view mercurial/pvec.py @ 46568:0d840b9d200d

copies-rust: track "overwrites" directly within CopySource Overwrite are "rare enough" that explicitly keeping track of them is going to be "cheap", or at least much cheaper that issuing many `is_ancestor` calls. Even a simple implementation using no specific optimisation (eg: using the generic HashSet type) yield good result in most cases. They are interesting optimization to can do on top of that. We will implement them in later changesets. We tried different approach to speed up the overwrite detection and this one seems the most promising. Without further optimization, we already see sizable speedup on various cases. Repo Case Source-Rev Dest-Rev # of revisions old time new time Difference Factor time per rev --------------------------------------------------------------------------------------------------------------------------------------------------------------- mozilla-try x00000_revs_x_added_0_copies 6a320851d377 1ebb79acd503 : 363753 revs, 5.138169 s, 4.482399 s, -0.655770 s, × 0.8724, 12 µs/rev mozilla-try x00000_revs_x_added_x_copies 5173c4b6f97c 95d83ee7242d : 362229 revs, 5.127809 s, 4.480366 s, -0.647443 s, × 0.8737, 12 µs/rev mozilla-try x00000_revs_x000_added_x_copies 9126823d0e9c ca82787bb23c : 359344 revs, 4.971136 s, 4.369070 s, -0.602066 s, × 0.8789, 12 µs/rev mozilla-try x00000_revs_x0000_added_x0000_copies 8d3fafa80d4b eb884023b810 : 192665 revs, 1.741678 s, 1.592506 s, -0.149172 s, × 0.9144, 8 µs/rev However, some of the case doing a lot of overwrite get significantly slower. The one with a really problematic slowdown are the special "head reducing" merge in mozilla-try so I am not too worried about them. In addition, further changeset are going to improve the performance of all this. Repo Case Source-Rev Dest-Rev # of revisions old time new time Difference Factor time per rev --------------------------------------------------------------------------------------------------------------------------------------------------------------- mozilla-try x0000_revs_xx000_added_x000_copies 89294cd501d9 7ccb2fc7ccb5 : 97052 revs, 1.343373 s, 2.119204 s, +0.775831 s, × 1.5775, 21 µs/rev mozilla-try x00000_revs_x00000_added_x0000_copies 1b661134e2ca 1ae03d022d6d : 228985 revs, 40.314822 s, 87.824489 s, +47.509667 s, × 2.1785, 383 µs/rev mozilla-try x00000_revs_x00000_added_x000_copies 9b2a99adc05e 8e29777b48e6 : 382065 revs, 20.048029 s, 43.304637 s, +23.256608 s, × 2.1600, 113 µs/rev Full benchmark below: Repo Case Source-Rev Dest-Rev # of revisions old time new time Difference Factor time per rev --------------------------------------------------------------------------------------------------------------------------------------------------------------- mercurial x_revs_x_added_0_copies ad6b123de1c7 39cfcef4f463 : 1 revs, 0.000042 s, 0.000043 s, +0.000001 s, × 1.0238, 43 µs/rev mercurial x_revs_x_added_x_copies 2b1c78674230 0c1d10351869 : 6 revs, 0.000110 s, 0.000114 s, +0.000004 s, × 1.0364, 19 µs/rev mercurial x000_revs_x000_added_x_copies 81f8ff2a9bf2 dd3267698d84 : 1032 revs, 0.004945 s, 0.004937 s, -0.000008 s, × 0.9984, 4 µs/rev pypy x_revs_x_added_0_copies aed021ee8ae8 099ed31b181b : 9 revs, 0.000192 s, 0.000339 s, +0.000147 s, × 1.7656, 37 µs/rev pypy x_revs_x000_added_0_copies 4aa4e1f8e19a 359343b9ac0e : 1 revs, 0.000049 s, 0.000049 s, +0.000000 s, × 1.0000, 49 µs/rev pypy x_revs_x_added_x_copies ac52eb7bbbb0 72e022663155 : 7 revs, 0.000112 s, 0.000202 s, +0.000090 s, × 1.8036, 28 µs/rev pypy x_revs_x00_added_x_copies c3b14617fbd7 ace7255d9a26 : 1 revs, 0.000323 s, 0.000409 