view tests/test-remotefilelog-histpack.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

#!/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:
            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__)