view tests/autodiff.py @ 47343:9f798c1b0d89 stable

cext: fix memory leak in phases computation

Without this a buffer whose size in bytes is the number of
changesets in the repository is leaked each time the repository is
opened and changeset phases are computed.

Impact: the current code in hgwebdir creates a new `localrepository`
instance for each HTTP request. Since any pull or push is made of several
requests, a team of 100 people can easily produce thousands of such
requests per day.

Being a low-level malloc, this leak can't be seen with the gc module and
tools relying on that, but was spotted by valgrind immediately.

Reproduction
------------

  for i in range(cl_args.iterations):
      repo = hg.repository(baseui, repo_path)
      rev = repo.revs(rev).first()
      ctx = repo[rev]

  del ctx
  del repo
  # avoid any pollution by other type of leak
  # (that should be fixed in 5.8)
  repoview._filteredrepotypes.clear()

  gc.collect()

Measurements
------------

Resident Set Size (RSS), taken on a clone of
mozilla-central for performance analysis (440 000
changesets).

before:
  5.8+hg19.5ac0f2a8ba72  1000 iterations: 1606MB
  5.8+hg19.5ac0f2a8ba72 10000 iterations: 5723MB
after:
  5.8+hg20.e2084d39e145  1000 iterations:  555MB
  5.8+hg20.e2084d39e145 10000 iterations:  555MB
         (double checked, not a copy/paste error)

(e2084d39e14 is the present changeset, before amendment
of the message to add the measurements)

# Extension dedicated to test patch.diff() upgrade modes

from __future__ import absolute_import

from mercurial import (
    error,
    patch,
    pycompat,
    registrar,
    scmutil,
)

cmdtable = {}
command = registrar.command(cmdtable)


@command(
    b'autodiff',
    [(b'', b'git', b'', b'git upgrade mode (yes/no/auto/warn/abort)')],
    b'[OPTION]... [FILE]...',
)
def autodiff(ui, repo, *pats, **opts):
    opts = pycompat.byteskwargs(opts)
    diffopts = patch.difffeatureopts(ui, opts)
    git = opts.get(b'git', b'no')
    brokenfiles = set()
    losedatafn = None
    if git in (b'yes', b'no'):
        diffopts.git = git == b'yes'
        diffopts.upgrade = False
    elif git == b'auto':
        diffopts.git = False
        diffopts.upgrade = True
    elif git == b'warn':
        diffopts.git = False
        diffopts.upgrade = True

        def losedatafn(fn=None, **kwargs):
            brokenfiles.add(fn)
            return True

    elif git == b'abort':
        diffopts.git = False
        diffopts.upgrade = True

        def losedatafn(fn=None, **kwargs):
            raise error.Abort(b'losing data for %s' % fn)

    else:
        raise error.Abort(b'--git must be yes, no or auto')

    ctx1, ctx2 = scmutil.revpair(repo, [])
    m = scmutil.match(ctx2, pats, opts)
    it = patch.diff(
        repo,
        ctx1.node(),
        ctx2.node(),
        match=m,
        opts=diffopts,
        losedatafn=losedatafn,
    )
    for chunk in it:
        ui.write(chunk)
    for fn in sorted(brokenfiles):
        ui.write((b'data lost for: %s\n' % fn))