##// END OF EJS Templates
upstream » mercurial-mirror
RSS

Clone from https://www.mercurial-scm.org/repo/hg-all

xdiff: add a preprocessing step that trims files...
xdiff: add a preprocessing step that trims files xdiff has a `xdl_trim_ends` step that removes common lines, unmatchable lines. That is in theory good, but happens too late - after splitting, hashing, and adjusting the hash values so they are unique. Those splitting, hashing and adjusting hash values steps could have noticeable overhead. Diffing two large files with minor (one-line-ish) changes are not uncommon. In that case, the raw performance of those preparation steps seriously matter. Even allocating an O(N) array and storing line offsets to it is expensive. Therefore my previous attempts [1] [2] cannot be good enough since they do not remove the O(N) array assignment. This patch adds a preprocessing step - `xdl_trim_files` that runs before other preprocessing steps. It counts common prefix and suffix and lines in them (needed for displaying line number), without doing anything else. Testing with a crafted large (169MB) file, with minor change: ``` open('a','w').write(''.join('%s\n' % (i % 100000) for i in xrange(30000000) if i != 6000000)) open('b','w').write(''.join('%s\n' % (i % 100000) for i in xrange(30000000) if i != 6003000)) ``` Running xdiff by a simple binary [3], this patch improves the xdiff perf by more than 10x for the above case: ``` # xdiff before this patch 2.41s user 1.13s system 98% cpu 3.592 total # xdiff after this patch 0.14s user 0.16s system 98% cpu 0.309 total # gnu diffutils 0.12s user 0.15s system 98% cpu 0.272 total # (best of 20 runs) ``` It's still slightly slower than GNU diffutils. But it's pretty close now. Testing with real repo data: For the whole repo, this patch makes xdiff 25% faster: ``` # hg perfbdiff --count 100 --alldata -c d334afc585e2 --blocks [--xdiff] # xdiff, after ! wall 0.058861 comb 0.050000 user 0.050000 sys 0.000000 (best of 100) # xdiff, before ! wall 0.077816 comb 0.080000 user 0.080000 sys 0.000000 (best of 91) # bdiff ! wall 0.117473 comb 0.120000 user 0.120000 sys 0.000000 (best of 67) ``` For files that are long (ex. commands.py), the speedup is more than 3x, very significant: ``` # hg perfbdiff --count 3000 --blocks commands.py.i 1 [--xdiff] # xdiff, after ! wall 0.690583 comb 0.690000 user 0.690000 sys 0.000000 (best of 12) # xdiff, before ! wall 2.240361 comb 2.210000 user 2.210000 sys 0.000000 (best of 4) # bdiff ! wall 2.469852 comb 2.440000 user 2.440000 sys 0.000000 (best of 4) ``` [1]: https://phab.mercurial-scm.org/D2631 [2]: https://phab.mercurial-scm.org/D2634 [3]: ``` // Code to run xdiff from command line. No proper error handling. #include <stdlib.h> #include <unistd.h> #include <sys/types.h> #include <sys/stat.h> #include <fcntl.h> #include "mercurial/thirdparty/xdiff/xdiff.h" #define ensure(x) if (!(x)) exit(255); mmfile_t readfile(const char *path) { struct stat st; int fd = open(path, O_RDONLY); fstat(fd, &st); mmfile_t file = { malloc(st.st_size), st.st_size }; ensure(read(fd, file.ptr, st.st_size) == st.st_size); close(fd); return file; } int main(int argc, char const *argv[]) { mmfile_t a = readfile(argv[1]), b = readfile(argv[2]); xpparam_t xpp = {0}; xdemitconf_t xecfg = {0}; xdemitcb_t ecb = {0}; xdl_diff(&a, &b, &xpp, &xecfg, &ecb); return 0; } ``` Differential Revision: https://phab.mercurial-scm.org/D2686
Pierre-Yves David - - Load All Authors

