annotate mercurial/copies.py @ 42162:31abb9d713af

copies: don't include copies that are not in source in directory move

I've been working on a rewrite of mergecopies(). I compared the output
of the rewritten version with the current version. I noticed that
between FIREFOX_NIGHTLY_59_END and FIREFOX_BETA_60_BASE in the
mozilla-unified repo, there were many copies that the current version
detected that the rewritten version did not. One example was
js/src/gc/Iteration.h -> js/src/gc/PublicIterators.h. Then I realized
that js/src/gc/Iteration.h doesn't even exist in
FIREFOX_NIGHTLY_59_END.

This patch adds a filtering step for the "fullcopy" dict. It turns out
that that change also affects the test for issue5020 in
test-merge-criss-cross.t. The 'dm' action no longer happens there. At
first I thought that the test case change meant that this patch was
broken, but I think it's actually correct tha the 'dm' action should
not happen there. The result of the bid merge is still the same.

I suspect this filtering is a better solution for the issue than
41f6af50c0d8 (merge: fix crash on criss cross merge with dir move and
delete (issue5020), 2017-01-31). I also suspect that it was broken
just a few months earlier by a005c33d0bd7 (mergecopies: add logic to
process incomplete data, 2016-10-04). Note that bid merge had been
enabled for a few years at that point, since 19903277f035 (merge: use
bid merge by default (BC), 2014-10-01).

This patch is still just a workaround. It will be cleaned up soon
(with the rewrite of mergecopies()). But doing this in a separate
patch makes later patches easier to understand and gives a place to
explain why this is changing.

Differential Revision: https://phab.mercurial-scm.org/D6244

# copies.py - copy detection for Mercurial
#
# Copyright 2008 Matt Mackall <[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 collections
import heapq
import os

from .i18n import _

from . import (
    match as matchmod,
    node,
    pathutil,
    util,
)
from .utils import (
    stringutil,
)

def _findlimit(repo, ctxa, ctxb):
    """
    Find the last revision that needs to be checked to ensure that a full
    transitive closure for file copies can be properly calculated.
    Generally, this means finding the earliest revision number that's an
    ancestor of a or b but not both, except when a or b is a direct descendent
    of the other, in which case we can return the minimum revnum of a and b.
    """

    # basic idea:
    # - mark a and b with different sides
    # - if a parent's children are all on the same side, the parent is
    #   on that side, otherwise it is on no side
    # - walk the graph in topological order with the help of a heap;
    #   - add unseen parents to side map
    #   - clear side of any parent that has children on different sides
    #   - track number of interesting revs that might still be on a side
    #   - track the lowest interesting rev seen
    #   - quit when interesting revs is zero

    cl = repo.changelog
    wdirparents = None
    a = ctxa.rev()
    b = ctxb.rev()
    if a is None:
        wdirparents = (ctxa.p1(), ctxa.p2())
        a = node.wdirrev
    if b is None:
        assert not wdirparents
        wdirparents = (ctxb.p1(), ctxb.p2())
        b = node.wdirrev

    side = {a: -1, b: 1}
    visit = [-a, -b]
    heapq.heapify(visit)
    interesting = len(visit)
    limit = node.wdirrev

    while interesting:
        r = -heapq.heappop(visit)
        if r == node.wdirrev:
            parents = [pctx.rev() for pctx in wdirparents]
        else:
            parents = cl.parentrevs(r)
        if parents[1] == node.nullrev:
            parents = parents[:1]
        for p in parents:
            if p not in side:
                # first time we see p; add it to visit
                side[p] = side[r]
                if side[p]:
                    interesting += 1
                heapq.heappush(visit, -p)
            elif side[p] and side[p] != side[r]:
                # p was interesting but now we know better
                side[p] = 0
                interesting -= 1
        if side[r]:
            limit = r # lowest rev visited
            interesting -= 1

    # Consider the following flow (see test-commit-amend.t under issue4405):
    # 1/ File 'a0' committed
    # 2/ File renamed from 'a0' to 'a1' in a new commit (call it 'a1')
    # 3/ Move back to first commit
    # 4/ Create a new commit via revert to contents of 'a1' (call it 'a1-amend')
    # 5/ Rename file from 'a1' to 'a2' and commit --amend 'a1-msg'
    #
    # During the amend in step five, we will be in this state:
    #
    # @  3 temporary amend commit for a1-amend
    # |
    # o  2 a1-amend
    # |
    # | o  1 a1
    # |/
    # o  0 a0
    #
    # When _findlimit is called, a and b are revs 3 and 0, so limit will be 2,
    # yet the filelog has the copy information in rev 1 and we will not look
    # back far enough unless we also look at the a and b as candidates.
    # This only occurs when a is a descendent of b or visa-versa.
    return min(limit, a, b)

