view hgext/convert/p4.py @ 40811:e13ab4acf555

rust: peek_mut optim for lazy ancestors

This is one of the two optimizations that are also
present in the Python code: replacing pairs of pop/push
on the BinaryHeap by single updates, hence having it
under the hood maintain its consistency (sift) only once.

On Mozilla central, the measured gain (see details below)
is around 7%.

Creating the PeekMut object by calling peek_mut() right away
instead of peek() first is less efficient (gain is only 4%, stats
not included).
Our interpretation is that its creation has a cost which is vasted
in the cases where it ends by droping the value (Peekmut::pop()
just does self.heap.pop() anyway). On the other hand, the immutable
peek() is very fast: it's just taking a reference in the
underlying vector.

The Python version still has another optimization:
if parent(current) == current-1, then the heap doesn't need
to maintain its consistency, since we already know that
it's bigger than all the others in the heap.
Rust's BinaryHeap doesn't allow us to mutate its biggest
element with no housekeeping, but we tried it anyway, with a
copy of the BinaryHeap implementation with a dedicaded added
method: it's not worth the technical debt in our opinion
(we measured only a further 1.6% improvement).
One possible explanation would be that the sift is really fast
anyway in that case, whereas it's not in the case of Python,
because it's at least partly done in slow Python code.
Still it's possible that replacing BinaryHeap by something more
dedicated to discrete ordered types could be faster.

Measurements on mozilla-central:

Three runs of 'hg perfancestors' on the parent changeset:

Moyenne des médianes: 0.100587
! wall 0.100062 comb 0.100000 user 0.100000 sys 0.000000 (best of 98)
! wall 0.135804 comb 0.130000 user 0.130000 sys 0.000000 (max of 98)
! wall 0.102864 comb 0.102755 user 0.099286 sys 0.003469 (avg of 98)
! wall 0.101486 comb 0.110000 user 0.110000 sys 0.000000 (median of 98)

! wall 0.096804 comb 0.090000 user 0.090000 sys 0.000000 (best of 100)
! wall 0.132235 comb 0.130000 user 0.120000 sys 0.010000 (max of 100)
! wall 0.100258 comb 0.100300 user 0.096000 sys 0.004300 (avg of 100)
! wall 0.098384 comb 0.100000 user 0.100000 sys 0.000000 (median of 100)

! wall 0.099925 comb 0.100000 user 0.100000 sys 0.000000 (best of 98)
! wall 0.133518 comb 0.140000 user 0.130000 sys 0.010000 (max of 98)
! wall 0.102381 comb 0.102449 user 0.098265 sys 0.004184 (avg of 98)
! wall 0.101891 comb 0.090000 user 0.090000 sys 0.000000 (median of 98)

Mean of the medians: 0.100587

On the present changeset:

! wall 0.091344 comb 0.090000 user 0.090000 sys 0.000000 (best of 100)
! wall 0.122728 comb 0.120000 user 0.110000 sys 0.010000 (max of 100)
! wall 0.093268 comb 0.093300 user 0.089300 sys 0.004000 (avg of 100)
! wall 0.092567 comb 0.100000 user 0.090000 sys 0.010000 (median of 100)

! wall 0.093294 comb 0.080000 user 0.080000 sys 0.000000 (best of 100)
! wall 0.144887 comb 0.150000 user 0.140000 sys 0.010000 (max of 100)
! wall 0.097708 comb 0.097700 user 0.093400 sys 0.004300 (avg of 100)
! wall 0.094980 comb 0.100000 user 0.090000 sys 0.010000 (median of 100)

! wall 0.091262 comb 0.090000 user 0.080000 sys 0.010000 (best of 100)
! wall 0.123772 comb 0.130000 user 0.120000 sys 0.010000 (max of 100)
! wall 0.093188 comb 0.093200 user 0.089300 sys 0.003900 (avg of 100)
! wall 0.092364 comb 0.100000 user 0.090000 sys 0.010000 (median of 100)

Mean of the medians is 0.0933

Differential Revision: https://phab.mercurial-scm.org/D5358

# Perforce source for convert extension.
#
# Copyright 2009, Frank Kingswood <[e-mail]>
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.
from __future__ import absolute_import

import marshal
import re

from mercurial.i18n import _
from mercurial import (
    error,
    util,
)
from mercurial.utils import (
    dateutil,
    procutil,
    stringutil,
)

from . import common

def loaditer(f):
    "Yield the dictionary objects generated by p4"
    try:
        while True:
            d = marshal.load(f)
            if not d:
                break
            yield d
    except EOFError:
        pass

def decodefilename(filename):
    """Perforce escapes special characters @, #, *, or %
    with %40, %23, %2A, or %25 respectively

