##// END OF EJS Templates
nodemap: introduce an option to use mmap to read the nodemap mapping...
nodemap: introduce an option to use mmap to read the nodemap mapping The performance and memory benefit is much greater if we don't have to copy all the data in memory for each information. So we introduce an option (on by default) to read the data using mmap. This changeset is the last one definition the API for index support nodemap data. (they have to be able to use the mmaping). Below are some benchmark comparing the best we currently have in 5.3 with the final step of this series (using the persistent nodemap implementation in Rust). The benchmark run `hg perfindex` with various revset and the following variants: Before: * do not use the persistent nodemap * use the CPython implementation of the index for nodemap * use mmapping of the changelog index After: * use the MixedIndex Rust code, with the NodeTree object for nodemap access (still in review) * use the persistent nodemap data from disk * access the persistent nodemap data through mmap * use mmapping of the changelog index The persistent nodemap greatly speed up most operation on very large repositories. Some of the previously very fast lookup end up a bit slower because the persistent nodemap has to be setup. However the absolute slowdown is very small and won't matters in the big picture. Here are some numbers (in seconds) for the reference copy of mozilla-try: Revset Before After abs-change speedup -10000: 0.004622 0.005532 0.000910 × 0.83 -10: 0.000050 0.000132 0.000082 × 0.37 tip 0.000052 0.000085 0.000033 × 0.61 0 + (-10000:) 0.028222 0.005337 -0.022885 × 5.29 0 0.023521 0.000084 -0.023437 × 280.01 (-10000:) + 0 0.235539 0.005308 -0.230231 × 44.37 (-10:) + :9 0.232883 0.000180 -0.232703 ×1293.79 (-10000:) + (:99) 0.238735 0.005358 -0.233377 × 44.55 :99 + (-10000:) 0.317942 0.005593 -0.312349 × 56.84 :9 + (-10:) 0.313372 0.000179 -0.313193 ×1750.68 :9 0.316450 0.000143 -0.316307 ×2212.93 On smaller repositories, the cost of nodemap related operation is not as big, so the win is much more modest. Yet it helps shaving a handful of millisecond here and there. Here are some numbers (in seconds) for the reference copy of mercurial: Revset Before After abs-change speedup -10: 0.000065 0.000097 0.000032 × 0.67 tip 0.000063 0.000078 0.000015 × 0.80 0 0.000561 0.000079 -0.000482 × 7.10 -10000: 0.004609 0.003648 -0.000961 × 1.26 0 + (-10000:) 0.005023 0.003715 -0.001307 × 1.35 (-10:) + :9 0.002187 0.000108 -0.002079 ×20.25 (-10000:) + 0 0.006252 0.003716 -0.002536 × 1.68 (-10000:) + (:99) 0.006367 0.003707 -0.002660 × 1.71 :9 + (-10:) 0.003846 0.000110 -0.003736 ×34.96 :9 0.003854 0.000099 -0.003755 ×38.92 :99 + (-10000:) 0.007644 0.003778 -0.003866 × 2.02 Differential Revision: https://phab.mercurial-scm.org/D7894

File last commit:

r44783:4caac36c default
r44843:f7459da7 default
Show More
files.rs
384 lines | 11.2 KiB | application/rls-services+xml | RustLexer
// files.rs
//
// Copyright 2019
// Raphaël Gomès <rgomes@octobus.net>,
// Yuya Nishihara <yuya@tcha.org>
//
// 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::ToOwned;
use std::fs::Metadata;
use std::iter::FusedIterator;
use std::ops::Deref;
use std::path::{Path, PathBuf};
pub fn get_path_from_bytes(bytes: &[u8]) -> &Path {
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>.
Path::new(os_str)
}
// 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> {
use std::os::unix::ffi::OsStrExt;
path.as_ref().as_os_str().as_bytes().to_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<'a>(path: &'a HgPath) -> Ancestors<'a> {
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<'a>(
path: &'a HgPath,
) -> AncestorsWithBase<'a> {
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())
}
#[derive(Eq, PartialEq, Ord, PartialOrd, Copy, Clone)]
pub struct HgMetadata {
pub st_dev: u64,
pub st_mode: u32,
pub st_nlink: u64,
pub st_size: u64,
pub st_mtime: i64,
pub st_ctime: i64,
}
// TODO support other plaforms
#[cfg(unix)]
impl HgMetadata {
pub fn from_metadata(metadata: Metadata) -> Self {
use std::os::unix::fs::MetadataExt;
Self {
st_dev: metadata.dev(),
st_mode: metadata.mode(),
st_nlink: metadata.nlink(),
st_size: metadata.size(),
st_mtime: metadata.mtime(),
st_ctime: metadata.ctime(),
}
}
}
/// 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)
} else {
name.to_owned()
};
let mut 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)?)?;
return Ok(name.to_owned());
} else if name == root {
return 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);
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)?)?;
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.to_owned(),
root: root.to_owned(),
})
}
}
#[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.to_owned(),
root: root.to_owned()
})
);
assert_eq!(
canonical_path(&root, Path::new(""), &under_repo_symlink),
Ok(PathBuf::from("d"))
);
}
}