s, +0.000086 s, × 1.2663, 409 µs/rev pypy x_revs_x000_added_x000_copies df6f7a526b60 a83dc6a2d56f : 6 revs, 0.010042 s, 0.011984 s, +0.001942 s, × 1.1934, 1997 µs/rev pypy x000_revs_xx00_added_0_copies 89a76aede314 2f22446ff07e : 4785 revs, 0.049813 s, 0.050820 s, +0.001007 s, × 1.0202, 10 µs/rev pypy x000_revs_x000_added_x_copies 8a3b5bfd266e 2c68e87c3efe : 6780 revs, 0.079937 s, 0.087953 s, +0.008016 s, × 1.1003, 12 µs/rev pypy x000_revs_x000_added_x000_copies 89a76aede314 7b3dda341c84 : 5441 revs, 0.059412 s, 0.062902 s, +0.003490 s, × 1.0587, 11 µs/rev pypy x0000_revs_x_added_0_copies d1defd0dc478 c9cb1334cc78 : 43645 revs, 0.533769 s, 0.679234 s, +0.145465 s, × 1.2725, 15 µs/rev pypy x0000_revs_xx000_added_0_copies bf2c629d0071 4ffed77c095c : 2 revs, 0.013147 s, 0.013095 s, -0.000052 s, × 0.9960, 6547 µs/rev pypy x0000_revs_xx000_added_x000_copies 08ea3258278e d9fa043f30c0 : 11316 revs, 0.110680 s, 0.120910 s, +0.010230 s, × 1.0924, 10 µs/rev netbeans x_revs_x_added_0_copies fb0955ffcbcd a01e9239f9e7 : 2 revs, 0.000085 s, 0.000087 s, +0.000002 s, × 1.0235, 43 µs/rev netbeans x_revs_x000_added_0_copies 6f360122949f 20eb231cc7d0 : 2 revs, 0.000107 s, 0.000107 s, +0.000000 s, × 1.0000, 53 µs/rev netbeans x_revs_x_added_x_copies 1ada3faf6fb6 5a39d12eecf4 : 3 revs, 0.000175 s, 0.000186 s, +0.000011 s, × 1.0629, 62 µs/rev netbeans x_revs_x00_added_x_copies 35be93ba1e2c 9eec5e90c05f : 9 revs, 0.000720 s, 0.000754 s, +0.000034 s, × 1.0472, 83 µs/rev netbeans x000_revs_xx00_added_0_copies eac3045b4fdd 51d4ae7f1290 : 1421 revs, 0.010019 s, 0.010443 s, +0.000424 s, × 1.0423, 7 µs/rev netbeans x000_revs_x000_added_x_copies e2063d266acd 6081d72689dc : 1533 revs, 0.015602 s, 0.015697 s, +0.000095 s, × 1.0061, 10 µs/rev netbeans x000_revs_x000_added_x000_copies ff453e9fee32 411350406ec2 : 5750 revs, 0.058759 s, 0.063528 s, +0.004769 s, × 1.0812, 11 µs/rev netbeans x0000_revs_xx000_added_x000_copies 588c2d1ced70 1aad62e59ddd : 66949 revs, 0.491550 s, 0.545515 s, +0.053965 s, × 1.1098, 8 µs/rev mozilla-central x_revs_x_added_0_copies 3697f962bb7b 7015fcdd43a2 : 2 revs, 0.000087 s, 0.000089 s, +0.000002 s, × 1.0230, 44 µs/rev mozilla-central x_revs_x000_added_0_copies dd390860c6c9 40d0c5bed75d : 8 revs, 0.000268 s, 0.000265 s, -0.000003 s, × 0.9888, 33 µs/rev mozilla-central x_revs_x_added_x_copies 8d198483ae3b 14207ffc2b2f : 9 revs, 0.000181 s, 0.000381 s, +0.000200 s, × 2.1050, 42 µs/rev mozilla-central x_revs_x00_added_x_copies 98cbc58cc6bc 446a150332c3 : 7 revs, 0.000661 s, 0.000672 s, +0.000011 s, × 1.0166, 96 µs/rev mozilla-central x_revs_x000_added_x000_copies 3c684b4b8f68 0a5e72d1b479 : 3 revs, 0.003256 s, 0.003497 s, +0.000241 s, × 1.0740, 1165 µs/rev mozilla-central x_revs_x0000_added_x0000_copies effb563bb7e5 c07a39dc4e80 : 6 revs, 0.066749 s, 0.073204 s, +0.006455 s, × 1.0967, 12200 µs/rev mozilla-central x000_revs_xx00_added_0_copies 6100d773079a 04a55431795e : 1593 revs, 0.006462 s, 0.006482 s, +0.000020 s, × 1.0031, 4 µs/rev mozilla-central x000_revs_x000_added_x_copies 9f17a6fc04f9 2d37b966abed : 41 revs, 0.004919 s, 0.005066 s, +0.000147 s, × 1.0299, 123 µs/rev mozilla-central x000_revs_x000_added_x000_copies 7c97034feb78 4407bd0c6330 : 7839 revs, 0.062421 s, 0.065707 s, +0.003286 s, × 1.0526, 8 µs/rev mozilla-central x0000_revs_xx000_added_0_copies 9eec5917337d 67118cc6dcad : 615 revs, 0.026633 s, 0.026800 s, +0.000167 s, × 1.0063, 43 µs/rev mozilla-central x0000_revs_xx000_added_x000_copies f78c615a656c 96a38b690156 : 30263 revs, 0.197792 s, 0.203856 s, +0.006064 s, × 1.0307, 6 µs/rev mozilla-central x00000_revs_x0000_added_x0000_copies 6832ae71433c 