view mercurial/mergestate.py @ 45892:06b64fabf91c

copies: cache the ancestor checking call when tracing copy A good share of the time spent in this function is spent doing ancestors checking. To avoid spending time in duplicated call, we cache the result of calls. In the slower case, this provide a quite significant performance boost. Below are the result for a set of selected pairs (many of them pathological): (And further down is another table that summarize the current state of filelog based vs changeset base copy tracing) The benchmark have been configured to be killed after 6 minutes of runtime, which mean that any detect slower than 2 minutes will be marked as "killed". This drop some useful information about how much slower these case are? but also prevent 99% of the benchmark time to be spent on case that can be labelled "very slow" anyway. Repo Case Source-Rev Dest-Rev Old-Time New-Time Difference Factor ------------------------------------------------------------------------------------------------------------------------------------ mercurial x_revs_x_added_0_copies ad6b123de1c7 39cfcef4f463 : 0.000044 s, 0.000044 s, +0.000000 s, ? 1.0000 mercurial x_revs_x_added_x_copies 2b1c78674230 0c1d10351869 : 0.000138 s, 0.000138 s, +0.000000 s, ? 1.0000 mercurial x000_revs_x000_added_x_copies 81f8ff2a9bf2 dd3267698d84 : 0.005067 s, 0.005052 s, -0.000015 s, ? 0.9970 pypy x_revs_x_added_0_copies aed021ee8ae8 099ed31b181b : 0.000218 s, 0.000219 s, +0.000001 s, ? 1.0046 pypy x_revs_x000_added_0_copies 4aa4e1f8e19a 359343b9ac0e : 0.000053 s, 0.000055 s, +0.000002 s, ? 1.0377 pypy x_revs_x_added_x_copies ac52eb7bbbb0 72e022663155 : 0.000125 s, 0.000128 s, +0.000003 s, ? 1.0240 pypy x_revs_x00_added_x_copies c3b14617fbd7 ace7255d9a26 : 0.001098 s, 0.001089 s, -0.000009 s, ? 0.9918 pypy x_revs_x000_added_x000_copies df6f7a526b60 a83dc6a2d56f : 0.017546 s, 0.017407 s, -0.000139 s, ? 0.9921 pypy x000_revs_xx00_added_0_copies 89a76aede314 2f22446ff07e : 0.096723 s, 0.094175 s, -0.002548 s, ? 0.9737 pypy x000_revs_x000_added_x_copies 8a3b5bfd266e 2c68e87c3efe : 0.271796 s, 0.238009 s, -0.033787 s, ? 0.8757 pypy x000_revs_x000_added_x000_copies 89a76aede314 7b3dda341c84 : 0.128602 s, 0.125876 s, -0.002726 s, ? 0.9788 pypy x0000_revs_x_added_0_copies d1defd0dc478 c9cb1334cc78 : 7.086742 s, 3.581556 s, -3.505186 s, ? 0.5054 pypy x0000_revs_xx000_added_0_copies bf2c629d0071 4ffed77c095c : 0.016634 s, 0.016721 s, +0.000087 s, ? 1.0052 pypy x0000_revs_xx000_added_x000_copies 08ea3258278e d9fa043f30c0 : 0.254225 s, 0.242367 s, -0.011858 s, ? 0.9534 netbeans x_revs_x_added_0_copies fb0955ffcbcd a01e9239f9e7 : 0.000166 s, 0.000165 s, -0.000001 s, ? 0.9940 netbeans x_revs_x000_added_0_copies 6f360122949f 20eb231cc7d0 : 0.000118 s, 0.000114 s, -0.000004 s, ? 0.9661 netbeans x_revs_x_added_x_copies 1ada3faf6fb6 5a39d12eecf4 : 0.000296 s, 0.000296 s, +0.000000 s, ? 1.0000 netbeans x_revs_x00_added_x_copies 35be93ba1e2c 9eec5e90c05f : 0.001137 s, 0.001124 s, -0.000013 s, ? 0.9886 netbeans x000_revs_xx00_added_0_copies eac3045b4fdd 51d4ae7f1290 : 0.014133 s, 0.013060 s, -0.001073 s, ? 0.9241 netbeans x000_revs_x000_added_x_copies e2063d266acd 6081d72689dc : 0.016988 s, 0.017112 s, +0.000124 s, ? 1.0073 netbeans x000_revs_x000_added_x000_copies ff453e9fee32 411350406ec2 : 0.676361 s, 0.660350 s, -0.016011 s, ? 0.9763 netbeans x0000_revs_xx000_added_x000_copies 588c2d1ced70 1aad62e59ddd : 12.515149 s, 10.032499 s, -2.482650 s, ? 0.8016 mozilla-central x_revs_x_added_0_copies 3697f962bb7b 7015fcdd43a2 : 0.000186 s, 0.000189 s, +0.000003 s, ? 1.0161 mozilla-central x_revs_x000_added_0_copies dd390860c6c9 40d0c5bed75d : 0.000459 s, 0.000462 s, +0.000003 s, ? 1.0065 mozilla-central x_revs_x_added_x_copies 8d198483ae3b 14207ffc2b2f : 0.000273 s, 0.000270 s, -0.000003 s, ? 0.9890 mozilla-central x_revs_x00_added_x_copies 98cbc58cc6bc 446a150332c3 : 0.001503 s, 0.001474 s, -0.000029 s, ? 0.9807 mozilla-central x_revs_x000_added_x000_copies 3c684b4b8f68 0a5e72d1b479 : 0.004862 s, 0.004806 s, -0.000056 s, ? 0.9885 mozilla-central x_revs_x0000_added_x0000_copies effb563bb7e5 c07a39dc4e80 : 0.088291 s, 0.085150 s, -0.003141 s, ? 0.9644 mozilla-central x000_revs_xx00_added_0_copies 6100d773079a 04a55431795e : 0.007113 s, 0.007064 s, -0.000049 s, ? 0.9931 mozilla-central x000_revs_x000_added_x_copies 9f17a6fc04f9 2d37b966abed : 0.004687 s, 0.004741 s, +0.000054 s, ? 1.0115 mozilla-central x000_revs_x000_added_x000_copies 7c97034feb78 4407bd0c6330 : 0.198710 s, 0.190133 s, -0.008577 s, ? 0.9568 mozilla-central x0000_revs_xx000_added_0_copies 9eec5917337d 67118cc6dcad : 0.036068 s, 0.035651 s, -0.000417 s, ? 0.9884 mozilla-central x0000_revs_xx000_added_x000_copies f78c615a656c 96a38b690156 : 0.465362 s, 0.440694 s, -0.024668 s, ? 0.9470 mozilla-central x00000_revs_x0000_added_x0000_copies 6832ae71433c 4c222a1d9a00 : 24.519684 s, 18.454163 s, -6.065521 s, ? 