[PATCH 1 of 5] convert: add support for --includerevs, --excluderevs
Yury Sulsky
yury.sulsky at gmail.com
Sat Nov 26 15:26:46 CST 2011
# HG changeset patch
# User Yury Sulsky <yury.sulsky at gmail.com>
# Date 1322282534 18000
# Branch stable
# Node ID 84c30a6c1760a6302f99daf45a992bf59be1a29a
# Parent 8a7f1722b28e50f8e1690870b0caa38e6a6def59
convert: add support for --includerevs, --excluderevs
diff -r 8a7f1722b28e -r 84c30a6c1760 hgext/convert/__init__.py
--- a/hgext/convert/__init__.py Sat Oct 15 14:31:29 2011 -0500
+++ b/hgext/convert/__init__.py Fri Nov 25 23:42:14 2011 -0500
@@ -306,6 +306,10 @@
_('splice synthesized history into place'), _('FILE')),
('', 'branchmap', '',
_('change branch names while converting'), _('FILE')),
+ ('', 'includerevs', '',
+ _('convert only these revisions'), _('FILE')),
+ ('', 'excluderevs', '',
+ _('exclude these revsisions from the conversion'), _('FILE')),
('', 'branchsort', None, _('try to sort changesets by branches')),
('', 'datesort', None, _('try to sort changesets by date')),
('', 'sourcesort', None, _('preserve source changesets order'))],
diff -r 8a7f1722b28e -r 84c30a6c1760 hgext/convert/convcmd.py
--- a/hgext/convert/convcmd.py Sat Oct 15 14:31:29 2011 -0500
+++ b/hgext/convert/convcmd.py Fri Nov 25 23:42:14 2011 -0500
@@ -15,7 +15,7 @@
from gnuarch import gnuarch_source
from bzr import bzr_source
from p4 import p4_source
-import filemap
+import filtermap
import os, shutil
from mercurial import hg, util, encoding
@@ -430,10 +430,15 @@
if sortmode == 'sourcesort' and not srcc.hasnativeorder():
raise util.Abort(_('--sourcesort is not supported by this data source'))
- fmap = opts.get('filemap')
- if fmap:
- srcc = filemap.filemap_source(ui, srcc, fmap)
- destc.setfilemapmode(True)
+ inrevs = opts.get('includerevs')
+ outrevs = opts.get('excluderevs')
+ filemap = opts.get('filemap')
+ if filemap or inrevs or outrevs:
+ srcc = filtermap.filtermap_source(ui, srcc,
+ inrevs = inrevs,
+ outrevs = outrevs,
+ filemap = filemap)
+ destc.setfilemapmode(filemap is not None)
if not revmapfile:
try:
diff -r 8a7f1722b28e -r 84c30a6c1760 hgext/convert/filemap.py
--- a/hgext/convert/filemap.py Sat Oct 15 14:31:29 2011 -0500
+++ b/hgext/convert/filemap.py Fri Nov 25 23:42:14 2011 -0500
@@ -7,7 +7,6 @@
import shlex
from mercurial.i18n import _
from mercurial import util
-from common import SKIPREV, converter_source
def rpairs(name):
e = len(name)
@@ -105,276 +104,3 @@
def active(self):
return bool(self.include or self.exclude or self.rename)
-
-# This class does two additional things compared to a regular source:
-#
-# - Filter and rename files. This is mostly wrapped by the filemapper
-# class above. We hide the original filename in the revision that is
-# returned by getchanges to be able to find things later in getfile.
-#
-# - Return only revisions that matter for the files we're interested in.
-# This involves rewriting the parents of the original revision to
-# create a graph that is restricted to those revisions.
-#
-# This set of revisions includes not only revisions that directly
-# touch files we're interested in, but also merges that merge two
-# or more interesting revisions.