File last commit:

r26587:56b2bcea default
r36838:f33a87cf default
Show More
treediscovery.py
159 lines | 5.2 KiB | text/x-python | PythonLexer
/ mercurial / treediscovery.py
History | Annotation | Raw |Copy content |Copy permalink
# discovery.py - protocol changeset discovery functions
#
# Copyright 2010 Matt Mackall <mpm@selenic.com>
#
# 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 collections
from .i18n import _
from .node import (
nullid,
short,
)
from . import (
error,
)
def findcommonincoming(repo, remote, heads=None, force=False):
"""Return a tuple (common, fetch, heads) used to identify the common
subset of nodes between repo and remote.
"common" is a list of (at least) the heads of the common subset.
"fetch" is a list of roots of the nodes that would be incoming, to be
supplied to changegroupsubset.
"heads" is either the supplied heads, or else the remote's heads.
"""
knownnode = repo.changelog.hasnode
search = []
fetch = set()
seen = set()
seenbranch = set()
base = set()
if not heads:
heads = remote.heads()
if repo.changelog.tip() == nullid:
base.add(nullid)
if heads != [nullid]:
return [nullid], [nullid], list(heads)
return [nullid], [], heads
# assume we're closer to the tip than the root
# and start by examining the heads
repo.ui.status(_("searching for changes\n"))
unknown = []
for h in heads:
if not knownnode(h):
unknown.append(h)
else:
base.add(h)
if not unknown:
return list(base), [], list(heads)
req = set(unknown)
reqcnt = 0
# search through remote branches
# a 'branch' here is a linear segment of history, with four parts:
# head, root, first parent, second parent
# (a branch always has two parents (or none) by definition)
unknown = collections.deque(remote.branches(unknown))
while unknown:
r = []
while unknown:
n = unknown.popleft()
if n[0] in seen:
continue
repo.ui.debug("examining %s:%s\n"
% (short(n[0]), short(n[1])))
if n[0] == nullid: # found the end of the branch
pass
elif n in seenbranch:
repo.ui.debug("branch already found\n")
continue
elif n[1] and knownnode(n[1]): # do we know the base?
repo.ui.debug("found incomplete branch %s:%s\n"
% (short(n[0]), short(n[1])))
search.append(n[0:2]) # schedule branch range for scanning
seenbranch.add(n)
else:
if n[1] not in seen and n[1] not in fetch:
if knownnode(n[2]) and knownnode(n[3]):
repo.ui.debug("found new changeset %s\n" %
short(n[1]))
fetch.add(n[1]) # earliest unknown
for p in n[2:4]:
if knownnode(p):
base.add(p) # latest known
for p in n[2:4]:
if p not in req and not knownnode(p):
r.append(p)
req.add(p)
seen.add(n[0])
if r:
reqcnt += 1
repo.ui.progress(_('searching'), reqcnt, unit=_('queries'))
repo.ui.debug("request %d: %s\n" %
(reqcnt, " ".join(map(short, r))))
for p in xrange(0, len(r), 10):
for b in remote.branches(r[p:p + 10]):
repo.ui.debug("received %s:%s\n" %
(short(b[0]), short(b[1])))
unknown.append(b)
# do binary search on the branches we found
while search:
newsearch = []
reqcnt += 1
repo.ui.progress(_('searching'), reqcnt, unit=_('queries'))
for n, l in zip(search, remote.between(search)):
l.append(n[1])
p = n[0]
f = 1
for i in l:
repo.ui.debug("narrowing %d:%d %s\n" % (f, len(l), short(i)))
if knownnode(i):
if f <= 2:
repo.ui.debug("found new branch changeset %s\n" %
short(p))
fetch.add(p)
base.add(i)
else:
repo.ui.debug("narrowed branch search to %s:%s\n"
% (short(p), short(i)))
newsearch.append((p, i))
break
p, f = i, f * 2
search = newsearch
# sanity check our fetch list
for f in fetch:
if knownnode(f):
raise error.RepoError(_("already have changeset ")
+ short(f[:4]))
base = list(base)
if base == [nullid]:
if force:
repo.ui.warn(_("warning: repository is unrelated\n"))
else:
raise error.Abort(_("repository is unrelated"))
repo.ui.debug("found new changesets starting at " +
" ".join([short(f) for f in fetch]) + "\n")
repo.ui.progress(_('searching'), None)
repo.ui.debug("%d total queries\n" % reqcnt)
return base, list(fetch), heads