Mercurial > hg
view tests/test-linelog.py @ 44363:f7459da77f23
nodemap: introduce an option to use mmap to read the nodemap mapping
The performance and memory benefit is much greater if we don't have to copy all
the data in memory for each information. So we introduce an option (on by
default) to read the data using mmap.
This changeset is the last one definition the API for index support nodemap
data. (they have to be able to use the mmaping).
Below are some benchmark comparing the best we currently have in 5.3 with the
final step of this series (using the persistent nodemap implementation in
Rust). The benchmark run `hg perfindex` with various revset and the following
variants:
Before:
* do not use the persistent nodemap
* use the CPython implementation of the index for nodemap
* use mmapping of the changelog index
After:
* use the MixedIndex Rust code, with the NodeTree object for nodemap access
(still in review)
* use the persistent nodemap data from disk
* access the persistent nodemap data through mmap
* use mmapping of the changelog index
The persistent nodemap greatly speed up most operation on very large
repositories. Some of the previously very fast lookup end up a bit slower because
the persistent nodemap has to be setup. However the absolute slowdown is very
small and won't matters in the big picture.
Here are some numbers (in seconds) for the reference copy of mozilla-try:
Revset Before After abs-change speedup
-10000: 0.004622 0.005532 0.000910 × 0.83
-10: 0.000050 0.000132 0.000082 × 0.37
tip 0.000052 0.000085 0.000033 × 0.61
0 + (-10000:) 0.028222 0.005337 -0.022885 × 5.29
0 0.023521 0.000084 -0.023437 × 280.01
(-10000:) + 0 0.235539 0.005308 -0.230231 × 44.37
(-10:) + :9 0.232883 0.000180 -0.232703 ×1293.79
(-10000:) + (:99) 0.238735 0.005358 -0.233377 × 44.55
:99 + (-10000:) 0.317942 0.005593 -0.312349 × 56.84
:9 + (-10:) 0.313372 0.000179 -0.313193 ×1750.68
:9 0.316450 0.000143 -0.316307 ×2212.93
On smaller repositories, the cost of nodemap related operation is not as big, so
the win is much more modest. Yet it helps shaving a handful of millisecond here
and there.
Here are some numbers (in seconds) for the reference copy of mercurial:
Revset Before After abs-change speedup
-10: 0.000065 0.000097 0.000032 × 0.67
tip 0.000063 0.000078 0.000015 × 0.80
0 0.000561 0.000079 -0.000482 × 7.10
-10000: 0.004609 0.003648 -0.000961 × 1.26
0 + (-10000:) 0.005023 0.003715 -0.001307 × 1.35
(-10:) + :9 0.002187 0.000108 -0.002079 ×20.25
(-10000:) + 0 0.006252 0.003716 -0.002536 × 1.68
(-10000:) + (:99) 0.006367 0.003707 -0.002660 × 1.71
:9 + (-10:) 0.003846 0.000110 -0.003736 ×34.96
:9 0.003854 0.000099 -0.003755 ×38.92
:99 + (-10000:) 0.007644 0.003778 -0.003866 × 2.02
Differential Revision: https://phab.mercurial-scm.org/D7894
author | Pierre-Yves David <pierre-yves.david@octobus.net> |
---|---|
date | Tue, 11 Feb 2020 11:18:52 +0100 |
parents | e52a9c85a7a8 |
children | 89a2afe31e82 |
line wrap: on
line source
from __future__ import absolute_import, print_function import difflib import random import unittest from mercurial import linelog vecratio = 3 # number of replacelines / number of replacelines_vec maxlinenum = 0xFFFFFF maxb1 = 0xFFFFFF maxdeltaa = 10 maxdeltab = 10 def _genedits(seed, endrev): lines = [] random.seed(seed) for rev in range(0, endrev): n = len(lines) a1 = random.randint(0, n) a2 = random.randint(a1, min(n, a1 + maxdeltaa)) b1 = random.randint(0, maxb1) b2 = random.randint(b1, b1 + maxdeltab) usevec = not bool(random.randint(0, vecratio)) if usevec: blines = [ (random.randint(0, rev), random.randint(0, maxlinenum)) for _ in range(b1, b2) ] else: blines = [(rev, bidx) for bidx in range(b1, b2)] lines[a1:a2] = blines yield lines, rev, a1, a2, b1, b2, blines, usevec class linelogtests(unittest.TestCase): def testlinelogencodedecode(self): program = [ linelog._eof(0, 0), linelog._jge(41, 42), linelog._jump(0, 43), linelog._eof(0, 0), linelog._jl(44, 45), linelog._line(46, 47), ] ll = linelog.linelog(program, maxrev=100) enc = ll.encode() # round-trips okay self.assertEqual(linelog.linelog.fromdata(enc)._program, ll._program) self.assertEqual(linelog.linelog.fromdata(enc), ll) # This