def _chain(src, dst, a, b):
    """chain two sets of copies a->b"""
    t = a.copy()
    for k, v in b.iteritems():
        if v in t:
            # found a chain
            if t[v] != k:
                # file wasn't renamed back to itself
                t[k] = t[v]
            if v not in dst:
                # chain was a rename, not a copy
                del t[v]
        if v in src:
            # file is a copy of an existing file
            t[k] = v

    for k, v in list(t.items()):
        # remove criss-crossed copies
        if k in src and v in dst:
            del t[k]
        # remove copies to files that were then removed
        elif k not in dst:
            del t[k]

    return t

def _tracefile(fctx, am, limit=node.nullrev):
    """return file context that is the ancestor of fctx present in ancestor
    manifest am, stopping after the first ancestor lower than limit"""

    for f in fctx.ancestors():
        if am.get(f.path(), None) == f.filenode():
            return f
        if limit >= 0 and not f.isintroducedafter(limit):
            return None

def _dirstatecopies(repo, match=None):
    ds = repo.dirstate
    c = ds.copies().copy()
    for k in list(c):
        if ds[k] not in 'anm' or (match and not match(k)):
            del c[k]
    return c

def _computeforwardmissing(a, b, match=None):
    """Computes which files are in b but not a.
    This is its own function so extensions can easily wrap this call to see what
    files _forwardcopies is about to process.
    """
    ma = a.manifest()
    mb = b.manifest()
    return mb.filesnotin(ma, match=match)

def usechangesetcentricalgo(repo):
    """Checks if we should use changeset-centric copy algorithms"""
    return (repo.ui.config('experimental', 'copies.read-from') in
            ('changeset-only', 'compatibility'))

def _committedforwardcopies(a, b, match):
    """Like _forwardcopies(), but b.rev() cannot be None (working copy)"""
    # files might have to be traced back to the fctx parent of the last
    # one-side-only changeset, but not further back than that
    repo = a._repo

    if usechangesetcentricalgo(repo):
        return _changesetforwardcopies(a, b, match)

    debug = repo.ui.debugflag and repo.ui.configbool('devel', 'debug.copies')
    dbg = repo.ui.debug
    if debug:
        dbg('debug.copies:    looking into rename from %s to %s\n'
            % (a, b))
    limit = _findlimit(repo, a, b)
    if debug:
        dbg('debug.copies:      search limit: %d\n' % limit)
    am = a.manifest()

    # find where new files came from
    # we currently don't try to find where old files went, too expensive
    # this means we can miss a case like 'hg rm b; hg cp a b'
    cm = {}

    # Computing the forward missing is quite expensive on large manifests, since
    # it compares the entire manifests. We can optimize it in the common use
    # case of computing what copies are in a commit versus its parent (like
    # during a rebase or histedit). Note, we exclude merge commits from this
    # optimization, since the ctx.files() for a merge commit is not correct for
    # this comparison.
    forwardmissingmatch = match
    if b.p1() == a and b.p2().node() == node.nullid:
        filesmatcher = matchmod.exact(b.files())
        forwardmissingmatch = matchmod.intersectmatchers(match, filesmatcher)
    missing = _computeforwardmissing(a, b, match=forwardmissingmatch)

    ancestrycontext = a._repo.changelog.ancestors([b.rev()], inclusive=True)

    if debug:
        dbg('debug.copies:      missing file to search: %d\n' % len(missing))

    for f in missing:
        if debug:
            dbg('debug.copies:        tracing file: %s\n' % f)
        fctx = b[f]
        fctx._ancestrycontext = ancestrycontext

        if debug:
            start = util.timer()
        ofctx = _tracefile(fctx, am, limit)
        if ofctx:
            if debug:
                dbg('debug.copies:          rename of: %s\n' % ofctx._path)
            cm[f] = ofctx.path()
        if debug:
            dbg('debug.copies:          time: %f seconds\n'
                % (util.timer() - start))
    return cm

def _changesetforwardcopies(a, b, match):
    if a.rev() == node.nullrev:
        return {}

    repo = a.repo()
    children = {}
    cl = repo.changelog
    missingrevs = cl.findmissingrevs(common=[a.rev()], heads=[b.rev()])
    for r in missingrevs:
        for p in cl.parentrevs(r):
            if p == node.nullrev:
                continue
            if p not in children:
                children[p] = [r]
            else:
                children[p].append(r)

