Mercurial > hg
view mercurial/logexchange.py @ 42745:4d20b1fe8a72
rust-discovery: using from Python code
As previously done in other topics, the Rust version is used if it's been
built.
The version fully in Rust of the partialdiscovery class has the performance
advantage over the Python version (actually using the Rust MissingAncestor) if
the undecided set is big enough. Otherwise no sampling occurs, and the
discovery is reasonably fast anyway.
Note: it's hard to predict the size of the initial undecided set, it can
depend on the kind of topological changes between the local and remote graphs.
The point of the Rust version is to make the bad cases acceptable.
More specifically, the performance advantages are:
- faster sampling, especially takefullsample()
- much faster addmissings() in almost all cases (see commit message in
grandparent of the present changeset)
- no conversion cost of the undecided set at the interface between Rust and
Python
== Measurements with big undecided sets
For an extreme example, discovery between mozilla-try and mozilla-unified
(over one million undecided revisions, same case as in dbd0fcca6dfc), we
get roughly a x2.5/x3 better performance:
Growing sample size (5% starting with 200): time goes down from
210 to 72 seconds.
Constant sample size of 200: time down from 1853 to 659 seconds.
With a sample size computed from number of roots and heads of the
undecided set (`respectsize` is `False`), here are perfdiscovery results:
Before ! wall 9.358729 comb 9.360000 user 9.310000 sys 0.050000 (median of 50)
After ! wall 3.793819 comb 3.790000 user 3.750000 sys 0.040000 (median of 50)
In that later case, the sample sizes are routinely in the hundreds of
thousands of revisions. While still faster, the Rust iteration in
addmissings has less of an advantage than with smaller sample sizes, but
one sees addcommons becoming faster, probably a consequence of not having
to copy big sets back and forth.
This example is not a goal in itself, but it showcases several different
areas in which the process can become slow, due to different factors, and
how this full Rust version can help.
== Measurements with small undecided sets
In cases the undecided set is small enough than no sampling occurs,
the Rust version has a disadvantage at init if `targetheads` is really big
(some time is lost in the translation to Rust data structures),
and that is compensated by the faster `addmissings()`.
On a private repository with over one million commits, we still get a minor
improvement, of 6.8%:
Before ! wall 0.593585 comb 0.590000 user 0.550000 sys 0.040000 (median of 50)
After ! wall 0.553035 comb 0.550000 user 0.520000 sys 0.030000 (median of 50)
What's interesting in that case is the first addinfo() at 180ms for Rust and
233ms for Python+C, mostly due to add_missings and the children cache
computation being done in less than 0.2ms on the Rust side vs over 40ms on the
Python side.
The worst case we have on hand is with mozilla-try, prepared with
discovery-helper.sh for 10 heads and depth 10, time goes up 2.2% on the median.
In this case `targetheads` is really huge with 165842 server heads.
Before ! wall 0.823884 comb 0.810000 user 0.790000 sys 0.020000 (median of 50)
After ! wall 0.842607 comb 0.840000 user 0.800000 sys 0.040000 (median of 50)
If that would be considered a problem, more adjustments can be made, which are
prematurate at this stage: cooking special variants of methods of the inner
MissingAncestors object, retrieving local heads directly from Rust to avoid
the cost of conversion. Effort would probably be better spent at this point
improving the surroundings if needed.
