view mercurial/grep.py @ 47010:76ae43d5b1db stable

repoview: fix memory leak of filtered repo classes

The leak occurs in long-running server processes with
extensions, and is measured at 110kB per request.

Before this change, the contents of the `_filteredrepotypes`
cache are not properly garbage collected, despite it begin
a `WeakKeyDictionary`.

Extensions have a tendency to generate a new repository class
for each `localrepo` instantiation. Server processes based
on `hgwebdir_mod` will instantiate a new `localrepo` for each
HTTP request that involves a repository.

As a result, with a testing process that repeatedly opens a
repository with several extensions activated
(`topic` notably among them), we see a steady increase in
resident memory of 110kB per repository instantiation before this
change. This is also true, if we call `gc.collect()` at each
instantiation, like `hgwebdir_mod` does, or not.

The cause of the leak is that the *values* aren't weak references.
This change uses `weakref.ref` for the values, and this makes
in our measurements the resident size increase drop to 5kB per
repository instantiation, with no explicit call of `gc.collect()`
at all.
There is currently no reason to believe that this remaining leak
of 5kB is related to or even due to Mercurial core.

We've also seen evidence that `ui.ui` instances weren't properly
garbage collected before the change (with the change, they are).
This could explain why the figures are relatively high.

In theory, the collection of weak references could lead to
much more misses in the cache, so we measured the impact on
the original case that was motivation for introducing that cache
in 7e89bd0cfb86 (see also issue5043): `hg convert` of the
mozilla-central repository. The bad news here is that there is a
major memory leak there, both with and without the present changeset.
There were no more cache misses, and we could see no
more memory leak with this change: the resident size after importing
roughly 100000 changesets was at 12.4GB before, and 12.5GB after.
The small increase is mentioned for completeness only, and we
believe that it should be ignored, at least as long as the main
leak isn't fixed. At less than 1% of the main leak, even finding out
whether it is merely noise would be wasteful.

Original context where this was spotted and first mitigated:
  https://foss.heptapod.net/heptapod/heptapod/-/issues/466

The leak reduction was also obtained in Heptapod inner HTTP server,
which amounts to the same as `hgwebdir_mod` for these questions.

The measurements done with Python 3.9, similar figures seen with 3.8.
More work on our side would be needed to give measurements with 2.7,
because of testing server process does not support it.

# grep.py - logic for history walk and grep
#
# Copyright 2005-2007 Olivia Mackall <[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.

from __future__ import absolute_import

import difflib
import errno

from .i18n import _

from . import (
    error,
    match as matchmod,
    pycompat,
    scmutil,
    util,
)


def matchlines(body, regexp):
    begin = 0
    linenum = 0
    while begin < len(body):
        match = regexp.search(body, begin)
        if not match:
            break
        mstart, mend = match.span()
        linenum += body.count(b'\n', begin, mstart) + 1
        lstart = body.rfind(b'\n', begin, mstart) + 1 or begin
        begin = body.find(b'\n', mend) + 1 or len(body) + 1
        lend = begin - 1
        yield linenum, mstart - lstart, mend - lstart, body[lstart:lend]


class linestate(object):
    def __init__(self, line, linenum, colstart, colend):
        self.line = line
        self.linenum = linenum
        self.colstart = colstart
        self.colend = colend

    def __hash__(self):
        return hash(self.line)

    def __eq__(self, other):
        return self.line == other.line

    def findpos(self, regexp):
        """Iterate all (start, end) indices of matches"""
        yield self.colstart, self.colend
        p = self.colend
        while p < len(self.line):
            m = regexp.search(self.line, p)
            if not m:
                break
            if m.end() == p:
                p += 1
            else:
                yield m.span()
                p = m.end()


def difflinestates(a, b):
    sm = difflib.SequenceMatcher(None, a, b)
    for tag, alo, ahi, blo, bhi in sm.get_opcodes():
        if tag == 'insert':
            for i in pycompat.xrange(blo, bhi):
                yield (b'+', b[i])
        elif tag == 'delete':
            for i in pycompat.xrange(alo, ahi):
                yield (b'-', a[i])
        elif tag == 'replace':
            for i in pycompat.xrange(alo, ahi):
                yield (b'-', a[i])
            for i in pycompat.xrange(blo, bhi):
                yield (b'+', b[i])


class grepsearcher(object):
    """Search files and revisions for lines matching the given pattern