    roots = set(children) - set(missingrevs)
    # 'work' contains 3-tuples of a (revision number, parent number, copies).
    # The parent number is only used for knowing which parent the copies dict
    # came from.
    work = [(r, 1, {}) for r in roots]
    heapq.heapify(work)
    while work:
        r, i1, copies1 = heapq.heappop(work)
        if work and work[0][0] == r:
            # We are tracing copies from both parents
            r, i2, copies2 = heapq.heappop(work)
            copies = {}
            ctx = repo[r]
            p1man, p2man = ctx.p1().manifest(), ctx.p2().manifest()
            allcopies = set(copies1) | set(copies2)
            # TODO: perhaps this filtering should be done as long as ctx
            # is merge, whether or not we're tracing from both parent.
            for dst in allcopies:
                if not match(dst):
                    continue
                if dst not in copies2:
                    # Copied on p1 side: mark as copy from p1 side if it didn't
                    # already exist on p2 side
                    if dst not in p2man:
                        copies[dst] = copies1[dst]
                elif dst not in copies1:
                    # Copied on p2 side: mark as copy from p2 side if it didn't
                    # already exist on p1 side
                    if dst not in p1man:
                        copies[dst] = copies2[dst]
                else:
                    # Copied on both sides: mark as copy from p1 side
                    copies[dst] = copies1[dst]
        else:
            copies = copies1
        if r == b.rev():
            return copies
        for c in children[r]:
            childctx = repo[c]
            if r == childctx.p1().rev():
                parent = 1
                childcopies = childctx.p1copies()
            else:
                assert r == childctx.p2().rev()
                parent = 2
                childcopies = childctx.p2copies()
            if not match.always():
                childcopies = {dst: src for dst, src in childcopies.items()
                               if match(dst)}
            childcopies = _chain(a, childctx, copies, childcopies)
            heapq.heappush(work, (c, parent, childcopies))
    assert False

def _forwardcopies(a, b, match=None):
    """find {dst@b: src@a} copy mapping where a is an ancestor of b"""

    match = a.repo().narrowmatch(match)
    # check for working copy
    if b.rev() is None:
        if a == b.p1():
            # short-circuit to avoid issues with merge states
            return _dirstatecopies(b._repo, match)

        cm = _committedforwardcopies(a, b.p1(), match)
        # combine copies from dirstate if necessary
        return _chain(a, b, cm, _dirstatecopies(b._repo, match))
    return _committedforwardcopies(a, b, match)

def _backwardrenames(a, b, match):
    if a._repo.ui.config('experimental', 'copytrace') == 'off':
        return {}

    # Even though we're not taking copies into account, 1:n rename situations
    # can still exist (e.g. hg cp a b; hg mv a c). In those cases we
    # arbitrarily pick one of the renames.
    # We don't want to pass in "match" here, since that would filter
    # the destination by it. Since we're reversing the copies, we want
    # to filter the source instead.
    f = _forwardcopies(b, a)
    r = {}
    for k, v in sorted(f.iteritems()):
        if match and not match(v):
            continue
        # remove copies
        if v in a:
            continue
        r[v] = k
    return r

def pathcopies(x, y, match=None):
    """find {dst@y: src@x} copy mapping for directed compare"""
    repo = x._repo
    debug = repo.ui.debugflag and repo.ui.configbool('devel', 'debug.copies')
    if debug:
        repo.ui.debug('debug.copies: searching copies from %s to %s\n'
                      % (x, y))
    if x == y or not x or not y:
        return {}
    a = y.ancestor(x)
    if a == x:
        if debug:
            repo.ui.debug('debug.copies: search mode: forward\n')
        return _forwardcopies(x, y, match=match)
    if a == y:
        if debug:
            repo.ui.debug('debug.copies: search mode: backward\n')
        return _backwardrenames(x, y, match=match)
    if debug:
        repo.ui.debug('debug.copies: search mode: combined\n')
    return _chain(x, y, _backwardrenames(x, a, match=match),
                  _forwardcopies(a, y, match=match))

def _computenonoverlap(repo, c1, c2, addedinm1, addedinm2, baselabel=''):
    """Computes, based on addedinm1 and addedinm2, the files exclusive to c1
    and c2. This is its own function so extensions can easily wrap this call
    to see what files mergecopies is about to process.

    Even though c1 and c2 are not used in this function, they are useful in
    other extensions for being able to read the file nodes of the changed files.

