rustdoc: fixed warnings about links

This is the minimal fix making those that actually were supposed
to be links to work (including in private items).

rust/hg-core/src/checkexec.rs

@@ -1,119 +1,119 b''
 use std::fs;
 use std::io;
 use std::os::unix::fs::{MetadataExt, PermissionsExt};
 use std::path::Path;
 
 const EXECFLAGS: u32 = 0o111;
 
 fn is_executable(path: impl AsRef<Path>) -> Result<bool, io::Error> {
     let metadata = fs::metadata(path)?;
     let mode = metadata.mode();
     Ok(mode & EXECFLAGS != 0)
 }
 
 fn make_executable(path: impl AsRef<Path>) -> Result<(), io::Error> {
     let mode = fs::metadata(path.as_ref())?.mode();
     fs::set_permissions(
         path,
         fs::Permissions::from_mode((mode & 0o777) | EXECFLAGS),
     )?;
     Ok(())
 }
 
 fn copy_mode(
     src: impl AsRef<Path>,
     dst: impl AsRef<Path>,
 ) -> Result<(), io::Error> {
     let mode = match fs::symlink_metadata(src) {
         Ok(metadata) => metadata.mode(),
         Err(e) if e.kind() == io::ErrorKind::NotFound =>
         // copymode in python has a more complicated handling of FileNotFound
         // error, which we don't need because all it does is applying
         // umask, which the OS already does when we mkdir.
         {
             return Ok(())
         }
         Err(e) => return Err(e),
     };
     fs::set_permissions(dst, fs::Permissions::from_mode(mode))?;
     Ok(())
 }
 
 fn check_exec_impl(path: impl AsRef<Path>) -> Result<bool, io::Error> {
     let basedir = path.as_ref().join(".hg");
     let cachedir = basedir.join("wcache");
     let storedir = basedir.join("store");
 
     if !cachedir.exists() {
         // we want to create the 'cache' directory, not the '.hg' one.
         // Automatically creating '.hg' directory could silently spawn
         // invalid Mercurial repositories. That seems like a bad idea.
         fs::create_dir(&cachedir)
             .and_then(|()| {
                 if storedir.exists() {
                     copy_mode(&storedir, &cachedir)
                 } else {
                     copy_mode(&basedir, &cachedir)
                 }
             })
             .ok();
     }
 
     let leave_file: bool;
     let checkdir: &Path;
     let checkisexec = cachedir.join("checkisexec");
     let checknoexec = cachedir.join("checknoexec");
     if cachedir.is_dir() {
         // Check if both files already exist in cache and have correct
         // permissions. if so, we assume that permissions work.
         // If not, we delete the files and try again.
         match is_executable(&checkisexec) {
             Err(e) if e.kind() == io::ErrorKind::NotFound => (),
             Err(e) => return Err(e),
             Ok(is_exec) => {
                 if is_exec {
                     let noexec_is_exec = match is_executable(&checknoexec) {
                         Err(e) if e.kind() == io::ErrorKind::NotFound => {
                             fs::write(&checknoexec, "")?;
                             is_executable(&checknoexec)?
                         }
                         Err(e) => return Err(e),
                         Ok(exec) => exec,
                     };
                     if !noexec_is_exec {
                         // check-exec is exec and check-no-exec is not exec
                         return Ok(true);
                     }
                     fs::remove_file(&checknoexec)?;
                 }
                 fs::remove_file(&checkisexec)?;
             }
         }
         checkdir = &cachedir;
         leave_file = true;
     } else {
         // no cache directory (probably because .hg doesn't exist):
         // check directly in `path` and don't leave the temp file behind
         checkdir = path.as_ref();
         leave_file = false;
     };
 
     let tmp_file = tempfile::NamedTempFile::new_in(checkdir)?;
     if !is_executable(tmp_file.path())? {
         make_executable(tmp_file.path())?;
         if is_executable(tmp_file.path())? {
             if leave_file {
                 tmp_file.persist(checkisexec).ok();
             }
             return Ok(true);
         }
     }
 
     Ok(false)
 }
 
-/// This function is a rust rewrite of [checkexec] function from [posix.py]
+/// This function is a rust rewrite of `checkexec` function from `posix.py`
 /// Returns true if the filesystem supports execute permissions.
 pub fn check_exec(path: impl AsRef<Path>) -> bool {
     check_exec_impl(path).unwrap_or(false)
 }

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

@@ -1,1069 +1,1069 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 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 {
     MultipleResults,
     /// A `Revision` stored in the nodemap could not be found in the index
     RevisionNotInIndex(Revision),
 }
 
 /// 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`] error is returned.
 ///
 /// In insert operations, the rule is thus that the `NodeMap` must always
 /// be updated after the `RevlogIndex`
 /// be updated first, and the `NodeMap` second.
 ///
 /// [`RevisionNotInIndex`]: enum.NodeMapError.html#variant.RevisionNotInIndex
 /// [`RevlogIndex`]: ../trait.RevlogIndex.html
 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`]
-    /// error 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`]
-    /// error is returned.
+    /// 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`, 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 [`Blocks`] elements
+/// 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 {
     Rev(Revision),
     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,
         })
     }
 }
 
 /// 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.
 
 const ELEMENTS_PER_BLOCK: usize = 16; // number of different values in a nybble
 
 #[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 `index`.
 fn has_prefix_or_none(
     idx: &impl RevlogIndex,
     prefix: NodePrefix,
     rev: Revision,
 ) -> Result<Option<Revision>, NodeMapError> {
     idx.node(rev)
         .ok_or(NodeMapError::RevisionNotInIndex(rev))
         .map(|node| {
             if prefix.is_prefix_of(node) {
                 Some(rev)
             } else {
                 None
             }
         })
 }
 
 /// 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<Revision>, 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 `buffer`,
     /// of which exactly `amount` bytes are used.
     ///
     /// - `buffer` could be derived from `PyBuffer` and `Mmap` objects.
     /// - `offset` allows for the final file format to include fixed data
     ///   (generation number, behavioural flags)
     /// - `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` 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 `Blocks` 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.
     ///
     /// The second returned value is the size of the smallest subprefix
     /// of `prefix` that would give the same result, i.e. not the
     /// `MultipleResults` error variant (again, using only the data of the
     /// `NodeTree`).
     fn lookup(
         &self,
         prefix: NodePrefix,
     ) -> Result<(Option<Revision>, 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
                 .node(old_rev)
                 .ok_or(NodeMapError::RevisionNotInIndex(old_rev))?;
             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));
                     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));
         }
 
         // 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
     // `Revision` that it may represent.
     //
     // If the item is not terminal, return `None`
     fn final_revision(&self) -> Option<Option<Revision>> {
         match self.element {
             Element::Block(_) => None,
             Element::Rev(r) => Some(Some(r)),
             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;
         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()))
     }
 }
 
 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)]
 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
             }
         )
     }
 
