(&mut [0; 3])
                 .bytes()
                 .collect()
         })
         .collect()
     };
 }
 
 fn hfs_ignore_clean(bytes: &[u8]) -> Vec<u8> {
     let mut buf = bytes.to_owned();
     let needs_escaping = bytes.iter().any(|b| *b == b'\xe2' || *b == b'\xef');
     if needs_escaping {
         for forbidden in IGNORED_CHARS.iter() {
             replace_slice(&mut buf, forbidden, &[])
         }
         buf
     } else {
         buf
     }
 }
 
 pub fn lower_clean(bytes: &[u8]) -> Vec<u8> {
     hfs_ignore_clean(&bytes.to_ascii_lowercase())
 }
 
 /// Returns the canonical path of `name`, given `cwd` and `root`
 pub fn canonical_path(
     root: impl AsRef<Path>,
     cwd: impl AsRef<Path>,
     name: impl AsRef<Path>,
 ) -> Result<PathBuf, HgPathError> {
     // TODO add missing normalization for other platforms
     let root = root.as_ref();
     let cwd = cwd.as_ref();
     let name = name.as_ref();
 
     let name = if !name.is_absolute() {
-        root.join(&cwd).join(&name)
+        root.join(cwd).join(name)
     } else {
         name.to_owned()
     };
-    let auditor = PathAuditor::new(&root);
-    if name != root && name.starts_with(&root) {
-        let name = name.strip_prefix(&root).unwrap();
+    let auditor = PathAuditor::new(root);
+    if name != root && name.starts_with(root) {
+        let name = name.strip_prefix(root).unwrap();
         auditor.audit_path(path_to_hg_path_buf(name)?)?;
         Ok(name.to_owned())
     } else if name == root {
         Ok("".into())
     } else {
         // Determine whether `name' is in the hierarchy at or beneath `root',
         // by iterating name=name.parent() until it returns `None` (can't
         // check name == '/', because that doesn't work on windows).
         let mut name = name.deref();
         let original_name = name.to_owned();
         loop {
-            let same = is_same_file(&name, &root).unwrap_or(false);
+            let same = is_same_file(name, root).unwrap_or(false);
             if same {
                 if name == original_name {
                     // `name` was actually the same as root (maybe a symlink)
                     return Ok("".into());
                 }
                 // `name` is a symlink to root, so `original_name` is under
                 // root
-                let rel_path = original_name.strip_prefix(&name).unwrap();
-                auditor.audit_path(path_to_hg_path_buf(&rel_path)?)?;
+                let rel_path = original_name.strip_prefix(name).unwrap();
+                auditor.audit_path(path_to_hg_path_buf(rel_path)?)?;
                 return Ok(rel_path.to_owned());
             }
             name = match name.parent() {
                 None => break,
                 Some(p) => p,
             };
         }
         // TODO hint to the user about using --cwd
         // Bubble up the responsibility to Python for now
         Err(HgPathError::NotUnderRoot {
             path: original_name,
             root: root.to_owned(),
         })
     }
 }
 
 /// Returns the representation of the path relative to the current working
 /// directory for display purposes.
 ///
 /// `cwd` is a `HgPath`, so it is considered relative to the root directory
 /// of the repository.
 ///
 /// # Examples
 ///
 /// ```
 /// use hg::utils::hg_path::HgPath;
 /// use hg::utils::files::relativize_path;
 /// use std::borrow::Cow;
 ///
 /// let file = HgPath::new(b"nested/file");
 /// let cwd = HgPath::new(b"");
 /// assert_eq!(relativize_path(file, cwd), Cow::Borrowed(b"nested/file"));
 ///
 /// let cwd = HgPath::new(b"nested");
 /// assert_eq!(relativize_path(file, cwd), Cow::Borrowed(b"file"));
 ///
 /// let cwd = HgPath::new(b"other");
 /// assert_eq!(relativize_path(file, cwd), Cow::Borrowed(b"../nested/file"));
 /// ```
 pub fn relativize_path(path: &HgPath, cwd: impl AsRef<HgPath>) -> Cow<[u8]> {
     if cwd.as_ref().is_empty() {
         Cow::Borrowed(path.as_bytes())
     } else {
         // This is not all accurate as to how large `res` will actually be, but
         // profiling `rhg files` on a large-ish repo shows it’s better than
         // starting from a zero-capacity `Vec` and letting `extend` reallocate
         // repeatedly.
         let guesstimate = path.as_bytes().len();
 
         let mut res: Vec<u8> = Vec::with_capacity(guesstimate);
         let mut path_iter = path.as_bytes().split(|b| *b == b'/').peekable();
         let mut cwd_iter =
             cwd.as_ref().as_bytes().split(|b| *b == b'/').peekable();
         loop {
             match (path_iter.peek(), cwd_iter.peek()) {
                 (Some(a), Some(b)) if a == b => (),
                 _ => break,
             }
             path_iter.next();
             cwd_iter.next();
         }
         let mut need_sep = false;
         for _ in cwd_iter {
             if need_sep {
                 res.extend(b"/")
             } else {
                 need_sep = true
             };
             res.extend(b"..");
         }
         for c in path_iter {
             if need_sep {
                 res.extend(b"/")
             } else {
                 need_sep = true
             };
             res.extend(c);
         }
         Cow::Owned(res)
     }
 }
 
 #[cfg(test)]
 mod tests {
     use super::*;
     use pretty_assertions::assert_eq;
 
     #[test]
     fn find_dirs_some() {
         let mut dirs = super::find_dirs(HgPath::new(b"foo/bar/baz"));
         assert_eq!(dirs.next(), Some(HgPath::new(b"foo/bar")));
         assert_eq!(dirs.next(), Some(HgPath::new(b"foo")));
         assert_eq!(dirs.next(), Some(HgPath::new(b"")));
         assert_eq!(dirs.next(), None);
         assert_eq!(dirs.next(), None);
     }
 
     #[test]
     fn find_dirs_empty() {
         // looks weird, but mercurial.pathutil.finddirs(b"") yields b""
         let mut dirs = super::find_dirs(HgPath::new(b""));
         assert_eq!(dirs.next(), Some(HgPath::new(b"")));
         assert_eq!(dirs.next(), None);
         assert_eq!(dirs.next(), None);
     }
 
     #[test]
     fn test_find_dirs_with_base_some() {
         let mut dirs = super::find_dirs_with_base(HgPath::new(b"foo/bar/baz"));
         assert_eq!(
             dirs.next(),
             Some((HgPath::new(b"foo/bar"), HgPath::new(b"baz")))
         );
         assert_eq!(
             dirs.next(),
             Some((HgPath::new(b"foo"), HgPath::new(b"bar")))
         );
         assert_eq!(dirs.next(), Some((HgPath::new(b""), HgPath::new(b"foo"))));
         assert_eq!(dirs.next(), None);
         assert_eq!(dirs.next(), None);
     }
 
     #[test]
     fn test_find_dirs_with_base_empty() {
         let mut dirs = super::find_dirs_with_base(HgPath::new(b""));
         assert_eq!(dirs.next(), Some((HgPath::new(b""), HgPath::new(b""))));
         assert_eq!(dirs.next(), None);
         assert_eq!(dirs.next(), None);
     }
 
     #[test]
     fn test_canonical_path() {
         let root = Path::new("/repo");
         let cwd = Path::new("/dir");
         let name = Path::new("filename");
         assert_eq!(
             canonical_path(root, cwd, name),
             Err(HgPathError::NotUnderRoot {
                 path: PathBuf::from("/dir/filename"),
                 root: root.to_path_buf()
             })
         );
 
         let root = Path::new("/repo");
         let cwd = Path::new("/");
         let name = Path::new("filename");
         assert_eq!(
             canonical_path(root, cwd, name),
             Err(HgPathError::NotUnderRoot {
                 path: PathBuf::from("/filename"),
                 root: root.to_path_buf()
             })
         );
 
         let root = Path::new("/repo");
         let cwd = Path::new("/");
         let name = Path::new("repo/filename");
         assert_eq!(
             canonical_path(root, cwd, name),
             Ok(PathBuf::from("filename"))
         );
 
         let root = Path::new("/repo");
         let cwd = Path::new("/repo");
         let name = Path::new("filename");
         assert_eq!(
             canonical_path(root, cwd, name),
             Ok(PathBuf::from("filename"))
         );
 
         let root = Path::new("/repo");
         let cwd = Path::new("/repo/subdir");
         let name = Path::new("filename");
         assert_eq!(
             canonical_path(root, cwd, name),
             Ok(PathBuf::from("subdir/filename"))
         );
     }
 
     #[test]
     fn test_canonical_path_not_rooted() {
         use std::fs::create_dir;
         use tempfile::tempdir;
 
         let base_dir = tempdir().unwrap();
         let base_dir_path = base_dir.path();
         let beneath_repo = base_dir_path.join("a");
         let root = base_dir_path.join("a/b");
         let out_of_repo = base_dir_path.join("c");
         let under_repo_symlink = out_of_repo.join("d");
 
         create_dir(&beneath_repo).unwrap();
         create_dir(&root).unwrap();
 
         // TODO make portable
         std::os::unix::fs::symlink(&root, &out_of_repo).unwrap();
 
         assert_eq!(
             canonical_path(&root, Path::new(""), out_of_repo),
             Ok(PathBuf::from(""))
         );
         assert_eq!(
             canonical_path(&root, Path::new(""), &beneath_repo),
             Err(HgPathError::NotUnderRoot {
                 path: beneath_repo,
                 root: root.to_owned()
             })
         );
         assert_eq!(
-            canonical_path(&root, Path::new(""), &under_repo_symlink),
+            canonical_path(&root, Path::new(""), under_repo_symlink),
             Ok(PathBuf::from("d"))
         );
     }
 }

rust/hg-core/src/utils/path_auditor.rs

@@ -1,223 +1,223 b''
 // path_auditor.rs
 //
 // Copyright 2020
 // Raphaël Gomès <[e-mail]>,
 //
 // This software may be used and distributed according to the terms of the
 // GNU General Public License version 2 or any later version.
 
 use crate::utils::{
     files::lower_clean,
     find_slice_in_slice,
     hg_path::{hg_path_to_path_buf, HgPath, HgPathBuf, HgPathError},
 };
 use std::collections::HashSet;
 use std::path::{Path, PathBuf};
 use std::sync::{Mutex, RwLock};
 
 /// Ensures that a path is valid for use in the repository i.e. does not use
 /// any banned components, does not traverse a symlink, etc.
 #[derive(Debug, Default)]
 pub struct PathAuditor {
     audited: Mutex<HashSet<HgPathBuf>>,
     audited_dirs: RwLock<HashSet<HgPathBuf>>,
     root: PathBuf,
 }
 
 impl PathAuditor {
     pub fn new(root: impl AsRef<Path>) -> Self {
         Self {
             root: root.as_ref().to_owned(),
             ..Default::default()
         }
     }
     pub fn audit_path(
         &self,
         path: impl AsRef<HgPath>,
     ) -> Result<(), HgPathError> {
         // TODO windows "localpath" normalization
         let path = path.as_ref();
         if path.is_empty() {
             return Ok(());
         }
         // TODO case normalization
         if self.audited.lock().unwrap().contains(path) {
             return Ok(());
         }
         // AIX ignores "/" at end of path, others raise EISDIR.
         let last_byte = path.as_bytes()[path.len() - 1];
         if last_byte == b'/' || last_byte == b'\\' {
             return Err(HgPathError::EndsWithSlash(path.to_owned()));
         }
         let parts: Vec<_> = path
             .as_bytes()
             .split(|b| std::path::is_separator(*b as char))
             .collect();
 
         let first_component = lower_clean(parts[0]);
         let first_component = first_component.as_slice();
         if !path.split_drive().0.is_empty()
             || (first_component == b".hg"
                 || first_component == b".hg."
                 || first_component == b"")
             || parts.iter().any(|c| c == b"..")
         {
             return Err(HgPathError::InsideDotHg(path.to_owned()));
         }
 