Here's another data point with a smaller repository, pypy, where performance
is almost identical
Before ! wall 0.015121 comb 0.030000 user 0.020000 sys 0.010000 (median of 186)
After ! wall 0.015009 comb 0.010000 user 0.010000 sys 0.000000 (median of 184)
Differential Revision: https://phab.mercurial-scm.org/D6430
author | Georges Racinet <georges.racinet@octobus.net> |
---|---|
date | Wed, 20 Feb 2019 09:04:54 +0100 |
parents | 876494fd967d |
children | 2372284d9457 |
line wrap: on
line source
# logexchange.py # # Copyright 2017 Augie Fackler <raf@durin42.com> # Copyright 2017 Sean Farley <sean@farley.io> # # 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 from .node import hex from . import ( util, vfs as vfsmod, ) # directory name in .hg/ in which remotenames files will be present remotenamedir = 'logexchange' def readremotenamefile(repo, filename): """ reads a file from .hg/logexchange/ directory and yields it's content filename: the file to be read yield a tuple (node, remotepath, name) """ vfs = vfsmod.vfs(repo.vfs.join(remotenamedir)) if not vfs.exists(filename): return f = vfs(filename) lineno = 0 for line in f: line = line.strip() if not line: continue # contains the version number if lineno == 0: lineno += 1 try: node, remote, rname = line.split('\0') yield node, remote, rname except ValueError: pass f.close() def readremotenames(repo): """ read the details about the remotenames stored in .hg/logexchange/ and yields a tuple (node, remotepath, name). It does not yields information about whether an entry yielded is branch or bookmark. To get that information, call the respective functions. """ for bmentry in readremotenamefile(repo, 'bookmarks'): yield bmentry for branchentry in readremotenamefile(repo, 'branches'): yield branchentry def writeremotenamefile(repo, remotepath, names, nametype): vfs = vfsmod.vfs(repo.vfs.join(remotenamedir)) f = vfs(nametype, 'w', atomictemp=True) # write the storage version info on top of file # version '0' represents the very initial version of the storage format f.write('0\n\n') olddata = set(readremotenamefile(repo, nametype)) # re-save the data from a different remote than this one. for node, oldpath, rname in sorted(olddata): if oldpath != remotepath: f.write('%s\0%s\0%s\n' % (node, oldpath, rname)) for name, node in sorted(names.iteritems()): if nametype == "branches": for n in node: f.write('%s\0%s\0%s\n' % (n, remotepath, name)) elif nametype == "bookmarks": if node: f.write('%s\0%s\0%s\n' % (node, remotepath, name)) f.close() def saveremotenames(repo, remotepath, branches=None, bookmarks=None): """ save remotenames i.e. remotebookmarks and remotebranches in their respective files under ".hg/logexchange/" directory. """ wlock = repo.wlock() try: if bookmarks: writeremotenamefile(repo, remotepath, bookmarks, 'bookmarks') if branches: writeremotenamefile(repo, remotepath, branches, 'branches') finally: wlock.release() def activepath(repo, remote): """returns remote path""" # is the remote a local peer local = remote.local() # determine the remote path from the repo, if possible; else just # use the string given to us rpath = remote if local: rpath = util.pconvert(remote._repo.root) elif not isinstance(remote, bytes): rpath = remote._url # represent the remotepath with user defined path name if exists for path, url in repo.ui.configitems('paths'): # remove auth info from user defined url noauthurl = util.removeauth(url) # Standardize on unix style paths, otherwise some {remotenames} end up # being an absolute path on Windows. url = util.pconvert(bytes(url)) noauthurl = util.pconvert(noauthurl) if url == rpath or noauthurl == rpath: rpath = path break return rpath def pullremotenames(localrepo, remoterepo): """ pulls bookmarks and branches information of the remote repo during a pull or clone operation. localrepo is our local repository remoterepo is the peer instance """ remotepath = activepath(localrepo, remoterepo) with remoterepo.commandexecutor() as e: bookmarks = e.callcommand('listkeys', { 'namespace': 'bookmarks', }).result() # on a push, we don't want to keep obsolete heads since # they won't show up as heads on the next pull, so we # remove them here otherwise we would require the user # to issue a pull to refresh the storage bmap = {} repo = localrepo.unfiltered() with remoterepo.commandexecutor() as e: branchmap = e.callcommand('branchmap', {}).result() for branch, nodes in branchmap.iteritems(): bmap[branch] = [] for node in nodes: if node in repo and not repo[node].obsolete(): bmap[branch].append(hex(node)) saveremotenames(localrepo, remotepath, bmap, bookmarks)