     #[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<Revision, Node>;
 
     impl RevlogIndex for TestIndex {
         fn node(&self, rev: Revision) -> Option<&Node> {
             self.get(&rev)
         }
 
         fn len(&self) -> usize {
             self.len()
         }
     }
 
     /// 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()
     }
 
     /// Pad hexadecimal Node prefix with zeros on the right, then insert
     fn pad_insert(idx: &mut TestIndex, rev: Revision, hex: &str) {
         idx.insert(rev, 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, 1, "1234deadcafe");
 
         let nt = NodeTree::from(vec![block! {1: Rev(1)}]);
         assert_eq!(nt.find_bin(&idx, hex("1"))?, Some(1));
         assert_eq!(nt.find_bin(&idx, hex("12"))?, Some(1));
         assert_eq!(nt.find_bin(&idx, hex("1234de"))?, Some(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).unwrap())?, Some(1));
         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, 9, "012");
         pad_insert(&mut idx, 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(9)));
         assert_eq!(nt.find_bin(&idx, hex("00")), Err(MultipleResults));
         assert_eq!(nt.find_bin(&idx, hex("00a")), Ok(Some(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, 9, "012");
         pad_insert(&mut idx, 0, "00a");
         pad_insert(&mut idx, 2, "cafe");
         pad_insert(&mut idx, 3, "15");
         pad_insert(&mut idx, 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(1));
         assert_eq!(nt.find_bin(&idx, hex("c"))?, Some(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(9));
         assert_eq!(nt.masked_readonly_blocks(), 2);
         Ok(())
     }
 
     struct TestNtIndex {
         index: TestIndex,
         nt: NodeTree,
     }
 
     impl TestNtIndex {
         fn new() -> Self {
             TestNtIndex {
                 index: HashMap::new(),
                 nt: NodeTree::default(),
             }
         }
 
         fn insert(
             &mut self,
             rev: Revision,
             hex: &str,
         ) -> Result<(), NodeMapError> {
             let node = pad_node(hex);
             self.index.insert(rev, node);
             self.nt.insert(&self.index, &node, rev)?;
             Ok(())
         }
 
         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),
             }
         }
     }
 
     #[test]
     fn test_insert_full_mutable() -> Result<(), NodeMapError> {
         let mut idx = TestNtIndex::new();
         idx.insert(0, "1234")?;
         assert_eq!(idx.find_hex("1")?, Some(0));
         assert_eq!(idx.find_hex("12")?, Some(0));
 
         // let's trigger a simple split
         idx.insert(1, "1a34")?;
         assert_eq!(idx.nt.growable.len(), 1);
         assert_eq!(idx.find_hex("12")?, Some(0));
         assert_eq!(idx.find_hex("1a")?, Some(1));
 
         // reinserting is a no_op
         idx.insert(1, "1a34")?;
         assert_eq!(idx.nt.growable.len(), 1);
         assert_eq!(idx.find_hex("12")?, Some(0));
         assert_eq!(idx.find_hex("1a")?, Some(1));
 
         idx.insert(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(0));
         assert_eq!(idx.find_hex("1a3")?, Some(1));
         assert_eq!(idx.find_hex("1a0")?, Some(2));
         assert_eq!(idx.find_hex("1a12")?, None);
 
         // now let's make it split and create more than one additional block
         idx.insert(3, "1a345")?;
         assert_eq!(idx.nt.growable.len(), 4);
         assert_eq!(idx.find_hex("1a340")?, Some(1));
         assert_eq!(idx.find_hex("1a345")?, Some(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(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(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 node1 = Node::from_hex(&node1_hex).unwrap();
 
         idx.insert(0, node0);
         nt.insert(idx, &node0, 0)?;
         idx.insert(1, node1);
         nt.insert(idx, &node1, 1)?;
 
         assert_eq!(nt.find_bin(idx, (&node0).into())?, Some(0));
         assert_eq!(nt.find_bin(idx, (&node1).into())?, Some(1));
         Ok(())
     }
 
     #[test]
     fn test_insert_partly_immutable() -> Result<(), NodeMapError> {
         let mut idx = TestNtIndex::new();
         idx.insert(0, "1234")?;
         idx.insert(1, "1235")?;
         idx.insert(2, "131")?;
         idx.insert(3, "cafe")?;
         let mut idx = idx.commit();
         assert_eq!(idx.find_hex("1234")?, Some(0));
         assert_eq!(idx.find_hex("1235")?, Some(1));
         assert_eq!(idx.find_hex("131")?, Some(2));
         assert_eq!(idx.find_hex("cafe")?, Some(3));
         // we did not add anything since init from readonly
         assert_eq!(idx.nt.masked_readonly_blocks(), 0);
 
         idx.insert(4, "123A")?;
         assert_eq!(idx.find_hex("1234")?, Some(0));
         assert_eq!(idx.find_hex("1235")?, Some(1));
         assert_eq!(idx.find_hex("131")?, Some(2));
         assert_eq!(idx.find_hex("cafe")?, Some(3));
         assert_eq!(idx.find_hex("123A")?, Some(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(5, "c0")?;
         assert_eq!(idx.find_hex("cafe")?, Some(3));
         assert_eq!(idx.find_hex("c0")?, Some(5));
         assert_eq!(idx.find_hex("c1")?, None);
         assert_eq!(idx.find_hex("1234")?, Some(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(0, "1234")?;
         idx.insert(1, "1235")?;
         idx.insert(2, "131")?;
         idx.insert(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(0, "1234")?;
         let mut idx = idx.commit();
         idx.insert(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/utils.rs

@@ -1,532 +1,532 b''
 // utils module
 //
 // 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.
 
 //! Contains useful functions, traits, structs, etc. for use in core.
 
 use crate::errors::{HgError, IoErrorContext};
 use crate::utils::hg_path::HgPath;
 use im_rc::ordmap::DiffItem;
 use im_rc::ordmap::OrdMap;
 use std::cell::Cell;
 use std::fmt;
 use std::{io::Write, ops::Deref};
 
 pub mod debug;
 pub mod files;
 pub mod hg_path;
 pub mod path_auditor;
 
 /// Useful until rust/issues/56345 is stable
 ///
 /// # Examples
 ///
 /// ```
 /// use crate::hg::utils::find_slice_in_slice;
 ///
 /// let haystack = b"This is the haystack".to_vec();
 /// assert_eq!(find_slice_in_slice(&haystack, b"the"), Some(8));
 /// assert_eq!(find_slice_in_slice(&haystack, b"not here"), None);
 /// ```
 pub fn find_slice_in_slice<T>(slice: &[T], needle: &[T]) -> Option<usize>
 where
     for<'a> &'a [T]: PartialEq,
 {
     slice
         .windows(needle.len())
         .position(|window| window == needle)
 }
 