    "baselabel" can be passed to help distinguish the multiple computations
    done in the graft case.
    """
    u1 = sorted(addedinm1 - addedinm2)
    u2 = sorted(addedinm2 - addedinm1)

    header = "  unmatched files in %s"
    if baselabel:
        header += ' (from %s)' % baselabel
    if u1:
        repo.ui.debug("%s:\n   %s\n" % (header % 'local', "\n   ".join(u1)))
    if u2:
        repo.ui.debug("%s:\n   %s\n" % (header % 'other', "\n   ".join(u2)))

    return u1, u2

def _makegetfctx(ctx):
    """return a 'getfctx' function suitable for _checkcopies usage

    We have to re-setup the function building 'filectx' for each
    '_checkcopies' to ensure the linkrev adjustment is properly setup for
    each. Linkrev adjustment is important to avoid bug in rename
    detection. Moreover, having a proper '_ancestrycontext' setup ensures
    the performance impact of this adjustment is kept limited. Without it,
    each file could do a full dag traversal making the time complexity of
    the operation explode (see issue4537).

    This function exists here mostly to limit the impact on stable. Feel
    free to refactor on default.
    """
    rev = ctx.rev()
    repo = ctx._repo
    ac = getattr(ctx, '_ancestrycontext', None)
    if ac is None:
        revs = [rev]
        if rev is None:
            revs = [p.rev() for p in ctx.parents()]
        ac = repo.changelog.ancestors(revs, inclusive=True)
        ctx._ancestrycontext = ac
    def makectx(f, n):
        if n in node.wdirfilenodeids:  # in a working context?
            if ctx.rev() is None:
                return ctx.filectx(f)
            return repo[None][f]
        fctx = repo.filectx(f, fileid=n)
        # setup only needed for filectx not create from a changectx
        fctx._ancestrycontext = ac
        fctx._descendantrev = rev
        return fctx
    return util.lrucachefunc(makectx)

def _combinecopies(copyfrom, copyto, finalcopy, diverge, incompletediverge):
    """combine partial copy paths"""
    remainder = {}
    for f in copyfrom:
        if f in copyto:
            finalcopy[copyto[f]] = copyfrom[f]
            del copyto[f]
    for f in incompletediverge:
        assert f not in diverge
        ic = incompletediverge[f]
        if ic[0] in copyto:
            diverge[f] = [copyto[ic[0]], ic[1]]
        else:
            remainder[f] = ic
    return remainder

def mergecopies(repo, c1, c2, base):
    """
    Finds moves and copies between context c1 and c2 that are relevant for
    merging. 'base' will be used as the merge base.

    Copytracing is used in commands like rebase, merge, unshelve, etc to merge
    files that were moved/ copied in one merge parent and modified in another.
    For example:

    o          ---> 4 another commit
    |
    |   o      ---> 3 commit that modifies a.txt
    |  /
    o /        ---> 2 commit that moves a.txt to b.txt
    |/
    o          ---> 1 merge base

    If we try to rebase revision 3 on revision 4, since there is no a.txt in
    revision 4, and if user have copytrace disabled, we prints the following
    message:

    ```other changed <file> which local deleted```

    Returns five dicts: "copy", "movewithdir", "diverge", "renamedelete" and
    "dirmove".

    "copy" is a mapping from destination name -> source name,
    where source is in c1 and destination is in c2 or vice-versa.

    "movewithdir" is a mapping from source name -> destination name,
    where the file at source present in one context but not the other
    needs to be moved to destination by the merge process, because the
    other context moved the directory it is in.

    "diverge" is a mapping of source name -> list of destination names
    for divergent renames.

    "renamedelete" is a mapping of source name -> list of destination
    names for files deleted in c1 that were renamed in c2 or vice-versa.

    "dirmove" is a mapping of detected source dir -> destination dir renames.
    This is needed for handling changes to new files previously grafted into
    renamed directories.

    This function calls different copytracing algorithms based on config.
    """
    # avoid silly behavior for update from empty dir
    if not c1 or not c2 or c1 == c2:
        return {}, {}, {}, {}, {}

    narrowmatch = c1.repo().narrowmatch()

    # avoid silly behavior for parent -> working dir
    if c2.node() is None and c1.node() == repo.dirstate.p1():
        return _dirstatecopies(repo, narrowmatch), {}, {}, {}, {}

    copytracing = repo.ui.config('experimental', 'copytrace')
    boolctrace = stringutil.parsebool(copytracing)

    # Copy trace disabling is explicitly below the node == p1 logic above
    # because the logic above is required for a simple copy to be kept across a
    # rebase.
    if copytracing == 'heuristics':
        # Do full copytracing if only non-public revisions are involved as
        # that will be fast enough and will also cover the copies which could
        # be missed by heuristics
        if _isfullcopytraceable(repo, c1, base):
            return _fullcopytracing(repo, c1, c2, base)
        return _heuristicscopytracing(repo, c1, c2, base)
    elif boolctrace is False:
        # stringutil.parsebool() returns None when it is unable to parse the
        # value, so we should rely on making sure copytracing is on such cases
        return {}, {}, {}, {}, {}
    else:
        return _fullcopytracing(repo, c1, c2, base)

def _isfullcopytraceable(repo, c1, base):
    """ Checks that if base, source and destination are all no-public branches,
    if yes let's use the full copytrace algorithm for increased capabilities
    since it will be fast enough.