-
-class filemap_source(converter_source):
- def __init__(self, ui, baseconverter, filemap):
- super(filemap_source, self).__init__(ui)
- self.base = baseconverter
- self.filemapper = filemapper(ui, filemap)
- self.commits = {}
- # if a revision rev has parent p in the original revision graph, then
- # rev will have parent self.parentmap[p] in the restricted graph.
- self.parentmap = {}
- # self.wantedancestors[rev] is the set of all ancestors of rev that
- # are in the restricted graph.
- self.wantedancestors = {}
- self.convertedorder = None
- self._rebuilt = False
- self.origparents = {}
- self.children = {}
- self.seenchildren = {}
-
- def before(self):
- self.base.before()
-
- def after(self):
- self.base.after()
-
- def setrevmap(self, revmap):
- # rebuild our state to make things restartable
- #
- # To avoid calling getcommit for every revision that has already
- # been converted, we rebuild only the parentmap, delaying the
- # rebuild of wantedancestors until we need it (i.e. until a
- # merge).
- #
- # We assume the order argument lists the revisions in
- # topological order, so that we can infer which revisions were
- # wanted by previous runs.
- self._rebuilt = not revmap
- seen = {SKIPREV: SKIPREV}
- dummyset = set()
- converted = []
- for rev in revmap.order:
- mapped = revmap[rev]
- wanted = mapped not in seen
- if wanted:
- seen[mapped] = rev
- self.parentmap[rev] = rev
- else:
- self.parentmap[rev] = seen[mapped]
- self.wantedancestors[rev] = dummyset
- arg = seen[mapped]
- if arg == SKIPREV:
- arg = None
- converted.append((rev, wanted, arg))
- self.convertedorder = converted
- return self.base.setrevmap(revmap)
-
- def rebuild(self):
- if self._rebuilt:
- return True
- self._rebuilt = True
- self.parentmap.clear()
- self.wantedancestors.clear()
- self.seenchildren.clear()
- for rev, wanted, arg in self.convertedorder:
- if rev not in self.origparents:
- self.origparents[rev] = self.getcommit(rev).parents
- if arg is not None:
- self.children[arg] = self.children.get(arg, 0) + 1
-
- for rev, wanted, arg in self.convertedorder:
- parents = self.origparents[rev]
- if wanted:
- self.mark_wanted(rev, parents)
- else:
- self.mark_not_wanted(rev, arg)
- self._discard(arg, *parents)
-
- return True
-
- def getheads(self):
- return self.base.getheads()
-
- def getcommit(self, rev):
- # We want to save a reference to the commit objects to be able
- # to rewrite their parents later on.
- c = self.commits[rev] = self.base.getcommit(rev)
- for p in c.parents:
- self.children[p] = self.children.get(p, 0) + 1
- return c
-
- def _cachedcommit(self, rev):
- if rev in self.commits:
- return self.commits[rev]
- return self.base.getcommit(rev)
-
- def _discard(self, *revs):
- for r in revs:
- if r is None:
- continue
- self.seenchildren[r] = self.seenchildren.get(r, 0) + 1
- if self.seenchildren[r] == self.children[r]:
- del self.wantedancestors[r]
- del self.parentmap[r]
- del self.seenchildren[r]
- if self._rebuilt:
- del self.children[r]
-
- def wanted(self, rev, i):
- # Return True if we're directly interested in rev.
- #
- # i is an index selecting one of the parents of rev (if rev
- # has no parents, i is None). getchangedfiles will give us
- # the list of files that are different in rev and in the parent
- # indicated by i. If we're interested in any of these files,
- # we're interested in rev.
- try:
- files = self.base.getchangedfiles(rev, i)
- except NotImplementedError:
- raise util.Abort(_("source repository doesn't support --filemap"))
- for f in files:
- if self.filemapper(f):
- return True
- return False
-
- def mark_not_wanted(self, rev, p):
- # Mark rev as not interesting and update data structures.