0.7526 mozilla-central x00000_revs_x00000_added_x000_copies 76caed42cf7c 1daa622bbe42 : 42.711897 s, 31.562719 s, -11.149178 s, ? 0.7390 mozilla-try x_revs_x_added_0_copies aaf6dde0deb8 9790f499805a : 0.001201 s, 0.001189 s, -0.000012 s, ? 0.9900 mozilla-try x_revs_x000_added_0_copies d8d0222927b4 5bb8ce8c7450 : 0.001216 s, 0.001204 s, -0.000012 s, ? 0.9901 mozilla-try x_revs_x_added_x_copies 092fcca11bdb 936255a0384a : 0.000595 s, 0.000586 s, -0.000009 s, ? 0.9849 mozilla-try x_revs_x00_added_x_copies b53d2fadbdb5 017afae788ec : 0.001856 s, 0.001845 s, -0.000011 s, ? 0.9941 mozilla-try x_revs_x000_added_x000_copies 20408ad61ce5 6f0ee96e21ad : 0.064936 s, 0.063822 s, -0.001114 s, ? 0.9828 mozilla-try x_revs_x0000_added_x0000_copies effb563bb7e5 c07a39dc4e80 : 0.090601 s, 0.088038 s, -0.002563 s, ? 0.9717 mozilla-try x000_revs_xx00_added_0_copies 6100d773079a 04a55431795e : 0.007510 s, 0.007389 s, -0.000121 s, ? 0.9839 mozilla-try x000_revs_x000_added_x_copies 9f17a6fc04f9 2d37b966abed : 0.004911 s, 0.004868 s, -0.000043 s, ? 0.9912 mozilla-try x000_revs_x000_added_x000_copies 1346fd0130e4 4c65cbdabc1f : 0.233231 s, 0.222450 s, -0.010781 s, ? 0.9538 mozilla-try x0000_revs_x_added_0_copies 63519bfd42ee a36a2a865d92 : 0.419989 s, 0.370675 s, -0.049314 s, ? 0.8826 mozilla-try x0000_revs_x_added_x_copies 9fe69ff0762d bcabf2a78927 : 0.401521 s, 0.358020 s, -0.043501 s, ? 0.8917 mozilla-try x0000_revs_xx000_added_x_copies 156f6e2674f2 4d0f2c178e66 : 0.179555 s, 0.145235 s, -0.034320 s, ? 0.8089 mozilla-try x0000_revs_xx000_added_0_copies 9eec5917337d 67118cc6dcad : 0.038004 s, 0.037606 s, -0.000398 s, ? 0.9895 mozilla-try x0000_revs_xx000_added_x000_copies 89294cd501d9 7ccb2fc7ccb5 : 52.838482 s, 7.382439 s, -45.456043 s, ? 0.1397 mozilla-try x0000_revs_x0000_added_x0000_copies e928c65095ed e951f4ad123a : 8.705874 s, 7.273506 s, -1.432368 s, ? 0.8355 mozilla-try x00000_revs_x00000_added_0_copies dc8a3ca7010e d16fde900c9c : 1.126708 s, 1.074593 s, -0.052115 s, ? 0.9537 mozilla-try x00000_revs_x0000_added_x0000_copies 8d3fafa80d4b eb884023b810 : 83.854020 s, 27.746195 s, -56.107825 s, ? 0.3309 Below is a table comparing the runtime of the current "filelog centric" algorithm, with the "changeset centric" one, we just modified. The changeset centric algorithm is a significant win in many scenario, but they are still various cases where it is quite slower. When many revision has to be considered the cost of retrieving the copy information, creating new dictionaries, merging dictionaries and checking if revision are ancestors of each other can slow things down. The rest of this series, will introduce a rust version of the copy tracing code to deal with most of theses issues. Repo Case Source-Rev Dest-Rev filelog sidedata Difference Factor --------------------------------------------------------------------------------------------------------------------------------------- mercurial x_revs_x_added_0_copies ad6b123de1c7 39cfcef4f463 : 0.000914 s, 0.000044 s, - 0.000870 s, ? 0.048140 mercurial x_revs_x_added_x_copies 2b1c78674230 0c1d10351869 : 0.001812 s, 0.000138 s, - 0.001674 s, ? 0.076159 mercurial x000_revs_x000_added_x_copies 81f8ff2a9bf2 dd3267698d84 : 0.017954 s, 0.005052 s, - 0.012902 s, ? 0.281386 pypy x_revs_x_added_0_copies aed021ee8ae8 099ed31b181b : 0.001509 s, 0.000219 s, - 0.001290 s, ? 0.145129 pypy x_revs_x000_added_0_copies 4aa4e1f8e19a 359343b9ac0e : 0.206881 s, 0.000055 s, - 0.206826 s, ? 0.000266 pypy x_revs_x_added_x_copies ac52eb7bbbb0 72e022663155 : 0.016951 s, 0.000128 s, - 0.016823 s, ? 0.007551 pypy x_revs_x00_added_x_copies c3b14617fbd7 ace7255d9a26 : 0.019096 s, 0.001089 s, - 0.018007 s, ? 0.057028 pypy x_revs_x000_added_x000_copies df6f7a526b60 a83dc6a2d56f : 0.762506 s, 0.017407 s, - 0.745099 s, ? 0.022829 pypy x000_revs_xx00_added_0_copies 89a76aede314 2f22446ff07e : 1.179211 s, 0.094175 s, - 1.085036 s, ? 0.079863 pypy x000_revs_x000_added_x_copies 8a3b5bfd266e 2c68e87c3efe : 1.249058 s, 0.238009 s, - 1.011049 s, ? 0.190551 pypy x000_revs_x000_added_x000_copies 89a76aede314 7b3dda341c84 : 1.614107 s, 0.125876 s, - 1.488231 s, ? 0.077985 pypy x0000_revs_x_added_0_copies d1defd0dc478 c9cb1334cc78 : 0.001064 s, 3.581556 s, + 3.580492 s, ? 3366.124060 pypy x0000_revs_xx000_added_0_copies bf2c629d0071 4ffed77c095c : 1.061275 s, 0.016721 s, - 1.044554 s, ? 0.015756 pypy x0000_revs_xx000_added_x000_copies 08ea3258278e d9fa043f30c0 : 1.341119 s, 0.242367 s, - 1.098752 s, ? 0.180720 netbeans x_revs_x_added_0_copies fb0955ffcbcd a01e9239f9e7 : 0.027803 s, 0.000165 s, - 0.027638 s, ? 0.005935 netbeans x_revs_x000_added_0_copies 6f360122949f 20eb231cc7d0 : 0.130014 s, 0.000114 s, - 0.129900 s, ? 0.000877 netbeans x_revs_x_added_x_copies 1ada3faf6fb6 5a39d12eecf4 : 0.024990 s, 0.000296 s, - 0.024694 s, ? 