Mercurial > hg
view mercurial/simplemerge.py @ 42044:bb271ec2fbfb
compression: introduce a `storage.revlog.zstd.level` configuration
This option control the zstd compression level used when compressing revlog
chunk. The usage of zstd for revlog compression has not graduated from
experimental yet, but we intend to fix that soon.
The option name for the compression level is more straight forward to pick, so
this changesets comes first. Having a dedicated option for each compression
engine is useful because they don't support the same range of values.
I ran the same measurement as for the zlib compression level (in the parent
changesets). The variation in repository size is stay mostly in the same (small)
range. The "read/write" performance see smallish variation, but are overall much
better than zlib. Write performance show the same tend of having better write
performance for when reaching high-end compression.
Again, we don't intend to change the default zstd compression level (currently:
3) in this series. However this is worth investigating in the future.
The Performance comparison of zlib vs zstd is quite impressive. The repository
size stay in the same range, but the performance are much better in all
situations.
Comparison summary
==================
We are looking at:
- performance range for zlib
- performance range for zstd
- comparison of default zstd (level-3) to default zlib (level 6)
- comparison of the slowest zstd time to the fastest zlib time
Read performance:
-----------------
| zlib | zstd | cmp | f2s
mercurial | 0.170159 - 0.189219 | 0.144127 - 0.149624 | 80% | 88%
pypy | 2.679217 - 2.768691 | 1.532317 - 1.705044 | 60% | 63%
netbeans | 122.477027 - 141.620281 | 72.996346 - 89.731560 | 58% | 73%
mozilla | 147.867662 - 170.572118 | 91.700995 - 105.853099 | 56% | 71%
Write performance:
------------------
| zlib | zstd | cmp | f2s
mercurial | 53.250304 - 56.2936129 | 40.877025 - 45.677286 | 75% | 86%
pypy | 460.721984 - 476.589918 | 270.545409 - 301.002219 | 63% | 65%
netbeans | 520.560316 - 715.930400 | 370.356311 - 428.329652 | 55% | 82%
mozilla | 739.803002 - 987.056093 | 505.152906 - 591.930683 | 57% | 80%
Raw data
--------
repo alg lvl .hg/store size 00manifest.d read write
mercurial zlib 1 49,402,813 5,963,475 0.170159 53.250304
mercurial zlib 6 47,197,397 5,875,730 0.182820 56.264320
mercurial zlib 9 47,121,596 5,849,781 0.189219 56.293612
mercurial zstd 1 49,737,084 5,966,355 0.144127 40.877025
mercurial zstd 3 48,961,867 5,895,208 0.146376 42.268142
mercurial zstd 5 48,200,592 5,938,676 0.149624 43.162875
mercurial zstd 10 47,833,520 5,913,353 0.145185 44.012489
mercurial zstd 15 47,314,604 5,728,679 0.147686 45.677286
mercurial zstd 20 47,330,502 5,830,539 0.145789 45.025407
mercurial zstd 22 47,330,076 5,830,539 0.143996 44.690460
pypy zlib 1 370,830,572 28,462,425 2.679217 460.721984
pypy zlib 6 340,112,317 27,648,747 2.768691 467.537158
pypy zlib 9 338,360,736 27,639,003 2.763495 476.589918
pypy zstd 1 362,377,479 27,916,214 1.532317 270.545409
pypy zstd 3 354,137,693 27,905,988 1.686718 294.951509
pypy zstd 5 342,640,043 27,655,774 1.705044 301.002219
pypy zstd 10 334,224,327 27,164,493 1.567287 285.186239
pypy zstd 15 329,000,363 26,645,965 1.637729 299.561332
pypy zstd 20 324,534,039 26,199,547 1.526813 302.149827
pypy zstd 22 324,530,595 26,198,932 1.525718 307.821218
netbeans zlib 1 1,281,847,810 165,495,457 122.477027 520.560316
netbeans zlib 6 1,205,284,353 159,161,207 139.876147 715.930400
netbeans zlib 9 1,197,135,671 155,034,586 141.620281 678.297064
netbeans zstd 1 1,259,581,737 160,840,613 72.996346 370.356311
netbeans zstd 3 1,232,978,122 157,691,551 81.622317 396.733087
netbeans zstd 5 1,208,034,075 160,246,880 83.080549 364.342626
netbeans zstd 10 1,188,624,176 156,083,417 79.323935 403.594602
netbeans zstd 15 1,176,973,589 153,859,477 89.731560 428.329652
netbeans zstd 20 1,162,958,258 151,147,535 82.842667 392.335349
netbeans zstd 22 1,162,707,029 151,150,220 82.565695 402.840655