         // Windows shortname aliases
         for part in parts.iter() {
             if part.contains(&b'~') {
                 let mut split = part.splitn(2, |b| *b == b'~');
                 let first =
                     split.next().unwrap().to_owned().to_ascii_uppercase();
                 let last = split.next().unwrap();
                 if last.iter().all(u8::is_ascii_digit)
                     && (first == b"HG" || first == b"HG8B6C")
                 {
                     return Err(HgPathError::ContainsIllegalComponent(
                         path.to_owned(),
                     ));
                 }
             }
         }
         let lower_path = lower_clean(path.as_bytes());
         if find_slice_in_slice(&lower_path, b".hg").is_some() {
             let lower_parts: Vec<_> = path
                 .as_bytes()
                 .split(|b| std::path::is_separator(*b as char))
                 .collect();
             for pattern in [b".hg".to_vec(), b".hg.".to_vec()].iter() {
                 if let Some(pos) = lower_parts[1..]
                     .iter()
                     .position(|part| part == &pattern.as_slice())
                 {
                     let base = lower_parts[..=pos]
                         .iter()
                         .fold(HgPathBuf::new(), |acc, p| {
                             acc.join(HgPath::new(p))
                         });
                     return Err(HgPathError::IsInsideNestedRepo {
                         path: path.to_owned(),
                         nested_repo: base,
                     });
                 }
             }
         }
 
         let parts = &parts[..parts.len().saturating_sub(1)];
 
         // We don't want to add "foo/bar/baz" to `audited_dirs` before checking
         // if there's a "foo/.hg" directory. This also means we won't
         // accidentally traverse a symlink into some other filesystem (which
         // is potentially expensive to access).
         for index in 0..parts.len() {
             let prefix = &parts[..=index].join(&b'/');
             let prefix = HgPath::new(prefix);
             if self.audited_dirs.read().unwrap().contains(prefix) {
                 continue;
             }
-            self.check_filesystem(&prefix, &path)?;
+            self.check_filesystem(prefix, path)?;
             self.audited_dirs.write().unwrap().insert(prefix.to_owned());
         }
 
         self.audited.lock().unwrap().insert(path.to_owned());
 
         Ok(())
     }
 
     pub fn check_filesystem(
         &self,
         prefix: impl AsRef<HgPath>,
         path: impl AsRef<HgPath>,
     ) -> Result<(), HgPathError> {
         let prefix = prefix.as_ref();
         let path = path.as_ref();
         let current_path = self.root.join(
             hg_path_to_path_buf(prefix)
                 .map_err(|_| HgPathError::NotFsCompliant(path.to_owned()))?,
         );
         match std::fs::symlink_metadata(&current_path) {
             Err(e) => {
                 // EINVAL can be raised as invalid path syntax under win32.
                 if e.kind() != std::io::ErrorKind::NotFound
                     && e.kind() != std::io::ErrorKind::InvalidInput
                     && e.raw_os_error() != Some(20)
                 {
                     // Rust does not yet have an `ErrorKind` for
                     // `NotADirectory` (errno 20)
                     // It happens if the dirstate contains `foo/bar` and
                     // foo is not a directory
                     return Err(HgPathError::NotFsCompliant(path.to_owned()));
                 }
             }
             Ok(meta) => {
                 if meta.file_type().is_symlink() {
                     return Err(HgPathError::TraversesSymbolicLink {
                         path: path.to_owned(),
                         symlink: prefix.to_owned(),
                     });
                 }
                 if meta.file_type().is_dir()
                     && current_path.join(".hg").is_dir()
                 {
                     return Err(HgPathError::IsInsideNestedRepo {
                         path: path.to_owned(),
                         nested_repo: prefix.to_owned(),
                     });
                 }
             }
         };
 
         Ok(())
     }
 
     pub fn check(&self, path: impl AsRef<HgPath>) -> bool {
         self.audit_path(path).is_ok()
     }
 }
 
 #[cfg(test)]
 mod tests {
     use super::*;
     use std::fs::{create_dir, File};
     use tempfile::tempdir;
 
     #[test]
     fn test_path_auditor() {
         let base_dir = tempdir().unwrap();
         let base_dir_path = base_dir.path();
         let auditor = PathAuditor::new(base_dir_path);
 
         let path = HgPath::new(b".hg/00changelog.i");
         assert_eq!(
             auditor.audit_path(path),
             Err(HgPathError::InsideDotHg(path.to_owned()))
         );
         let path = HgPath::new(b"this/is/nested/.hg/thing.txt");
         assert_eq!(
             auditor.audit_path(path),
             Err(HgPathError::IsInsideNestedRepo {
                 path: path.to_owned(),
                 nested_repo: HgPathBuf::from_bytes(b"this/is/nested")
             })
         );
 
-        create_dir(&base_dir_path.join("realdir")).unwrap();
-        File::create(&base_dir_path.join("realdir/realfile")).unwrap();
+        create_dir(base_dir_path.join("realdir")).unwrap();
+        File::create(base_dir_path.join("realdir/realfile")).unwrap();
         // TODO make portable
         std::os::unix::fs::symlink(
-            &base_dir_path.join("realdir"),
-            &base_dir_path.join("symlink"),
+            base_dir_path.join("realdir"),
+            base_dir_path.join("symlink"),
         )
         .unwrap();
         let path = HgPath::new(b"symlink/realfile");
         assert_eq!(
             auditor.audit_path(path),
             Err(HgPathError::TraversesSymbolicLink {
                 path: path.to_owned(),
                 symlink: HgPathBuf::from_bytes(b"symlink"),
             })
         );
     }
 }

rust/hg-cpython/src/dirstate/status.rs

@@ -1,307 +1,307 b''
 // status.rs
 //
 // Copyright 2019, Raphaël Gomès <[e-mail]>
 //
 // This software may be used and distributed according to the terms of the
 // GNU General Public License version 2 or any later version.
 
 //! Bindings for the `hg::status` module provided by the
 //! `hg-core` crate. From Python, this will be seen as
 //! `rustext.dirstate.status`.
 
 use crate::{dirstate::DirstateMap, exceptions::FallbackError};
 use cpython::{
     exc::ValueError, ObjectProtocol, PyBytes, PyErr, PyList, PyObject,
     PyResult, PyTuple, Python, PythonObject, ToPyObject,
 };
 use hg::dirstate::status::StatusPath;
 use hg::matchers::{
     DifferenceMatcher, IntersectionMatcher, Matcher, NeverMatcher,
     UnionMatcher,
 };
 use hg::{
     matchers::{AlwaysMatcher, FileMatcher, IncludeMatcher},
     parse_pattern_syntax,
     utils::{
         files::{get_bytes_from_path, get_path_from_bytes},
         hg_path::{HgPath, HgPathBuf},
     },
     BadMatch, DirstateStatus, IgnorePattern, PatternFileWarning, StatusError,
     StatusOptions,
 };
 use std::borrow::Borrow;
 
 fn collect_status_path_list(py: Python, paths: &[StatusPath<'_>]) -> PyList {
     collect_pybytes_list(py, paths.iter().map(|item| &*item.path))
 }
 
 /// This will be useless once trait impls for collection are added to `PyBytes`
 /// upstream.
 fn collect_pybytes_list(
     py: Python,
     iter: impl Iterator<Item = impl AsRef<HgPath>>,
 ) -> PyList {
     let list = PyList::new(py, &[]);
 
     for path in iter {
         list.append(
             py,
             PyBytes::new(py, path.as_ref().as_bytes()).into_object(),
         )
     }
 
     list
 }
 
 fn collect_bad_matches(
     py: Python,
     collection: &[(impl AsRef<HgPath>, BadMatch)],
 ) -> PyResult<PyList> {
     let list = PyList::new(py, &[]);
 
     let os = py.import("os")?;
     let get_error_message = |code: i32| -> PyResult<_> {
         os.call(
             py,
             "strerror",
             PyTuple::new(py, &[code.to_py_object(py).into_object()]),
             None,
         )
     };
 
     for (path, bad_match) in collection.iter() {
         let message = match bad_match {
             BadMatch::OsError(code) => get_error_message(*code)?,
             BadMatch::BadType(bad_type) => {
                 format!("unsupported file type (type is {})", bad_type)
                     .to_py_object(py)
                     .into_object()
             }
         };
         list.append(
             py,
             (PyBytes::new(py, path.as_ref().as_bytes()), message)
                 .to_py_object(py)
                 .into_object(),
         )
     }
 
     Ok(list)
 }
 
 fn handle_fallback(py: Python, err: StatusError) -> PyErr {
     match err {
         StatusError::Pattern(e) => {
             let as_string = e.to_string();
             log::trace!("Rust status fallback: `{}`", &as_string);
 
             PyErr::new::<FallbackError, _>(py, &as_string)
         }
         e => PyErr::new::<ValueError, _>(py, e.to_string()),
     }
 }
 
 pub fn status_wrapper(
     py: Python,
     dmap: DirstateMap,
     matcher: PyObject,
     root_dir: PyObject,
     ignore_files: PyList,
     check_exec: bool,
     list_clean: bool,
     list_ignored: bool,
     list_unknown: bool,
     collect_traversed_dirs: bool,
 ) -> PyResult<PyTuple> {
     let bytes = root_dir.extract::<PyBytes>(py)?;
     let root_dir = get_path_from_bytes(bytes.data(py));
 
     let dmap: DirstateMap = dmap.to_py_object(py);
     let mut dmap = dmap.get_inner_mut(py);
 
     let ignore_files: PyResult<Vec<_>> = ignore_files
         .iter(py)
         .map(|b| {
             let file = b.extract::<PyBytes>(py)?;
             Ok(get_path_from_bytes(file.data(py)).to_owned())
         })
         .collect();
     let ignore_files = ignore_files?;
     // The caller may call `copymap.items()` separately
     let list_copies = false;
 
     let after_status = |res: Result<(DirstateStatus<'_>, _), StatusError>| {
         let (status_res, warnings) =
             res.map_err(|e| handle_fallback(py, e))?;
         build_response(py, status_res, warnings)
     };
 
     let matcher = extract_matcher(py, matcher)?;
     dmap.with_status(
         &*matcher,
         root_dir.to_path_buf(),
         ignore_files,
         StatusOptions {
             check_exec,
             list_clean,
             list_ignored,
             list_unknown,
             list_copies,
             collect_traversed_dirs,
         },
         after_status,
     )
 }
 