0.011845 netbeans x_revs_x00_added_x_copies 35be93ba1e2c 9eec5e90c05f : 0.052201 s, 0.001124 s, - 0.051077 s, ? 0.021532 netbeans x000_revs_xx00_added_0_copies eac3045b4fdd 51d4ae7f1290 : 0.037642 s, 0.013060 s, - 0.024582 s, ? 0.346953 netbeans x000_revs_x000_added_x_copies e2063d266acd 6081d72689dc : 0.197086 s, 0.017112 s, - 0.179974 s, ? 0.086825 netbeans x000_revs_x000_added_x000_copies ff453e9fee32 411350406ec2 : 0.935148 s, 0.660350 s, - 0.274798 s, ? 0.706145 netbeans x0000_revs_xx000_added_x000_copies 588c2d1ced70 1aad62e59ddd : 3.920674 s, 10.032499 s, + 6.111825 s, ? 2.558871 mozilla-central x_revs_x_added_0_copies 3697f962bb7b 7015fcdd43a2 : 0.024232 s, 0.000189 s, - 0.024043 s, ? 0.007800 mozilla-central x_revs_x000_added_0_copies dd390860c6c9 40d0c5bed75d : 0.141483 s, 0.000462 s, - 0.141021 s, ? 0.003265 mozilla-central x_revs_x_added_x_copies 8d198483ae3b 14207ffc2b2f : 0.025775 s, 0.000270 s, - 0.025505 s, ? 0.010475 mozilla-central x_revs_x00_added_x_copies 98cbc58cc6bc 446a150332c3 : 0.084922 s, 0.001474 s, - 0.083448 s, ? 0.017357 mozilla-central x_revs_x000_added_x000_copies 3c684b4b8f68 0a5e72d1b479 : 0.194784 s, 0.004806 s, - 0.189978 s, ? 0.024673 mozilla-central x_revs_x0000_added_x0000_copies effb563bb7e5 c07a39dc4e80 : 2.161103 s, 0.085150 s, - 2.075953 s, ? 0.039401 mozilla-central x000_revs_xx00_added_0_copies 6100d773079a 04a55431795e : 0.089347 s, 0.007064 s, - 0.082283 s, ? 0.079063 mozilla-central x000_revs_x000_added_x_copies 9f17a6fc04f9 2d37b966abed : 0.732171 s, 0.004741 s, - 0.727430 s, ? 0.006475 mozilla-central x000_revs_x000_added_x000_copies 7c97034feb78 4407bd0c6330 : 1.157287 s, 0.190133 s, - 0.967154 s, ? 0.164292 mozilla-central x0000_revs_xx000_added_0_copies 9eec5917337d 67118cc6dcad : 6.726568 s, 0.035651 s, - 6.690917 s, ? 0.005300 mozilla-central x0000_revs_xx000_added_x000_copies f78c615a656c 96a38b690156 : 3.266229 s, 0.440694 s, - 2.825535 s, ? 0.134924 mozilla-central x00000_revs_x0000_added_x0000_copies 6832ae71433c 4c222a1d9a00 : 15.860534 s, 18.454163 s, + 2.593629 s, ? 1.163527 mozilla-central x00000_revs_x00000_added_x000_copies 76caed42cf7c 1daa622bbe42 : 20.450475 s, 31.562719 s, +11.112244 s, ? 1.543373 mozilla-try x_revs_x_added_0_copies aaf6dde0deb8 9790f499805a : 0.080442 s, 0.001189 s, - 0.079253 s, ? 0.014781 mozilla-try x_revs_x000_added_0_copies d8d0222927b4 5bb8ce8c7450 : 0.497672 s, 0.001204 s, - 0.496468 s, ? 0.002419 mozilla-try x_revs_x_added_x_copies 092fcca11bdb 936255a0384a : 0.021183 s, 0.000586 s, - 0.020597 s, ? 0.027664 mozilla-try x_revs_x00_added_x_copies b53d2fadbdb5 017afae788ec : 0.230991 s, 0.001845 s, - 0.229146 s, ? 0.007987 mozilla-try x_revs_x000_added_x000_copies 20408ad61ce5 6f0ee96e21ad : 1.118461 s, 0.063822 s, - 1.054639 s, ? 0.057062 mozilla-try x_revs_x0000_added_x0000_copies effb563bb7e5 c07a39dc4e80 : 2.206083 s, 0.088038 s, - 2.118045 s, ? 0.039907 mozilla-try x000_revs_xx00_added_0_copies 6100d773079a 04a55431795e : 0.089404 s, 0.007389 s, - 0.082015 s, ? 0.082647 mozilla-try x000_revs_x000_added_x_copies 9f17a6fc04f9 2d37b966abed : 0.733043 s, 0.004868 s, - 0.728175 s, ? 0.006641 mozilla-try x000_revs_x000_added_x000_copies 1346fd0130e4 4c65cbdabc1f : 1.163367 s, 0.222450 s, - 0.940917 s, ? 0.191212 mozilla-try x0000_revs_x_added_0_copies 63519bfd42ee a36a2a865d92 : 0.085456 s, 0.370675 s, + 0.285219 s, ? 4.337612 mozilla-try x0000_revs_x_added_x_copies 9fe69ff0762d bcabf2a78927 : 0.083601 s, 0.358020 s, + 0.274419 s, ? 4.282485 mozilla-try x0000_revs_xx000_added_x_copies 156f6e2674f2 4d0f2c178e66 : 7.366614 s, 0.145235 s, - 7.221379 s, ? 0.019715 mozilla-try x0000_revs_xx000_added_0_copies 9eec5917337d 67118cc6dcad : 6.664464 s, 0.037606 s, - 6.626858 s, ? 0.005643 mozilla-try x0000_revs_xx000_added_x000_copies 89294cd501d9 7ccb2fc7ccb5 : 7.467836 s, 7.382439 s, - 0.085397 s, ? 0.988565 mozilla-try x0000_revs_x0000_added_x0000_copies e928c65095ed e951f4ad123a : 9.801294 s, 7.273506 s, - 2.527788 s, ? 0.742097 mozilla-try x00000_revs_x_added_0_copies 6a320851d377 1ebb79acd503 : 0.091886 s, killed mozilla-try x00000_revs_x00000_added_0_copies dc8a3ca7010e d16fde900c9c : 26.491140 s, 1.074593 s, -25.416547 s, ? 0.040564 mozilla-try x00000_revs_x_added_x_copies 5173c4b6f97c 95d83ee7242d : 0.092863 s, killed mozilla-try x00000_revs_x000_added_x_copies 9126823d0e9c ca82787bb23c : 0.226823 s, killed mozilla-try x00000_revs_x0000_added_x0000_copies 8d3fafa80d4b eb884023b810 : 18.914630 s, 27.746195 s, + 8.831565 s, ? 1.466917 mozilla-try x00000_revs_x00000_added_x0000_copies 1b661134e2ca 1ae03d022d6d : 21.198903 s, killed mozilla-try x00000_revs_x00000_added_x000_copies 9b2a99adc05e 8e29777b48e6 : 24.952268 s, killed Differential Revision: https://phab.mercurial-scm.org/D9296
author Pierre-Yves David <pierre-yves.david@octobus.net>
date Mon, 02 Nov 2020 11:03:56 +0100
parents 0428978bca22
children 89a2afe31e82
line wrap: on
line source

