hgext/blackbox.py

commit: improve the files field of changelog for merges

Currently, the files list of merge commits repeats all the deletions
(either actual deletions, or files that got renamed) that happened
between base and p2 of the merge. If p2 is the main branch, the list
can easily be much bigger than the change being merged.

This results in various problems worth improving:
- changelog is bigger than necessary
- `hg log directory` lists many unrelated merge commits, and `hg log
  -v -r commit` frequently fills multiple screens worth of files
- it possibly slows down adjustlinkrev, by forcing it to read more
  manifests, and that function can certainly be a bottleneck
- the server side of pulls can waste a lot of time simply opening the
  filelogs for pointless files (the constant factors for opening even
  a tiny filelog is apparently pretty bad)

So stop listing such files as described in the code.  Impacted merge
commits and their descendants get a different hash than they would
have without this. This doesn't seem problematic, except for
convert. The previous commit helped with that in the hg->hg case (but
if you do svn->hg twice from scratch, hashes can still change).

The rest of the description is numbers. I don't have much to report,
because recreating the files list of existing repositories is not
easy:
- debugupgradeformat and bundle/unbundle don't recreate the list
- export/import tends to choke quickly applying patches or on
  description that contain diffs,
- merge commits from the convert extension don't have the right files
  list for reasons orthogonal to the current commit
- replaying the merge with hg update/hg merge/hg revert --all/hg
  commit can end up failing in hg revert
- I wasn't sure that using debugsetparents + debugrebuilddirstate
  would really build the right thing

I measured commit time before and after this change, in a case with no
files filtered out, several files filtered out (no difference) and 5k
files filtered out (+1% time).

Recreating the 100 more recent merges in a private repo, the
concatenated uncompressed files lists goes from 1.12MB to
0.52MB. Excluding 3 merges that are not representative, then the size
goes from 570k to 15k.
I converted part of mozilla-central, and observed file list shrinking
quite a bit too, starting at the very first merge, 733641d9feaf, going
from 550 files to 10 files (although they have relatively few merges,
so they probably wouldn't care).

Differential Revision: https://phab.mercurial-scm.org/D6613

# blackbox.py - log repository events to a file for post-mortem debugging
#
# Copyright 2010 Nicolas Dumazet
# Copyright 2013 Facebook, Inc.
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.

"""log repository events to a blackbox for debugging

Logs event information to .hg/blackbox.log to help debug and diagnose problems.
The events that get logged can be configured via the blackbox.track and
blackbox.ignore config keys.

Examples::

  [blackbox]
  track = *
  ignore = pythonhook
  # dirty is *EXPENSIVE* (slow);
  # each log entry indicates `+` if the repository is dirty, like :hg:`id`.
  dirty = True
  # record the source of log messages
  logsource = True

  [blackbox]
  track = command, commandfinish, commandexception, exthook, pythonhook

  [blackbox]
  track = incoming

  [blackbox]
  # limit the size of a log file
  maxsize = 1.5 MB
  # rotate up to N log files when the current one gets too big
  maxfiles = 3

  [blackbox]
  # Include nanoseconds in log entries with %f (see Python function
  # datetime.datetime.strftime)
  date-format = '%Y-%m-%d @ %H:%M:%S.%f'

"""

from __future__ import absolute_import

import re

from mercurial.i18n import _
from mercurial.node import hex

from mercurial import (
    encoding,
    loggingutil,
    registrar,
)
from mercurial.utils import (
    dateutil,
    procutil,
)

# Note for extension authors: ONLY specify testedwith = 'ships-with-hg-core' for
# extensions which SHIP WITH MERCURIAL. Non-mainline extensions should
# be specifying the version(s) of Mercurial they are tested with, or
# leave the attribute unspecified.
testedwith = 'ships-with-hg-core'

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

configtable = {}
configitem = registrar.configitem(configtable)

configitem('blackbox', 'dirty',
    default=False,
)
configitem('blackbox', 'maxsize',
    default='1 MB',
)
configitem('blackbox', 'logsource',
    default=False,
)
configitem('blackbox', 'maxfiles',
    default=7,
)
configitem('blackbox', 'track',
    default=lambda: ['*'],
)
configitem('blackbox', 'ignore',
    default=lambda: ['chgserver', 'cmdserver', 'extension'],
)
configitem('blackbox', 'date-format',
    default='%Y/%m/%d %H:%M:%S',
)