 /// Transform a Python matcher into a Rust matcher.
 fn extract_matcher(
     py: Python,
     matcher: PyObject,
 ) -> PyResult<Box<dyn Matcher + Sync>> {
     match matcher.get_type(py).name(py).borrow() {
         "alwaysmatcher" => Ok(Box::new(AlwaysMatcher)),
         "nevermatcher" => Ok(Box::new(NeverMatcher)),
         "exactmatcher" => {
             let files = matcher.call_method(
                 py,
                 "files",
                 PyTuple::new(py, &[]),
                 None,
             )?;
             let files: PyList = files.cast_into(py)?;
             let files: PyResult<Vec<HgPathBuf>> = files
                 .iter(py)
                 .map(|f| {
                     Ok(HgPathBuf::from_bytes(
                         f.extract::<PyBytes>(py)?.data(py),
                     ))
                 })
                 .collect();
 
             let files = files?;
             let file_matcher = FileMatcher::new(files)
                 .map_err(|e| PyErr::new::<ValueError, _>(py, e.to_string()))?;
             Ok(Box::new(file_matcher))
         }
         "includematcher" => {
             // Get the patterns from Python even though most of them are
             // redundant with those we will parse later on, as they include
             // those passed from the command line.
             let ignore_patterns: PyResult<Vec<_>> = matcher
                 .getattr(py, "_kindpats")?
                 .iter(py)?
                 .map(|k| {
                     let k = k?;
                     let syntax = parse_pattern_syntax(
                         &[
                             k.get_item(py, 0)?
                                 .extract::<PyBytes>(py)?
                                 .data(py),
                             &b":"[..],
                         ]
                         .concat(),
                     )
                     .map_err(|e| {
                         handle_fallback(py, StatusError::Pattern(e))
                     })?;
                     let pattern = k.get_item(py, 1)?.extract::<PyBytes>(py)?;
                     let pattern = pattern.data(py);
                     let source = k.get_item(py, 2)?.extract::<PyBytes>(py)?;
                     let source = get_path_from_bytes(source.data(py));
                     let new = IgnorePattern::new(syntax, pattern, source);
                     Ok(new)
                 })
                 .collect();
 
             let ignore_patterns = ignore_patterns?;
 
             let matcher = IncludeMatcher::new(ignore_patterns)
                 .map_err(|e| handle_fallback(py, e.into()))?;
 
             Ok(Box::new(matcher))
         }
         "unionmatcher" => {
             let matchers: PyResult<Vec<_>> = matcher
                 .getattr(py, "_matchers")?
                 .iter(py)?
                 .map(|py_matcher| extract_matcher(py, py_matcher?))
                 .collect();
 
             Ok(Box::new(UnionMatcher::new(matchers?)))
         }
         "intersectionmatcher" => {
             let m1 = extract_matcher(py, matcher.getattr(py, "_m1")?)?;
             let m2 = extract_matcher(py, matcher.getattr(py, "_m2")?)?;
 
             Ok(Box::new(IntersectionMatcher::new(m1, m2)))
         }
         "differencematcher" => {
             let m1 = extract_matcher(py, matcher.getattr(py, "_m1")?)?;
             let m2 = extract_matcher(py, matcher.getattr(py, "_m2")?)?;
 
             Ok(Box::new(DifferenceMatcher::new(m1, m2)))
         }
         e => Err(PyErr::new::<FallbackError, _>(
             py,
             format!("Unsupported matcher {}", e),
         )),
     }
 }
 
 fn build_response(
     py: Python,
     status_res: DirstateStatus,
     warnings: Vec<PatternFileWarning>,
 ) -> PyResult<PyTuple> {
     let modified = collect_status_path_list(py, &status_res.modified);
     let added = collect_status_path_list(py, &status_res.added);
     let removed = collect_status_path_list(py, &status_res.removed);
     let deleted = collect_status_path_list(py, &status_res.deleted);
     let clean = collect_status_path_list(py, &status_res.clean);
     let ignored = collect_status_path_list(py, &status_res.ignored);
     let unknown = collect_status_path_list(py, &status_res.unknown);
     let unsure = collect_status_path_list(py, &status_res.unsure);
     let bad = collect_bad_matches(py, &status_res.bad)?;
     let traversed = collect_pybytes_list(py, status_res.traversed.iter());
     let dirty = status_res.dirty.to_py_object(py);
     let py_warnings = PyList::new(py, &[]);
     for warning in warnings.iter() {
         // We use duck-typing on the Python side for dispatch, good enough for
         // now.
         match warning {
             PatternFileWarning::InvalidSyntax(file, syn) => {
                 py_warnings.append(
                     py,
                     (
-                        PyBytes::new(py, &get_bytes_from_path(&file)),
+                        PyBytes::new(py, &get_bytes_from_path(file)),
                         PyBytes::new(py, syn),
                     )
                         .to_py_object(py)
                         .into_object(),
                 );
             }
             PatternFileWarning::NoSuchFile(file) => py_warnings.append(
                 py,
-                PyBytes::new(py, &get_bytes_from_path(&file)).into_object(),
+                PyBytes::new(py, &get_bytes_from_path(file)).into_object(),
             ),
         }
     }
 
     Ok(PyTuple::new(
         py,
         &[
             unsure.into_object(),
             modified.into_object(),
             added.into_object(),
             removed.into_object(),
             deleted.into_object(),
             clean.into_object(),
             ignored.into_object(),
             unknown.into_object(),
             py_warnings.into_object(),
             bad.into_object(),
             traversed.into_object(),
             dirty.into_object(),
         ][..],
     ))
 }

rust/hg-cpython/src/revlog.rs

@@ -1,524 +1,524 b''
 // revlog.rs
 //
 // Copyright 2019-2020 Georges Racinet <[e-mail]>
 //
 // This software may be used and distributed according to the terms of the
 // GNU General Public License version 2 or any later version.
 
 use crate::{
     cindex,
     utils::{node_from_py_bytes, node_from_py_object},
     PyRevision,
 };
 use cpython::{
     buffer::{Element, PyBuffer},
     exc::{IndexError, ValueError},
     ObjectProtocol, PyBytes, PyClone, PyDict, PyErr, PyInt, PyModule,
     PyObject, PyResult, PyString, PyTuple, Python, PythonObject, ToPyObject,
 };
 use hg::{
     nodemap::{Block, NodeMapError, NodeTree},
     revlog::{nodemap::NodeMap, NodePrefix, RevlogIndex},
     BaseRevision, Revision, UncheckedRevision,
 };
 use std::cell::RefCell;
 
 /// Return a Struct implementing the Graph trait
 pub(crate) fn pyindex_to_graph(
     py: Python,
     index: PyObject,
 ) -> PyResult<cindex::Index> {
     match index.extract::<MixedIndex>(py) {
         Ok(midx) => Ok(midx.clone_cindex(py)),
         Err(_) => cindex::Index::new(py, index),
     }
 }
 
 py_class!(pub class MixedIndex |py| {
     data cindex: RefCell<cindex::Index>;
     data nt: RefCell<Option<NodeTree>>;
     data docket: RefCell<Option<PyObject>>;
     // Holds a reference to the mmap'ed persistent nodemap data
     data mmap: RefCell<Option<PyBuffer>>;
 
     def __new__(_cls, cindex: PyObject) -> PyResult<MixedIndex> {
         Self::new(py, cindex)
     }
 
     /// Compatibility layer used for Python consumers needing access to the C index
     ///
     /// Only use case so far is `scmutil.shortesthexnodeidprefix`,
     /// that may need to build a custom `nodetree`, based on a specified revset.
     /// With a Rust implementation of the nodemap, we will be able to get rid of
     /// this, by exposing our own standalone nodemap class,
     /// ready to accept `MixedIndex`.
     def get_cindex(&self) -> PyResult<PyObject> {
         Ok(self.cindex(py).borrow().inner().clone_ref(py))
     }
 
     // Index API involving nodemap, as defined in mercurial/pure/parsers.py
 
     /// Return Revision if found, raises a bare `error.RevlogError`
     /// in case of ambiguity, same as C version does
     def get_rev(&self, node: PyBytes) -> PyResult<Option<PyRevision>> {
         let opt = self.get_nodetree(py)?.borrow();
         let nt = opt.as_ref().unwrap();
         let idx = &*self.cindex(py).borrow();
         let node = node_from_py_bytes(py, &node)?;
         let res = nt.find_bin(idx, node.into());
         Ok(res.map_err(|e| nodemap_error(py, e))?.map(Into::into))
     }
 
     /// same as `get_rev()` but raises a bare `error.RevlogError` if node
     /// is not found.
     ///
     /// No need to repeat `node` in the exception, `mercurial/revlog.py`
     /// will catch and rewrap with it
     def rev(&self, node: PyBytes) -> PyResult<PyRevision> {
         self.get_rev(py, node)?.ok_or_else(|| revlog_error(py))
     }
 
     /// return True if the node exist in the index
     def has_node(&self, node: PyBytes) -> PyResult<bool> {
         self.get_rev(py, node).map(|opt| opt.is_some())
     }
 
     /// find length of shortest hex nodeid of a binary ID
     def shortest(&self, node: PyBytes) -> PyResult<usize> {
         let opt = self.get_nodetree(py)?.borrow();
         let nt = opt.as_ref().unwrap();
         let idx = &*self.cindex(py).borrow();
         match nt.unique_prefix_len_node(idx, &node_from_py_bytes(py, &node)?)
         {
             Ok(Some(l)) => Ok(l),
             Ok(None) => Err(revlog_error(py)),
             Err(e) => Err(nodemap_error(py, e)),
         }
     }
 
     def partialmatch(&self, node: PyObject) -> PyResult<Option<PyBytes>> {
         let opt = self.get_nodetree(py)?.borrow();
         let nt = opt.as_ref().unwrap();
         let idx = &*self.cindex(py).borrow();
 
         let node_as_string = if cfg!(feature = "python3-sys") {
             node.cast_as::<PyString>(py)?.to_string(py)?.to_string()
         }
         else {
             let node = node.extract::<PyBytes>(py)?;
             String::from_utf8_lossy(node.data(py)).to_string()
         };
 
         let prefix = NodePrefix::from_hex(&node_as_string)
             .map_err(|_| PyErr::new::<ValueError, _>(
                 py, format!("Invalid node or prefix '{}'", node_as_string))
             )?;
 
         nt.find_bin(idx, prefix)
             // TODO make an inner API returning the node directly
             .map(|opt| opt.map(
                 |rev| PyBytes::new(py, idx.node(rev).unwrap().as_bytes())))
             .map_err(|e| nodemap_error(py, e))
 
     }
 
     /// append an index entry
     def append(&self, tup: PyTuple) -> PyResult<PyObject> {
         if tup.len(py) < 8 {
             // this is better than the panic promised by tup.get_item()
             return Err(
                 PyErr::new::<IndexError, _>(py, "tuple index out of range"))
         }
         let node_bytes = tup.get_item(py, 7).extract(py)?;
         let node = node_from_py_object(py, &node_bytes)?;
 
         let mut idx = self.cindex(py).borrow_mut();
 
         // This is ok since we will just add the revision to the index
         let rev = Revision(idx.len() as BaseRevision);
         idx.append(py, tup)?;
 
         self.get_nodetree(py)?.borrow_mut().as_mut().unwrap()
             .insert(&*idx, &node, rev)
             .map_err(|e| nodemap_error(py, e))?;
         Ok(py.None())
     }
 
     def __delitem__(&self, key: PyObject) -> PyResult<()> {
         // __delitem__ is both for `del idx[r]` and `del idx[r1:r2]`
         self.cindex(py).borrow().inner().del_item(py, key)?;
         let mut opt = self.get_nodetree(py)?.borrow_mut();
         let nt = opt.as_mut().unwrap();
         nt.invalidate_all();
         self.fill_nodemap(py, nt)?;
         Ok(())
     }
 
