view tests/test-remotefilelog-histpack.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%)

#!/usr/bin/env python
from __future__ import absolute_import

import hashlib
import os
import random
import shutil
import stat
import struct
import sys
import tempfile
import unittest

import silenttestrunner

from mercurial.node import nullid
from mercurial import (
    pycompat,
    ui as uimod,
)

# Load the local remotefilelog, not the system one
sys.path[0:0] = [os.path.join(os.path.dirname(__file__), '..')]
from hgext.remotefilelog import (
    basepack,
    historypack,
)


class histpacktests(unittest.TestCase):
    def setUp(self):
        self.tempdirs = []

    def tearDown(self):
        for d in self.tempdirs:
            shutil.rmtree(d)

    def makeTempDir(self):
        tempdir = tempfile.mkdtemp()
        self.tempdirs.append(tempdir)
        return pycompat.fsencode(tempdir)

    def getHash(self, content):
        return hashlib.sha1(content).digest()

    def getFakeHash(self):
        return b''.join(
            pycompat.bytechr(random.randint(0, 255)) for _ in range(20)
        )

    def createPack(self, revisions=None):
        """Creates and returns a historypack containing the specified revisions.

        `revisions` is a list of tuples, where each tuple contains a filanem,
        node, p1node, p2node, and linknode.
        """
        if revisions is None:
            revisions = [
                (
                    b"filename",
                    self.getFakeHash(),
                    nullid,
                    nullid,
                    self.getFakeHash(),
                    None,
                )
            ]

        packdir = pycompat.fsencode(self.makeTempDir())
        packer = historypack.mutablehistorypack(uimod.ui(), packdir, version=2)

        for filename, node, p1, p2, linknode, copyfrom in revisions:
            packer.add(filename, node, p1, p2, linknode, copyfrom)

        path = packer.close()
        return historypack.historypack(path)

    def testAddSingle(self):
        """Test putting a single entry into a pack and reading it out.
        """
        filename = b"foo"
        node = self.getFakeHash()
        p1 = self.getFakeHash()
        p2 = self.getFakeHash()
        linknode = self.getFakeHash()

        revisions = [(filename, node, p1, p2, linknode, None)]
        pack = self.createPack(revisions)

        actual = pack.getancestors(filename, node)[node]
        self.assertEqual(p1, actual[0])
        self.assertEqual(p2, actual[1])
        self.assertEqual(linknode, actual[2])

    def testAddMultiple(self):
        """Test putting multiple unrelated revisions into a pack and reading
        them out.
        """
        revisions = []
        for i in range(10):
            filename = b"foo-%d" % i
            node = self.getFakeHash()
            p1 = self.getFakeHash()
            p2 = self.getFakeHash()
            linknode = self.getFakeHash()
            revisions.append((filename, node, p1, p2, linknode, None))

        pack = self.createPack(revisions)

        for filename, node, p1, p2, linknode, copyfrom in revisions:
            actual = pack.getancestors(filename, node)[node]
            self.assertEqual(p1, actual[0])
            self.assertEqual(p2, actual[1])
            self.assertEqual(linknode, actual[2])
            self.assertEqual(copyfrom, actual[3])

    def testAddAncestorChain(self):
        """Test putting multiple revisions in into a pack and read the ancestor
        chain.
        """
        revisions = []
        filename = b"foo"
        lastnode = nullid
        for i in range(10):
            node = self.getFakeHash()
            revisions.append((filename, node, lastnode, nullid, nullid, None))
            lastnode = node

        # revisions must be added in topological order, newest first
        revisions = list(reversed(revisions))
        pack = self.createPack(revisions)

        # Test that the chain has all the entries
        ancestors = pack.getancestors(revisions[0][0], revisions[0][1])
        for filename, node, p1, p2, linknode, copyfrom in revisions:
            ap1, ap2, alinknode, acopyfrom = ancestors[node]
            self.assertEqual(ap1, p1)
            self.assertEqual(ap2, p2)
            self.assertEqual(alinknode, linknode)
            self.assertEqual(acopyfrom, copyfrom)

    def testPackMany(self):
        """Pack many related and unrelated ancestors.
        """
        # Build a random pack file
        allentries = {}
        ancestorcounts = {}
        revisions = []
        random.seed(0)
        for i in range(100):
            filename = b"filename-%d" % i
            entries = []
            p2 = nullid
            linknode = nullid
            for j in range(random.randint(1, 100)):
                node = self.getFakeHash()
                p1 = nullid
                if len(entries) > 0:
                    p1 = entries[random.randint(0, len(entries) - 1)]
                entries.append(node)
                revisions.append((filename, node, p1, p2, linknode, None))
                allentries[(filename, node)] = (p1, p2, linknode)
                if p1 == nullid:
                    ancestorcounts[(filename, node)] = 1
                else:
                    newcount = ancestorcounts[(filename, p1)] + 1
                    ancestorcounts[(filename, node)] = newcount