_lastlogger = loggingutil.proxylogger()

class blackboxlogger(object):
    def __init__(self, ui, repo):
        self._repo = repo
        self._trackedevents = set(ui.configlist('blackbox', 'track'))
        self._ignoredevents = set(ui.configlist('blackbox', 'ignore'))
        self._maxfiles = ui.configint('blackbox', 'maxfiles')
        self._maxsize = ui.configbytes('blackbox', 'maxsize')
        self._inlog = False

    def tracked(self, event):
        return ((b'*' in self._trackedevents
                 and event not in self._ignoredevents)
                or event in self._trackedevents)

    def log(self, ui, event, msg, opts):
        # self._log() -> ctx.dirty() may create new subrepo instance, which
        # ui is derived from baseui. So the recursion guard in ui.log()
        # doesn't work as it's local to the ui instance.
        if self._inlog:
            return
        self._inlog = True
        try:
            self._log(ui, event, msg, opts)
        finally:
            self._inlog = False

    def _log(self, ui, event, msg, opts):
        default = ui.configdate('devel', 'default-date')
        date = dateutil.datestr(default, ui.config('blackbox', 'date-format'))
        user = procutil.getuser()
        pid = '%d' % procutil.getpid()
        changed = ''
        ctx = self._repo[None]
        parents = ctx.parents()
        rev = ('+'.join([hex(p.node()) for p in parents]))
        if (ui.configbool('blackbox', 'dirty') and
            ctx.dirty(missing=True, merge=False, branch=False)):
            changed = '+'
        if ui.configbool('blackbox', 'logsource'):
            src = ' [%s]' % event
        else:
            src = ''
        try:
            fmt = '%s %s @%s%s (%s)%s> %s'
            args = (date, user, rev, changed, pid, src, msg)
            with loggingutil.openlogfile(
                    ui, self._repo.vfs, name='blackbox.log',
                    maxfiles=self._maxfiles, maxsize=self._maxsize) as fp:
                fp.write(fmt % args)
        except (IOError, OSError) as err:
            # deactivate this to avoid failed logging again
            self._trackedevents.clear()
            ui.debug('warning: cannot write to blackbox.log: %s\n' %
                     encoding.strtolocal(err.strerror))
            return
        _lastlogger.logger = self

def uipopulate(ui):
    ui.setlogger(b'blackbox', _lastlogger)

def reposetup(ui, repo):
    # During 'hg pull' a httppeer repo is created to represent the remote repo.
    # It doesn't have a .hg directory to put a blackbox in, so we don't do
    # the blackbox setup for it.
    if not repo.local():
        return

    # Since blackbox.log is stored in the repo directory, the logger should be
    # instantiated per repository.
    logger = blackboxlogger(ui, repo)
    ui.setlogger(b'blackbox', logger)

    # Set _lastlogger even if ui.log is not called. This gives blackbox a
    # fallback place to log
    if _lastlogger.logger is None:
        _lastlogger.logger = logger

    repo._wlockfreeprefix.add('blackbox.log')

@command('blackbox',
    [('l', 'limit', 10, _('the number of events to show')),
    ],
    _('hg blackbox [OPTION]...'),
    helpcategory=command.CATEGORY_MAINTENANCE,
    helpbasic=True)
def blackbox(ui, repo, *revs, **opts):
    '''view the recent repository events
    '''

    if not repo.vfs.exists('blackbox.log'):
        return

    limit = opts.get(r'limit')
    fp = repo.vfs('blackbox.log', 'r')
    lines = fp.read().split('\n')

    count = 0
    output = []
    for line in reversed(lines):
        if count >= limit:
            break

        # count the commands by matching lines like: 2013/01/23 19:13:36 root>
        if re.match(br'^\d{4}/\d{2}/\d{2} \d{2}:\d{2}:\d{2} .*> .*', line):
            count += 1
        output.append(line)

    ui.status('\n'.join(reversed(output)))