     //
     // Reforwarded C index API
     //
 
     // index_methods (tp_methods). Same ordering as in revlog.c
 
     /// return the gca set of the given revs
     def ancestors(&self, *args, **kw) -> PyResult<PyObject> {
         self.call_cindex(py, "ancestors", args, kw)
     }
 
     /// return the heads of the common ancestors of the given revs
     def commonancestorsheads(&self, *args, **kw) -> PyResult<PyObject> {
         self.call_cindex(py, "commonancestorsheads", args, kw)
     }
 
     /// Clear the index caches and inner py_class data.
     /// It is Python's responsibility to call `update_nodemap_data` again.
     def clearcaches(&self, *args, **kw) -> PyResult<PyObject> {
         self.nt(py).borrow_mut().take();
         self.docket(py).borrow_mut().take();
         self.mmap(py).borrow_mut().take();
         self.call_cindex(py, "clearcaches", args, kw)
     }
 
     /// return the raw binary string representing a revision
     def entry_binary(&self, *args, **kw) -> PyResult<PyObject> {
         self.call_cindex(py, "entry_binary", args, kw)
     }
 
     /// return a binary packed version of the header
     def pack_header(&self, *args, **kw) -> PyResult<PyObject> {
         self.call_cindex(py, "pack_header", args, kw)
     }
 
     /// get an index entry
     def get(&self, *args, **kw) -> PyResult<PyObject> {
         self.call_cindex(py, "get", args, kw)
     }
 
     /// compute phases
     def computephasesmapsets(&self, *args, **kw) -> PyResult<PyObject> {
         self.call_cindex(py, "computephasesmapsets", args, kw)
     }
 
     /// reachableroots
     def reachableroots2(&self, *args, **kw) -> PyResult<PyObject> {
         self.call_cindex(py, "reachableroots2", args, kw)
     }
 
     /// get head revisions
     def headrevs(&self, *args, **kw) -> PyResult<PyObject> {
         self.call_cindex(py, "headrevs", args, kw)
     }
 
     /// get filtered head revisions
     def headrevsfiltered(&self, *args, **kw) -> PyResult<PyObject> {
         self.call_cindex(py, "headrevsfiltered", args, kw)
     }
 
     /// True if the object is a snapshot
     def issnapshot(&self, *args, **kw) -> PyResult<PyObject> {
         self.call_cindex(py, "issnapshot", args, kw)
     }
 
     /// Gather snapshot data in a cache dict
     def findsnapshots(&self, *args, **kw) -> PyResult<PyObject> {
         self.call_cindex(py, "findsnapshots", args, kw)
     }
 
     /// determine revisions with deltas to reconstruct fulltext
     def deltachain(&self, *args, **kw) -> PyResult<PyObject> {
         self.call_cindex(py, "deltachain", args, kw)
     }
 
     /// slice planned chunk read to reach a density threshold
     def slicechunktodensity(&self, *args, **kw) -> PyResult<PyObject> {
         self.call_cindex(py, "slicechunktodensity", args, kw)
     }
 
     /// stats for the index
     def stats(&self, *args, **kw) -> PyResult<PyObject> {
         self.call_cindex(py, "stats", args, kw)
     }
 
     // index_sequence_methods and index_mapping_methods.
     //
     // Since we call back through the high level Python API,
     // there's no point making a distinction between index_get
     // and index_getitem.
 
     def __len__(&self) -> PyResult<usize> {
         self.cindex(py).borrow().inner().len(py)
     }
 
     def __getitem__(&self, key: PyObject) -> PyResult<PyObject> {
         // this conversion seems needless, but that's actually because
         // `index_getitem` does not handle conversion from PyLong,
         // which expressions such as [e for e in index] internally use.
         // Note that we don't seem to have a direct way to call
         // PySequence_GetItem (does the job), which would possibly be better
         // for performance
         let key = match key.extract::<i32>(py) {
             Ok(rev) => rev.to_py_object(py).into_object(),
             Err(_) => key,
         };
         self.cindex(py).borrow().inner().get_item(py, key)
     }
 
     def __setitem__(&self, key: PyObject, value: PyObject) -> PyResult<()> {
         self.cindex(py).borrow().inner().set_item(py, key, value)
     }
 
     def __contains__(&self, item: PyObject) -> PyResult<bool> {
         // ObjectProtocol does not seem to provide contains(), so
         // this is an equivalent implementation of the index_contains()
         // defined in revlog.c
         let cindex = self.cindex(py).borrow();
         match item.extract::<i32>(py) {
             Ok(rev) => {
                 Ok(rev >= -1 && rev < cindex.inner().len(py)? as BaseRevision)
             }
             Err(_) => {
                 cindex.inner().call_method(
                     py,
                     "has_node",
                     PyTuple::new(py, &[item]),
                     None)?
                 .extract(py)
             }
         }
     }
 
     def nodemap_data_all(&self) -> PyResult<PyBytes> {
         self.inner_nodemap_data_all(py)
     }
 
     def nodemap_data_incremental(&self) -> PyResult<PyObject> {
         self.inner_nodemap_data_incremental(py)
     }
     def update_nodemap_data(
         &self,
         docket: PyObject,
         nm_data: PyObject
     ) -> PyResult<PyObject> {
         self.inner_update_nodemap_data(py, docket, nm_data)
     }
 
     @property
     def entry_size(&self) -> PyResult<PyInt> {
         self.cindex(py).borrow().inner().getattr(py, "entry_size")?.extract::<PyInt>(py)
     }
 
     @property
     def rust_ext_compat(&self) -> PyResult<PyInt> {
         self.cindex(py).borrow().inner().getattr(py, "rust_ext_compat")?.extract::<PyInt>(py)
     }
 
 });
 
 impl MixedIndex {
     fn new(py: Python, cindex: PyObject) -> PyResult<MixedIndex> {
         Self::create_instance(
             py,
             RefCell::new(cindex::Index::new(py, cindex)?),
             RefCell::new(None),
             RefCell::new(None),
             RefCell::new(None),
         )
     }
 
     /// This is scaffolding at this point, but it could also become
     /// a way to start a persistent nodemap or perform a
     /// vacuum / repack operation
     fn fill_nodemap(
         &self,
         py: Python,
         nt: &mut NodeTree,
     ) -> PyResult<PyObject> {
         let index = self.cindex(py).borrow();
         for r in 0..index.len() {
             let rev = Revision(r as BaseRevision);
             // in this case node() won't ever return None
             nt.insert(&*index, index.node(rev).unwrap(), rev)
                 .map_err(|e| nodemap_error(py, e))?
         }
         Ok(py.None())
     }
 
     fn get_nodetree<'a>(
         &'a self,
         py: Python<'a>,
     ) -> PyResult<&'a RefCell<Option<NodeTree>>> {
         if self.nt(py).borrow().is_none() {
-            let readonly = Box::new(Vec::new());
+            let readonly = Box::<Vec<_>>::default();
             let mut nt = NodeTree::load_bytes(readonly, 0);
             self.fill_nodemap(py, &mut nt)?;
             self.nt(py).borrow_mut().replace(nt);
         }
         Ok(self.nt(py))
     }
 
     /// forward a method call to the underlying C index
     fn call_cindex(
         &self,
         py: Python,
         name: &str,
         args: &PyTuple,
         kwargs: Option<&PyDict>,
     ) -> PyResult<PyObject> {
         self.cindex(py)
             .borrow()
             .inner()
             .call_method(py, name, args, kwargs)
     }
 
     pub fn clone_cindex(&self, py: Python) -> cindex::Index {
         self.cindex(py).borrow().clone_ref(py)
     }
 
     /// Returns the full nodemap bytes to be written as-is to disk
     fn inner_nodemap_data_all(&self, py: Python) -> PyResult<PyBytes> {
         let nodemap = self.get_nodetree(py)?.borrow_mut().take().unwrap();
         let (readonly, bytes) = nodemap.into_readonly_and_added_bytes();
 
         // If there's anything readonly, we need to build the data again from
         // scratch
         let bytes = if readonly.len() > 0 {
-            let mut nt = NodeTree::load_bytes(Box::new(vec![]), 0);
+            let mut nt = NodeTree::load_bytes(Box::<Vec<_>>::default(), 0);
             self.fill_nodemap(py, &mut nt)?;
 
             let (readonly, bytes) = nt.into_readonly_and_added_bytes();
             assert_eq!(readonly.len(), 0);
 
             bytes
         } else {
             bytes
         };
 
         let bytes = PyBytes::new(py, &bytes);
         Ok(bytes)
     }
 
     /// Returns the last saved docket along with the size of any changed data
     /// (in number of blocks), and said data as bytes.
     fn inner_nodemap_data_incremental(
         &self,
         py: Python,
     ) -> PyResult<PyObject> {
         let docket = self.docket(py).borrow();
         let docket = match docket.as_ref() {
             Some(d) => d,
             None => return Ok(py.None()),
         };
 
         let node_tree = self.get_nodetree(py)?.borrow_mut().take().unwrap();
         let masked_blocks = node_tree.masked_readonly_blocks();
         let (_, data) = node_tree.into_readonly_and_added_bytes();
         let changed = masked_blocks * std::mem::size_of::<Block>();
 
         Ok((docket, changed, PyBytes::new(py, &data))
             .to_py_object(py)
             .into_object())
     }
 
     /// Update the nodemap from the new (mmaped) data.
     /// The docket is kept as a reference for later incremental calls.
     fn inner_update_nodemap_data(
         &self,
         py: Python,
         docket: PyObject,
         nm_data: PyObject,
     ) -> PyResult<PyObject> {
         let buf = PyBuffer::get(py, &nm_data)?;
         let len = buf.item_count();
 
         // Build a slice from the mmap'ed buffer data
         let cbuf = buf.buf_ptr();
         let bytes = if std::mem::size_of::<u8>() == buf.item_size()
             && buf.is_c_contiguous()
             && u8::is_compatible_format(buf.format())
         {
             unsafe { std::slice::from_raw_parts(cbuf as *const u8, len) }
         } else {
             return Err(PyErr::new::<ValueError, _>(
                 py,
                 "Nodemap data buffer has an invalid memory representation"
                     .to_string(),
             ));
         };
 
         // Keep a reference to the mmap'ed buffer, otherwise we get a dangling
         // pointer.
         self.mmap(py).borrow_mut().replace(buf);
 
         let mut nt = NodeTree::load_bytes(Box::new(bytes), len);
 
         let data_tip = docket
             .getattr(py, "tip_rev")?
             .extract::<BaseRevision>(py)?
             .into();
         self.docket(py).borrow_mut().replace(docket.clone_ref(py));
         let idx = self.cindex(py).borrow();
         let data_tip = idx.check_revision(data_tip).ok_or_else(|| {
             nodemap_error(py, NodeMapError::RevisionNotInIndex(data_tip))
         })?;
         let current_tip = idx.len();
 
         for r in (data_tip.0 + 1)..current_tip as BaseRevision {
             let rev = Revision(r);
             // in this case node() won't ever return None
             nt.insert(&*idx, idx.node(rev).unwrap(), rev)
                 .map_err(|e| nodemap_error(py, e))?
         }
 
