view mercurial/unionrepo.py @ 23702:c48924787eaa

filectx.parents: enforce changeid of parent to be in own changectx ancestors

Because of the way filenodes are computed, you can have multiple changesets
"introducing" the same file revision. For example, in the changeset graph
below, changeset 2 and 3 both change a file -to- and -from- the same content.

  o 3: content = new
  |
  | o 2: content = new
  |/
  o 1: content = old

In such cases, the file revision is create once, when 2 is added, and just reused
for 3. So the file change in '3' (from "old" to "new)" has no linkrev pointing
to it).  We'll call this situation "linkrev-shadowing". As the linkrev is used for
optimization purposes when walking a file history, the linkrev-shadowing
results in an unexpected jump to another branch during such a walk.. This leads to
multiple bugs with log, annotate and rename detection.

One element to fix such bugs is to ensure that walking the file history sticks on
the same topology as the changeset's history. For this purpose, we extend the
logic in 'basefilectx.parents' so that it always defines the proper changeset
to associate the parent file revision with. This "proper" changeset has to be an
ancestor of the changeset associated with the child file revision.

This logic is performed in the '_adjustlinkrev' function. This function is
given the starting changeset and all the information regarding the parent file
revision. If the linkrev for the file revision is an ancestor of the starting
changeset, the linkrev is valid and will be used. If it is not, we detected a
topological jump caused by linkrev shadowing, we are going to walk the
ancestors of the starting changeset until we find one setting the file to the
revision we are trying to create.

The performance impact appears acceptable:

- We are walking the changelog once for each filelog traversal (as there should
  be no overlap between searches),

- changelog traversal itself is fairly cheap, compared to what is likely going
  to be perform on the result on the filelog traversal,

- We only touch the manifest for ancestors touching the file, And such
  changesets are likely to be the one introducing the file. (except in
  pathological cases involving merge),

- We use manifest diff instead of full manifest unpacking to check manifest
  content, so it does not involve applying multiple diffs in most case.

- linkrev shadowing is not the common case.

Tests for fixed issues in log, annotate and rename detection have been
added.

But this changeset does not solve all problems. It fixes -ancestry-
computation, but if the linkrev-shadowed changesets is the starting one, we'll
still get things wrong. We'll have to fix the bootstrapping of such operations
in a later changeset. Also, the usage of `hg log FILE`  without --follow still
has issues with linkrev pointing to hidden changesets, because it relies on the
`filelog` revset which implement its own traversal logic that is still to be
fixed.

Thanks goes to:
- Matt Mackall: for nudging me in the right direction
- Julien Cristau and Rémi Cardona: for keep telling me linkrev bug were an
  evolution show stopper for 3 years.
- Durham Goode: for finding a new linkrev issue every few weeks
- Mads Kiilerich: for that last rename bug who raise this topic over my
  anoyance limit.

# unionrepo.py - repository class for viewing union of repository changesets
#
# Derived from bundlerepo.py
# Copyright 2006, 2007 Benoit Boissinot <[e-mail]>
# Copyright 2013 Unity Technologies, Mads Kiilerich <[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.

"""Repository class for "in-memory pull" of one local repository to another,
allowing operations like diff and log with revsets.
"""

from node import nullid
from i18n import _
import os
import util, mdiff, cmdutil, scmutil
import localrepo, changelog, manifest, filelog, revlog

class unionrevlog(revlog.revlog):
    def __init__(self, opener, indexfile, revlog2, linkmapper):
        # How it works:
        # To retrieve a revision, we just need to know the node id so we can
        # look it up in revlog2.
        #
        # To differentiate a rev in the second revlog from a rev in the revlog,
        # we check revision against repotiprev.
        opener = scmutil.readonlyvfs(opener)
        revlog.revlog.__init__(self, opener, indexfile)
        self.revlog2 = revlog2

        n = len(self)
        self.repotiprev = n - 1
        self.bundlerevs = set() # used by 'bundle()' revset expression
        for rev2 in self.revlog2:
            rev = self.revlog2.index[rev2]
            # rev numbers - in revlog2, very different from self.rev
            _start, _csize, _rsize, _base, linkrev, p1rev, p2rev, node = rev

            if linkmapper is None: # link is to same revlog
                assert linkrev == rev2 # we never link back
                link = n
            else: # rev must be mapped from repo2 cl to unified cl by linkmapper
                link = linkmapper(linkrev)

            if node in self.nodemap:
                # this happens for the common revlog revisions
                self.bundlerevs.add(self.nodemap[node])
                continue

            p1node = self.revlog2.node(p1rev)
            p2node = self.revlog2.node(p2rev)

            e = (None, None, None, None,
                 link, self.rev(p1node), self.rev(p2node), node)
            self.index.insert(-1, e)
            self.nodemap[node] = n
            self.bundlerevs.add(n)
            n += 1

    def _chunk(self, rev):
        if rev <= self.repotiprev:
            return revlog.revlog._chunk(self, rev)
        return self.revlog2._chunk(self.node(rev))

    def revdiff(self, rev1, rev2):
        """return or calculate a delta between two revisions"""
        if rev1 > self.repotiprev and rev2 > self.repotiprev:
            return self.revlog2.revdiff(
                self.revlog2.rev(self.node(rev1)),
                self.revlog2.rev(self.node(rev2)))
        elif rev1 <= self.repotiprev and rev2 <= self.repotiprev:
            return self.baserevdiff(rev1, rev2)