mozilla zlib 1 2,775,497,186 298,527,987 147.867662 751.263721
mozilla zlib 6 2,596,856,420 286,597,671 170.572118 987.056093
mozilla zlib 9 2,587,542,494 287,018,264 163.622338 739.803002
mozilla zstd 1 2,723,159,348 286,617,532 91.700995 570.042751
mozilla zstd 3 2,665,055,001 286,152,013 95.240155 561.412805
mozilla zstd 5 2,607,819,817 288,060,030 101.978048 505.152906
mozilla zstd 10 2,558,761,085 283,967,648 104.113481 497.771202
mozilla zstd 15 2,526,216,060 275,581,300 105.853099 591.930683
mozilla zstd 20 2,485,114,806 266,478,859 95.268795 576.515389
mozilla zstd 22 2,484,869,080 266,456,505 94.429282 572.785537
author | Pierre-Yves David <pierre-yves.david@octobus.net> |
---|---|
date | Wed, 27 Mar 2019 18:35:59 +0100 |
parents | aaad36b88298 |
children | 2372284d9457 |
line wrap: on
line source
# Copyright (C) 2004, 2005 Canonical Ltd # # This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 2 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # You should have received a copy of the GNU General Public License # along with this program; if not, see <http://www.gnu.org/licenses/>. # mbp: "you know that thing where cvs gives you conflict markers?" # s: "i hate that." from __future__ import absolute_import from .i18n import _ from . import ( error, mdiff, pycompat, ) from .utils import ( stringutil, ) class CantReprocessAndShowBase(Exception): pass def intersect(ra, rb): """Given two ranges return the range where they intersect or None. >>> intersect((0, 10), (0, 6)) (0, 6) >>> intersect((0, 10), (5, 15)) (5, 10) >>> intersect((0, 10), (10, 15)) >>> intersect((0, 9), (10, 15)) >>> intersect((0, 9), (7, 15)) (7, 9) """ assert ra[0] <= ra[1] assert rb[0] <= rb[1] sa = max(ra[0], rb[0]) sb = min(ra[1], rb[1]) if sa < sb: return sa, sb else: return None def compare_range(a, astart, aend, b, bstart, bend): """Compare a[astart:aend] == b[bstart:bend], without slicing. """ if (aend - astart) != (bend - bstart): return False for ia, ib in zip(pycompat.xrange(astart, aend), pycompat.xrange(bstart, bend)): if a[ia] != b[ib]: return False else: return True class Merge3Text(object): """3-way merge of texts. Given strings BASE, OTHER, THIS, tries to produce a combined text incorporating the changes from both BASE->OTHER and BASE->THIS.""" def __init__(self, basetext, atext, btext, base=None, a=None, b=None): self.basetext = basetext self.atext = atext self.btext = btext if base is None: base = mdiff.splitnewlines(basetext) if a is None: a = mdiff.splitnewlines(atext) if b is None: b = mdiff.splitnewlines(btext) self.base = base self.a = a self.b = b def merge_lines(self, name_a=None, name_b=None, name_base=None, start_marker='<<<<<<<', mid_marker='=======', end_marker='>>>>>>>', base_marker=None, localorother=None, minimize=False): """Return merge in cvs-like form. """ self.conflicts = False newline = '\n' if len(self.a) > 0: if self.a[0].endswith('\r\n'): newline = '\r\n' elif self.a[0].endswith('\r'): newline = '\r' if name_a and start_marker: start_marker = start_marker + ' ' + name_a if name_b and end_marker: end_marker = end_marker + ' ' + name_b if name_base and base_marker: base_marker = base_marker + ' ' + name_base merge_regions = self.merge_regions() if minimize: merge_regions = self.minimize(merge_regions) for t in merge_regions: what = t[0] if what == 'unchanged': for i in range(t[1], t[2]): yield self.base[i] elif what == 'a' or what == 'same': for i in range(t[1], t[2]): yield self.a[i] elif what == 'b': for i in range(t[1], t[2]): yield self.b[i] elif what == 'conflict': if localorother == 'local': for i in range(t[3], t[4]): yield self.a[i] elif localorother == 'other': for i in range(t[5], t[6]): yield self.b[i] else: self.conflicts = True if start_marker is not None: yield start_marker + newline for i in range(t[3], t[4]): yield self.a[i] if base_marker is not None: yield base_marker + newline for i in range(t[1], t[2]): yield self.base[i] if mid_marker is not None: yield mid_marker + newline for i