         *self.nt(py).borrow_mut() = Some(nt);
 
         Ok(py.None())
     }
 }
 
 fn revlog_error(py: Python) -> PyErr {
     match py
         .import("mercurial.error")
         .and_then(|m| m.get(py, "RevlogError"))
     {
         Err(e) => e,
         Ok(cls) => PyErr::from_instance(
             py,
             cls.call(py, (py.None(),), None).ok().into_py_object(py),
         ),
     }
 }
 
 fn rev_not_in_index(py: Python, rev: UncheckedRevision) -> PyErr {
     PyErr::new::<ValueError, _>(
         py,
         format!(
             "Inconsistency: Revision {} found in nodemap \
              is not in revlog index",
             rev
         ),
     )
 }
 
 /// Standard treatment of NodeMapError
 fn nodemap_error(py: Python, err: NodeMapError) -> PyErr {
     match err {
         NodeMapError::MultipleResults => revlog_error(py),
         NodeMapError::RevisionNotInIndex(r) => rev_not_in_index(py, r),
     }
 }
 
 /// Create the module, with __package__ given from parent
 pub fn init_module(py: Python, package: &str) -> PyResult<PyModule> {
     let dotted_name = &format!("{}.revlog", package);
     let m = PyModule::new(py, dotted_name)?;
     m.add(py, "__package__", package)?;
     m.add(py, "__doc__", "RevLog - Rust implementations")?;
 
     m.add_class::<MixedIndex>(py)?;
 
     let sys = PyModule::import(py, "sys")?;
     let sys_modules: PyDict = sys.get(py, "modules")?.extract(py)?;
     sys_modules.set_item(py, dotted_name, &m)?;
 
     Ok(m)
 }

rust/rhg/src/commands/cat.rs

@@ -1,116 +1,116 b''
 use crate::error::CommandError;
 use clap::Arg;
 use format_bytes::format_bytes;
 use hg::operations::cat;
 use hg::utils::hg_path::HgPathBuf;
 use std::ffi::OsString;
 use std::os::unix::prelude::OsStrExt;
 
 pub const HELP_TEXT: &str = "
 Output the current or given revision of files
 ";
 
 pub fn args() -> clap::Command {
     clap::command!("cat")
         .arg(
             Arg::new("rev")
                 .help("search the repository as it is in REV")
                 .short('r')
                 .long("rev")
                 .value_name("REV"),
         )
         .arg(
             clap::Arg::new("files")
                 .required(true)
                 .num_args(1..)
                 .value_name("FILE")
                 .value_parser(clap::value_parser!(std::ffi::OsString))
                 .help("Files to output"),
         )
         .about(HELP_TEXT)
 }
 
 #[logging_timer::time("trace")]
 pub fn run(invocation: &crate::CliInvocation) -> Result<(), CommandError> {
     let cat_enabled = invocation.config.get_bool(b"rhg", b"cat")?;
     if !cat_enabled {
         return Err(CommandError::unsupported(
             "cat is disabled in rhg (enable it with 'rhg.cat = true' \
             or enable fallback with 'rhg.on-unsupported = fallback')",
         ));
     }
 
     let rev = invocation.subcommand_args.get_one::<String>("rev");
     let file_args =
         match invocation.subcommand_args.get_many::<OsString>("files") {
             Some(files) => files
                 .filter(|s| !s.is_empty())
                 .map(|s| s.as_os_str())
                 .collect(),
             None => vec![],
         };
 
     let repo = invocation.repo?;
     let cwd = hg::utils::current_dir()?;
     let working_directory = repo.working_directory_path();
     let working_directory = cwd.join(working_directory); // Make it absolute
 
     let mut files = vec![];
     for file in file_args {
         if file.as_bytes().starts_with(b"set:") {
             let message = "fileset";
             return Err(CommandError::unsupported(message));
         }
 
-        let normalized = cwd.join(&file);
+        let normalized = cwd.join(file);
         // TODO: actually normalize `..` path segments etc?
         let dotted = normalized.components().any(|c| c.as_os_str() == "..");
         if file.as_bytes() == b"." || dotted {
             let message = "`..` or `.` path segment";
             return Err(CommandError::unsupported(message));
         }
         let relative_path = working_directory
             .strip_prefix(&cwd)
             .unwrap_or(&working_directory);
         let stripped = normalized
             .strip_prefix(&working_directory)
             .map_err(|_| {
                 CommandError::abort(format!(
                     "abort: {} not under root '{}'\n(consider using '--cwd {}')",
                     String::from_utf8_lossy(file.as_bytes()),
                     working_directory.display(),
                     relative_path.display(),
                 ))
             })?;
         let hg_file = HgPathBuf::try_from(stripped.to_path_buf())
             .map_err(|e| CommandError::abort(e.to_string()))?;
         files.push(hg_file);
     }
     let files = files.iter().map(|file| file.as_ref()).collect();
     // TODO probably move this to a util function like `repo.default_rev` or
     // something when it's used somewhere else
     let rev = match rev {
         Some(r) => r.to_string(),
         None => format!("{:x}", repo.dirstate_parents()?.p1),
     };
 
     let output = cat(repo, &rev, files).map_err(|e| (e, rev.as_str()))?;
     for (_file, contents) in output.results {
         invocation.ui.write_stdout(&contents)?;
     }
     if !output.missing.is_empty() {
         let short = format!("{:x}", output.node.short()).into_bytes();
         for path in &output.missing {
             invocation.ui.write_stderr(&format_bytes!(
                 b"{}: no such file in rev {}\n",
                 path.as_bytes(),
                 short
             ))?;
         }
     }
     if output.found_any {
         Ok(())
     } else {
         Err(CommandError::Unsuccessful)
     }
 }

rust/rhg/src/commands/debugdata.rs

@@ -1,71 +1,71 b''
 use crate::error::CommandError;
 use clap::Arg;
 use clap::ArgGroup;
 use hg::operations::{debug_data, DebugDataKind};
 
 pub const HELP_TEXT: &str = "
 Dump the contents of a data file revision
 ";
 
 pub fn args() -> clap::Command {
     clap::command!("debugdata")
         .arg(
             Arg::new("changelog")
                 .help("open changelog")
                 .short('c')
                 .action(clap::ArgAction::SetTrue),
         )
         .arg(
             Arg::new("manifest")
                 .help("open manifest")
                 .short('m')
                 .action(clap::ArgAction::SetTrue),
         )
         .group(
             ArgGroup::new("revlog")
-                .args(&["changelog", "manifest"])
+                .args(["changelog", "manifest"])
                 .required(true),
         )
         .arg(
             Arg::new("rev")
                 .help("revision")
                 .required(true)
                 .value_name("REV"),
         )
         .about(HELP_TEXT)
 }
 
 #[logging_timer::time("trace")]
 pub fn run(invocation: &crate::CliInvocation) -> Result<(), CommandError> {
     let args = invocation.subcommand_args;
     let rev = args
         .get_one::<String>("rev")
         .expect("rev should be a required argument");
     let kind = match (
         args.get_one::<bool>("changelog").unwrap(),
         args.get_one::<bool>("manifest").unwrap(),
     ) {
         (true, false) => DebugDataKind::Changelog,
         (false, true) => DebugDataKind::Manifest,
         (true, true) => {
             unreachable!("Should not happen since options are exclusive")
         }
         (false, false) => {
             unreachable!("Should not happen since options are required")
         }
     };
 
     let repo = invocation.repo?;
     if repo.has_narrow() {
         return Err(CommandError::unsupported(
             "support for ellipsis nodes is missing and repo has narrow enabled",
         ));
     }
     let data = debug_data(repo, rev, kind).map_err(|e| (e, rev.as_ref()))?;
 
     let mut stdout = invocation.ui.stdout_buffer();
     stdout.write_all(&data)?;
     stdout.flush()?;
 
     Ok(())
 }

rust/rhg/src/commands/root.rs

@@ -1,28 +1,28 b''
 use crate::error::CommandError;
 use format_bytes::format_bytes;
 use hg::errors::{IoErrorContext, IoResultExt};
 use hg::utils::files::get_bytes_from_path;
 
 pub const HELP_TEXT: &str = "
 Print the root directory of the current repository.
 
 Returns 0 on success.
 ";
 
 pub fn args() -> clap::Command {
     clap::command!("root").about(HELP_TEXT)
 }
 
 pub fn run(invocation: &crate::CliInvocation) -> Result<(), CommandError> {
     let repo = invocation.repo?;
     let working_directory = repo.working_directory_path();
     let working_directory = std::fs::canonicalize(working_directory)
         .with_context(|| {
             IoErrorContext::CanonicalizingPath(working_directory.to_owned())
         })?;
-    let bytes = get_bytes_from_path(&working_directory);
+    let bytes = get_bytes_from_path(working_directory);
     invocation
         .ui
         .write_stdout(&format_bytes!(b"{}\n", bytes.as_slice()))?;
     Ok(())
 }

rust/rhg/src/main.rs

@@ -1,842 +1,842 b''
 extern crate log;
 use crate::error::CommandError;
 use crate::ui::{local_to_utf8, Ui};
 use clap::{command, Arg, ArgMatches};
 use format_bytes::{format_bytes, join};
 use hg::config::{Config, ConfigSource, PlainInfo};
 use hg::repo::{Repo, RepoError};
 use hg::utils::files::{get_bytes_from_os_str, get_path_from_bytes};
 use hg::utils::SliceExt;
 use hg::{exit_codes, requirements};
 use std::borrow::Cow;
 use std::collections::HashSet;
 use std::ffi::OsString;
 use std::os::unix::prelude::CommandExt;
 use std::path::PathBuf;
 use std::process::Command;
 
 mod blackbox;
 mod color;
 mod error;
 mod ui;
 pub mod utils {
     pub mod path_utils;
 }
 
 fn main_with_result(
     argv: Vec<OsString>,
     process_start_time: &blackbox::ProcessStartTime,
     ui: &ui::Ui,
     repo: Result<&Repo, &NoRepoInCwdError>,
     config: &Config,
 ) -> Result<(), CommandError> {
     check_unsupported(config, repo)?;
 
     let app = command!()
         .subcommand_required(true)
         .arg(
             Arg::new("repository")
                 .help("repository root directory")
                 .short('R')
                 .value_name("REPO")
                 // Both ok: `hg -R ./foo log` or `hg log -R ./foo`
                 .global(true),
         )
         .arg(
             Arg::new("config")
                 .help("set/override config option (use 'section.name=value')")
                 .value_name("CONFIG")
                 .global(true)
                 .long("config")
                 // Ok: `--config section.key1=val --config section.key2=val2`
                 // Not ok: `--config section.key1=val section.key2=val2`
                 .action(clap::ArgAction::Append),
         )
         .arg(
             Arg::new("cwd")
                 .help("change working directory")
                 .value_name("DIR")
                 .long("cwd")
                 .global(true),
         )
         .arg(
             Arg::new("color")
                 .help("when to colorize (boolean, always, auto, never, or debug)")
                 .value_name("TYPE")
                 .long("color")
                 .global(true),
         )
         .version("0.0.1");
     let app = add_subcommand_args(app);
 
     let matches = app.try_get_matches_from(argv.iter())?;
 
     let (subcommand_name, subcommand_args) =
         matches.subcommand().expect("subcommand required");
 