 /// Replaces the `from` slice with the `to` slice inside the `buf` slice.
 ///
 /// # Examples
 ///
 /// ```
 /// use crate::hg::utils::replace_slice;
 /// let mut line = b"I hate writing tests!".to_vec();
 /// replace_slice(&mut line, b"hate", b"love");
 /// assert_eq!(
 ///     line,
 ///     b"I love writing tests!".to_vec()
 /// );
 /// ```
 pub fn replace_slice<T>(buf: &mut [T], from: &[T], to: &[T])
 where
     T: Clone + PartialEq,
 {
     if buf.len() < from.len() || from.len() != to.len() {
         return;
     }
     for i in 0..=buf.len() - from.len() {
         if buf[i..].starts_with(from) {
             buf[i..(i + from.len())].clone_from_slice(to);
         }
     }
 }
 
 pub trait SliceExt {
     fn trim_end(&self) -> &Self;
     fn trim_start(&self) -> &Self;
     fn trim_end_matches(&self, f: impl FnMut(u8) -> bool) -> &Self;
     fn trim_start_matches(&self, f: impl FnMut(u8) -> bool) -> &Self;
     fn trim(&self) -> &Self;
     fn drop_prefix(&self, needle: &Self) -> Option<&Self>;
     fn split_2(&self, separator: u8) -> Option<(&[u8], &[u8])>;
     fn split_2_by_slice(&self, separator: &[u8]) -> Option<(&[u8], &[u8])>;
 }
 
 impl SliceExt for [u8] {
     fn trim_end(&self) -> &[u8] {
         self.trim_end_matches(|byte| byte.is_ascii_whitespace())
     }
 
     fn trim_start(&self) -> &[u8] {
         self.trim_start_matches(|byte| byte.is_ascii_whitespace())
     }
 
     fn trim_end_matches(&self, mut f: impl FnMut(u8) -> bool) -> &Self {
         if let Some(last) = self.iter().rposition(|&byte| !f(byte)) {
             &self[..=last]
         } else {
             &[]
         }
     }
 
     fn trim_start_matches(&self, mut f: impl FnMut(u8) -> bool) -> &Self {
         if let Some(first) = self.iter().position(|&byte| !f(byte)) {
             &self[first..]
         } else {
             &[]
         }
     }
 
     /// ```
     /// use hg::utils::SliceExt;
     /// assert_eq!(
     ///     b"  to trim  ".trim(),
     ///     b"to trim"
     /// );
     /// assert_eq!(
     ///     b"to trim  ".trim(),
     ///     b"to trim"
     /// );
     /// assert_eq!(
     ///     b"  to trim".trim(),
     ///     b"to trim"
     /// );
     /// ```
     fn trim(&self) -> &[u8] {
         self.trim_start().trim_end()
     }
 
     fn drop_prefix(&self, needle: &Self) -> Option<&Self> {
         if self.starts_with(needle) {
             Some(&self[needle.len()..])
         } else {
             None
         }
     }
 
     fn split_2(&self, separator: u8) -> Option<(&[u8], &[u8])> {
         let mut iter = self.splitn(2, |&byte| byte == separator);
         let a = iter.next()?;
         let b = iter.next()?;
         Some((a, b))
     }
 
     fn split_2_by_slice(&self, separator: &[u8]) -> Option<(&[u8], &[u8])> {
         find_slice_in_slice(self, separator)
             .map(|pos| (&self[..pos], &self[pos + separator.len()..]))
     }
 }
 
 pub trait Escaped {
     /// Return bytes escaped for display to the user
     fn escaped_bytes(&self) -> Vec<u8>;
 }
 
 impl Escaped for u8 {
     fn escaped_bytes(&self) -> Vec<u8> {
         let mut acc = vec![];
         match self {
             c @ b'\'' | c @ b'\\' => {
                 acc.push(b'\\');
                 acc.push(*c);
             }
             b'\t' => {
                 acc.extend(br"\\t");
             }
             b'\n' => {
                 acc.extend(br"\\n");
             }
             b'\r' => {
                 acc.extend(br"\\r");
             }
             c if (*c < b' ' || *c >= 127) => {
                 write!(acc, "\\x{:x}", self).unwrap();
             }
             c => {
                 acc.push(*c);
             }
         }
         acc
     }
 }
 
 impl<'a, T: Escaped> Escaped for &'a [T] {
     fn escaped_bytes(&self) -> Vec<u8> {
         self.iter().flat_map(Escaped::escaped_bytes).collect()
     }
 }
 
 impl<T: Escaped> Escaped for Vec<T> {
     fn escaped_bytes(&self) -> Vec<u8> {
         self.deref().escaped_bytes()
     }
 }
 
 impl<'a> Escaped for &'a HgPath {
     fn escaped_bytes(&self) -> Vec<u8> {
         self.as_bytes().escaped_bytes()
     }
 }
 
 #[cfg(unix)]
 pub fn shell_quote(value: &[u8]) -> Vec<u8> {
     if value.iter().all(|&byte| {
         matches!(
             byte,
             b'a'..=b'z'
             | b'A'..=b'Z'
             | b'0'..=b'9'
             | b'.'
             | b'_'
             | b'/'
             | b'+'
             | b'-'
         )
     }) {
         value.to_owned()
     } else {
         let mut quoted = Vec::with_capacity(value.len() + 2);
         quoted.push(b'\'');
         for &byte in value {
             if byte == b'\'' {
                 quoted.push(b'\\');
             }
             quoted.push(byte);
         }
         quoted.push(b'\'');
         quoted
     }
 }
 
 pub fn current_dir() -> Result<std::path::PathBuf, HgError> {
     std::env::current_dir().map_err(|error| HgError::IoError {
         error,
         context: IoErrorContext::CurrentDir,
     })
 }
 
 pub fn current_exe() -> Result<std::path::PathBuf, HgError> {
     std::env::current_exe().map_err(|error| HgError::IoError {
         error,
         context: IoErrorContext::CurrentExe,
     })
 }
 
 /// Expand `$FOO` and `${FOO}` environment variables in the given byte string
 pub fn expand_vars(s: &[u8]) -> std::borrow::Cow<[u8]> {
     lazy_static::lazy_static! {
         /// https://github.com/python/cpython/blob/3.9/Lib/posixpath.py#L301
         /// The `x` makes whitespace ignored.
         /// `-u` disables the Unicode flag, which makes `\w` like Python with the ASCII flag.
         static ref VAR_RE: regex::bytes::Regex =
             regex::bytes::Regex::new(r"(?x-u)
                 \$
                 (?:
                     (\w+)
                     |
                     \{
                         ([^}]*)
                     \}
                 )
             ").unwrap();
     }
     VAR_RE.replace_all(s, |captures: &regex::bytes::Captures| {
         let var_name = files::get_os_str_from_bytes(
             captures
                 .get(1)
                 .or_else(|| captures.get(2))
                 .expect("either side of `|` must participate in match")
                 .as_bytes(),
         );
         std::env::var_os(var_name)
             .map(files::get_bytes_from_os_str)
             .unwrap_or_else(|| {
                 // Referencing an environment variable that does not exist.
                 // Leave the $FOO reference as-is.
                 captures[0].to_owned()
             })
     })
 }
 