    `experimental.copytrace.sourcecommitlimit` can be used to set a limit for
    number of changesets from c1 to base such that if number of changesets are
    more than the limit, full copytracing algorithm won't be used.
    """
    if c1.rev() is None:
        c1 = c1.p1()
    if c1.mutable() and base.mutable():
        sourcecommitlimit = repo.ui.configint('experimental',
                                              'copytrace.sourcecommitlimit')
        commits = len(repo.revs('%d::%d', base.rev(), c1.rev()))
        return commits < sourcecommitlimit
    return False

def _fullcopytracing(repo, c1, c2, base):
    """ The full copytracing algorithm which finds all the new files that were
    added from merge base up to the top commit and for each file it checks if
    this file was copied from another file.

    This is pretty slow when a lot of changesets are involved but will track all
    the copies.
    """
    # In certain scenarios (e.g. graft, update or rebase), base can be
    # overridden We still need to know a real common ancestor in this case We
    # can't just compute _c1.ancestor(_c2) and compare it to ca, because there
    # can be multiple common ancestors, e.g. in case of bidmerge.  Because our
    # caller may not know if the revision passed in lieu of the CA is a genuine
    # common ancestor or not without explicitly checking it, it's better to
    # determine that here.
    #
    # base.isancestorof(wc) is False, work around that
    _c1 = c1.p1() if c1.rev() is None else c1
    _c2 = c2.p1() if c2.rev() is None else c2
    # an endpoint is "dirty" if it isn't a descendant of the merge base
    # if we have a dirty endpoint, we need to trigger graft logic, and also
    # keep track of which endpoint is dirty
    dirtyc1 = not base.isancestorof(_c1)
    dirtyc2 = not base.isancestorof(_c2)
    graft = dirtyc1 or dirtyc2
    tca = base
    if graft:
        tca = _c1.ancestor(_c2)

    limit = _findlimit(repo, c1, c2)
    repo.ui.debug("  searching for copies back to rev %d\n" % limit)

    m1 = c1.manifest()
    m2 = c2.manifest()
    mb = base.manifest()

    # gather data from _checkcopies:
    # - diverge = record all diverges in this dict
    # - copy = record all non-divergent copies in this dict
    # - fullcopy = record all copies in this dict
    # - incomplete = record non-divergent partial copies here
    # - incompletediverge = record divergent partial copies here
    diverge = {} # divergence data is shared
    incompletediverge  = {}
    data1 = {'copy': {},
             'fullcopy': {},
             'incomplete': {},
             'diverge': diverge,
             'incompletediverge': incompletediverge,
            }
    data2 = {'copy': {},
             'fullcopy': {},
             'incomplete': {},
             'diverge': diverge,
             'incompletediverge': incompletediverge,
            }

    # find interesting file sets from manifests
    addedinm1 = m1.filesnotin(mb, repo.narrowmatch())
    addedinm2 = m2.filesnotin(mb, repo.narrowmatch())
    bothnew = sorted(addedinm1 & addedinm2)
    if tca == base:
        # unmatched file from base
        u1r, u2r = _computenonoverlap(repo, c1, c2, addedinm1, addedinm2)
        u1u, u2u = u1r, u2r
    else:
        # unmatched file from base (DAG rotation in the graft case)
        u1r, u2r = _computenonoverlap(repo, c1, c2, addedinm1, addedinm2,
                                      baselabel='base')
        # unmatched file from topological common ancestors (no DAG rotation)
        # need to recompute this for directory move handling when grafting
        mta = tca.manifest()
        u1u, u2u = _computenonoverlap(repo, c1, c2,
                                      m1.filesnotin(mta, repo.narrowmatch()),
                                      m2.filesnotin(mta, repo.narrowmatch()),
                                      baselabel='topological common ancestor')

    for f in u1u:
        _checkcopies(c1, c2, f, base, tca, dirtyc1, limit, data1)

    for f in u2u:
        _checkcopies(c2, c1, f, base, tca, dirtyc2, limit, data2)

    copy = dict(data1['copy'])
    copy.update(data2['copy'])
    fullcopy = dict(data1['fullcopy'])
    fullcopy.update(data2['fullcopy'])

    if dirtyc1:
        _combinecopies(data2['incomplete'], data1['incomplete'], copy, diverge,
                       incompletediverge)
    if dirtyc2:
        _combinecopies(data1['incomplete'], data2['incomplete'], copy, diverge,
                       incompletediverge)