-
- if p is None:
- # A root revision. Use SKIPREV to indicate that it doesn't
- # map to any revision in the restricted graph. Put SKIPREV
- # in the set of wanted ancestors to simplify code elsewhere
- self.parentmap[rev] = SKIPREV
- self.wantedancestors[rev] = set((SKIPREV,))
- return
-
- # Reuse the data from our parent.
- self.parentmap[rev] = self.parentmap[p]
- self.wantedancestors[rev] = self.wantedancestors[p]
-
- def mark_wanted(self, rev, parents):
- # Mark rev ss wanted and update data structures.
-
- # rev will be in the restricted graph, so children of rev in
- # the original graph should still have rev as a parent in the
- # restricted graph.
- self.parentmap[rev] = rev
-
- # The set of wanted ancestors of rev is the union of the sets
- # of wanted ancestors of its parents. Plus rev itself.
- wrev = set()
- for p in parents:
- wrev.update(self.wantedancestors[p])
- wrev.add(rev)
- self.wantedancestors[rev] = wrev
-
- def getchanges(self, rev):
- parents = self.commits[rev].parents
- if len(parents) > 1:
- self.rebuild()
-
- # To decide whether we're interested in rev we:
- #
- # - calculate what parents rev will have if it turns out we're
- # interested in it. If it's going to have more than 1 parent,
- # we're interested in it.
- #
- # - otherwise, we'll compare it with the single parent we found.
- # If any of the files we're interested in is different in the
- # the two revisions, we're interested in rev.
-
- # A parent p is interesting if its mapped version (self.parentmap[p]):
- # - is not SKIPREV
- # - is still not in the list of parents (we don't want duplicates)
- # - is not an ancestor of the mapped versions of the other parents
- mparents = []
- wp = None
- for i, p1 in enumerate(parents):
- mp1 = self.parentmap[p1]
- if mp1 == SKIPREV or mp1 in mparents:
- continue
- for p2 in parents:
- if p1 == p2 or mp1 == self.parentmap[p2]:
- continue
- if mp1 in self.wantedancestors[p2]:
- break
- else:
- mparents.append(mp1)
- wp = i
-
- if wp is None and parents:
- wp = 0
-
- self.origparents[rev] = parents
-
- closed = False
- if 'close' in self.commits[rev].extra:
- # A branch closing revision is only useful if one of its
- # parents belong to the branch being closed
- branch = self.commits[rev].branch
- pbranches = [self._cachedcommit(p).branch for p in mparents]
- if branch in pbranches:
- closed = True
-
- if len(mparents) < 2 and not closed and not self.wanted(rev, wp):
- # We don't want this revision.
- # Update our state and tell the convert process to map this
- # revision to the same revision its parent as mapped to.
- p = None
- if parents:
- p = parents[wp]
- self.mark_not_wanted(rev, p)
- self.convertedorder.append((rev, False, p))
- self._discard(*parents)
- return self.parentmap[rev]
-
- # We want this revision.
- # Rewrite the parents of the commit object
- self.commits[rev].parents = mparents
- self.mark_wanted(rev, parents)
- self.convertedorder.append((rev, True, None))
- self._discard(*parents)
-
- # Get the real changes and do the filtering/mapping. To be
- # able to get the files later on in getfile, we hide the
- # original filename in the rev part of the return value.