in range(t[5], t[6]): yield self.b[i] if end_marker is not None: yield end_marker + newline else: raise ValueError(what) def merge_groups(self): """Yield sequence of line groups. Each one is a tuple: 'unchanged', lines Lines unchanged from base 'a', lines Lines taken from a 'same', lines Lines taken from a (and equal to b) 'b', lines Lines taken from b 'conflict', base_lines, a_lines, b_lines Lines from base were changed to either a or b and conflict. """ for t in self.merge_regions(): what = t[0] if what == 'unchanged': yield what, self.base[t[1]:t[2]] elif what == 'a' or what == 'same': yield what, self.a[t[1]:t[2]] elif what == 'b': yield what, self.b[t[1]:t[2]] elif what == 'conflict': yield (what, self.base[t[1]:t[2]], self.a[t[3]:t[4]], self.b[t[5]:t[6]]) else: raise ValueError(what) def merge_regions(self): """Return sequences of matching and conflicting regions. This returns tuples, where the first value says what kind we have: 'unchanged', start, end Take a region of base[start:end] 'same', astart, aend b and a are different from base but give the same result 'a', start, end Non-clashing insertion from a[start:end] 'conflict', zstart, zend, astart, aend, bstart, bend Conflict between a and b, with z as common ancestor Method is as follows: The two sequences align only on regions which match the base and both descendants. These are found by doing a two-way diff of each one against the base, and then finding the intersections between those regions. These "sync regions" are by definition unchanged in both and easily dealt with. The regions in between can be in any of three cases: conflicted, or changed on only one side. """ # section a[0:ia] has been disposed of, etc iz = ia = ib = 0 for region in self.find_sync_regions(): zmatch, zend, amatch, aend, bmatch, bend = region #print 'match base [%d:%d]' % (zmatch, zend) matchlen = zend - zmatch assert matchlen >= 0 assert matchlen == (aend - amatch) assert matchlen == (bend - bmatch) len_a = amatch - ia len_b = bmatch - ib len_base = zmatch - iz assert len_a >= 0 assert len_b >= 0 assert len_base >= 0 #print 'unmatched a=%d, b=%d' % (len_a, len_b) if len_a or len_b: # try to avoid actually slicing the lists equal_a = compare_range(self.a, ia, amatch, self.base, iz, zmatch) equal_b = compare_range(self.b, ib, bmatch, self.base, iz, zmatch) same = compare_range(self.a, ia, amatch, self.b, ib, bmatch) if same: yield 'same', ia, amatch elif equal_a and not equal_b: yield 'b', ib, bmatch elif equal_b and not equal_a: yield 'a', ia, amatch elif not equal_a and not equal_b: yield 'conflict', iz, zmatch, ia, amatch, ib, bmatch else: raise AssertionError("can't handle a=b=base but unmatched") ia = amatch ib = bmatch iz = zmatch # if the same part of the base was deleted on both sides # that's OK, we can just skip it. if matchlen > 0: assert ia == amatch assert ib == bmatch assert iz == zmatch yield 'unchanged', zmatch, zend iz = zend ia = aend ib = bend def minimize(self, merge_regions): """Trim conflict regions of lines where A and B sides match. Lines where both A and B have made the same changes at the beginning or the end of each merge region are eliminated from the conflict region and are instead considered the same. """ for region in merge_regions: if region[0] != "conflict": yield region continue issue, z1, z2, a1, a2, b1, b2 = region alen = a2 - a1 blen = b2 - b1 # find matches at the front ii = 0 while (ii < alen and ii < blen and self.a[a1 + ii] == self.b[b1 + ii]): ii += 1 startmatches = ii # find matches at the end ii = 0 while (ii < alen and ii < blen and self.a[a2 - ii - 1] == self.b[b2 - ii - 1]): ii += 1 endmatches = ii if startmatches > 0: yield 'same', a1, a1 + startmatches yield ('conflict', z1, z2, a1 + startmatches, a2 - endmatches, b1 + startmatches, b2 - endmatches) if endmatches > 0: yield 'same', a2 - endmatches, a2 def find_sync_regions(self): """Return a list of sync regions, where both descendants match the base. Generates a list of (base1, base2, a1, a2, b1, b2). There is always a zero-length sync region at the end of all the files. """ ia = ib = 0 amatches = mdiff.get_matching_blocks(self.basetext, self.atext) bmatches = mdiff.get_matching_blocks(self.basetext, self.btext) len_a = len(amatches) len_b = len(bmatches) sl = [] while ia < len_a and ib < len_b: abase, amatch, alen = amatches[ia] bbase, bmatch, blen = bmatches[ib] # there is an unconflicted block at i; how long does it # extend? until whichever one ends earlier. i = intersect((abase, abase + alen), (bbase, bbase + blen)) if i: intbase = i[0] intend = i[1] intlen = intend - intbase # found a match of base[i[0], i[1]]; this may be less than # the region that matches in either one assert intlen <= alen assert intlen <= blen assert abase <= intbase assert bbase <= intbase asub = amatch + (intbase - abase) bsub = bmatch + (intbase - bbase) aend = asub + intlen bend = bsub + intlen assert self.base[intbase:intend] == self.a[asub:aend], ( (self.base[intbase:intend], self.a[asub:aend])) assert self.base[intbase:intend] == self.b[bsub:bend] sl.append((intbase, intend, asub, aend, bsub, bend)) # advance whichever one ends first in the base text if (abase + alen) < (bbase + blen): ia += 1 else: ib += 1 intbase = len(self.base) abase = len(self.a) bbase = len(self.b) sl.append((intbase, intbase, abase, abase, bbase, bbase)) return sl def find_unconflicted(self): """Return a list of ranges in base that are not conflicted.""" am = mdiff.get_matching_blocks(self.basetext, self.atext) bm = mdiff.get_matching_blocks(self.basetext, self.btext) unc = [] while am and bm: # there is an unconflicted block at i; how long does it # extend? until whichever one ends earlier. a1 = am[0][0] a2 = a1 + am[0][2] b1 = bm[0][0] b2 = b1 + bm[0][2] i = intersect((a1, a2), (b1, b2)) if i: unc.append(i) if a2 < b2: del am[0] else: del bm[0] return unc def _verifytext(text, path, ui, opts): """verifies that text is non-binary (unless opts[text] is passed, then we just warn)""" if stringutil.binary(text): msg = _("%s looks like a binary file.") % path if not opts.get('quiet'): ui.warn(_('warning: %s\n') % msg) if not opts.get('text'): raise error.Abort(msg) return text def _picklabels(defaults, overrides): if len(overrides) > 3: raise error.Abort(_("can only specify three labels.")) result = defaults[:] for i, override in enumerate(overrides): result[i] = override return result def simplemerge(ui, localctx, basectx, otherctx, **opts): """Performs the simplemerge algorithm. The merged result is written into `localctx`. """ opts = pycompat.byteskwargs(opts) def readctx(ctx): # Merges were always run in the working copy before, which means # they used decoded data, if the user defined any repository # filters. # # Maintain that behavior today for BC, though perhaps in the future # it'd be worth considering whether merging encoded data (what the # repository usually sees) might be more useful. return _verifytext(ctx.decodeddata(), ctx.path(), ui, opts) mode = opts.get('mode','merge') name_a, name_b, name_base = None, None, None if mode != 'union': name_a, name_b, name_base = _picklabels([localctx.path(), otherctx.path(), None], opts.get('label', [])) try: localtext = readctx(localctx) basetext = readctx(basectx) othertext = readctx(otherctx) except error.Abort: return 1 m3 = Merge3Text(basetext, localtext, othertext) extrakwargs = { "localorother": opts.get("localorother", None), 'minimize': True, } if mode == 'union': extrakwargs['start_marker'] = None extrakwargs['mid_marker'] = None extrakwargs['end_marker'] = None elif name_base is not None: extrakwargs['base_marker'] = '|||||||' extrakwargs['name_base'] = name_base extrakwargs['minimize'] = False mergedtext = "" for line in m3.merge_lines(name_a=name_a, name_b=name_b, **pycompat.strkwargs(extrakwargs)): if opts.get('print'): ui.fout.write(line) else: mergedtext += line if not opts.get('print'): localctx.write(mergedtext, localctx.flags()) if m3.conflicts and not mode == 'union': return 1