     // Mercurial allows users to define "defaults" for commands, fallback
     // if a default is detected for the current command
     let defaults = config.get_str(b"defaults", subcommand_name.as_bytes())?;
     match defaults {
         // Programmatic usage might set defaults to an empty string to unset
         // it; allow that
         None | Some("") => {}
         Some(_) => {
             let msg = "`defaults` config set";
             return Err(CommandError::unsupported(msg));
         }
     }
 
     for prefix in ["pre", "post", "fail"].iter() {
         // Mercurial allows users to define generic hooks for commands,
         // fallback if any are detected
         let item = format!("{}-{}", prefix, subcommand_name);
         let hook_for_command =
             config.get_str_no_default(b"hooks", item.as_bytes())?;
         if hook_for_command.is_some() {
             let msg = format!("{}-{} hook defined", prefix, subcommand_name);
             return Err(CommandError::unsupported(msg));
         }
     }
     let run = subcommand_run_fn(subcommand_name)
         .expect("unknown subcommand name from clap despite Command::subcommand_required");
 
     let invocation = CliInvocation {
         ui,
         subcommand_args,
         config,
         repo,
     };
 
     if let Ok(repo) = repo {
         // We don't support subrepos, fallback if the subrepos file is present
         if repo.working_directory_vfs().join(".hgsub").exists() {
             let msg = "subrepos (.hgsub is present)";
             return Err(CommandError::unsupported(msg));
         }
     }
 
     if config.is_extension_enabled(b"blackbox") {
         let blackbox =
             blackbox::Blackbox::new(&invocation, process_start_time)?;
         blackbox.log_command_start(argv.iter());
         let result = run(&invocation);
         blackbox.log_command_end(
             argv.iter(),
             exit_code(
                 &result,
                 // TODO: show a warning or combine with original error if
                 // `get_bool` returns an error
                 config
                     .get_bool(b"ui", b"detailed-exit-code")
                     .unwrap_or(false),
             ),
         );
         result
     } else {
         run(&invocation)
     }
 }
 
 fn rhg_main(argv: Vec<OsString>) -> ! {
     // Run this first, before we find out if the blackbox extension is even
     // enabled, in order to include everything in-between in the duration
     // measurements. Reading config files can be slow if they’re on NFS.
     let process_start_time = blackbox::ProcessStartTime::now();
 
     env_logger::init();
 
     // Make sure nothing in a future version of `rhg` sets the global
     // threadpool before we can cap default threads. (This is also called
     // in core because Python uses the same code path, we're adding a
     // redundant check.)
     hg::utils::cap_default_rayon_threads()
         .expect("Rayon threadpool already initialized");
 
     let early_args = EarlyArgs::parse(&argv);
 
     let initial_current_dir = early_args.cwd.map(|cwd| {
         let cwd = get_path_from_bytes(&cwd);
         std::env::current_dir()
             .and_then(|initial| {
                 std::env::set_current_dir(cwd)?;
                 Ok(initial)
             })
             .unwrap_or_else(|error| {
                 exit(
                     &argv,
                     &None,
                     &Ui::new_infallible(&Config::empty()),
                     OnUnsupported::Abort,
                     Err(CommandError::abort(format!(
                         "abort: {}: '{}'",
                         error,
                         cwd.display()
                     ))),
                     false,
                 )
             })
     });
 
     let mut non_repo_config =
         Config::load_non_repo().unwrap_or_else(|error| {
             // Normally this is decided based on config, but we don’t have that
             // available. As of this writing config loading never returns an
             // "unsupported" error but that is not enforced by the type system.
             let on_unsupported = OnUnsupported::Abort;
 
             exit(
                 &argv,
                 &initial_current_dir,
                 &Ui::new_infallible(&Config::empty()),
                 on_unsupported,
                 Err(error.into()),
                 false,
             )
         });
 
     non_repo_config
         .load_cli_args(early_args.config, early_args.color)
         .unwrap_or_else(|error| {
             exit(
                 &argv,
                 &initial_current_dir,
                 &Ui::new_infallible(&non_repo_config),
                 OnUnsupported::from_config(&non_repo_config),
                 Err(error.into()),
                 non_repo_config
                     .get_bool(b"ui", b"detailed-exit-code")
                     .unwrap_or(false),
             )
         });
 
     if let Some(repo_path_bytes) = &early_args.repo {
         lazy_static::lazy_static! {
             static ref SCHEME_RE: regex::bytes::Regex =
                 // Same as `_matchscheme` in `mercurial/util.py`
                 regex::bytes::Regex::new("^[a-zA-Z0-9+.\\-]+:").unwrap();
         }
         if SCHEME_RE.is_match(repo_path_bytes) {
             exit(
                 &argv,
                 &initial_current_dir,
                 &Ui::new_infallible(&non_repo_config),
                 OnUnsupported::from_config(&non_repo_config),
                 Err(CommandError::UnsupportedFeature {
                     message: format_bytes!(
                         b"URL-like --repository {}",
                         repo_path_bytes
                     ),
                 }),
                 // TODO: show a warning or combine with original error if
                 // `get_bool` returns an error
                 non_repo_config
                     .get_bool(b"ui", b"detailed-exit-code")
                     .unwrap_or(false),
             )
         }
     }
     let repo_arg = early_args.repo.unwrap_or_default();
     let repo_path: Option<PathBuf> = {
         if repo_arg.is_empty() {
             None
         } else {
             let local_config = {
                 if std::env::var_os("HGRCSKIPREPO").is_none() {
                     // TODO: handle errors from find_repo_root
                     if let Ok(current_dir_path) = Repo::find_repo_root() {
                         let config_files = vec![
                             ConfigSource::AbsPath(
                                 current_dir_path.join(".hg/hgrc"),
                             ),
                             ConfigSource::AbsPath(
                                 current_dir_path.join(".hg/hgrc-not-shared"),
                             ),
                         ];
                         // TODO: handle errors from
                         // `load_from_explicit_sources`
                         Config::load_from_explicit_sources(config_files).ok()
                     } else {
                         None
                     }
                 } else {
                     None
                 }
             };
 
             let non_repo_config_val = {
                 let non_repo_val = non_repo_config.get(b"paths", &repo_arg);
                 match &non_repo_val {
                     Some(val) if !val.is_empty() => home::home_dir()
                         .unwrap_or_else(|| PathBuf::from("~"))
                         .join(get_path_from_bytes(val))
                         .canonicalize()
                         // TODO: handle error and make it similar to python
                         // implementation maybe?
                         .ok(),
                     _ => None,
                 }
             };
 
             let config_val = match &local_config {
                 None => non_repo_config_val,
                 Some(val) => {
                     let local_config_val = val.get(b"paths", &repo_arg);
                     match &local_config_val {
                         Some(val) if !val.is_empty() => {
                             // presence of a local_config assures that
                             // current_dir
                             // wont result in an Error
                             let canpath = hg::utils::current_dir()
                                 .unwrap()
                                 .join(get_path_from_bytes(val))
                                 .canonicalize();
                             canpath.ok().or(non_repo_config_val)
                         }
                         _ => non_repo_config_val,
                     }
                 }
             };
             config_val
                 .or_else(|| Some(get_path_from_bytes(&repo_arg).to_path_buf()))
         }
     };
 
     let simple_exit =
         |ui: &Ui, config: &Config, result: Result<(), CommandError>| -> ! {
             exit(
                 &argv,
                 &initial_current_dir,
                 ui,
                 OnUnsupported::from_config(config),
                 result,
                 // TODO: show a warning or combine with original error if
                 // `get_bool` returns an error
                 non_repo_config
                     .get_bool(b"ui", b"detailed-exit-code")
                     .unwrap_or(false),
             )
         };
     let early_exit = |config: &Config, error: CommandError| -> ! {
         simple_exit(&Ui::new_infallible(config), config, Err(error))
     };
     let repo_result = match Repo::find(&non_repo_config, repo_path.to_owned())
     {
         Ok(repo) => Ok(repo),
         Err(RepoError::NotFound { at }) if repo_path.is_none() => {
             // Not finding a repo is not fatal yet, if `-R` was not given
             Err(NoRepoInCwdError { cwd: at })
         }
         Err(error) => early_exit(&non_repo_config, error.into()),
     };
 
     let config = if let Ok(repo) = &repo_result {
         repo.config()
     } else {
         &non_repo_config
     };
 
     let mut config_cow = Cow::Borrowed(config);
     config_cow.to_mut().apply_plain(PlainInfo::from_env());
     if !ui::plain(Some("tweakdefaults"))
         && config_cow
             .as_ref()
             .get_bool(b"ui", b"tweakdefaults")
             .unwrap_or_else(|error| early_exit(config, error.into()))
     {
         config_cow.to_mut().tweakdefaults()
     };
     let config = config_cow.as_ref();
     let ui = Ui::new(config)
         .unwrap_or_else(|error| early_exit(config, error.into()));
 
     if let Ok(true) = config.get_bool(b"rhg", b"fallback-immediately") {
         exit(
             &argv,
             &initial_current_dir,
             &ui,
             OnUnsupported::fallback(config),
             Err(CommandError::unsupported(
                 "`rhg.fallback-immediately is true`",
             )),
             false,
         )
     }
 
     let result = main_with_result(
         argv.iter().map(|s| s.to_owned()).collect(),
         &process_start_time,
         &ui,
         repo_result.as_ref(),
         config,
     );
     simple_exit(&ui, config, result)
 }
 
 fn main() -> ! {
     rhg_main(std::env::args_os().collect())
 }
 
 fn exit_code(
     result: &Result<(), CommandError>,
     use_detailed_exit_code: bool,
 ) -> i32 {
     match result {
         Ok(()) => exit_codes::OK,
         Err(CommandError::Abort {
             detailed_exit_code, ..
         }) => {
             if use_detailed_exit_code {
                 *detailed_exit_code
             } else {
                 exit_codes::ABORT
             }
         }
         Err(CommandError::Unsuccessful) => exit_codes::UNSUCCESSFUL,
         // Exit with a specific code and no error message to let a potential
         // wrapper script fallback to Python-based Mercurial.
         Err(CommandError::UnsupportedFeature { .. }) => {
             exit_codes::UNIMPLEMENTED
         }
         Err(CommandError::InvalidFallback { .. }) => {
             exit_codes::INVALID_FALLBACK
         }
     }
 }
 