    >>> decodefilename(b'portable-net45%252Bnetcore45%252Bwp8%252BMonoAndroid')
    'portable-net45%2Bnetcore45%2Bwp8%2BMonoAndroid'
    >>> decodefilename(b'//Depot/Directory/%2525/%2523/%23%40.%2A')
    '//Depot/Directory/%25/%23/#@.*'
    """
    replacements = [('%2A', '*'), ('%23', '#'), ('%40', '@'), ('%25', '%')]
    for k, v in replacements:
        filename = filename.replace(k, v)
    return filename

class p4_source(common.converter_source):
    def __init__(self, ui, repotype, path, revs=None):
        # avoid import cycle
        from . import convcmd

        super(p4_source, self).__init__(ui, repotype, path, revs=revs)

        if "/" in path and not path.startswith('//'):
            raise common.NoRepo(_('%s does not look like a P4 repository') %
                                path)

        common.checktool('p4', abort=False)

        self.revmap = {}
        self.encoding = self.ui.config('convert', 'p4.encoding',
                                       convcmd.orig_encoding)
        self.re_type = re.compile(
            "([a-z]+)?(text|binary|symlink|apple|resource|unicode|utf\d+)"
            "(\+\w+)?$")
        self.re_keywords = re.compile(
            r"\$(Id|Header|Date|DateTime|Change|File|Revision|Author)"
            r":[^$\n]*\$")
        self.re_keywords_old = re.compile("\$(Id|Header):[^$\n]*\$")

        if revs and len(revs) > 1:
            raise error.Abort(_("p4 source does not support specifying "
                               "multiple revisions"))

    def setrevmap(self, revmap):
        """Sets the parsed revmap dictionary.

        Revmap stores mappings from a source revision to a target revision.
        It is set in convertcmd.convert and provided by the user as a file
        on the commandline.

        Revisions in the map are considered beeing present in the
        repository and ignored during _parse(). This allows for incremental
        imports if a revmap is provided.
        """
        self.revmap = revmap

    def _parse_view(self, path):
        "Read changes affecting the path"
        cmd = 'p4 -G changes -s submitted %s' % procutil.shellquote(path)
        stdout = procutil.popen(cmd, mode='rb')
        p4changes = {}
        for d in loaditer(stdout):
            c = d.get("change", None)
            if c:
                p4changes[c] = True
        return p4changes

    def _parse(self, ui, path):
        "Prepare list of P4 filenames and revisions to import"
        p4changes = {}
        changeset = {}
        files_map = {}
        copies_map = {}
        localname = {}
        depotname = {}
        heads = []

        ui.status(_('reading p4 views\n'))

        # read client spec or view
        if "/" in path:
            p4changes.update(self._parse_view(path))
            if path.startswith("//") and path.endswith("/..."):
                views = {path[:-3]:""}
            else:
                views = {"//": ""}
        else:
            cmd = 'p4 -G client -o %s' % procutil.shellquote(path)
            clientspec = marshal.load(procutil.popen(cmd, mode='rb'))

            views = {}
            for client in clientspec:
                if client.startswith("View"):
                    sview, cview = clientspec[client].split()
                    p4changes.update(self._parse_view(sview))
                    if sview.endswith("...") and cview.endswith("..."):
                        sview = sview[:-3]
                        cview = cview[:-3]
                    cview = cview[2:]
                    cview = cview[cview.find("/") + 1:]
                    views[sview] = cview

        # list of changes that affect our source files
        p4changes = p4changes.keys()
        p4changes.sort(key=int)

        # list with depot pathnames, longest first
        vieworder = views.keys()
        vieworder.sort(key=len, reverse=True)

        # handle revision limiting
        startrev = self.ui.config('convert', 'p4.startrev')

        # now read the full changelists to get the list of file revisions
        ui.status(_('collecting p4 changelists\n'))
        lastid = None
        for change in p4changes:
            if startrev and int(change) < int(startrev):
                continue
            if self.revs and int(change) > int(self.revs[0]):
                continue
            if change in self.revmap:
                # Ignore already present revisions, but set the parent pointer.
                lastid = change
                continue

            if lastid:
                parents = [lastid]
            else:
                parents = []

            d = self._fetch_revision(change)
            c = self._construct_commit(d, parents)

            descarr = c.desc.splitlines(True)
            if len(descarr) > 0:
                shortdesc = descarr[0].rstrip('\r\n')
            else:
                shortdesc = '**empty changelist description**'

            t = '%s %s' % (c.rev, repr(shortdesc)[1:-1])
            ui.status(stringutil.ellipsis(t, 80) + '\n')

            files = []
            copies = {}
            copiedfiles = []
            i = 0
            while ("depotFile%d" % i) in d and ("rev%d" % i) in d:
                oldname = d["depotFile%d" % i]
                filename = None
                for v in vieworder:
                    if oldname.lower().startswith(v.lower()):
                        filename = decodefilename(views[v] + oldname[len(v):])
                        break
                if filename:
                    files.append((filename, d["rev%d" % i]))
                    depotname[filename] = oldname
                    if (d.get("action%d" % i) == "move/add"):
                        copiedfiles.append(filename)
                    localname[oldname] = filename
                i += 1