    renamedelete = {}
    renamedeleteset = set()
    divergeset = set()
    for of, fl in list(diverge.items()):
        if len(fl) == 1 or of in c1 or of in c2:
            del diverge[of] # not actually divergent, or not a rename
            if of not in c1 and of not in c2:
                # renamed on one side, deleted on the other side, but filter
                # out files that have been renamed and then deleted
                renamedelete[of] = [f for f in fl if f in c1 or f in c2]
                renamedeleteset.update(fl) # reverse map for below
        else:
            divergeset.update(fl) # reverse map for below

    if bothnew:
        repo.ui.debug("  unmatched files new in both:\n   %s\n"
                      % "\n   ".join(bothnew))
    bothdiverge = {}
    bothincompletediverge = {}
    remainder = {}
    both1 = {'copy': {},
             'fullcopy': {},
             'incomplete': {},
             'diverge': bothdiverge,
             'incompletediverge': bothincompletediverge
            }
    both2 = {'copy': {},
             'fullcopy': {},
             'incomplete': {},
             'diverge': bothdiverge,
             'incompletediverge': bothincompletediverge
            }
    for f in bothnew:
        _checkcopies(c1, c2, f, base, tca, dirtyc1, limit, both1)
        _checkcopies(c2, c1, f, base, tca, dirtyc2, limit, both2)
    if dirtyc1 and dirtyc2:
        remainder = _combinecopies(both2['incomplete'], both1['incomplete'],
                                   copy, bothdiverge, bothincompletediverge)
        remainder1 = _combinecopies(both1['incomplete'], both2['incomplete'],
                                   copy, bothdiverge, bothincompletediverge)
        remainder.update(remainder1)
    elif dirtyc1:
        # incomplete copies may only be found on the "dirty" side for bothnew
        assert not both2['incomplete']
        remainder = _combinecopies({}, both1['incomplete'], copy, bothdiverge,
                                   bothincompletediverge)
    elif dirtyc2:
        assert not both1['incomplete']
        remainder = _combinecopies({}, both2['incomplete'], copy, bothdiverge,
                                   bothincompletediverge)
    else:
        # incomplete copies and divergences can't happen outside grafts
        assert not both1['incomplete']
        assert not both2['incomplete']
        assert not bothincompletediverge
    for f in remainder:
        assert f not in bothdiverge
        ic = remainder[f]
        if ic[0] in (m1 if dirtyc1 else m2):
            # backed-out rename on one side, but watch out for deleted files
            bothdiverge[f] = ic
    for of, fl in bothdiverge.items():
        if len(fl) == 2 and fl[0] == fl[1]:
            copy[fl[0]] = of # not actually divergent, just matching renames

    # Sometimes we get invalid copies here (the "and not remotebase" in
    # _checkcopies() seems suspicious). Filter them out.
    for dst, src in fullcopy.copy().items():
        if src not in mb:
            del fullcopy[dst]
    if fullcopy and repo.ui.debugflag:
        repo.ui.debug("  all copies found (* = to merge, ! = divergent, "
                      "% = renamed and deleted):\n")
        for f in sorted(fullcopy):
            note = ""
            if f in copy:
                note += "*"
            if f in divergeset:
                note += "!"
            if f in renamedeleteset:
                note += "%"
            repo.ui.debug("   src: '%s' -> dst: '%s' %s\n" % (fullcopy[f], f,
                                                              note))
    del divergeset

    if not fullcopy:
        return copy, {}, diverge, renamedelete, {}

    repo.ui.debug("  checking for directory renames\n")

    # generate a directory move map
    d1, d2 = c1.dirs(), c2.dirs()
    # Hack for adding '', which is not otherwise added, to d1 and d2
    d1.addpath('/')
    d2.addpath('/')
    invalid = set()
    dirmove = {}

    # examine each file copy for a potential directory move, which is
    # when all the files in a directory are moved to a new directory
    for dst, src in fullcopy.iteritems():
        dsrc, ddst = pathutil.dirname(src), pathutil.dirname(dst)
        if dsrc in invalid:
            # already seen to be uninteresting
            continue
        elif dsrc in d1 and ddst in d1:
            # directory wasn't entirely moved locally
            invalid.add(dsrc)
        elif dsrc in d2 and ddst in d2:
            # directory wasn't entirely moved remotely
            invalid.add(dsrc)
        elif dsrc in dirmove and dirmove[dsrc] != ddst:
            # files from the same directory moved to two different places
            invalid.add(dsrc)
        else:
            # looks good so far
            dirmove[dsrc] = ddst

    for i in invalid:
        if i in dirmove:
            del dirmove[i]
    del d1, d2, invalid