        # Must add file entries in reverse topological order
        revisions = list(reversed(revisions))
        pack = self.createPack(revisions)

        # Verify the pack contents
        for (filename, node) in allentries:
            ancestors = pack.getancestors(filename, node)
            self.assertEqual(ancestorcounts[(filename, node)], len(ancestors))
            for anode, (ap1, ap2, alinknode, copyfrom) in ancestors.items():
                ep1, ep2, elinknode = allentries[(filename, anode)]
                self.assertEqual(ap1, ep1)
                self.assertEqual(ap2, ep2)
                self.assertEqual(alinknode, elinknode)
                self.assertEqual(copyfrom, None)

    def testGetNodeInfo(self):
        revisions = []
        filename = b"foo"
        lastnode = nullid
        for i in range(10):
            node = self.getFakeHash()
            revisions.append((filename, node, lastnode, nullid, nullid, None))
            lastnode = node

        pack = self.createPack(revisions)

        # Test that getnodeinfo returns the expected results
        for filename, node, p1, p2, linknode, copyfrom in revisions:
            ap1, ap2, alinknode, acopyfrom = pack.getnodeinfo(filename, node)
            self.assertEqual(ap1, p1)
            self.assertEqual(ap2, p2)
            self.assertEqual(alinknode, linknode)
            self.assertEqual(acopyfrom, copyfrom)

    def testGetMissing(self):
        """Test the getmissing() api.
        """
        revisions = []
        filename = b"foo"
        for i in range(10):
            node = self.getFakeHash()
            p1 = self.getFakeHash()
            p2 = self.getFakeHash()
            linknode = self.getFakeHash()
            revisions.append((filename, node, p1, p2, linknode, None))

        pack = self.createPack(revisions)

        missing = pack.getmissing([(filename, revisions[0][1])])
        self.assertFalse(missing)

        missing = pack.getmissing(
            [(filename, revisions[0][1]), (filename, revisions[1][1])]
        )
        self.assertFalse(missing)

        fakenode = self.getFakeHash()
        missing = pack.getmissing(
            [(filename, revisions[0][1]), (filename, fakenode)]
        )
        self.assertEqual(missing, [(filename, fakenode)])

        # Test getmissing on a non-existant filename
        missing = pack.getmissing([(b"bar", fakenode)])
        self.assertEqual(missing, [(b"bar", fakenode)])

    def testAddThrows(self):
        pack = self.createPack()

        try:
            pack.add(b'filename', nullid, nullid, nullid, nullid, None)
            self.assertTrue(False, "historypack.add should throw")
        except RuntimeError:
            pass

    def testBadVersionThrows(self):
        pack = self.createPack()
        path = pack.path + b'.histpack'
        with open(path, 'rb') as f:
            raw = f.read()
        raw = struct.pack('!B', 255) + raw[1:]
        os.chmod(path, os.stat(path).st_mode | stat.S_IWRITE)
        with open(path, 'wb+') as f:
            f.write(raw)

        try:
            pack = historypack.historypack(pack.path)
            self.assertTrue(False, "bad version number should have thrown")
        except RuntimeError:
            pass

    def testLargePack(self):
        """Test creating and reading from a large pack with over X entries.
        This causes it to use a 2^16 fanout table instead."""
        total = basepack.SMALLFANOUTCUTOFF + 1
        revisions = []
        for i in pycompat.xrange(total):
            filename = b"foo-%d" % i
            node = self.getFakeHash()
            p1 = self.getFakeHash()
            p2 = self.getFakeHash()
            linknode = self.getFakeHash()
            revisions.append((filename, node, p1, p2, linknode, None))

        pack = self.createPack(revisions)
        self.assertEqual(pack.params.fanoutprefix, basepack.LARGEFANOUTPREFIX)

        for filename, node, p1, p2, linknode, copyfrom in revisions:
            actual = pack.getancestors(filename, node)[node]
            self.assertEqual(p1, actual[0])
            self.assertEqual(p2, actual[1])
            self.assertEqual(linknode, actual[2])
            self.assertEqual(copyfrom, actual[3])


# TODO:
# histpack store:
# - repack two packs into one

if __name__ == '__main__':
    if pycompat.iswindows:
        sys.exit(80)  # Skip on Windows
    silenttestrunner.main(__name__)