 #[test]
 fn test_expand_vars() {
     // Modifying process-global state in a test isn’t great,
     // but hopefully this won’t collide with anything.
     std::env::set_var("TEST_EXPAND_VAR", "1");
     assert_eq!(
         expand_vars(b"before/$TEST_EXPAND_VAR/after"),
         &b"before/1/after"[..]
     );
     assert_eq!(
         expand_vars(b"before${TEST_EXPAND_VAR}${TEST_EXPAND_VAR}${TEST_EXPAND_VAR}after"),
         &b"before111after"[..]
     );
     let s = b"before $SOME_LONG_NAME_THAT_WE_ASSUME_IS_NOT_AN_ACTUAL_ENV_VAR after";
     assert_eq!(expand_vars(s), &s[..]);
 }
 
 pub(crate) enum MergeResult<V> {
     Left,
     Right,
     New(V),
 }
 
 /// Return the union of the two given maps,
 /// calling `merge(key, left_value, right_value)` to resolve keys that exist in
 /// both.
 ///
-/// CC https://github.com/bodil/im-rs/issues/166
+/// CC <https://github.com/bodil/im-rs/issues/166>
 pub(crate) fn ordmap_union_with_merge<K, V>(
     left: OrdMap<K, V>,
     right: OrdMap<K, V>,
     mut merge: impl FnMut(&K, &V, &V) -> MergeResult<V>,
 ) -> OrdMap<K, V>
 where
     K: Clone + Ord,
     V: Clone + PartialEq,
 {
     if left.ptr_eq(&right) {
         // One of the two maps is an unmodified clone of the other
         left
     } else if left.len() / 2 > right.len() {
         // When two maps have different sizes,
         // their size difference is a lower bound on
         // how many keys of the larger map are not also in the smaller map.
         // This in turn is a lower bound on the number of differences in
         // `OrdMap::diff` and the "amount of work" that would be done
         // by `ordmap_union_with_merge_by_diff`.
         //
         // Here `left` is more than twice the size of `right`,
         // so the number of differences is more than the total size of
         // `right`. Therefore an algorithm based on iterating `right`
         // is more efficient.
         //
         // This helps a lot when a tiny (or empty) map is merged
         // with a large one.
         ordmap_union_with_merge_by_iter(left, right, merge)
     } else if left.len() < right.len() / 2 {
         // Same as above but with `left` and `right` swapped
         ordmap_union_with_merge_by_iter(right, left, |key, a, b| {
             // Also swapped in `merge` arguments:
             match merge(key, b, a) {
                 MergeResult::New(v) => MergeResult::New(v),
                 // … and swap back in `merge` result:
                 MergeResult::Left => MergeResult::Right,
                 MergeResult::Right => MergeResult::Left,
             }
         })
     } else {
         // For maps of similar size, use the algorithm based on `OrdMap::diff`
         ordmap_union_with_merge_by_diff(left, right, merge)
     }
 }
 
 /// Efficient if `right` is much smaller than `left`
 fn ordmap_union_with_merge_by_iter<K, V>(
     mut left: OrdMap<K, V>,
     right: OrdMap<K, V>,
     mut merge: impl FnMut(&K, &V, &V) -> MergeResult<V>,
 ) -> OrdMap<K, V>
 where
     K: Clone + Ord,
     V: Clone,
 {
     for (key, right_value) in right {
         match left.get(&key) {
             None => {
                 left.insert(key, right_value);
             }
             Some(left_value) => match merge(&key, left_value, &right_value) {
                 MergeResult::Left => {}
                 MergeResult::Right => {
                     left.insert(key, right_value);
                 }
                 MergeResult::New(new_value) => {
                     left.insert(key, new_value);
                 }
             },
         }
     }
     left
 }
 
 /// Fallback when both maps are of similar size
 fn ordmap_union_with_merge_by_diff<K, V>(
     mut left: OrdMap<K, V>,
     mut right: OrdMap<K, V>,
     mut merge: impl FnMut(&K, &V, &V) -> MergeResult<V>,
 ) -> OrdMap<K, V>
 where
     K: Clone + Ord,
     V: Clone + PartialEq,
 {
     // (key, value) pairs that would need to be inserted in either map
     // in order to turn it into the union.
     //
     // TODO: if/when https://github.com/bodil/im-rs/pull/168 is accepted,
     // change these from `Vec<(K, V)>` to `Vec<(&K, Cow<V>)>`
     // with `left_updates` only borrowing from `right` and `right_updates` from
     // `left`, and with `Cow::Owned` used for `MergeResult::New`.
     //
     // This would allow moving all `.clone()` calls to after we’ve decided
     // which of `right_updates` or `left_updates` to use
     // (value ones becoming `Cow::into_owned`),
     // and avoid making clones we don’t end up using.
     let mut left_updates = Vec::new();
     let mut right_updates = Vec::new();
 
     for difference in left.diff(&right) {
         match difference {
             DiffItem::Add(key, value) => {
                 left_updates.push((key.clone(), value.clone()))
             }
             DiffItem::Remove(key, value) => {
                 right_updates.push((key.clone(), value.clone()))
             }
             DiffItem::Update {
                 old: (key, left_value),
                 new: (_, right_value),
             } => match merge(key, left_value, right_value) {
                 MergeResult::Left => {
                     right_updates.push((key.clone(), left_value.clone()))
                 }
                 MergeResult::Right => {
                     left_updates.push((key.clone(), right_value.clone()))
                 }
                 MergeResult::New(new_value) => {
                     left_updates.push((key.clone(), new_value.clone()));
                     right_updates.push((key.clone(), new_value))
                 }
             },
         }
     }
     if left_updates.len() < right_updates.len() {
         for (key, value) in left_updates {
             left.insert(key, value);
         }
         left
     } else {
         for (key, value) in right_updates {
             right.insert(key, value);
         }
         right
     }
 }
 
 /// Join items of the iterable with the given separator, similar to Python’s
 /// `separator.join(iter)`.
 ///
 /// Formatting the return value consumes the iterator.
 /// Formatting it again will produce an empty string.
 pub fn join_display(
     iter: impl IntoIterator<Item = impl fmt::Display>,
     separator: impl fmt::Display,
 ) -> impl fmt::Display {
     JoinDisplay {
         iter: Cell::new(Some(iter.into_iter())),
         separator,
     }
 }
 
 struct JoinDisplay<I, S> {
     iter: Cell<Option<I>>,
     separator: S,
 }
 
 impl<I, T, S> fmt::Display for JoinDisplay<I, S>
 where
     I: Iterator<Item = T>,
     T: fmt::Display,
     S: fmt::Display,
 {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
         if let Some(mut iter) = self.iter.take() {
             if let Some(first) = iter.next() {
                 first.fmt(f)?;
             }
             for value in iter {
                 self.separator.fmt(f)?;
                 value.fmt(f)?;
             }
         }
         Ok(())
     }
 }
 
 /// Like `Iterator::filter_map`, but over a fallible iterator of `Result`s.
 ///
 /// The callback is only called for incoming `Ok` values. Errors are passed
 /// through as-is. In order to let it use the `?` operator the callback is
 /// expected to return a `Result` of `Option`, instead of an `Option` of
 /// `Result`.
 pub fn filter_map_results<'a, I, F, A, B, E>(
     iter: I,
     f: F,
 ) -> impl Iterator<Item = Result<B, E>> + 'a
 where
     I: Iterator<Item = Result<A, E>> + 'a,
     F: Fn(A) -> Result<Option<B>, E> + 'a,
 {
     iter.filter_map(move |result| match result {
         Ok(node) => f(node).transpose(),
         Err(e) => Some(Err(e)),
     })
 }
 