4c222a1d9a00 : 153721 revs, 1.259970 s, 1.293394 s, +0.033424 s, × 1.0265, 8 µs/rev mozilla-central x00000_revs_x00000_added_x000_copies 76caed42cf7c 1daa622bbe42 : 204976 revs, 1.689184 s, 1.698239 s, +0.009055 s, × 1.0054, 8 µs/rev mozilla-try x_revs_x_added_0_copies aaf6dde0deb8 9790f499805a : 2 revs, 0.000865 s, 0.000875 s, +0.000010 s, × 1.0116, 437 µs/rev mozilla-try x_revs_x000_added_0_copies d8d0222927b4 5bb8ce8c7450 : 2 revs, 0.000893 s, 0.000891 s, -0.000002 s, × 0.9978, 445 µs/rev mozilla-try x_revs_x_added_x_copies 092fcca11bdb 936255a0384a : 4 revs, 0.000172 s, 0.000292 s, +0.000120 s, × 1.6977, 73 µs/rev mozilla-try x_revs_x00_added_x_copies b53d2fadbdb5 017afae788ec : 2 revs, 0.001159 s, 0.003939 s, +0.002780 s, × 3.3986, 1969 µs/rev mozilla-try x_revs_x000_added_x000_copies 20408ad61ce5 6f0ee96e21ad : 1 revs, 0.031621 s, 0.033027 s, +0.001406 s, × 1.0445, 33027 µs/rev mozilla-try x_revs_x0000_added_x0000_copies effb563bb7e5 c07a39dc4e80 : 6 revs, 0.068571 s, 0.073703 s, +0.005132 s, × 1.0748, 12283 µs/rev mozilla-try x000_revs_xx00_added_0_copies 6100d773079a 04a55431795e : 1593 revs, 0.006452 s, 0.006469 s, +0.000017 s, × 1.0026, 4 µs/rev mozilla-try x000_revs_x000_added_x_copies 9f17a6fc04f9 2d37b966abed : 41 revs, 0.005443 s, 0.005278 s, -0.000165 s, × 0.9697, 128 µs/rev mozilla-try x000_revs_x000_added_x000_copies 1346fd0130e4 4c65cbdabc1f : 6657 revs, 0.063180 s, 0.064995 s, +0.001815 s, × 1.0287, 9 µs/rev mozilla-try x0000_revs_x_added_0_copies 63519bfd42ee a36a2a865d92 : 40314 revs, 0.293564 s, 0.301041 s, +0.007477 s, × 1.0255, 7 µs/rev mozilla-try x0000_revs_x_added_x_copies 9fe69ff0762d bcabf2a78927 : 38690 revs, 0.286595 s, 0.285575 s, -0.001020 s, × 0.9964, 7 µs/rev mozilla-try x0000_revs_xx000_added_x_copies 156f6e2674f2 4d0f2c178e66 : 8598 revs, 0.083256 s, 0.085597 s, +0.002341 s, × 1.0281, 9 µs/rev mozilla-try x0000_revs_xx000_added_0_copies 9eec5917337d 67118cc6dcad : 615 revs, 0.027282 s, 0.027118 s, -0.000164 s, × 0.9940, 44 µs/rev mozilla-try x0000_revs_xx000_added_x000_copies 89294cd501d9 7ccb2fc7ccb5 : 97052 revs, 1.343373 s, 2.119204 s, +0.775831 s, × 1.5775, 21 µs/rev mozilla-try x0000_revs_x0000_added_x0000_copies e928c65095ed e951f4ad123a : 52031 revs, 0.665737 s, 0.701479 s, +0.035742 s, × 1.0537, 13 µs/rev mozilla-try x00000_revs_x_added_0_copies 6a320851d377 1ebb79acd503 : 363753 revs, 5.138169 s, 4.482399 s, -0.655770 s, × 0.8724, 12 µs/rev mozilla-try x00000_revs_x00000_added_0_copies dc8a3ca7010e d16fde900c9c : 34414 revs, 0.573276 s, 0.574082 s, +0.000806 s, × 1.0014, 16 µs/rev mozilla-try x00000_revs_x_added_x_copies 5173c4b6f97c 95d83ee7242d : 362229 revs, 5.127809 s, 4.480366 s, -0.647443 s, × 0.8737, 12 µs/rev mozilla-try x00000_revs_x000_added_x_copies 9126823d0e9c ca82787bb23c : 359344 revs, 4.971136 s, 4.369070 s, -0.602066 s, × 0.8789, 12 µs/rev mozilla-try x00000_revs_x0000_added_x0000_copies 8d3fafa80d4b eb884023b810 : 192665 revs, 1.741678 s, 1.592506 s, -0.149172 s, × 0.9144, 8 µs/rev mozilla-try x00000_revs_x00000_added_x0000_copies 1b661134e2ca 1ae03d022d6d : 228985 revs, 40.314822 s, 87.824489 s, +47.509667 s, × 2.1785, 383 µs/rev mozilla-try x00000_revs_x00000_added_x000_copies 9b2a99adc05e 8e29777b48e6 : 382065 revs, 20.048029 s, 43.304637 s, +23.256608 s, × 2.1600, 113 µs/rev private : 459513 revs, 37.179470 s, 33.853687 s, -3.325783 s, × 0.9105, 73 µs/rev Differential Revision: https://phab.mercurial-scm.org/D9644
author Pierre-Yves David <pierre-yves.david@octobus.net>
date Tue, 15 Dec 2020 18:04:23 +0100
parents a89aa2d7b34d
children d4ba4d51f85f
line wrap: on
line source