from __future__ import absolute_import

import collections
import errno
import shutil
import struct

from .i18n import _
from .node import (
    bin,
    hex,
    nullhex,
    nullid,
)
from . import (
    error,
    filemerge,
    pycompat,
    util,
)
from .utils import hashutil

_pack = struct.pack
_unpack = struct.unpack


def _droponode(data):
    # used for compatibility for v1
    bits = data.split(b'\0')
    bits = bits[:-2] + bits[-1:]
    return b'\0'.join(bits)


def _filectxorabsent(hexnode, ctx, f):
    if hexnode == nullhex:
        return filemerge.absentfilectx(ctx, f)
    else:
        return ctx[f]


# Merge state record types. See ``mergestate`` docs for more.

####
# merge records which records metadata about a current merge
# exists only once in a mergestate
#####
RECORD_LOCAL = b'L'
RECORD_OTHER = b'O'
# record merge labels
RECORD_LABELS = b'l'

#####
# record extra information about files, with one entry containing info about one
# file. Hence, multiple of them can exists
#####
RECORD_FILE_VALUES = b'f'

#####
# merge records which represents state of individual merges of files/folders
# These are top level records for each entry containing merge related info.
# Each record of these has info about one file. Hence multiple of them can
# exists
#####
RECORD_MERGED = b'F'
RECORD_CHANGEDELETE_CONFLICT = b'C'
# the path was dir on one side of merge and file on another
RECORD_PATH_CONFLICT = b'P'

#####
# possible state which a merge entry can have. These are stored inside top-level
# merge records mentioned just above.
#####
MERGE_RECORD_UNRESOLVED = b'u'
MERGE_RECORD_RESOLVED = b'r'
MERGE_RECORD_UNRESOLVED_PATH = b'pu'
MERGE_RECORD_RESOLVED_PATH = b'pr'
# represents that the file was automatically merged in favor
# of other version. This info is used on commit.
# This is now deprecated and commit related information is now
# stored in RECORD_FILE_VALUES
MERGE_RECORD_MERGED_OTHER = b'o'

#####
# top level record which stores other unknown records. Multiple of these can
# exists
#####
RECORD_OVERRIDE = b't'

#####
# legacy records which are no longer used but kept to prevent breaking BC
#####
# This record was release in 5.4 and usage was removed in 5.5
LEGACY_RECORD_RESOLVED_OTHER = b'R'
# This record was release in 3.7 and usage was removed in 5.6
LEGACY_RECORD_DRIVER_RESOLVED = b'd'
# This record was release in 3.7 and usage was removed in 5.6
LEGACY_MERGE_DRIVER_STATE = b'm'
# This record was release in 3.7 and usage was removed in 5.6
LEGACY_MERGE_DRIVER_MERGE = b'D'