    if not dirmove:
        return copy, {}, diverge, renamedelete, {}

    dirmove = {k + "/": v + "/" for k, v in dirmove.iteritems()}

    for d in dirmove:
        repo.ui.debug("   discovered dir src: '%s' -> dst: '%s'\n" %
                      (d, dirmove[d]))

    movewithdir = {}
    # check unaccounted nonoverlapping files against directory moves
    for f in u1r + u2r:
        if f not in fullcopy:
            for d in dirmove:
                if f.startswith(d):
                    # new file added in a directory that was moved, move it
                    df = dirmove[d] + f[len(d):]
                    if df not in copy:
                        movewithdir[f] = df
                        repo.ui.debug(("   pending file src: '%s' -> "
                                       "dst: '%s'\n") % (f, df))
                    break

    return copy, movewithdir, diverge, renamedelete, dirmove

def _heuristicscopytracing(repo, c1, c2, base):
    """ Fast copytracing using filename heuristics

    Assumes that moves or renames are of following two types:

    1) Inside a directory only (same directory name but different filenames)
    2) Move from one directory to another
                    (same filenames but different directory names)

    Works only when there are no merge commits in the "source branch".
    Source branch is commits from base up to c2 not including base.

    If merge is involved it fallbacks to _fullcopytracing().

    Can be used by setting the following config:

        [experimental]
        copytrace = heuristics

    In some cases the copy/move candidates found by heuristics can be very large
    in number and that will make the algorithm slow. The number of possible
    candidates to check can be limited by using the config
    `experimental.copytrace.movecandidateslimit` which defaults to 100.
    """

    if c1.rev() is None:
        c1 = c1.p1()
    if c2.rev() is None:
        c2 = c2.p1()

    copies = {}

    changedfiles = set()
    m1 = c1.manifest()
    if not repo.revs('%d::%d', base.rev(), c2.rev()):
        # If base is not in c2 branch, we switch to fullcopytracing
        repo.ui.debug("switching to full copytracing as base is not "
                      "an ancestor of c2\n")
        return _fullcopytracing(repo, c1, c2, base)

    ctx = c2
    while ctx != base:
        if len(ctx.parents()) == 2:
            # To keep things simple let's not handle merges
            repo.ui.debug("switching to full copytracing because of merges\n")
            return _fullcopytracing(repo, c1, c2, base)
        changedfiles.update(ctx.files())
        ctx = ctx.p1()

    cp = _forwardcopies(base, c2)
    for dst, src in cp.iteritems():
        if src in m1:
            copies[dst] = src

    # file is missing if it isn't present in the destination, but is present in
    # the base and present in the source.
    # Presence in the base is important to exclude added files, presence in the
    # source is important to exclude removed files.
    filt = lambda f: f not in m1 and f in base and f in c2
    missingfiles = [f for f in changedfiles if filt(f)]

    if missingfiles:
        basenametofilename = collections.defaultdict(list)
        dirnametofilename = collections.defaultdict(list)

        for f in m1.filesnotin(base.manifest()):
            basename = os.path.basename(f)
            dirname = os.path.dirname(f)
            basenametofilename[basename].append(f)
            dirnametofilename[dirname].append(f)

        for f in missingfiles:
            basename = os.path.basename(f)
            dirname = os.path.dirname(f)
            samebasename = basenametofilename[basename]
            samedirname = dirnametofilename[dirname]
            movecandidates = samebasename + samedirname
            # f is guaranteed to be present in c2, that's why
            # c2.filectx(f) won't fail
            f2 = c2.filectx(f)
            # we can have a lot of candidates which can slow down the heuristics
            # config value to limit the number of candidates moves to check
            maxcandidates = repo.ui.configint('experimental',
                                              'copytrace.movecandidateslimit')

            if len(movecandidates) > maxcandidates:
                repo.ui.status(_("skipping copytracing for '%s', more "
                                 "candidates than the limit: %d\n")
                               % (f, len(movecandidates)))
                continue

            for candidate in movecandidates:
                f1 = c1.filectx(candidate)
                if _related(f1, f2):
                    # if there are a few related copies then we'll merge
                    # changes into all of them. This matches the behaviour
                    # of upstream copytracing
                    copies[candidate] = f

    return copies, {}, {}, {}, {}

def _related(f1, f2):
    """return True if f1 and f2 filectx have a common ancestor