-fn exit<'a>(
-    original_args: &'a [OsString],
+fn exit(
+    original_args: &[OsString],
     initial_current_dir: &Option<PathBuf>,
     ui: &Ui,
     mut on_unsupported: OnUnsupported,
     result: Result<(), CommandError>,
     use_detailed_exit_code: bool,
 ) -> ! {
     if let (
         OnUnsupported::Fallback { executable },
         Err(CommandError::UnsupportedFeature { message }),
     ) = (&on_unsupported, &result)
     {
         let mut args = original_args.iter();
         let executable = match executable {
             None => {
                 exit_no_fallback(
                     ui,
                     OnUnsupported::Abort,
                     Err(CommandError::abort(
                         "abort: 'rhg.on-unsupported=fallback' without \
                                 'rhg.fallback-executable' set.",
                     )),
                     false,
                 );
             }
             Some(executable) => executable,
         };
         let executable_path = get_path_from_bytes(executable);
         let this_executable = args.next().expect("exepcted argv[0] to exist");
         if executable_path == *this_executable {
             // Avoid spawning infinitely many processes until resource
             // exhaustion.
             let _ = ui.write_stderr(&format_bytes!(
                 b"Blocking recursive fallback. The 'rhg.fallback-executable = {}' config \
                 points to `rhg` itself.\n",
                 executable
             ));
             on_unsupported = OnUnsupported::Abort
         } else {
             log::debug!("falling back (see trace-level log)");
             log::trace!("{}", local_to_utf8(message));
             if let Err(err) = which::which(executable_path) {
                 exit_no_fallback(
                     ui,
                     OnUnsupported::Abort,
                     Err(CommandError::InvalidFallback {
                         path: executable.to_owned(),
                         err: err.to_string(),
                     }),
                     use_detailed_exit_code,
                 )
             }
             // `args` is now `argv[1..]` since we’ve already consumed
             // `argv[0]`
             let mut command = Command::new(executable_path);
             command.args(args);
             if let Some(initial) = initial_current_dir {
                 command.current_dir(initial);
             }
             // We don't use subprocess because proper signal handling is harder
             // and we don't want to keep `rhg` around after a fallback anyway.
             // For example, if `rhg` is run in the background and falls back to
             // `hg` which, in turn, waits for a signal, we'll get stuck if
             // we're doing plain subprocess.
             //
             // If `exec` returns, we can only assume our process is very broken
             // (see its documentation), so only try to forward the error code
             // when exiting.
             let err = command.exec();
             std::process::exit(
                 err.raw_os_error().unwrap_or(exit_codes::ABORT),
             );
         }
     }
     exit_no_fallback(ui, on_unsupported, result, use_detailed_exit_code)
 }
 
 fn exit_no_fallback(
     ui: &Ui,
     on_unsupported: OnUnsupported,
     result: Result<(), CommandError>,
     use_detailed_exit_code: bool,
 ) -> ! {
     match &result {
         Ok(_) => {}
         Err(CommandError::Unsuccessful) => {}
         Err(CommandError::Abort { message, hint, .. }) => {
             // Ignore errors when writing to stderr, we’re already exiting
             // with failure code so there’s not much more we can do.
             if !message.is_empty() {
                 let _ = ui.write_stderr(&format_bytes!(b"{}\n", message));
             }
             if let Some(hint) = hint {
                 let _ = ui.write_stderr(&format_bytes!(b"({})\n", hint));
             }
         }
         Err(CommandError::UnsupportedFeature { message }) => {
             match on_unsupported {
                 OnUnsupported::Abort => {
                     let _ = ui.write_stderr(&format_bytes!(
                         b"unsupported feature: {}\n",
                         message
                     ));
                 }
                 OnUnsupported::AbortSilent => {}
                 OnUnsupported::Fallback { .. } => unreachable!(),
             }
         }
         Err(CommandError::InvalidFallback { path, err }) => {
             let _ = ui.write_stderr(&format_bytes!(
                 b"abort: invalid fallback '{}': {}\n",
                 path,
                 err.as_bytes(),
             ));
         }
     }
     std::process::exit(exit_code(&result, use_detailed_exit_code))
 }
 
 macro_rules! subcommands {
     ($( $command: ident )+) => {
         mod commands {
             $(
                 pub mod $command;
             )+
         }
 
         fn add_subcommand_args(app: clap::Command) -> clap::Command {
             app
             $(
                 .subcommand(commands::$command::args())
             )+
         }
 
         pub type RunFn = fn(&CliInvocation) -> Result<(), CommandError>;
 
         fn subcommand_run_fn(name: &str) -> Option<RunFn> {
             match name {
                 $(
                     stringify!($command) => Some(commands::$command::run),
                 )+
                 _ => None,
             }
         }
     };
 }
 
 subcommands! {
     cat
     debugdata
     debugrequirements
     debugignorerhg
     debugrhgsparse
     files
     root
     config
     status
 }
 
 pub struct CliInvocation<'a> {
     ui: &'a Ui,
     subcommand_args: &'a ArgMatches,
     config: &'a Config,
     /// References inside `Result` is a bit peculiar but allow
     /// `invocation.repo?` to work out with `&CliInvocation` since this
     /// `Result` type is `Copy`.
     repo: Result<&'a Repo, &'a NoRepoInCwdError>,
 }
 
 struct NoRepoInCwdError {
     cwd: PathBuf,
 }
 
 /// CLI arguments to be parsed "early" in order to be able to read
 /// configuration before using Clap. Ideally we would also use Clap for this,
 /// see <https://github.com/clap-rs/clap/discussions/2366>.
 ///
 /// These arguments are still declared when we do use Clap later, so that Clap
 /// does not return an error for their presence.
 struct EarlyArgs {
     /// Values of all `--config` arguments. (Possibly none)
     config: Vec<Vec<u8>>,
     /// Value of all the `--color` argument, if any.
     color: Option<Vec<u8>>,
     /// Value of the `-R` or `--repository` argument, if any.
     repo: Option<Vec<u8>>,
     /// Value of the `--cwd` argument, if any.
     cwd: Option<Vec<u8>>,
 }
 
 impl EarlyArgs {
     fn parse<'a>(args: impl IntoIterator<Item = &'a OsString>) -> Self {
         let mut args = args.into_iter().map(get_bytes_from_os_str);
         let mut config = Vec::new();
         let mut color = None;
         let mut repo = None;
         let mut cwd = None;
         // Use `while let` instead of `for` so that we can also call
         // `args.next()` inside the loop.
         while let Some(arg) = args.next() {
             if arg == b"--config" {
                 if let Some(value) = args.next() {
                     config.push(value)
                 }
             } else if let Some(value) = arg.drop_prefix(b"--config=") {
                 config.push(value.to_owned())
             }
 
             if arg == b"--color" {
                 if let Some(value) = args.next() {
                     color = Some(value)
                 }
             } else if let Some(value) = arg.drop_prefix(b"--color=") {
                 color = Some(value.to_owned())
             }
 
             if arg == b"--cwd" {
                 if let Some(value) = args.next() {
                     cwd = Some(value)
                 }
             } else if let Some(value) = arg.drop_prefix(b"--cwd=") {
                 cwd = Some(value.to_owned())
             }
 
             if arg == b"--repository" || arg == b"-R" {
                 if let Some(value) = args.next() {
                     repo = Some(value)
                 }
             } else if let Some(value) = arg.drop_prefix(b"--repository=") {
                 repo = Some(value.to_owned())
             } else if let Some(value) = arg.drop_prefix(b"-R") {
                 repo = Some(value.to_owned())
             }
         }
         Self {
             config,
             color,
             repo,
             cwd,
         }
     }
 }
 
 /// What to do when encountering some unsupported feature.
 ///
 /// See `HgError::UnsupportedFeature` and `CommandError::UnsupportedFeature`.
 enum OnUnsupported {
     /// Print an error message describing what feature is not supported,
     /// and exit with code 252.
     Abort,
     /// Silently exit with code 252.
     AbortSilent,
     /// Try running a Python implementation
     Fallback { executable: Option<Vec<u8>> },
 }
 
 impl OnUnsupported {
     const DEFAULT: Self = OnUnsupported::Abort;
 
     fn fallback_executable(config: &Config) -> Option<Vec<u8>> {
         config
             .get(b"rhg", b"fallback-executable")
             .map(|x| x.to_owned())
     }
 
     fn fallback(config: &Config) -> Self {
         OnUnsupported::Fallback {
             executable: Self::fallback_executable(config),
         }
     }
 
     fn from_config(config: &Config) -> Self {
         match config
             .get(b"rhg", b"on-unsupported")
             .map(|value| value.to_ascii_lowercase())
             .as_deref()
         {
             Some(b"abort") => OnUnsupported::Abort,
             Some(b"abort-silent") => OnUnsupported::AbortSilent,
             Some(b"fallback") => Self::fallback(config),
             None => Self::DEFAULT,
             Some(_) => {
                 // TODO: warn about unknown config value
                 Self::DEFAULT
             }
         }
     }
 }
 
 /// The `*` extension is an edge-case for config sub-options that apply to all
 /// extensions. For now, only `:required` exists, but that may change in the
 /// future.
 const SUPPORTED_EXTENSIONS: &[&[u8]] = &[
     b"blackbox",
     b"share",
     b"sparse",
     b"narrow",
     b"*",
     b"strip",
     b"rebase",
 ];
 
 fn check_extensions(config: &Config) -> Result<(), CommandError> {
     if let Some(b"*") = config.get(b"rhg", b"ignored-extensions") {
         // All extensions are to be ignored, nothing to do here
         return Ok(());
     }
 
     let enabled: HashSet<&[u8]> = config
         .iter_section(b"extensions")
         .filter_map(|(extension, value)| {
             if value == b"!" {
                 // Filter out disabled extensions
                 return None;
             }
             // Ignore extension suboptions. Only `required` exists for now.
             // `rhg` either supports an extension or doesn't, so it doesn't
             // make sense to consider the loading of an extension.
             let actual_extension =
                 extension.split_2(b':').unwrap_or((extension, b"")).0;
             Some(actual_extension)
         })
         .collect();
 
     let mut unsupported = enabled;
     for supported in SUPPORTED_EXTENSIONS {
         unsupported.remove(supported);
     }
 
     if let Some(ignored_list) = config.get_list(b"rhg", b"ignored-extensions")
     {
         for ignored in ignored_list {
             unsupported.remove(ignored.as_slice());
         }
     }
 
     if unsupported.is_empty() {
         Ok(())
     } else {
         let mut unsupported: Vec<_> = unsupported.into_iter().collect();
         // Sort the extensions to get a stable output
         unsupported.sort();
         Err(CommandError::UnsupportedFeature {
             message: format_bytes!(
                 b"extensions: {} (consider adding them to 'rhg.ignored-extensions' config)",
                 join(unsupported, b", ")
             ),
         })
     }
 }
 
 /// Array of tuples of (auto upgrade conf, feature conf, local requirement)
 #[allow(clippy::type_complexity)]
 const AUTO_UPGRADES: &[((&str, &str), (&str, &str), &str)] = &[
     (
         ("format", "use-share-safe.automatic-upgrade-of-mismatching-repositories"),
         ("format", "use-share-safe"),
         requirements::SHARESAFE_REQUIREMENT,
     ),
     (
         ("format", "use-dirstate-tracked-hint.automatic-upgrade-of-mismatching-repositories"),
         ("format", "use-dirstate-tracked-hint"),
         requirements::DIRSTATE_TRACKED_HINT_V1,
     ),
     (
         ("format", "use-dirstate-v2.automatic-upgrade-of-mismatching-repositories"),
         ("format", "use-dirstate-v2"),
         requirements::DIRSTATE_V2_REQUIREMENT,
     ),
 ];
 