ACTION_FORGET = b'f'
ACTION_REMOVE = b'r'
ACTION_ADD = b'a'
ACTION_GET = b'g'
ACTION_PATH_CONFLICT = b'p'
ACTION_PATH_CONFLICT_RESOLVE = b'pr'
ACTION_ADD_MODIFIED = b'am'
ACTION_CREATED = b'c'
ACTION_DELETED_CHANGED = b'dc'
ACTION_CHANGED_DELETED = b'cd'
ACTION_MERGE = b'm'
ACTION_LOCAL_DIR_RENAME_GET = b'dg'
ACTION_DIR_RENAME_MOVE_LOCAL = b'dm'
ACTION_KEEP = b'k'
# the file was absent on local side before merge and we should
# keep it absent (absent means file not present, it can be a result
# of file deletion, rename etc.)
ACTION_KEEP_ABSENT = b'ka'
# the file is absent on the ancestor and remote side of the merge
# hence this file is new and we should keep it
ACTION_KEEP_NEW = b'kn'
ACTION_EXEC = b'e'
ACTION_CREATED_MERGE = b'cm'

# actions which are no op
NO_OP_ACTIONS = (
    ACTION_KEEP,
    ACTION_KEEP_ABSENT,
    ACTION_KEEP_NEW,
)


class _mergestate_base(object):
    '''track 3-way merge state of individual files

    The merge state is stored on disk when needed. Two files are used: one with
    an old format (version 1), and one with a new format (version 2). Version 2
    stores a superset of the data in version 1, including new kinds of records
    in the future. For more about the new format, see the documentation for
    `_readrecordsv2`.

    Each record can contain arbitrary content, and has an associated type. This
    `type` should be a letter. If `type` is uppercase, the record is mandatory:
    versions of Mercurial that don't support it should abort. If `type` is
    lowercase, the record can be safely ignored.

    Currently known records:

    L: the node of the "local" part of the merge (hexified version)
    O: the node of the "other" part of the merge (hexified version)
    F: a file to be merged entry
    C: a change/delete or delete/change conflict
    P: a path conflict (file vs directory)
    f: a (filename, dictionary) tuple of optional values for a given file
    l: the labels for the parts of the merge.

    Merge record states (stored in self._state, indexed by filename):
    u: unresolved conflict
    r: resolved conflict
    pu: unresolved path conflict (file conflicts with directory)
    pr: resolved path conflict
    o: file was merged in favor of other parent of merge (DEPRECATED)

    The resolve command transitions between 'u' and 'r' for conflicts and
    'pu' and 'pr' for path conflicts.
    '''

    def __init__(self, repo):
        """Initialize the merge state.

        Do not use this directly! Instead call read() or clean()."""
        self._repo = repo
        self._state = {}
        self._stateextras = collections.defaultdict(dict)
        self._local = None
        self._other = None
        self._labels = None
        # contains a mapping of form:
        # {filename : (merge_return_value, action_to_be_performed}
        # these are results of re-running merge process
        # this dict is used to perform actions on dirstate caused by re-running
        # the merge
        self._results = {}
        self._dirty = False

    def reset(self):
        pass

    def start(self, node, other, labels=None):
        self._local = node
        self._other = other
        self._labels = labels

    @util.propertycache
    def local(self):
        if self._local is None:
            msg = b"local accessed but self._local isn't set"
            raise error.ProgrammingError(msg)
        return self._local

    @util.propertycache
    def localctx(self):
        return self._repo[self.local]

    @util.propertycache
    def other(self):
        if self._other is None:
            msg = b"other accessed but self._other isn't set"
            raise error.ProgrammingError(msg)
        return self._other

    @util.propertycache
    def otherctx(self):
        return self._repo[self.other]

    def active(self):
        """Whether mergestate is active.

        Returns True if there appears to be mergestate. This is a rough proxy
        for "is a merge in progress."
        """
        return bool(self._local) or bool(self._state)

    def commit(self):
        """Write current state on disk (if necessary)"""

    @staticmethod
    def getlocalkey(path):
        """hash the path of a local file context for storage in the .hg/merge
        directory."""

        return hex(hashutil.sha1(path).digest())

    def _make_backup(self, fctx, localkey):
        raise NotImplementedError()

    def _restore_backup(self, fctx, localkey, flags):
        raise NotImplementedError()

    def add(self, fcl, fco, fca, fd):
        """add a new (potentially?) conflicting file the merge state
        fcl: file context for local,
        fco: file context for remote,
        fca: file context for ancestors,
        fd:  file path of the resulting merge.

        note: also write the local version to the `.hg/merge` directory.
        """
        if fcl.isabsent():
            localkey = nullhex
        else:
            localkey = mergestate.getlocalkey(fcl.path())
            self._make_backup(fcl, localkey)
        self._state[fd] = [
            MERGE_RECORD_UNRESOLVED,
            localkey,
            fcl.path(),
            fca.path(),
            hex(fca.filenode()),
            fco.path(),
            hex(fco.filenode()),
            fcl.flags(),
        ]
        self._stateextras[fd][b'ancestorlinknode'] = hex(fca.node())
        self._dirty = True

    def addpathconflict(self, path, frename, forigin):
        """add a new conflicting path to the merge state
        path:    the path that conflicts
        frename: the filename the conflicting file was renamed to
        forigin: origin of the file ('l' or 'r' for local/remote)
        """
        self._state[path] = [MERGE_RECORD_UNRESOLVED_PATH, frename, forigin]
        self._dirty = True

    def addcommitinfo(self, path, data):
        """ stores information which is required at commit
        into _stateextras """
        self._stateextras[path].update(data)
        self._dirty = True

    def __contains__(self, dfile):
        return dfile in self._state

    def __getitem__(self, dfile):
        return self._state[dfile][0]

    def __iter__(self):
        return iter(sorted(self._state))

    def files(self):
        return self._state.keys()

    def mark(self, dfile, state):
        self._state[dfile][0] = state
        self._dirty = True

    def unresolved(self):
        """Obtain the paths of unresolved files."""