    Walk back to common ancestor to see if the two files originate
    from the same file. Since workingfilectx's rev() is None it messes
    up the integer comparison logic, hence the pre-step check for
    None (f1 and f2 can only be workingfilectx's initially).
    """

    if f1 == f2:
        return True # a match

    g1, g2 = f1.ancestors(), f2.ancestors()
    try:
        f1r, f2r = f1.linkrev(), f2.linkrev()

        if f1r is None:
            f1 = next(g1)
        if f2r is None:
            f2 = next(g2)

        while True:
            f1r, f2r = f1.linkrev(), f2.linkrev()
            if f1r > f2r:
                f1 = next(g1)
            elif f2r > f1r:
                f2 = next(g2)
            else: # f1 and f2 point to files in the same linkrev
                return f1 == f2 # true if they point to the same file
    except StopIteration:
        return False

def _checkcopies(srcctx, dstctx, f, base, tca, remotebase, limit, data):
    """
    check possible copies of f from msrc to mdst

    srcctx = starting context for f in msrc
    dstctx = destination context for f in mdst
    f = the filename to check (as in msrc)
    base = the changectx used as a merge base
    tca = topological common ancestor for graft-like scenarios
    remotebase = True if base is outside tca::srcctx, False otherwise
    limit = the rev number to not search beyond
    data = dictionary of dictionary to store copy data. (see mergecopies)

    note: limit is only an optimization, and provides no guarantee that
    irrelevant revisions will not be visited
    there is no easy way to make this algorithm stop in a guaranteed way
    once it "goes behind a certain revision".
    """

    msrc = srcctx.manifest()
    mdst = dstctx.manifest()
    mb = base.manifest()
    mta = tca.manifest()
    # Might be true if this call is about finding backward renames,
    # This happens in the case of grafts because the DAG is then rotated.
    # If the file exists in both the base and the source, we are not looking
    # for a rename on the source side, but on the part of the DAG that is
    # traversed backwards.
    #
    # In the case there is both backward and forward renames (before and after
    # the base) this is more complicated as we must detect a divergence.
    # We use 'backwards = False' in that case.
    backwards = not remotebase and base != tca and f in mb
    getsrcfctx = _makegetfctx(srcctx)
    getdstfctx = _makegetfctx(dstctx)

    if msrc[f] == mb.get(f) and not remotebase:
        # Nothing to merge
        return

    of = None
    seen = {f}
    for oc in getsrcfctx(f, msrc[f]).ancestors():
        of = oc.path()
        if of in seen:
            # check limit late - grab last rename before
            if oc.linkrev() < limit:
                break
            continue
        seen.add(of)

        # remember for dir rename detection
        if backwards:
            data['fullcopy'][of] = f # grafting backwards through renames
        else:
            data['fullcopy'][f] = of
        if of not in mdst:
            continue # no match, keep looking
        if mdst[of] == mb.get(of):
            return # no merge needed, quit early
        c2 = getdstfctx(of, mdst[of])
        # c2 might be a plain new file on added on destination side that is
        # unrelated to the droids we are looking for.
        cr = _related(oc, c2)
        if cr and (of == f or of == c2.path()): # non-divergent
            if backwards:
                data['copy'][of] = f
            elif of in mb:
                data['copy'][f] = of
            elif remotebase: # special case: a <- b <- a -> b "ping-pong" rename
                data['copy'][of] = f
                del data['fullcopy'][f]
                data['fullcopy'][of] = f
            else: # divergence w.r.t. graft CA on one side of topological CA
                for sf in seen:
                    if sf in mb:
                        assert sf not in data['diverge']
                        data['diverge'][sf] = [f, of]
                        break
            return

    if of in mta:
        if backwards or remotebase:
            data['incomplete'][of] = f
        else:
            for sf in seen:
                if sf in mb:
                    if tca == base:
                        data['diverge'].setdefault(sf, []).append(f)
                    else:
                        data['incompletediverge'][sf] = [of, f]
                    return

def duplicatecopies(repo, wctx, rev, fromrev, skiprev=None):
    """reproduce copies from fromrev to rev in the dirstate

    If skiprev is specified, it's a revision that should be used to
    filter copy records. Any copies that occur between fromrev and
    skiprev will not be duplicated, even if they appear in the set of
    copies between fromrev and rev.
    """
    exclude = {}
    ctraceconfig = repo.ui.config('experimental', 'copytrace')
    bctrace = stringutil.parsebool(ctraceconfig)
    if (skiprev is not None and
        (ctraceconfig == 'heuristics' or bctrace or bctrace is None)):
        # copytrace='off' skips this line, but not the entire function because
        # the line below is O(size of the repo) during a rebase, while the rest
        # of the function is much faster (and is required for carrying copy
        # metadata across the rebase anyway).
        exclude = pathcopies(repo[fromrev], repo[skiprev])
    for dst, src in pathcopies(repo[fromrev], repo[rev]).iteritems():
        # copies.pathcopies returns backward renames, so dst might not
        # actually be in the dirstate
        if dst in exclude:
            continue
        wctx[dst].markcopied(src)