view tests/test-linelog.py @ 43198:c16fe77e340a

pathcopies: give up any optimization based on `introrev`

Between 8a0136f69027 and d98fb3f42f33, we sped up the search for the
introduction revision during path copies. However, further checking show that
finding the introduction revision is still expensive and that we are better off
without it. So we simply drop it and only rely on the linkrev optimisation.

I ran `perfpathcopies` on 6989 pair of revision in the pypy
repository (`hg perfhelper-pathcopies`. The result is massively in favor of
dropping this condition. The result of the copy tracing are unchanged.

Attempt to use a smaller changes preserving linkrev usage were unsuccessful, it
can return wrong result. The following changesets broke test-mv-cp-st-diff.t

    -        if not f.isintroducedafter(limit):
    +        if limit >= 0 and f.linkrev() < limit:
                 return None


Here are various numbers (before this changeset/after this changesets)

            source       destination  before      after    saved-time  ratio
worth cases e66f24650daf 695dfb0f493b 1.062843 1.246369 -0.183526 1.172675
            c979853a3b6a 8d60fe293e79 1.036985 1.196414 -0.159429 1.153743
            22349fa2fc33 fbb1c9fd86c0 0.879926 1.038682 -0.158756 1.180420
            682b98f3e672 a4878080a536 0.909952 1.063801 -0.153849 1.169074
            5adabc9b9848 920958a93997 0.993622 1.147452 -0.153830 1.154817
worse  1%   dbfbfcf077e9 aea8f2fd3593 1.016595 1.082999 -0.066404 1.065320
worse  5%   c95f1ced15f2 7d29d5e39734 0.453694 0.471156 -0.017462 1.038488
worse 10%   3e144ed1d5b7 2aef0e942480 0.035140 0.037535 -0.002395 1.068156
worse 25%   321fc60db035 801748ba582a 0.009267 0.009325 -0.000058 1.006259
median      2088ce763fc2 e6991321d78b 0.000665 0.000651 0.000014 0.978947
best 25%    915631a97de6 385b31354be6 0.040743 0.040363 0.000380 0.990673
best 10%    ad495c36a765 19c10384d3e7 0.431658 0.411490 0.020168 0.953278
best  5%    d13ae7d283ae 813c99f810ac 1.141404 1.075346 0.066058 0.942126
best  1%    81593cb4a496 99ae11866969 1.833297 0.063823 1.769474 0.034813
best cases  c3b14617fbd7 743a0fcaa4eb 1101.811740 2.735970 1099.075770 0.002483
            c3b14617fbd7 9ba6ab77fd29 1116.753953 2.800729 1113.953224 0.002508
            058b99d6e81f 57e249b7a3ea 1246.128485 3.042762 1243.085723 0.002442
            9a8c361aab49 0354a250d371 1253.111894 3.085796 1250.026098 0.002463
            442dbbc53c68 3ec1002a818c 1261.786294 3.138607 1258.647687 0.002487


As one can see, the average case is not really impacted. However, the worth case
we get after this changeset are much better than the one we had before it. We
have 30 pairs where improvements are above 10 minutes.

This reflect in the combined time for all pairs
before: 26256s
after:   1300s (-95%)

If we remove these pathological 30 cases, we still see a significant improvements:
before:  1631s
after:   1245s (-24%)

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)
            ar = 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__)