    Options:
    - all_files to search unchanged files at that revision.
    - diff to search files in the parent revision so diffs can be generated.
    - follow to skip files across copies and renames.
    """

    def __init__(
        self, ui, repo, regexp, all_files=False, diff=False, follow=False
    ):
        self._ui = ui
        self._repo = repo
        self._regexp = regexp
        self._all_files = all_files
        self._diff = diff
        self._follow = follow

        self._getfile = util.lrucachefunc(repo.file)
        self._getrenamed = scmutil.getrenamedfn(repo)

        self._matches = {}
        self._copies = {}
        self._skip = set()
        self._revfiles = {}

    def skipfile(self, fn, rev):
        """Exclude the given file (and the copy at the specified revision)
        from future search"""
        copy = self._copies.get(rev, {}).get(fn)
        self._skip.add(fn)
        if copy:
            self._skip.add(copy)

    def searchfiles(self, revs, makefilematcher):
        """Walk files and revisions to yield (fn, ctx, pstates, states)
        matches

        states is a list of linestate objects. pstates may be empty unless
        diff is True.
        """
        for ctx in scmutil.walkchangerevs(
            self._repo, revs, makefilematcher, self._prep
        ):
            rev = ctx.rev()
            parent = ctx.p1().rev()
            for fn in sorted(self._revfiles.get(rev, [])):
                states = self._matches[rev][fn]
                copy = self._copies.get(rev, {}).get(fn)
                if fn in self._skip:
                    if copy:
                        self._skip.add(copy)
                    continue
                pstates = self._matches.get(parent, {}).get(copy or fn, [])
                if pstates or states:
                    yield fn, ctx, pstates, states
            del self._revfiles[rev]
            # We will keep the matches dict for the duration of the window
            # clear the matches dict once the window is over
            if not self._revfiles:
                self._matches.clear()

    def _grepbody(self, fn, rev, body):
        self._matches[rev].setdefault(fn, [])
        m = self._matches[rev][fn]
        if body is None:
            return

        for lnum, cstart, cend, line in matchlines(body, self._regexp):
            s = linestate(line, lnum, cstart, cend)
            m.append(s)

    def _readfile(self, ctx, fn):
        rev = ctx.rev()
        if rev is None:
            fctx = ctx[fn]
            try:
                return fctx.data()
            except IOError as e:
                if e.errno != errno.ENOENT:
                    raise
        else:
            flog = self._getfile(fn)
            fnode = ctx.filenode(fn)
            try:
                return flog.read(fnode)
            except error.CensoredNodeError:
                self._ui.warn(
                    _(
                        b'cannot search in censored file: '
                        b'%(filename)s:%(revnum)s\n'
                    )
                    % {b'filename': fn, b'revnum': pycompat.bytestr(rev)}
                )

    def _prep(self, ctx, fmatch):
        rev = ctx.rev()
        pctx = ctx.p1()
        self._matches.setdefault(rev, {})
        if self._diff:
            parent = pctx.rev()
            self._matches.setdefault(parent, {})
        files = self._revfiles.setdefault(rev, [])
        if rev is None:
            # in `hg grep pattern`, 2/3 of the time is spent is spent in
            # pathauditor checks without this in mozilla-central
            contextmanager = self._repo.wvfs.audit.cached
        else:
            contextmanager = util.nullcontextmanager
        with contextmanager():
            # TODO: maybe better to warn missing files?
            if self._all_files:
                fmatch = matchmod.badmatch(fmatch, lambda f, msg: None)
                filenames = ctx.matches(fmatch)
            else:
                filenames = (f for f in ctx.files() if fmatch(f))
            for fn in filenames:
                # fn might not exist in the revision (could be a file removed by
                # the revision). We could check `fn not in ctx` even when rev is
                # None, but it's less racy to protect againt that in readfile.
                if rev is not None and fn not in ctx:
                    continue

                copy = None
                if self._follow:
                    copy = self._getrenamed(fn, rev)
                    if copy:
                        self._copies.setdefault(rev, {})[fn] = copy
                        if fn in self._skip:
                            self._skip.add(copy)
                if fn in self._skip:
                    continue
                files.append(fn)

                if fn not in self._matches[rev]:
                    self._grepbody(fn, rev, self._readfile(ctx, fn))

                if self._diff:
                    pfn = copy or fn
                    if pfn not in self._matches[parent] and pfn in pctx:
                        self._grepbody(pfn, parent, self._readfile(pctx, pfn))