 /// Force the global rayon threadpool to not exceed 16 concurrent threads
 /// unless the user has specified a value.
 /// This is a stop-gap measure until we figure out why using more than 16
 /// threads makes `status` slower for each additional thread.
 ///
 /// TODO find the underlying cause and fix it, then remove this.
 ///
 /// # Errors
 ///
 /// Returns an error if the global threadpool has already been initialized if
 /// we try to initialize it.
 pub fn cap_default_rayon_threads() -> Result<(), rayon::ThreadPoolBuildError> {
     const THREAD_CAP: usize = 16;
 
     if std::env::var("RAYON_NUM_THREADS").is_err() {
         let available_parallelism = std::thread::available_parallelism()
             .map(usize::from)
             .unwrap_or(1);
         let new_thread_count = THREAD_CAP.min(available_parallelism);
         let res = rayon::ThreadPoolBuilder::new()
             .num_threads(new_thread_count)
             .build_global();
         if res.is_ok() {
             log::trace!(
                 "Capped the rayon threadpool to {new_thread_count} threads",
             );
         }
         return res;
     }
     Ok(())
 }

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

@@ -1,812 +1,812 b''
 // hg_path.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.
 
 use crate::utils::SliceExt;
 use std::borrow::Borrow;
 use std::borrow::Cow;
 use std::ffi::{OsStr, OsString};
 use std::fmt;
 use std::ops::Deref;
 use std::path::{Path, PathBuf};
 
 #[derive(Debug, Eq, PartialEq)]
 pub enum HgPathError {
     /// Bytes from the invalid `HgPath`
     LeadingSlash(Vec<u8>),
     ConsecutiveSlashes {
         bytes: Vec<u8>,
         second_slash_index: usize,
     },
     ContainsNullByte {
         bytes: Vec<u8>,
         null_byte_index: usize,
     },
     /// Bytes
     DecodeError(Vec<u8>),
     /// The rest come from audit errors
     EndsWithSlash(HgPathBuf),
     ContainsIllegalComponent(HgPathBuf),
     /// Path is inside the `.hg` folder
     InsideDotHg(HgPathBuf),
     IsInsideNestedRepo {
         path: HgPathBuf,
         nested_repo: HgPathBuf,
     },
     TraversesSymbolicLink {
         path: HgPathBuf,
         symlink: HgPathBuf,
     },
     NotFsCompliant(HgPathBuf),
     /// `path` is the smallest invalid path
     NotUnderRoot {
         path: PathBuf,
         root: PathBuf,
     },
 }
 
 impl fmt::Display for HgPathError {
     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
         match self {
             HgPathError::LeadingSlash(bytes) => {
                 write!(f, "Invalid HgPath '{:?}': has a leading slash.", bytes)
             }
             HgPathError::ConsecutiveSlashes {
                 bytes,
                 second_slash_index: pos,
             } => write!(
                 f,
                 "Invalid HgPath '{:?}': consecutive slashes at pos {}.",
                 bytes, pos
             ),
             HgPathError::ContainsNullByte {
                 bytes,
                 null_byte_index: pos,
             } => write!(
                 f,
                 "Invalid HgPath '{:?}': contains null byte at pos {}.",
                 bytes, pos
             ),
             HgPathError::DecodeError(bytes) => write!(
                 f,
                 "Invalid HgPath '{:?}': could not be decoded.",
                 bytes
             ),
             HgPathError::EndsWithSlash(path) => {
                 write!(f, "Audit failed for '{}': ends with a slash.", path)
             }
             HgPathError::ContainsIllegalComponent(path) => write!(
                 f,
                 "Audit failed for '{}': contains an illegal component.",
                 path
             ),
             HgPathError::InsideDotHg(path) => write!(
                 f,
                 "Audit failed for '{}': is inside the '.hg' folder.",
                 path
             ),
             HgPathError::IsInsideNestedRepo {
                 path,
                 nested_repo: nested,
             } => {
                 write!(f,
                 "Audit failed for '{}': is inside a nested repository '{}'.",
                 path, nested
             )
             }
             HgPathError::TraversesSymbolicLink { path, symlink } => write!(
                 f,
                 "Audit failed for '{}': traverses symbolic link '{}'.",
                 path, symlink
             ),
             HgPathError::NotFsCompliant(path) => write!(
                 f,
                 "Audit failed for '{}': cannot be turned into a \
                  filesystem path.",
                 path
             ),
             HgPathError::NotUnderRoot { path, root } => write!(
                 f,
                 "Audit failed for '{}': not under root {}.",
                 path.display(),
                 root.display()
             ),
         }
     }
 }
 
 impl From<HgPathError> for std::io::Error {
     fn from(e: HgPathError) -> Self {
         std::io::Error::new(std::io::ErrorKind::InvalidData, e.to_string())
     }
 }
 
 /// This is a repository-relative path (or canonical path):
 ///     - no null characters
 ///     - `/` separates directories
 ///     - no consecutive slashes
 ///     - no leading slash,
 ///     - no `.` nor `..` of special meaning
 ///     - stored in repository and shared across platforms
 ///
 /// Note: there is no guarantee of any `HgPath` being well-formed at any point
 /// in its lifetime for performance reasons and to ease ergonomics. It is
 /// however checked using the `check_state` method before any file-system
 /// operation.
 ///
 /// This allows us to be encoding-transparent as much as possible, until really
 /// needed; `HgPath` can be transformed into a platform-specific path (`OsStr`
 /// or `Path`) whenever more complex operations are needed:
 /// On Unix, it's just byte-to-byte conversion. On Windows, it has to be
 /// decoded from MBCS to WTF-8. If WindowsUTF8Plan is implemented, the source
 /// character encoding will be determined on a per-repository basis.
 #[derive(Eq, Ord, PartialEq, PartialOrd, Hash)]
 #[repr(transparent)]
 pub struct HgPath {
     inner: [u8],
 }
 
 impl HgPath {
     pub fn new<S: AsRef<[u8]> + ?Sized>(s: &S) -> &Self {
         unsafe { &*(s.as_ref() as *const [u8] as *const Self) }
     }
     pub fn is_empty(&self) -> bool {
         self.inner.is_empty()
     }
     pub fn len(&self) -> usize {
         self.inner.len()
     }
     fn to_hg_path_buf(&self) -> HgPathBuf {
         HgPathBuf {
             inner: self.inner.to_owned(),
         }
     }
     pub fn bytes(&self) -> std::slice::Iter<u8> {
         self.inner.iter()
     }
     pub fn to_ascii_uppercase(&self) -> HgPathBuf {
         HgPathBuf::from(self.inner.to_ascii_uppercase())
     }
     pub fn to_ascii_lowercase(&self) -> HgPathBuf {
         HgPathBuf::from(self.inner.to_ascii_lowercase())
     }
     pub fn as_bytes(&self) -> &[u8] {
         &self.inner
     }
     pub fn contains(&self, other: u8) -> bool {
         self.inner.contains(&other)
     }
     pub fn starts_with(&self, needle: impl AsRef<Self>) -> bool {
         self.inner.starts_with(needle.as_ref().as_bytes())
     }
     pub fn trim_trailing_slash(&self) -> &Self {
         Self::new(if self.inner.last() == Some(&b'/') {
             &self.inner[..self.inner.len() - 1]
         } else {
             &self.inner[..]
         })
     }
     /// Returns a tuple of slices `(base, filename)` resulting from the split
     /// at the rightmost `/`, if any.
     ///
     /// # Examples:
     ///
     /// ```
     /// use hg::utils::hg_path::HgPath;
     ///
     /// let path = HgPath::new(b"cool/hg/path").split_filename();
     /// assert_eq!(path, (HgPath::new(b"cool/hg"), HgPath::new(b"path")));
     ///
     /// let path = HgPath::new(b"pathwithoutsep").split_filename();
     /// assert_eq!(path, (HgPath::new(b""), HgPath::new(b"pathwithoutsep")));
     /// ```
     pub fn split_filename(&self) -> (&Self, &Self) {
         match &self.inner.iter().rposition(|c| *c == b'/') {
             None => (HgPath::new(""), self),
             Some(size) => (
                 HgPath::new(&self.inner[..*size]),
                 HgPath::new(&self.inner[*size + 1..]),
             ),
         }
     }
 