# pvec.py - probabilistic vector clocks for Mercurial
#
# Copyright 2012 Matt Mackall <mpm@selenic.com>
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.

'''
A "pvec" is a changeset property based on the theory of vector clocks
that can be compared to discover relatedness without consulting a
graph. This can be useful for tasks like determining how a
disconnected patch relates to a repository.

Currently a pvec consist of 448 bits, of which 24 are 'depth' and the
remainder are a bit vector. It is represented as a 70-character base85
string.

Construction:

- a root changeset has a depth of 0 and a bit vector based on its hash
- a normal commit has a changeset where depth is increased by one and
  one bit vector bit is flipped based on its hash
- a merge changeset pvec is constructed by copying changes from one pvec into
  the other to balance its depth

Properties:

- for linear changes, difference in depth is always <= hamming distance
- otherwise, changes are probably divergent
- when hamming distance is < 200, we can reliably detect when pvecs are near

Issues:

- hamming distance ceases to work over distances of ~ 200
- detecting divergence is less accurate when the common ancestor is very close
  to either revision or total distance is high
- this could probably be improved by modeling the relation between
  delta and hdist

Uses:

- a patch pvec can be used to locate the nearest available common ancestor for
  resolving conflicts
- ordering of patches can be established without a DAG
- two head pvecs can be compared to determine whether push/pull/merge is needed
  and approximately how many changesets are involved
- can be used to find a heuristic divergence measure between changesets on
  different branches
'''

from __future__ import absolute_import

from .node import nullrev
from . import (
    pycompat,
    util,
)