- changes, copies = self.base.getchanges(rev)
- newnames = {}
- files = []
- for f, r in changes:
- newf = self.filemapper(f)
- if newf:
- files.append((newf, (f, r)))
- newnames[f] = newf
-
- ncopies = {}
- for c in copies:
- newc = self.filemapper(c)
- if newc:
- newsource = self.filemapper(copies[c])
- if newsource:
- ncopies[newc] = newsource
-
- return files, ncopies
-
- def getfile(self, name, rev):
- realname, realrev = rev
- return self.base.getfile(realname, realrev)
-
- def gettags(self):
- return self.base.gettags()
-
- def hasnativeorder(self):
- return self.base.hasnativeorder()
-
- def lookuprev(self, rev):
- return self.base.lookuprev(rev)
-
- def getbookmarks(self):
- return self.base.getbookmarks()
diff -r 8a7f1722b28e -r 84c30a6c1760 hgext/convert/filtermap.py
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/hgext/convert/filtermap.py Fri Nov 25 23:42:14 2011 -0500
@@ -0,0 +1,340 @@
+# Copyright 2007 Bryan O'Sullivan <bos at serpentine.com>
+# Copyright 2007 Alexis S. L. Carvalho <alexis at cecm.usp.br>
+#
+# This software may be used and distributed according to the terms of the
+# GNU General Public License version 2 or any later version.
+
+from mercurial.i18n import _
+from mercurial import util
+from common import SKIPREV, converter_source
+from filemap import filemapper
+
+# This class serves as the base for filtering and mapping revisions
+# between the source and the sink. A subclass should override the
+# "wanted" function to determine if it cares about a certain revision
+# (see "filtermap_source" below).
+#
+# All such revisions, along with merge points or revisions that close
+# branches are transferred to the converter sink.
+
+class filtermap_base(converter_source):
+
+ def wanted(self, rev, i):
+ """Return True if we want to keep this revision"""
+ return True
+
+ def __init__(self, ui, baseconverter):
+ super(filtermap_base, self).__init__(ui)
+ self.base = baseconverter
+ self.commits = {}
+ # if a revision rev has parent p in the original revision graph, then
+ # rev will have parent self.parentmap[p] in the restricted graph.
+ self.parentmap = {}
+ # self.wantedancestors[rev] is the set of all ancestors of rev that
+ # are in the restricted graph.
+ self.wantedancestors = {}
+ self.convertedorder = None
+ self._rebuilt = False
+ self.origparents = {}
+ self.children = {}
+ self.seenchildren = {}
+
+ def before(self):
+ self.base.before()
+
+ def after(self):
+ self.base.after()
+
+ def setrevmap(self, revmap):
+ # rebuild our state to make things restartable
+ #
+ # To avoid calling getcommit for every revision that has already
+ # been converted, we rebuild only the parentmap, delaying the
+ # rebuild of wantedancestors until we need it (i.e. until a
+ # merge).
+ #
+ # We assume the order argument lists the revisions in
+ # topological order, so that we can infer which revisions were
+ # wanted by previous runs.
+ self._rebuilt = not revmap
+ seen = {SKIPREV: SKIPREV}
+ dummyset = set()
+ converted = []
+ for rev in revmap.order:
+ mapped = revmap[rev]
+ wanted = mapped not in seen
+ if wanted:
+ seen[mapped] = rev
+ self.parentmap[rev] = rev
+ else:
+ self.parentmap[rev] = seen[mapped]
+ self.wantedancestors[rev] = dummyset
+ arg = seen[mapped]
+ if arg == SKIPREV:
+ arg = None
+ converted.append((rev, wanted, arg))
+ self.convertedorder = converted
+ return self.base.setrevmap(revmap)
+
+ def rebuild(self):
+ if self._rebuilt:
+ return True
+ self._rebuilt = True
+ self.parentmap.clear()
+ self.wantedancestors.clear()
+ self.seenchildren.clear()
+ for rev, wanted, arg in self.convertedorder:
+ if rev not in self.origparents:
+ self.origparents[rev] = self.getcommit(rev).parents
+ if arg is not None:
+ self.children[arg] = self.children.get(arg, 0) + 1
+
+ for rev, wanted, arg in self.convertedorder:
+ parents = self.origparents[rev]
+ if wanted:
+ self.mark_wanted(rev, parents)
+ else:
+ self.mark_not_wanted(rev, arg)
+ self._discard(arg, *parents)
+
+ return True
+
+ def getheads(self):
+ return self.base.getheads()
+
+ def getcommit(self, rev):
+ # We want to save a reference to the commit objects to be able
+ # to rewrite their parents later on.