     pub fn join(&self, path: &HgPath) -> HgPathBuf {
         let mut buf = self.to_owned();
         buf.push(path);
         buf
     }
 
     pub fn components(&self) -> impl Iterator<Item = &HgPath> {
         self.inner.split(|&byte| byte == b'/').map(HgPath::new)
     }
 
     /// Returns the first (that is "root-most") slash-separated component of
     /// the path, and the rest after the first slash if there is one.
     pub fn split_first_component(&self) -> (&HgPath, Option<&HgPath>) {
         match self.inner.split_2(b'/') {
             Some((a, b)) => (HgPath::new(a), Some(HgPath::new(b))),
             None => (self, None),
         }
     }
 
     pub fn parent(&self) -> &Self {
         let inner = self.as_bytes();
         HgPath::new(match inner.iter().rposition(|b| *b == b'/') {
             Some(pos) => &inner[..pos],
             None => &[],
         })
     }
     /// Given a base directory, returns the slice of `self` relative to the
     /// base directory. If `base` is not a directory (does not end with a
     /// `b'/'`), returns `None`.
     pub fn relative_to(&self, base: impl AsRef<Self>) -> Option<&Self> {
         let base = base.as_ref();
         if base.is_empty() {
             return Some(self);
         }
         let is_dir = base.as_bytes().ends_with(b"/");
         if is_dir && self.starts_with(base) {
             Some(Self::new(&self.inner[base.len()..]))
         } else {
             None
         }
     }
 
     #[cfg(windows)]
     /// Copied from the Python stdlib's `os.path.splitdrive` implementation.
     ///
     /// Split a pathname into drive/UNC sharepoint and relative path
     /// specifiers. Returns a 2-tuple (drive_or_unc, path); either part may
     /// be empty.
     ///
     /// If you assign
     ///  result = split_drive(p)
     /// It is always true that:
     ///  result[0] + result[1] == p
     ///
     /// If the path contained a drive letter, drive_or_unc will contain
     /// everything up to and including the colon.
     /// e.g. split_drive("c:/dir") returns ("c:", "/dir")
     ///
     /// If the path contained a UNC path, the drive_or_unc will contain the
     /// host name and share up to but not including the fourth directory
     /// separator character.
     /// e.g. split_drive("//host/computer/dir") returns ("//host/computer",
     /// "/dir")
     ///
     /// Paths cannot contain both a drive letter and a UNC path.
     pub fn split_drive<'a>(&self) -> (&HgPath, &HgPath) {
         let bytes = self.as_bytes();
         let is_sep = |b| std::path::is_separator(b as char);
 
         if self.len() < 2 {
             (HgPath::new(b""), &self)
         } else if is_sep(bytes[0])
             && is_sep(bytes[1])
             && (self.len() == 2 || !is_sep(bytes[2]))
         {
             // Is a UNC path:
             // vvvvvvvvvvvvvvvvvvvv drive letter or UNC path
             // \\machine\mountpoint\directory\etc\...
             //           directory ^^^^^^^^^^^^^^^
 
             let machine_end_index = bytes[2..].iter().position(|b| is_sep(*b));
             let mountpoint_start_index = if let Some(i) = machine_end_index {
                 i + 2
             } else {
                 return (HgPath::new(b""), &self);
             };
 
             match bytes[mountpoint_start_index + 1..]
                 .iter()
                 .position(|b| is_sep(*b))
             {
                 // A UNC path can't have two slashes in a row
                 // (after the initial two)
                 Some(0) => (HgPath::new(b""), &self),
                 Some(i) => {
                     let (a, b) =
                         bytes.split_at(mountpoint_start_index + 1 + i);
                     (HgPath::new(a), HgPath::new(b))
                 }
                 None => (&self, HgPath::new(b"")),
             }
         } else if bytes[1] == b':' {
             // Drive path c:\directory
             let (a, b) = bytes.split_at(2);
             (HgPath::new(a), HgPath::new(b))
         } else {
             (HgPath::new(b""), &self)
         }
     }
 
     #[cfg(unix)]
     /// Split a pathname into drive and path. On Posix, drive is always empty.
     pub fn split_drive(&self) -> (&HgPath, &HgPath) {
         (HgPath::new(b""), self)
     }
 
     /// Checks for errors in the path, short-circuiting at the first one.
     /// This generates fine-grained errors useful for debugging.
     /// To simply check if the path is valid during tests, use `is_valid`.
     pub fn check_state(&self) -> Result<(), HgPathError> {
         if self.is_empty() {
             return Ok(());
         }
         let bytes = self.as_bytes();
         let mut previous_byte = None;
 
         if bytes[0] == b'/' {
             return Err(HgPathError::LeadingSlash(bytes.to_vec()));
         }
         for (index, byte) in bytes.iter().enumerate() {
             match byte {
                 0 => {
                     return Err(HgPathError::ContainsNullByte {
                         bytes: bytes.to_vec(),
                         null_byte_index: index,
                     })
                 }
                 b'/' => {
                     if previous_byte.is_some() && previous_byte == Some(b'/') {
                         return Err(HgPathError::ConsecutiveSlashes {
                             bytes: bytes.to_vec(),
                             second_slash_index: index,
                         });
                     }
                 }
                 _ => (),
             };
             previous_byte = Some(*byte);
         }
         Ok(())
     }
 
     #[cfg(test)]
     /// Only usable during tests to force developers to handle invalid states
     fn is_valid(&self) -> bool {
         self.check_state().is_ok()
     }
 }
 
 impl fmt::Debug for HgPath {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
         write!(f, "HgPath({:?})", String::from_utf8_lossy(&self.inner))
     }
 }
 
 impl fmt::Display for HgPath {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
         write!(f, "{}", String::from_utf8_lossy(&self.inner))
     }
 }
 