_size = 448  # 70 chars b85-encoded
_bytes = _size // 8
_depthbits = 24
_depthbytes = _depthbits // 8
_vecbytes = _bytes - _depthbytes
_vecbits = _vecbytes * 8
_radius = (_vecbits - 30) // 2  # high probability vectors are related


def _bin(bs):
    '''convert a bytestring to a long'''
    v = 0
    for b in bs:
        v = v * 256 + ord(b)
    return v


def _str(v, l):
    # type: (int, int) -> bytes
    bs = b""
    for p in pycompat.xrange(l):
        bs = pycompat.bytechr(v & 255) + bs
        v >>= 8
    return bs


def _split(b):
    '''depth and bitvec'''
    return _bin(b[:_depthbytes]), _bin(b[_depthbytes:])


def _join(depth, bitvec):
    return _str(depth, _depthbytes) + _str(bitvec, _vecbytes)


def _hweight(x):
    c = 0
    while x:
        if x & 1:
            c += 1
        x >>= 1
    return c


_htab = [_hweight(x) for x in pycompat.xrange(256)]


def _hamming(a, b):
    '''find the hamming distance between two longs'''
    d = a ^ b
    c = 0
    while d:
        c += _htab[d & 0xFF]
        d >>= 8
    return c


def _mergevec(x, y, c):
    # Ideally, this function would be x ^ y ^ ancestor, but finding
    # ancestors is a nuisance. So instead we find the minimal number
    # of changes to balance the depth and hamming distance

    d1, v1 = x
    d2, v2 = y
    if d1 < d2:
        d1, d2, v1, v2 = d2, d1, v2, v1

    hdist = _hamming(v1, v2)
    ddist = d1 - d2
    v = v1
    m = v1 ^ v2  # mask of different bits
    i = 1

    if hdist > ddist:
        # if delta = 10 and hdist = 100, then we need to go up 55 steps
        # to the ancestor and down 45
        changes = (hdist - ddist + 1) // 2
    else:
        # must make at least one change
        changes = 1
    depth = d1 + changes

    # copy changes from v2
    if m:
        while changes:
            if m & i:
                v ^= i
                changes -= 1
            i <<= 1
    else:
        v = _flipbit(v, c)

    return depth, v


def _flipbit(v, node):
    # converting bit strings to longs is slow
    bit = (hash(node) & 0xFFFFFFFF) % _vecbits
    return v ^ (1 << bit)


def ctxpvec(ctx):
    '''construct a pvec for ctx while filling in the cache'''
    r = ctx.repo()
    if not util.safehasattr(r, "_pveccache"):
        r._pveccache = {}
    pvc = r._pveccache
    if ctx.rev() not in pvc:
        cl = r.changelog
        for n in pycompat.xrange(ctx.rev() + 1):
            if n not in pvc:
                node = cl.node(n)
                p1, p2 = cl.parentrevs(n)
                if p1 == nullrev:
                    # start with a 'random' vector at root
                    pvc[n] = (0, _bin((node * 3)[:_vecbytes]))
                elif p2 == nullrev:
                    d, v = pvc[p1]
                    pvc[n] = (d + 1, _flipbit(v, node))
                else:
                    pvc[n] = _mergevec(pvc[p1], pvc[p2], node)
    bs = _join(*pvc[ctx.rev()])
    return pvec(util.b85encode(bs))


class pvec(object):
    def __init__(self, hashorctx):
        if isinstance(hashorctx, bytes):
            self._bs = hashorctx
            self._depth, self._vec = _split(util.b85decode(hashorctx))
        else:
            self._vec = ctxpvec(hashorctx)

    def __str__(self):
        return self._bs

    def __eq__(self, b):
        return self._vec == b._vec and self._depth == b._depth

    def __lt__(self, b):
        delta = b._depth - self._depth
        if delta < 0:
            return False  # always correct
        if _hamming(self._vec, b._vec) > delta:
            return False
        return True

    def __gt__(self, b):
        return b < self

    def __or__(self, b):
        delta = abs(b._depth - self._depth)
        if _hamming(self._vec, b._vec) <= delta:
            return False
        return True

    def __sub__(self, b):
        if self | b:
            raise ValueError(b"concurrent pvecs")
        return self._depth - b._depth

    def distance(self, b):
        d = abs(b._depth - self._depth)
        h = _hamming(self._vec, b._vec)
        return max(d, h)

    def near(self, b):
        dist = abs(b.depth - self._depth)
        if dist > _radius or _hamming(self._vec, b._vec) > _radius:
            return False