        for f, entry in pycompat.iteritems(self._state):
            if entry[0] in (
                MERGE_RECORD_UNRESOLVED,
                MERGE_RECORD_UNRESOLVED_PATH,
            ):
                yield f

    def allextras(self):
        """ return all extras information stored with the mergestate """
        return self._stateextras

    def extras(self, filename):
        """ return extras stored with the mergestate for the given filename """
        return self._stateextras[filename]

    def _resolve(self, preresolve, dfile, wctx):
        """rerun merge process for file path `dfile`.
        Returns whether the merge was completed and the return value of merge
        obtained from filemerge._filemerge().
        """
        if self[dfile] in (
            MERGE_RECORD_RESOLVED,
            LEGACY_RECORD_DRIVER_RESOLVED,
        ):
            return True, 0
        stateentry = self._state[dfile]
        state, localkey, lfile, afile, anode, ofile, onode, flags = stateentry
        octx = self._repo[self._other]
        extras = self.extras(dfile)
        anccommitnode = extras.get(b'ancestorlinknode')
        if anccommitnode:
            actx = self._repo[anccommitnode]
        else:
            actx = None
        fcd = _filectxorabsent(localkey, wctx, dfile)
        fco = _filectxorabsent(onode, octx, ofile)
        # TODO: move this to filectxorabsent
        fca = self._repo.filectx(afile, fileid=anode, changectx=actx)
        # "premerge" x flags
        flo = fco.flags()
        fla = fca.flags()
        if b'x' in flags + flo + fla and b'l' not in flags + flo + fla:
            if fca.node() == nullid and flags != flo:
                if preresolve:
                    self._repo.ui.warn(
                        _(
                            b'warning: cannot merge flags for %s '
                            b'without common ancestor - keeping local flags\n'
                        )
                        % afile
                    )
            elif flags == fla:
                flags = flo
        if preresolve:
            # restore local
            if localkey != nullhex:
                self._restore_backup(wctx[dfile], localkey, flags)
            else:
                wctx[dfile].remove(ignoremissing=True)
            complete, merge_ret, deleted = filemerge.premerge(
                self._repo,
                wctx,
                self._local,
                lfile,
                fcd,
                fco,
                fca,
                labels=self._labels,
            )
        else:
            complete, merge_ret, deleted = filemerge.filemerge(
                self._repo,
                wctx,
                self._local,
                lfile,
                fcd,
                fco,
                fca,
                labels=self._labels,
            )
        if merge_ret is None:
            # If return value of merge is None, then there are no real conflict
            del self._state[dfile]
            self._stateextras.pop(dfile, None)
            self._dirty = True
        elif not merge_ret:
            self.mark(dfile, MERGE_RECORD_RESOLVED)

        if complete:
            action = None
            if deleted:
                if fcd.isabsent():
                    # dc: local picked. Need to drop if present, which may
                    # happen on re-resolves.
                    action = ACTION_FORGET
                else:
                    # cd: remote picked (or otherwise deleted)
                    action = ACTION_REMOVE
            else:
                if fcd.isabsent():  # dc: remote picked
                    action = ACTION_GET
                elif fco.isabsent():  # cd: local picked
                    if dfile in self.localctx:
                        action = ACTION_ADD_MODIFIED
                    else:
                        action = ACTION_ADD
                # else: regular merges (no action necessary)
            self._results[dfile] = merge_ret, action

        return complete, merge_ret

    def preresolve(self, dfile, wctx):
        """run premerge process for dfile

        Returns whether the merge is complete, and the exit code."""
        return self._resolve(True, dfile, wctx)

    def resolve(self, dfile, wctx):
        """run merge process (assuming premerge was run) for dfile

        Returns the exit code of the merge."""
        return self._resolve(False, dfile, wctx)[1]

    def counts(self):
        """return counts for updated, merged and removed files in this
        session"""
        updated, merged, removed = 0, 0, 0
        for r, action in pycompat.itervalues(self._results):
            if r is None:
                updated += 1
            elif r == 0:
                if action == ACTION_REMOVE:
                    removed += 1
                else:
                    merged += 1
        return updated, merged, removed

    def unresolvedcount(self):
        """get unresolved count for this merge (persistent)"""
        return len(list(self.unresolved()))

    def actions(self):
        """return lists of actions to perform on the dirstate"""
        actions = {
            ACTION_REMOVE: [],
            ACTION_FORGET: [],
            ACTION_ADD: [],
            ACTION_ADD_MODIFIED: [],
            ACTION_GET: [],
        }
        for f, (r, action) in pycompat.iteritems(self._results):
            if action is not None:
                actions[action].append((f, None, b"merge result"))
        return actions


class mergestate(_mergestate_base):

    statepathv1 = b'merge/state'
    statepathv2 = b'merge/state2'

    @staticmethod
    def clean(repo):
        """Initialize a brand new merge state, removing any existing state on
        disk."""
        ms = mergestate(repo)
        ms.reset()
        return ms

    @staticmethod
    def read(repo):
        """Initialize the merge state, reading it from disk."""
        ms = mergestate(repo)
        ms._read()
        return ms

    def _read(self):
        """Analyse each record content to restore a serialized state from disk