 #[derive(
     Default, Eq, Ord, Clone, PartialEq, PartialOrd, Hash, derive_more::From,
 )]
 pub struct HgPathBuf {
     inner: Vec<u8>,
 }
 
 impl HgPathBuf {
     pub fn new() -> Self {
         Default::default()
     }
 
     pub fn push<T: ?Sized + AsRef<HgPath>>(&mut self, other: &T) {
         if !self.inner.is_empty() && self.inner.last() != Some(&b'/') {
             self.inner.push(b'/');
         }
         self.inner.extend(other.as_ref().bytes())
     }
 
     pub fn push_byte(&mut self, byte: u8) {
         self.inner.push(byte);
     }
     pub fn from_bytes(s: &[u8]) -> HgPathBuf {
         HgPath::new(s).to_owned()
     }
     pub fn into_vec(self) -> Vec<u8> {
         self.inner
     }
 }
 
 impl fmt::Debug for HgPathBuf {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
         write!(f, "HgPathBuf({:?})", String::from_utf8_lossy(&self.inner))
     }
 }
 
 impl fmt::Display for HgPathBuf {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
         write!(f, "{}", String::from_utf8_lossy(&self.inner))
     }
 }
 
 impl Deref for HgPathBuf {
     type Target = HgPath;
 
     #[inline]
     fn deref(&self) -> &HgPath {
         HgPath::new(&self.inner)
     }
 }
 
 impl<T: ?Sized + AsRef<HgPath>> From<&T> for HgPathBuf {
     fn from(s: &T) -> HgPathBuf {
         s.as_ref().to_owned()
     }
 }
 
 impl From<HgPathBuf> for Vec<u8> {
     fn from(val: HgPathBuf) -> Self {
         val.inner
     }
 }
 
 impl Borrow<HgPath> for HgPathBuf {
     fn borrow(&self) -> &HgPath {
         HgPath::new(self.as_bytes())
     }
 }
 
 impl ToOwned for HgPath {
     type Owned = HgPathBuf;
 
     fn to_owned(&self) -> HgPathBuf {
         self.to_hg_path_buf()
     }
 }
 
 impl AsRef<HgPath> for HgPath {
     fn as_ref(&self) -> &HgPath {
         self
     }
 }
 
 impl AsRef<HgPath> for HgPathBuf {
     fn as_ref(&self) -> &HgPath {
         self
     }
 }
 
 impl Extend<u8> for HgPathBuf {
     fn extend<T: IntoIterator<Item = u8>>(&mut self, iter: T) {
         self.inner.extend(iter);
     }
 }
 
-/// TODO: Once https://www.mercurial-scm.org/wiki/WindowsUTF8Plan is
+/// TODO: Once <https://www.mercurial-scm.org/wiki/WindowsUTF8Plan> is
 /// implemented, these conversion utils will have to work differently depending
 /// on the repository encoding: either `UTF-8` or `MBCS`.
 
 pub fn hg_path_to_os_string<P: AsRef<HgPath>>(
     hg_path: P,
 ) -> Result<OsString, HgPathError> {
     hg_path.as_ref().check_state()?;
     let os_str;
     #[cfg(unix)]
     {
         use std::os::unix::ffi::OsStrExt;
         os_str = std::ffi::OsStr::from_bytes(hg_path.as_ref().as_bytes());
     }
     // TODO Handle other platforms
     // TODO: convert from WTF8 to Windows MBCS (ANSI encoding).
     Ok(os_str.to_os_string())
 }
 
 pub fn hg_path_to_path_buf<P: AsRef<HgPath>>(
     hg_path: P,
 ) -> Result<PathBuf, HgPathError> {
     Ok(Path::new(&hg_path_to_os_string(hg_path)?).to_path_buf())
 }
 
 pub fn os_string_to_hg_path_buf<S: AsRef<OsStr>>(
     os_string: S,
 ) -> Result<HgPathBuf, HgPathError> {
     let buf;
     #[cfg(unix)]
     {
         use std::os::unix::ffi::OsStrExt;
         buf = HgPathBuf::from_bytes(os_string.as_ref().as_bytes());
     }
     // TODO Handle other platforms
     // TODO: convert from WTF8 to Windows MBCS (ANSI encoding).
 
     buf.check_state()?;
     Ok(buf)
 }
 
 pub fn path_to_hg_path_buf<P: AsRef<Path>>(
     path: P,
 ) -> Result<HgPathBuf, HgPathError> {
     let buf;
     let os_str = path.as_ref().as_os_str();
     #[cfg(unix)]
     {
         use std::os::unix::ffi::OsStrExt;
         buf = HgPathBuf::from_bytes(os_str.as_bytes());
     }
     // TODO Handle other platforms
     // TODO: convert from WTF8 to Windows MBCS (ANSI encoding).
 
     buf.check_state()?;
     Ok(buf)
 }
 
 impl TryFrom<PathBuf> for HgPathBuf {
     type Error = HgPathError;
     fn try_from(path: PathBuf) -> Result<Self, Self::Error> {
         path_to_hg_path_buf(path)
     }
 }
 
 impl From<HgPathBuf> for Cow<'_, HgPath> {
     fn from(path: HgPathBuf) -> Self {
         Cow::Owned(path)
     }
 }
 
 impl<'a> From<&'a HgPath> for Cow<'a, HgPath> {
     fn from(path: &'a HgPath) -> Self {
         Cow::Borrowed(path)
     }
 }
 
 impl<'a> From<&'a HgPathBuf> for Cow<'a, HgPath> {
     fn from(path: &'a HgPathBuf) -> Self {
         Cow::Borrowed(&**path)
     }
 }
 
 #[cfg(test)]
 mod tests {
     use super::*;
     use pretty_assertions::assert_eq;
 
     #[test]
     fn test_path_states() {
         assert_eq!(
             Err(HgPathError::LeadingSlash(b"/".to_vec())),
             HgPath::new(b"/").check_state()
         );
         assert_eq!(
             Err(HgPathError::ConsecutiveSlashes {
                 bytes: b"a/b//c".to_vec(),
                 second_slash_index: 4
             }),
             HgPath::new(b"a/b//c").check_state()
         );
         assert_eq!(
             Err(HgPathError::ContainsNullByte {
                 bytes: b"a/b/\0c".to_vec(),
                 null_byte_index: 4
             }),
             HgPath::new(b"a/b/\0c").check_state()
         );
         // TODO test HgPathError::DecodeError for the Windows implementation.
         assert_eq!(true, HgPath::new(b"").is_valid());
         assert_eq!(true, HgPath::new(b"a/b/c").is_valid());
         // Backslashes in paths are not significant, but allowed
         assert_eq!(true, HgPath::new(br"a\b/c").is_valid());
         // Dots in paths are not significant, but allowed
         assert_eq!(true, HgPath::new(b"a/b/../c/").is_valid());
         assert_eq!(true, HgPath::new(b"./a/b/../c/").is_valid());
     }
 