            # Collect information about copied files
            for filename in copiedfiles:
                oldname = depotname[filename]

                flcmd = 'p4 -G filelog %s' \
                      % procutil.shellquote(oldname)
                flstdout = procutil.popen(flcmd, mode='rb')

                copiedfilename = None
                for d in loaditer(flstdout):
                    copiedoldname = None

                    i = 0
                    while ("change%d" % i) in d:
                        if (d["change%d" % i] == change and
                            d["action%d" % i] == "move/add"):
                            j = 0
                            while ("file%d,%d" % (i, j)) in d:
                                if d["how%d,%d" % (i, j)] == "moved from":
                                    copiedoldname = d["file%d,%d" % (i, j)]
                                    break
                                j += 1
                        i += 1

                    if copiedoldname and copiedoldname in localname:
                        copiedfilename = localname[copiedoldname]
                        break

                if copiedfilename:
                    copies[filename] = copiedfilename
                else:
                    ui.warn(_("cannot find source for copied file: %s@%s\n")
                            % (filename, change))

            changeset[change] = c
            files_map[change] = files
            copies_map[change] = copies
            lastid = change

        if lastid and len(changeset) > 0:
            heads = [lastid]

        return {
            'changeset': changeset,
            'files': files_map,
            'copies': copies_map,
            'heads': heads,
            'depotname': depotname,
        }

    @util.propertycache
    def _parse_once(self):
        return self._parse(self.ui, self.path)

    @util.propertycache
    def copies(self):
        return self._parse_once['copies']

    @util.propertycache
    def files(self):
        return self._parse_once['files']

    @util.propertycache
    def changeset(self):
        return self._parse_once['changeset']

    @util.propertycache
    def heads(self):
        return self._parse_once['heads']

    @util.propertycache
    def depotname(self):
        return self._parse_once['depotname']

    def getheads(self):
        return self.heads

    def getfile(self, name, rev):
        cmd = 'p4 -G print %s' \
            % procutil.shellquote("%s#%s" % (self.depotname[name], rev))

        lasterror = None
        while True:
            stdout = procutil.popen(cmd, mode='rb')

            mode = None
            contents = []
            keywords = None

            for d in loaditer(stdout):
                code = d["code"]
                data = d.get("data")

                if code == "error":
                    # if this is the first time error happened
                    # re-attempt getting the file
                    if not lasterror:
                        lasterror = IOError(d["generic"], data)
                        # this will exit inner-most for-loop
                        break
                    else:
                        raise lasterror

                elif code == "stat":
                    action = d.get("action")
                    if action in ["purge", "delete", "move/delete"]:
                        return None, None
                    p4type = self.re_type.match(d["type"])
                    if p4type:
                        mode = ""
                        flags = ((p4type.group(1) or "")
                               + (p4type.group(3) or ""))
                        if "x" in flags:
                            mode = "x"
                        if p4type.group(2) == "symlink":
                            mode = "l"
                        if "ko" in flags:
                            keywords = self.re_keywords_old
                        elif "k" in flags:
                            keywords = self.re_keywords

                elif code == "text" or code == "binary":
                    contents.append(data)

                lasterror = None

            if not lasterror:
                break

        if mode is None:
            return None, None

        contents = ''.join(contents)

        if keywords:
            contents = keywords.sub("$\\1$", contents)
        if mode == "l" and contents.endswith("\n"):
            contents = contents[:-1]

        return contents, mode

    def getchanges(self, rev, full):
        if full:
            raise error.Abort(_("convert from p4 does not support --full"))
        return self.files[rev], self.copies[rev], set()

    def _construct_commit(self, obj, parents=None):
        """
        Constructs a common.commit object from an unmarshalled
        `p4 describe` output
        """
        desc = self.recode(obj.get("desc", ""))
        date = (int(obj["time"]), 0)     # timezone not set
        if parents is None:
            parents = []

        return common.commit(author=self.recode(obj["user"]),
            date=dateutil.datestr(date, '%Y-%m-%d %H:%M:%S %1%2'),
            parents=parents, desc=desc, branch=None, rev=obj['change'],
            extra={"p4": obj['change'], "convert_revision": obj['change']})

    def _fetch_revision(self, rev):
        """Return an output of `p4 describe` including author, commit date as
        a dictionary."""
        cmd = "p4 -G describe -s %s" % rev
        stdout = procutil.popen(cmd, mode='rb')
        return marshal.load(stdout)

    def getcommit(self, rev):
        if rev in self.changeset:
            return self.changeset[rev]
        elif rev in self.revmap:
            d = self._fetch_revision(rev)
            return self._construct_commit(d, parents=None)
        raise error.Abort(
            _("cannot find %s in the revmap or parsed changesets") % rev)

    def gettags(self):
        return {}

    def getchangedfiles(self, rev, i):
        return sorted([x[0] for x in self.files[rev]])