 /// Mercurial allows users to automatically upgrade their repository.
 /// `rhg` does not have the ability to upgrade yet, so fallback if an upgrade
 /// is needed.
 fn check_auto_upgrade(
     config: &Config,
     reqs: &HashSet<String>,
 ) -> Result<(), CommandError> {
     for (upgrade_conf, feature_conf, local_req) in AUTO_UPGRADES.iter() {
         let auto_upgrade = config
             .get_bool(upgrade_conf.0.as_bytes(), upgrade_conf.1.as_bytes())?;
 
         if auto_upgrade {
             let want_it = config.get_bool(
                 feature_conf.0.as_bytes(),
                 feature_conf.1.as_bytes(),
             )?;
             let have_it = reqs.contains(*local_req);
 
             let action = match (want_it, have_it) {
                 (true, false) => Some("upgrade"),
                 (false, true) => Some("downgrade"),
                 _ => None,
             };
             if let Some(action) = action {
                 let message = format!(
                     "automatic {} {}.{}",
                     action, upgrade_conf.0, upgrade_conf.1
                 );
                 return Err(CommandError::unsupported(message));
             }
         }
     }
     Ok(())
 }
 
 fn check_unsupported(
     config: &Config,
     repo: Result<&Repo, &NoRepoInCwdError>,
 ) -> Result<(), CommandError> {
     check_extensions(config)?;
 
     if std::env::var_os("HG_PENDING").is_some() {
         // TODO: only if the value is `== repo.working_directory`?
         // What about relative v.s. absolute paths?
         Err(CommandError::unsupported("$HG_PENDING"))?
     }
 
     if let Ok(repo) = repo {
         if repo.has_subrepos()? {
             Err(CommandError::unsupported("sub-repositories"))?
         }
         check_auto_upgrade(config, repo.requirements())?;
     }
 
     if config.has_non_empty_section(b"encode") {
         Err(CommandError::unsupported("[encode] config"))?
     }
 
     if config.has_non_empty_section(b"decode") {
         Err(CommandError::unsupported("[decode] config"))?
     }
 
     Ok(())
 }

rust/rhg/src/ui.rs

@@ -1,307 +1,307 b''
 use crate::color::ColorConfig;
 use crate::color::Effect;
 use crate::error::CommandError;
 use format_bytes::format_bytes;
 use format_bytes::write_bytes;
 use hg::config::Config;
 use hg::config::PlainInfo;
 use hg::errors::HgError;
 use hg::repo::Repo;
 use hg::sparse;
 use hg::utils::files::get_bytes_from_path;
 use hg::PatternFileWarning;
 use std::borrow::Cow;
 use std::io;
 use std::io::{ErrorKind, Write};
 
 pub struct Ui {
     stdout: std::io::Stdout,
     stderr: std::io::Stderr,
     colors: Option<ColorConfig>,
 }
 
 /// The kind of user interface error
 pub enum UiError {
     /// The standard output stream cannot be written to
     StdoutError(io::Error),
     /// The standard error stream cannot be written to
     StderrError(io::Error),
 }
 
 /// The commandline user interface
 impl Ui {
     pub fn new(config: &Config) -> Result<Self, HgError> {
         Ok(Ui {
             // If using something else, also adapt `isatty()` below.
             stdout: std::io::stdout(),
 
             stderr: std::io::stderr(),
             colors: ColorConfig::new(config)?,
         })
     }
 
     /// Default to no color if color configuration errors.
     ///
     /// Useful when we’re already handling another error.
     pub fn new_infallible(config: &Config) -> Self {
         Ui {
             // If using something else, also adapt `isatty()` below.
             stdout: std::io::stdout(),
 
             stderr: std::io::stderr(),
             colors: ColorConfig::new(config).unwrap_or(None),
         }
     }
 
     /// Returns a buffered handle on stdout for faster batch printing
     /// operations.
     pub fn stdout_buffer(&self) -> StdoutBuffer<std::io::StdoutLock> {
         StdoutBuffer::new(self.stdout.lock())
     }
 
     /// Write bytes to stdout
     pub fn write_stdout(&self, bytes: &[u8]) -> Result<(), UiError> {
         let mut stdout = self.stdout.lock();
 
         stdout.write_all(bytes).or_else(handle_stdout_error)?;
 
         stdout.flush().or_else(handle_stdout_error)
     }
 
     /// Write bytes to stderr
     pub fn write_stderr(&self, bytes: &[u8]) -> Result<(), UiError> {
         let mut stderr = self.stderr.lock();
 
         stderr.write_all(bytes).or_else(handle_stderr_error)?;
 
         stderr.flush().or_else(handle_stderr_error)
     }
 
     /// Write bytes to stdout with the given label
     ///
     /// Like the optional `label` parameter in `mercurial/ui.py`,
     /// this label influences the color used for this output.
     pub fn write_stdout_labelled(
         &self,
         bytes: &[u8],
         label: &str,
     ) -> Result<(), UiError> {
         if let Some(colors) = &self.colors {
             if let Some(effects) = colors.styles.get(label.as_bytes()) {
                 if !effects.is_empty() {
                     return self
                         .write_stdout_with_effects(bytes, effects)
                         .or_else(handle_stdout_error);
                 }
             }
         }
         self.write_stdout(bytes)
     }
 
     fn write_stdout_with_effects(
         &self,
         bytes: &[u8],
         effects: &[Effect],
     ) -> io::Result<()> {
         let stdout = &mut self.stdout.lock();
         let mut write_line = |line: &[u8], first: bool| {
             // `line` does not include the newline delimiter
             if !first {
                 stdout.write_all(b"\n")?;
             }
             if line.is_empty() {
                 return Ok(());
             }
             /// 0x1B == 27 == 0o33
             const ASCII_ESCAPE: &[u8] = b"\x1b";
             write_bytes!(stdout, b"{}[0", ASCII_ESCAPE)?;
             for effect in effects {
                 write_bytes!(stdout, b";{}", effect)?;
             }
             write_bytes!(stdout, b"m")?;
             stdout.write_all(line)?;
             write_bytes!(stdout, b"{}[0m", ASCII_ESCAPE)
         };
         let mut lines = bytes.split(|&byte| byte == b'\n');
         if let Some(first) = lines.next() {
             write_line(first, true)?;
             for line in lines {
                 write_line(line, false)?
             }
         }
         stdout.flush()
     }
 }
 
 // TODO: pass the PlainInfo to call sites directly and
 // delete this function
 pub fn plain(opt_feature: Option<&str>) -> bool {
     let plain_info = PlainInfo::from_env();
     match opt_feature {
         None => plain_info.is_plain(),
         Some(feature) => plain_info.is_feature_plain(feature),
     }
 }
 
 /// A buffered stdout writer for faster batch printing operations.
 pub struct StdoutBuffer<W: Write> {
     buf: io::BufWriter<W>,
 }
 
 impl<W: Write> StdoutBuffer<W> {
     pub fn new(writer: W) -> Self {
         let buf = io::BufWriter::new(writer);
         Self { buf }
     }
 
     /// Write bytes to stdout buffer
     pub fn write_all(&mut self, bytes: &[u8]) -> Result<(), UiError> {
         self.buf.write_all(bytes).or_else(handle_stdout_error)
     }
 
     /// Flush bytes to stdout
     pub fn flush(&mut self) -> Result<(), UiError> {
         self.buf.flush().or_else(handle_stdout_error)
     }
 }
 
 /// Sometimes writing to stdout is not possible, try writing to stderr to
 /// signal that failure, otherwise just bail.
 fn handle_stdout_error(error: io::Error) -> Result<(), UiError> {
     if let ErrorKind::BrokenPipe = error.kind() {
         // This makes `| head` work for example
         return Ok(());
     }
     let mut stderr = io::stderr();
 
     stderr
         .write_all(&format_bytes!(
             b"abort: {}\n",
             error.to_string().as_bytes()
         ))
         .map_err(UiError::StderrError)?;
 
     stderr.flush().map_err(UiError::StderrError)?;
 
     Err(UiError::StdoutError(error))
 }
 
 /// Sometimes writing to stderr is not possible.
 fn handle_stderr_error(error: io::Error) -> Result<(), UiError> {
     // A broken pipe should not result in a error
     // like with `| head` for example
     if let ErrorKind::BrokenPipe = error.kind() {
         return Ok(());
     }
     Err(UiError::StdoutError(error))
 }
 
 /// Encode rust strings according to the user system.
 pub fn utf8_to_local(s: &str) -> Cow<[u8]> {
     // TODO encode for the user's system //
     let bytes = s.as_bytes();
     Cow::Borrowed(bytes)
 }
 
 /// Decode user system bytes to Rust string.
 pub fn local_to_utf8(s: &[u8]) -> Cow<str> {
     // TODO decode from the user's system
     String::from_utf8_lossy(s)
 }
 
 /// Should formatted output be used?
 ///
 /// Note: rhg does not have the formatter mechanism yet,
 /// but this is also used when deciding whether to use color.
 pub fn formatted(config: &Config) -> Result<bool, HgError> {
     if let Some(formatted) = config.get_option(b"ui", b"formatted")? {
         Ok(formatted)
     } else {
         isatty(config)
     }
 }
 
 pub enum RelativePaths {
     Legacy,
     Bool(bool),
 }
 
 pub fn relative_paths(config: &Config) -> Result<RelativePaths, HgError> {
     Ok(match config.get(b"ui", b"relative-paths") {
         None | Some(b"legacy") => RelativePaths::Legacy,
         _ => RelativePaths::Bool(config.get_bool(b"ui", b"relative-paths")?),
     })
 }
 
 fn isatty(config: &Config) -> Result<bool, HgError> {
     Ok(if config.get_bool(b"ui", b"nontty")? {
         false
     } else {
         atty::is(atty::Stream::Stdout)
     })
 }
 
 /// Return the formatted bytestring corresponding to a pattern file warning,
 /// as expected by the CLI.
 pub(crate) fn format_pattern_file_warning(
     warning: &PatternFileWarning,
     repo: &Repo,
 ) -> Vec<u8> {
     match warning {
         PatternFileWarning::InvalidSyntax(path, syntax) => format_bytes!(
             b"{}: ignoring invalid syntax '{}'\n",
             get_bytes_from_path(path),
-            &*syntax
+            syntax
         ),
         PatternFileWarning::NoSuchFile(path) => {
             let path = if let Ok(relative) =
                 path.strip_prefix(repo.working_directory_path())
             {
                 relative
             } else {
-                &*path
+                path
             };
             format_bytes!(
                 b"skipping unreadable pattern file '{}': \
                     No such file or directory\n",
                 get_bytes_from_path(path),
             )
         }
     }
 }
 
 /// Print with `Ui` the formatted bytestring corresponding to a
 /// sparse/narrow warning, as expected by the CLI.
 pub(crate) fn print_narrow_sparse_warnings(
     narrow_warnings: &[sparse::SparseWarning],
     sparse_warnings: &[sparse::SparseWarning],
     ui: &Ui,
     repo: &Repo,
 ) -> Result<(), CommandError> {
     for warning in narrow_warnings.iter().chain(sparse_warnings) {
         match &warning {
             sparse::SparseWarning::RootWarning { context, line } => {
                 let msg = format_bytes!(
                     b"warning: {} profile cannot use paths \"
                 starting with /, ignoring {}\n",
                     context,
                     line
                 );
                 ui.write_stderr(&msg)?;
             }
             sparse::SparseWarning::ProfileNotFound { profile, rev } => {
                 let msg = format_bytes!(
                     b"warning: sparse profile '{}' not found \"
                 in rev {} - ignoring it\n",
                     profile,
                     rev
                 );
                 ui.write_stderr(&msg)?;
             }
             sparse::SparseWarning::Pattern(e) => {
                 ui.write_stderr(&format_pattern_file_warning(e, repo))?;
             }
         }
     }
     Ok(())
 }