     #[test]
     fn test_iter() {
         let path = HgPath::new(b"a");
         let mut iter = path.bytes();
         assert_eq!(Some(&b'a'), iter.next());
         assert_eq!(None, iter.next_back());
         assert_eq!(None, iter.next());
 
         let path = HgPath::new(b"a");
         let mut iter = path.bytes();
         assert_eq!(Some(&b'a'), iter.next_back());
         assert_eq!(None, iter.next_back());
         assert_eq!(None, iter.next());
 
         let path = HgPath::new(b"abc");
         let mut iter = path.bytes();
         assert_eq!(Some(&b'a'), iter.next());
         assert_eq!(Some(&b'c'), iter.next_back());
         assert_eq!(Some(&b'b'), iter.next_back());
         assert_eq!(None, iter.next_back());
         assert_eq!(None, iter.next());
 
         let path = HgPath::new(b"abc");
         let mut iter = path.bytes();
         assert_eq!(Some(&b'a'), iter.next());
         assert_eq!(Some(&b'b'), iter.next());
         assert_eq!(Some(&b'c'), iter.next());
         assert_eq!(None, iter.next_back());
         assert_eq!(None, iter.next());
 
         let path = HgPath::new(b"abc");
         let iter = path.bytes();
         let mut vec = Vec::new();
         vec.extend(iter);
         assert_eq!(vec![b'a', b'b', b'c'], vec);
 
         let path = HgPath::new(b"abc");
         let mut iter = path.bytes();
         assert_eq!(Some(2), iter.rposition(|c| *c == b'c'));
 
         let path = HgPath::new(b"abc");
         let mut iter = path.bytes();
         assert_eq!(None, iter.rposition(|c| *c == b'd'));
     }
 
     #[test]
     fn test_join() {
         let path = HgPathBuf::from_bytes(b"a").join(HgPath::new(b"b"));
         assert_eq!(b"a/b", path.as_bytes());
 
         let path = HgPathBuf::from_bytes(b"a/").join(HgPath::new(b"b/c"));
         assert_eq!(b"a/b/c", path.as_bytes());
 
         // No leading slash if empty before join
         let path = HgPathBuf::new().join(HgPath::new(b"b/c"));
         assert_eq!(b"b/c", path.as_bytes());
 
         // The leading slash is an invalid representation of an `HgPath`, but
         // it can happen. This creates another invalid representation of
         // consecutive bytes.
         // TODO What should be done in this case? Should we silently remove
         // the extra slash? Should we change the signature to a problematic
         // `Result<HgPathBuf, HgPathError>`, or should we just keep it so and
         // let the error happen upon filesystem interaction?
         let path = HgPathBuf::from_bytes(b"a/").join(HgPath::new(b"/b"));
         assert_eq!(b"a//b", path.as_bytes());
         let path = HgPathBuf::from_bytes(b"a").join(HgPath::new(b"/b"));
         assert_eq!(b"a//b", path.as_bytes());
     }
 
     #[test]
     fn test_relative_to() {
         let path = HgPath::new(b"");
         let base = HgPath::new(b"");
         assert_eq!(Some(path), path.relative_to(base));
 
         let path = HgPath::new(b"path");
         let base = HgPath::new(b"");
         assert_eq!(Some(path), path.relative_to(base));
 
         let path = HgPath::new(b"a");
         let base = HgPath::new(b"b");
         assert_eq!(None, path.relative_to(base));
 
         let path = HgPath::new(b"a/b");
         let base = HgPath::new(b"a");
         assert_eq!(None, path.relative_to(base));
 
         let path = HgPath::new(b"a/b");
         let base = HgPath::new(b"a/");
         assert_eq!(Some(HgPath::new(b"b")), path.relative_to(base));
 
         let path = HgPath::new(b"nested/path/to/b");
         let base = HgPath::new(b"nested/path/");
         assert_eq!(Some(HgPath::new(b"to/b")), path.relative_to(base));
 
         let path = HgPath::new(b"ends/with/dir/");
         let base = HgPath::new(b"ends/");
         assert_eq!(Some(HgPath::new(b"with/dir/")), path.relative_to(base));
     }
 
     #[test]
     #[cfg(unix)]
     fn test_split_drive() {
         // Taken from the Python stdlib's tests
         assert_eq!(
             HgPath::new(br"/foo/bar").split_drive(),
             (HgPath::new(b""), HgPath::new(br"/foo/bar"))
         );
         assert_eq!(
             HgPath::new(br"foo:bar").split_drive(),
             (HgPath::new(b""), HgPath::new(br"foo:bar"))
         );
         assert_eq!(
             HgPath::new(br":foo:bar").split_drive(),
             (HgPath::new(b""), HgPath::new(br":foo:bar"))
         );
         // Also try NT paths; should not split them
         assert_eq!(
             HgPath::new(br"c:\foo\bar").split_drive(),
             (HgPath::new(b""), HgPath::new(br"c:\foo\bar"))
         );
         assert_eq!(
             HgPath::new(b"c:/foo/bar").split_drive(),
             (HgPath::new(b""), HgPath::new(br"c:/foo/bar"))
         );
         assert_eq!(
             HgPath::new(br"\\conky\mountpoint\foo\bar").split_drive(),
             (
                 HgPath::new(b""),
                 HgPath::new(br"\\conky\mountpoint\foo\bar")
             )
         );
     }
 
     #[test]
     #[cfg(windows)]
     fn test_split_drive() {
         assert_eq!(
             HgPath::new(br"c:\foo\bar").split_drive(),
             (HgPath::new(br"c:"), HgPath::new(br"\foo\bar"))
         );
         assert_eq!(
             HgPath::new(b"c:/foo/bar").split_drive(),
             (HgPath::new(br"c:"), HgPath::new(br"/foo/bar"))
         );
         assert_eq!(
             HgPath::new(br"\\conky\mountpoint\foo\bar").split_drive(),
             (
                 HgPath::new(br"\\conky\mountpoint"),
                 HgPath::new(br"\foo\bar")
             )
         );
         assert_eq!(
             HgPath::new(br"//conky/mountpoint/foo/bar").split_drive(),
             (
                 HgPath::new(br"//conky/mountpoint"),
                 HgPath::new(br"/foo/bar")
             )
         );
         assert_eq!(
             HgPath::new(br"\\\conky\mountpoint\foo\bar").split_drive(),
             (
                 HgPath::new(br""),
                 HgPath::new(br"\\\conky\mountpoint\foo\bar")
             )
         );
         assert_eq!(
             HgPath::new(br"///conky/mountpoint/foo/bar").split_drive(),
             (
                 HgPath::new(br""),
                 HgPath::new(br"///conky/mountpoint/foo/bar")
             )
         );
         assert_eq!(
             HgPath::new(br"\\conky\\mountpoint\foo\bar").split_drive(),
             (
                 HgPath::new(br""),
                 HgPath::new(br"\\conky\\mountpoint\foo\bar")
             )
         );
         assert_eq!(
             HgPath::new(br"//conky//mountpoint/foo/bar").split_drive(),
             (
                 HgPath::new(br""),
                 HgPath::new(br"//conky//mountpoint/foo/bar")
             )
         );
         // UNC part containing U+0130
         assert_eq!(
             HgPath::new(b"//conky/MOUNTPO\xc4\xb0NT/foo/bar").split_drive(),
             (
                 HgPath::new(b"//conky/MOUNTPO\xc4\xb0NT"),
                 HgPath::new(br"/foo/bar")
             )
         );
     }
 
     #[test]
     fn test_parent() {
         let path = HgPath::new(b"");
         assert_eq!(path.parent(), path);
 
         let path = HgPath::new(b"a");
         assert_eq!(path.parent(), HgPath::new(b""));
 
         let path = HgPath::new(b"a/b");
         assert_eq!(path.parent(), HgPath::new(b"a"));
 
         let path = HgPath::new(b"a/other/b");
         assert_eq!(path.parent(), HgPath::new(b"a/other"));
     }
 }