Mercurial > hg
view hgext/logtoprocess.py @ 29830:92ac2baaea86
revlog: use an LRU cache for delta chain bases
Profiling using statprof revealed a hotspot during changegroup
application calculating delta chain bases on generaldelta repos.
Essentially, revlog._addrevision() was performing a lot of redundant
work tracing the delta chain as part of determining when the chain
distance was acceptable. This was most pronounced when adding
revisions to manifests, which can have delta chains thousands of
revisions long.
There was a delta chain base cache on revlogs before, but it only
captured a single revision. This was acceptable before generaldelta,
when _addrevision would build deltas from the previous revision and
thus we'd pretty much guarantee a cache hit when resolving the delta
chain base on a subsequent _addrevision call. However, it isn't
suitable for generaldelta because parent revisions aren't necessarily
the last processed revision.
This patch converts the delta chain base cache to an LRU dict cache.
The cache can hold multiple entries, so generaldelta repos have a
higher chance of getting a cache hit.
The impact of this change when processing changegroup additions is
significant. On a generaldelta conversion of the "mozilla-unified"
repo (which contains heads of the main Firefox repositories in
chronological order - this means there are lots of transitions between
heads in revlog order), this change has the following impact when
performing an `hg unbundle` of an uncompressed bundle of the repo:
before: 5:42 CPU time
after: 4:34 CPU time
Most of this time is saved when applying the changelog and manifest
revlogs:
before: 2:30 CPU time
after: 1:17 CPU time
That nearly a 50% reduction in CPU time applying changesets and
manifests!
Applying a gzipped bundle of the same repo (effectively simulating a
`hg clone` over HTTP) showed a similar speedup:
before: 5:53 CPU time
after: 4:46 CPU time
Wall time improvements were basically the same as CPU time.
I didn't measure explicitly, but it feels like most of the time
is saved when processing manifests. This makes sense, as large
manifests tend to have very long delta chains and thus benefit the
most from this cache.
So, this change effectively makes changegroup application (which is
used by `hg unbundle`, `hg clone`, `hg pull`, `hg unshelve`, and
various other commands) significantly faster when delta chains are
long (which can happen on repos with large numbers of files and thus
large manifests).
In theory, this change can result in more memory utilization. However,
we're caching a dict of ints. At most we have 200 ints + Python object
overhead per revlog. And, the cache is really only populated when
performing read-heavy operations, such as adding changegroups or
scanning an individual revlog. For memory bloat to be an issue, we'd
need to scan/read several revisions from several revlogs all while
having active references to several revlogs. I don't think there are
many operations that do this, so I don't think memory bloat from the
cache will be an issue.
author | Gregory Szorc <gregory.szorc@gmail.com> |
---|---|
date | Mon, 22 Aug 2016 21:48:50 -0700 |
parents | 4cf0542dcbe7 |
children | d5883fd055c6 |
line wrap: on
line source
# logtoprocess.py - send ui.log() data to a subprocess # # Copyright 2016 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. """Send ui.log() data to a subprocess (EXPERIMENTAL) This extension lets you specify a shell command per ui.log() event, sending all remaining arguments to as environment variables to that command. Each positional argument to the method results in a `MSG[N]` key in the environment, starting at 1 (so `MSG1`, `MSG2`, etc.). Each keyword argument is set as a `OPT_UPPERCASE_KEY` variable (so the key is uppercased, and prefixed with `OPT_`). The original event name is passed in the `EVENT` environment variable, and the process ID of mercurial is given in `HGPID`. So given a call `ui.log('foo', 'bar', 'baz', spam='eggs'), a script configured for the `foo` event can expect an environment with `MSG1=bar`, `MSG2=baz`, and `OPT_SPAM=eggs`. Scripts are configured in the `[logtoprocess]` section, each key an event name. For example:: [logtoprocess] commandexception = echo "$MSG2$MSG3" > /var/log/mercurial_exceptions.log would log the warning message and traceback of any failed command dispatch. Scripts are run asychronously as detached daemon processes; mercurial will not ensure that they exit cleanly. """ from __future__ import absolute_import import itertools import os import platform import subprocess import sys # Note for extension authors: ONLY specify testedwith = 'internal' 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 = 'internal' def uisetup(ui): if platform.system() == 'Windows': # no fork on Windows, but we can create a detached process # https://msdn.microsoft.com/en-us/library/windows/desktop/ms684863.aspx # No stdlib constant exists for this value DETACHED_PROCESS = 0x00000008 _creationflags = DETACHED_PROCESS | subprocess.CREATE_NEW_PROCESS_GROUP def runshellcommand(script, env): # we can't use close_fds *and* redirect stdin. I'm not sure that we # need to because the detached process has no console connection. subprocess.Popen( script, shell=True, env=env, close_fds=True, creationflags=_creationflags) else: def runshellcommand(script, env): # double-fork to completely detach from the parent process # based on http://code.activestate.com/recipes/278731 pid = os.fork() if pid: # parent return # subprocess.Popen() forks again, all we need to add is # flag the new process as a new session. if sys.version_info < (3, 2): newsession = {'preexec_fn': os.setsid} else: newsession = {'start_new_session': True} try: # connect stdin to devnull to make sure the subprocess can't # muck up that stream for mercurial. subprocess.Popen( script, shell=True, stdin=open(os.devnull, 'r'), env=env, close_fds=True, **newsession) finally: # mission accomplished, this child needs to exit and not # continue the hg process here. os._exit(0) class logtoprocessui(ui.__class__): def log(self, event, *msg, **opts): """Map log events to external commands Arguments are passed on as environment variables. """ script = self.config('logtoprocess', event) if script: if msg: # try to format the log message given the remaining # arguments try: # Python string formatting with % either uses a # dictionary *or* tuple, but not both. If we have # keyword options, assume we need a mapping. formatted = msg[0] % (opts or msg[1:]) except (TypeError, KeyError): # Failed to apply the arguments, ignore formatted = msg[0] messages = (formatted,) + msg[1:] else: messages = msg # positional arguments are listed as MSG[N] keys in the # environment msgpairs = ( ('MSG{0:d}'.format(i), str(m)) for i, m in enumerate(messages, 1)) # keyword arguments get prefixed with OPT_ and uppercased optpairs = ( ('OPT_{0}'.format(key.upper()), str(value)) for key, value in opts.iteritems()) env = dict(itertools.chain(os.environ.items(), msgpairs, optpairs), EVENT=event, HGPID=str(os.getpid())) # Connect stdin to /dev/null to prevent child processes messing # with mercurial's stdin. runshellcommand(script, env) return super(logtoprocessui, self).log(event, *msg, **opts) # Replace the class for this instance and all clones created from it: ui.__class__ = logtoprocessui