encoding matches the encoding used by hg-experimental's # linelog file, or is supposed to if it doesn't. self.assertEqual( enc, ( b'\x00\x00\x01\x90\x00\x00\x00\x06' b'\x00\x00\x00\xa4\x00\x00\x00*' b'\x00\x00\x00\x00\x00\x00\x00+' b'\x00\x00\x00\x00\x00\x00\x00\x00' b'\x00\x00\x00\xb1\x00\x00\x00-' b'\x00\x00\x00\xba\x00\x00\x00/' ), ) def testsimpleedits(self): ll = linelog.linelog() # Initial revision: add lines 0, 1, and 2 ll.replacelines(1, 0, 0, 0, 3) self.assertEqual( [(l.rev, l.linenum) for l in ll.annotate(1)], [(1, 0), (1, 1), (1, 2),], ) # Replace line 1 with a new line ll.replacelines(2, 1, 2, 1, 2) self.assertEqual( [(l.rev, l.linenum) for l in ll.annotate(2)], [(1, 0), (2, 1), (1, 2),], ) # delete a line out of 2 ll.replacelines(3, 1, 2, 0, 0) self.assertEqual( [(l.rev, l.linenum) for l in ll.annotate(3)], [(1, 0), (1, 2),] ) # annotation of 1 is unchanged self.assertEqual( [(l.rev, l.linenum) for l in ll.annotate(1)], [(1, 0), (1, 1), (1, 2),], ) ll.annotate(3) # set internal state to revision 3 start = ll.getoffset(0) end = ll.getoffset(1) self.assertEqual(ll.getalllines(start, end), [(1, 0), (2, 1), (1, 1),]) self.assertEqual(ll.getalllines(), [(1, 0), (2, 1), (1, 1), (1, 2),]) def testparseclinelogfile(self): # This data is what the replacements in testsimpleedits # produce when fed to the original linelog.c implementation. data = ( b'\x00\x00\x00\x0c\x00\x00\x00\x0f' b'\x00\x00\x00\x00\x00\x00\x00\x02' b'\x00\x00\x00\x05\x00\x00\x00\x06' b'\x00\x00\x00\x06\x00\x00\x00\x00' b'\x00\x00\x00\x00\x00\x00\x00\x07' b'\x00\x00\x00\x06\x00\x00\x00\x02' b'\x00\x00\x00\x00\x00\x00\x00\x00' b'\x00\x00\x00\t\x00\x00\x00\t' b'\x00\x00\x00\x00\x00\x00\x00\x0c' b'\x00\x00\x00\x08\x00\x00\x00\x05' b'\x00\x00\x00\x06\x00\x00\x00\x01' b'\x00\x00\x00\x00\x00\x00\x00\x05' b'\x00\x00\x00\x0c\x00\x00\x00\x05' b'\x00\x00\x00\n\x00\x00\x00\x01' b'\x00\x00\x00\x00\x00\x00\x00\t' ) llc = linelog.linelog.fromdata(data) self.assertEqual( [(l.rev, l.linenum) for l in llc.annotate(1)], [(1, 0), (1, 1), (1, 2),], ) self.assertEqual( [(l.rev, l.linenum) for l in llc.annotate(2)], [(1, 0), (2, 1), (1, 2),], ) self.assertEqual( [(l.rev, l.linenum) for l in llc.annotate(3)], [(1, 0), (1, 2),] ) # Check we emit the same bytecode. ll = linelog.linelog() # Initial revision: add lines 0, 1, and 2 ll.replacelines(1, 0, 0, 0, 3) # Replace line 1 with a new line ll.replacelines(2, 1, 2, 1, 2) # delete a line out of 2 ll.replacelines(3, 1, 2, 0, 0) diff = '\n ' + '\n '.join( difflib.unified_diff( ll.debugstr().splitlines(), llc.debugstr().splitlines(), 'python', 'c', lineterm='', ) ) self.assertEqual(ll._program, llc._program, 'Program mismatch: ' + diff) # Done as a secondary step so we get a better result if the # program is where the mismatch is. self.assertEqual(ll, llc) self.assertEqual(ll.encode(), data) def testanothersimplecase(self): ll = linelog.linelog() ll.replacelines(3, 0, 0, 0, 2) ll.replacelines(4, 0, 2, 0, 0) self.assertEqual([(l.rev, l.linenum) for l in ll.annotate(4)], []) self.assertEqual( [(l.rev, l.linenum) for l in ll.annotate(3)], [(3, 0), (3, 1)] ) # rev 2 is empty because contents were only ever introduced in rev 3 self.assertEqual([(l.rev, l.linenum) for l in ll.annotate(2)], []) def testrandomedits(self): # Inspired by original linelog tests. seed = random.random() numrevs = 2000 ll = linelog.linelog() # Populate linelog for lines, rev, a1, a2, b1, b2, blines, usevec in _genedits( seed, numrevs ): if usevec: ll.replacelines_vec(rev, a1, a2, blines) else: ll.replacelines(rev, a1, a2, b1, b2) ll.annotate(rev) self.assertEqual(ll.annotateresult, lines) # Verify we can get back these states by annotating each rev for lines, rev, a1, a2, b1, b2, blines, usevec in _genedits( seed, numrevs ): ar = ll.annotate(rev) self.assertEqual([(l.rev, l.linenum) for l in ar], lines) def testinfinitebadprogram(self): ll = linelog.linelog.fromdata( b'\x00\x00\x00\x00\x00\x00\x00\x02' # header b'\x00\x00\x00\x00\x00\x00\x00\x01' # JUMP to self ) with self.assertRaises(linelog.LineLogError): # should not be an infinite loop and raise ll.annotate(1) if __name__ == '__main__': import silenttestrunner silenttestrunner.main(__name__)