        This function process "record" entry produced by the de-serialization
        of on disk file.
        """
        unsupported = set()
        records = self._readrecords()
        for rtype, record in records:
            if rtype == RECORD_LOCAL:
                self._local = bin(record)
            elif rtype == RECORD_OTHER:
                self._other = bin(record)
            elif rtype == LEGACY_MERGE_DRIVER_STATE:
                pass
            elif rtype in (
                RECORD_MERGED,
                RECORD_CHANGEDELETE_CONFLICT,
                RECORD_PATH_CONFLICT,
                LEGACY_MERGE_DRIVER_MERGE,
                LEGACY_RECORD_RESOLVED_OTHER,
            ):
                bits = record.split(b'\0')
                # merge entry type MERGE_RECORD_MERGED_OTHER is deprecated
                # and we now store related information in _stateextras, so
                # lets write to _stateextras directly
                if bits[1] == MERGE_RECORD_MERGED_OTHER:
                    self._stateextras[bits[0]][b'filenode-source'] = b'other'
                else:
                    self._state[bits[0]] = bits[1:]
            elif rtype == RECORD_FILE_VALUES:
                filename, rawextras = record.split(b'\0', 1)
                extraparts = rawextras.split(b'\0')
                extras = {}
                i = 0
                while i < len(extraparts):
                    extras[extraparts[i]] = extraparts[i + 1]
                    i += 2

                self._stateextras[filename] = extras
            elif rtype == RECORD_LABELS:
                labels = record.split(b'\0', 2)
                self._labels = [l for l in labels if len(l) > 0]
            elif not rtype.islower():
                unsupported.add(rtype)

        if unsupported:
            raise error.UnsupportedMergeRecords(unsupported)

    def _readrecords(self):
        """Read merge state from disk and return a list of record (TYPE, data)

        We read data from both v1 and v2 files and decide which one to use.

        V1 has been used by version prior to 2.9.1 and contains less data than
        v2. We read both versions and check if no data in v2 contradicts
        v1. If there is not contradiction we can safely assume that both v1
        and v2 were written at the same time and use the extract data in v2. If
        there is contradiction we ignore v2 content as we assume an old version
        of Mercurial has overwritten the mergestate file and left an old v2
        file around.

        returns list of record [(TYPE, data), ...]"""
        v1records = self._readrecordsv1()
        v2records = self._readrecordsv2()
        if self._v1v2match(v1records, v2records):
            return v2records
        else:
            # v1 file is newer than v2 file, use it
            # we have to infer the "other" changeset of the merge
            # we cannot do better than that with v1 of the format
            mctx = self._repo[None].parents()[-1]
            v1records.append((RECORD_OTHER, mctx.hex()))
            # add place holder "other" file node information
            # nobody is using it yet so we do no need to fetch the data
            # if mctx was wrong `mctx[bits[-2]]` may fails.
            for idx, r in enumerate(v1records):
                if r[0] == RECORD_MERGED:
                    bits = r[1].split(b'\0')
                    bits.insert(-2, b'')
                    v1records[idx] = (r[0], b'\0'.join(bits))
            return v1records

    def _v1v2match(self, v1records, v2records):
        oldv2 = set()  # old format version of v2 record
        for rec in v2records:
            if rec[0] == RECORD_LOCAL:
                oldv2.add(rec)
            elif rec[0] == RECORD_MERGED:
                # drop the onode data (not contained in v1)
                oldv2.add((RECORD_MERGED, _droponode(rec[1])))
        for rec in v1records:
            if rec not in oldv2:
                return False
        else:
            return True

    def _readrecordsv1(self):
        """read on disk merge state for version 1 file

        returns list of record [(TYPE, data), ...]

        Note: the "F" data from this file are one entry short
              (no "other file node" entry)
        """
        records = []
        try:
            f = self._repo.vfs(self.statepathv1)
            for i, l in enumerate(f):
                if i == 0:
                    records.append((RECORD_LOCAL, l[:-1]))
                else:
                    records.append((RECORD_MERGED, l[:-1]))
            f.close()
        except IOError as err:
            if err.errno != errno.ENOENT:
                raise
        return records

    def _readrecordsv2(self):
        """read on disk merge state for version 2 file

        This format is a list of arbitrary records of the form:

          [type][length][content]

        `type` is a single character, `length` is a 4 byte integer, and
        `content` is an arbitrary byte sequence of length `length`.

        Mercurial versions prior to 3.7 have a bug where if there are
        unsupported mandatory merge records, attempting to clear out the merge
        state with hg update --clean or similar aborts. The 't' record type
        works around that by writing out what those versions treat as an
        advisory record, but later versions interpret as special: the first
        character is the 'real' record type and everything onwards is the data.

        Returns list of records [(TYPE, data), ...]."""
        records = []
        try:
            f = self._repo.vfs(self.statepathv2)
            data = f.read()
            off = 0
            end = len(data)
            while off < end:
                rtype = data[off : off + 1]
                off += 1
                length = _unpack(b'>I', data[off : (off + 4)])[0]
                off += 4
                record = data[off : (off + length)]
                off += length
                if rtype == RECORD_OVERRIDE:
                    rtype, record = record[0:1], record[1:]
                records.append((rtype, record))
            f.close()
        except IOError as err:
            if err.errno != errno.ENOENT:
                raise
        return records

    def commit(self):
        if self._dirty:
            records = self._makerecords()
            self._writerecords(records)
            self._dirty = False

    def _makerecords(self):
        records = []
        records.append((RECORD_LOCAL, hex(self._local)))
        records.append((RECORD_OTHER, hex(self._other)))
        # Write out state items. In all cases, the value of the state map entry
        # is written as the contents of the record. The record type depends on
        # the type of state that is stored, and capital-letter records are used
        # to prevent older versions of Mercurial that do not support the feature
        # from loading them.
        for filename, v in pycompat.iteritems(self._state):
            if v[0] in (
                MERGE_RECORD_UNRESOLVED_PATH,
                MERGE_RECORD_RESOLVED_PATH,
            ):
                # Path conflicts. These are stored in 'P' records.  The current
                # resolution state ('pu' or 'pr') is stored within the record.
                records.append(
                    (RECORD_PATH_CONFLICT, b'\0'.join([filename] + v))
                )
            elif v[1] == nullhex or v[6] == nullhex:
                # Change/Delete or Delete/Change conflicts. These are stored in
                # 'C' records. v[1] is the local file, and is nullhex when the
                # file is deleted locally ('dc'). v[6] is the remote file, and
                # is nullhex when the file is deleted remotely ('cd').
                records.append(
                    (RECORD_CHANGEDELETE_CONFLICT, b'\0'.join([filename] + v))
                )
            else:
                # Normal files.  These are stored in 'F' records.
                records.append((RECORD_MERGED, b'\0'.join([filename] + v)))
        for filename, extras in sorted(pycompat.iteritems(self._stateextras)):
            rawextras = b'\0'.join(
                b'%s\0%s' % (k, v) for k, v in pycompat.iteritems(extras)
            )
            records.append(
                (RECORD_FILE_VALUES, b'%s\0%s' % (filename, rawextras))
            )
        if self._labels is not None:
            labels = b'\0'.join(self._labels)
            records.append((RECORD_LABELS, labels))
        return records