        return mdiff.textdiff(self.revision(self.node(rev1)),
                              self.revision(self.node(rev2)))

    def revision(self, nodeorrev):
        """return an uncompressed revision of a given node or revision
        number.
        """
        if isinstance(nodeorrev, int):
            rev = nodeorrev
            node = self.node(rev)
        else:
            node = nodeorrev
            rev = self.rev(node)

        if node == nullid:
            return ""

        if rev > self.repotiprev:
            text = self.revlog2.revision(node)
            self._cache = (node, rev, text)
        else:
            text = self.baserevision(rev)
            # already cached
        return text

    def baserevision(self, nodeorrev):
        # Revlog subclasses may override 'revision' method to modify format of
        # content retrieved from revlog. To use unionrevlog with such class one
        # needs to override 'baserevision' and make more specific call here.
        return revlog.revlog.revision(self, nodeorrev)

    def baserevdiff(self, rev1, rev2):
        # Exists for the same purpose as baserevision.
        return revlog.revlog.revdiff(self, rev1, rev2)

    def addrevision(self, text, transaction, link, p1=None, p2=None, d=None):
        raise NotImplementedError
    def addgroup(self, revs, linkmapper, transaction):
        raise NotImplementedError
    def strip(self, rev, minlink):
        raise NotImplementedError
    def checksize(self):
        raise NotImplementedError

class unionchangelog(unionrevlog, changelog.changelog):
    def __init__(self, opener, opener2):
        changelog.changelog.__init__(self, opener)
        linkmapper = None
        changelog2 = changelog.changelog(opener2)
        unionrevlog.__init__(self, opener, self.indexfile, changelog2,
                             linkmapper)

    def baserevision(self, nodeorrev):
        # Although changelog doesn't override 'revision' method, some extensions
        # may replace this class with another that does. Same story with
        # manifest and filelog classes.
        return changelog.changelog.revision(self, nodeorrev)

    def baserevdiff(self, rev1, rev2):
        return changelog.changelog.revdiff(self, rev1, rev2)

class unionmanifest(unionrevlog, manifest.manifest):
    def __init__(self, opener, opener2, linkmapper):
        manifest.manifest.__init__(self, opener)
        manifest2 = manifest.manifest(opener2)
        unionrevlog.__init__(self, opener, self.indexfile, manifest2,
                             linkmapper)

    def baserevision(self, nodeorrev):
        return manifest.manifest.revision(self, nodeorrev)

    def baserevdiff(self, rev1, rev2):
        return manifest.manifest.revdiff(self, rev1, rev2)

class unionfilelog(unionrevlog, filelog.filelog):
    def __init__(self, opener, path, opener2, linkmapper, repo):
        filelog.filelog.__init__(self, opener, path)
        filelog2 = filelog.filelog(opener2, path)
        unionrevlog.__init__(self, opener, self.indexfile, filelog2,
                             linkmapper)
        self._repo = repo

    def baserevision(self, nodeorrev):
        return filelog.filelog.revision(self, nodeorrev)

    def baserevdiff(self, rev1, rev2):
        return filelog.filelog.revdiff(self, rev1, rev2)

    def _file(self, f):
        self._repo.file(f)

class unionpeer(localrepo.localpeer):
    def canpush(self):
        return False

class unionrepository(localrepo.localrepository):
    def __init__(self, ui, path, path2):
        localrepo.localrepository.__init__(self, ui, path)
        self.ui.setconfig('phases', 'publish', False, 'unionrepo')

        self._url = 'union:%s+%s' % (util.expandpath(path),
                                     util.expandpath(path2))
        self.repo2 = localrepo.localrepository(ui, path2)

    @localrepo.unfilteredpropertycache
    def changelog(self):
        return unionchangelog(self.sopener, self.repo2.sopener)

    def _clrev(self, rev2):
        """map from repo2 changelog rev to temporary rev in self.changelog"""
        node = self.repo2.changelog.node(rev2)
        return self.changelog.rev(node)

    @localrepo.unfilteredpropertycache
    def manifest(self):
        return unionmanifest(self.sopener, self.repo2.sopener,
                             self._clrev)

    def url(self):
        return self._url

    def file(self, f):
        return unionfilelog(self.sopener, f, self.repo2.sopener,
                            self._clrev, self)

    def close(self):
        self.repo2.close()

    def cancopy(self):
        return False

    def peer(self):
        return unionpeer(self)

    def getcwd(self):
        return os.getcwd() # always outside the repo

def instance(ui, path, create):
    if create:
        raise util.Abort(_('cannot create new union repository'))
    parentpath = ui.config("bundle", "mainreporoot", "")
    if not parentpath:
        # try to find the correct path to the working directory repo
        parentpath = cmdutil.findrepo(os.getcwd())
        if parentpath is None:
            parentpath = ''
    if parentpath:
        # Try to make the full path relative so we get a nice, short URL.
        # In particular, we don't want temp dir names in test outputs.
        cwd = os.getcwd()
        if parentpath == cwd:
            parentpath = ''
        else:
            cwd = os.path.join(cwd,'')
            if parentpath.startswith(cwd):
                parentpath = parentpath[len(cwd):]
    if path.startswith('union:'):
        s = path.split(":", 1)[1].split("+", 1)
        if len(s) == 1:
            repopath, repopath2 = parentpath, s[0]
        else:
            repopath, repopath2 = s
    else:
        repopath, repopath2 = parentpath, path
    return unionrepository(ui, repopath, repopath2)