    def _writerecords(self, records):
        """Write current state on disk (both v1 and v2)"""
        self._writerecordsv1(records)
        self._writerecordsv2(records)

    def _writerecordsv1(self, records):
        """Write current state on disk in a version 1 file"""
        f = self._repo.vfs(self.statepathv1, b'wb')
        irecords = iter(records)
        lrecords = next(irecords)
        assert lrecords[0] == RECORD_LOCAL
        f.write(hex(self._local) + b'\n')
        for rtype, data in irecords:
            if rtype == RECORD_MERGED:
                f.write(b'%s\n' % _droponode(data))
        f.close()

    def _writerecordsv2(self, records):
        """Write current state on disk in a version 2 file

        See the docstring for _readrecordsv2 for why we use 't'."""
        # these are the records that all version 2 clients can read
        allowlist = (RECORD_LOCAL, RECORD_OTHER, RECORD_MERGED)
        f = self._repo.vfs(self.statepathv2, b'wb')
        for key, data in records:
            assert len(key) == 1
            if key not in allowlist:
                key, data = RECORD_OVERRIDE, b'%s%s' % (key, data)
            format = b'>sI%is' % len(data)
            f.write(_pack(format, key, len(data), data))
        f.close()

    def _make_backup(self, fctx, localkey):
        self._repo.vfs.write(b'merge/' + localkey, fctx.data())

    def _restore_backup(self, fctx, localkey, flags):
        with self._repo.vfs(b'merge/' + localkey) as f:
            fctx.write(f.read(), flags)

    def reset(self):
        shutil.rmtree(self._repo.vfs.join(b'merge'), True)


class memmergestate(_mergestate_base):
    def __init__(self, repo):
        super(memmergestate, self).__init__(repo)
        self._backups = {}

    def _make_backup(self, fctx, localkey):
        self._backups[localkey] = fctx.data()

    def _restore_backup(self, fctx, localkey, flags):
        fctx.write(self._backups[localkey], flags)


def recordupdates(repo, actions, branchmerge, getfiledata):
    """record merge actions to the dirstate"""
    # remove (must come first)
    for f, args, msg in actions.get(ACTION_REMOVE, []):
        if branchmerge:
            repo.dirstate.remove(f)
        else:
            repo.dirstate.drop(f)

    # forget (must come first)
    for f, args, msg in actions.get(ACTION_FORGET, []):
        repo.dirstate.drop(f)

    # resolve path conflicts
    for f, args, msg in actions.get(ACTION_PATH_CONFLICT_RESOLVE, []):
        (f0, origf0) = args
        repo.dirstate.add(f)
        repo.dirstate.copy(origf0, f)
        if f0 == origf0:
            repo.dirstate.remove(f0)
        else:
            repo.dirstate.drop(f0)

    # re-add
    for f, args, msg in actions.get(ACTION_ADD, []):
        repo.dirstate.add(f)

    # re-add/mark as modified
    for f, args, msg in actions.get(ACTION_ADD_MODIFIED, []):
        if branchmerge:
            repo.dirstate.normallookup(f)
        else:
            repo.dirstate.add(f)

    # exec change
    for f, args, msg in actions.get(ACTION_EXEC, []):
        repo.dirstate.normallookup(f)

    # keep
    for f, args, msg in actions.get(ACTION_KEEP, []):
        pass

    # keep deleted
    for f, args, msg in actions.get(ACTION_KEEP_ABSENT, []):
        pass

    # keep new
    for f, args, msg in actions.get(ACTION_KEEP_NEW, []):
        pass

    # get
    for f, args, msg in actions.get(ACTION_GET, []):
        if branchmerge:
            repo.dirstate.otherparent(f)
        else:
            parentfiledata = getfiledata[f] if getfiledata else None
            repo.dirstate.normal(f, parentfiledata=parentfiledata)

    # merge
    for f, args, msg in actions.get(ACTION_MERGE, []):
        f1, f2, fa, move, anc = args
        if branchmerge:
            # We've done a branch merge, mark this file as merged
            # so that we properly record the merger later
            repo.dirstate.merge(f)
            if f1 != f2:  # copy/rename
                if move:
                    repo.dirstate.remove(f1)
                if f1 != f:
                    repo.dirstate.copy(f1, f)
                else:
                    repo.dirstate.copy(f2, f)
        else:
            # We've update-merged a locally modified file, so
            # we set the dirstate to emulate a normal checkout
            # of that file some time in the past. Thus our
            # merge will appear as a normal local file
            # modification.
            if f2 == f:  # file not locally copied/moved
                repo.dirstate.normallookup(f)
            if move:
                repo.dirstate.drop(f1)

    # directory rename, move local
    for f, args, msg in actions.get(ACTION_DIR_RENAME_MOVE_LOCAL, []):
        f0, flag = args
        if branchmerge:
            repo.dirstate.add(f)
            repo.dirstate.remove(f0)
            repo.dirstate.copy(f0, f)
        else:
            repo.dirstate.normal(f)
            repo.dirstate.drop(f0)

    # directory rename, get
    for f, args, msg in actions.get(ACTION_LOCAL_DIR_RENAME_GET, []):
        f0, flag = args
        if branchmerge:
            repo.dirstate.add(f)
            repo.dirstate.copy(f0, f)
        else:
            repo.dirstate.normal(f)