+ c = self.commits[rev] = self.base.getcommit(rev)
+ for p in c.parents:
+ self.children[p] = self.children.get(p, 0) + 1
+ return c
+
+ def _cachedcommit(self, rev):
+ if rev in self.commits:
+ return self.commits[rev]
+ return self.base.getcommit(rev)
+
+ def _discard(self, *revs):
+ for r in revs:
+ if r is None:
+ continue
+ self.seenchildren[r] = self.seenchildren.get(r, 0) + 1
+ if self.seenchildren[r] == self.children[r]:
+ del self.wantedancestors[r]
+ del self.parentmap[r]
+ del self.seenchildren[r]
+ if self._rebuilt:
+ del self.children[r]
+
+ def mark_not_wanted(self, rev, p):
+ # Mark rev as not interesting and update data structures.
+
+ if p is None:
+ # A root revision. Use SKIPREV to indicate that it doesn't
+ # map to any revision in the restricted graph. Put SKIPREV
+ # in the set of wanted ancestors to simplify code elsewhere
+ self.parentmap[rev] = SKIPREV
+ self.wantedancestors[rev] = set((SKIPREV,))
+ return
+
+ # Reuse the data from our parent.
+ self.parentmap[rev] = self.parentmap[p]
+ self.wantedancestors[rev] = self.wantedancestors[p]
+
+ def mark_wanted(self, rev, parents):
+ # Mark rev ss wanted and update data structures.
+
+ # rev will be in the restricted graph, so children of rev in
+ # the original graph should still have rev as a parent in the
+ # restricted graph.
+ self.parentmap[rev] = rev
+
+ # The set of wanted ancestors of rev is the union of the sets
+ # of wanted ancestors of its parents. Plus rev itself.
+ wrev = set()
+ for p in parents:
+ wrev.update(self.wantedancestors[p])
+ wrev.add(rev)
+ self.wantedancestors[rev] = wrev
+
+ def getchanges(self, rev):
+ parents = self.commits[rev].parents
+ if len(parents) > 1:
+ self.rebuild()
+
+ # To decide whether we're interested in rev we:
+ #
+ # - calculate what parents rev will have if it turns out we're
+ # interested in it. If it's going to have more than 1 parent,
+ # we're interested in it.
+ #
+ # - otherwise, we call the "wanted" function to tell us if we're
+ # still interested.
+
+ # A parent p is interesting if its mapped version (self.parentmap[p]):
+ # - is not SKIPREV
+ # - is still not in the list of parents (we don't want duplicates)
+ # - is not an ancestor of the mapped versions of the other parents
+ mparents = []
+ wp = None
+ for i, p1 in enumerate(parents):
+ mp1 = self.parentmap[p1]
+ if mp1 == SKIPREV or mp1 in mparents:
+ continue
+ for p2 in parents:
+ if p1 == p2 or mp1 == self.parentmap[p2]:
+ continue
+ if mp1 in self.wantedancestors[p2]:
+ break
+ else:
+ mparents.append(mp1)
+ wp = i
+
+ if wp is None and parents:
+ wp = 0
+
+ self.origparents[rev] = parents
+
+ closed = False
+ if 'close' in self.commits[rev].extra:
+ # A branch closing revision is only useful if one of its
+ # parents belong to the branch being closed
+ branch = self.commits[rev].branch
+ pbranches = [self._cachedcommit(p).branch for p in mparents]
+ if branch in pbranches:
+ closed = True
+
+ if len(mparents) < 2 and not closed and not self.wanted(rev, wp):
+ # We don't want this revision.
+ # Update our state and tell the convert process to map this
+ # revision to the same revision its parent as mapped to.
+ p = None
+ if parents:
+ p = parents[wp]
+ self.mark_not_wanted(rev, p)
+ self.convertedorder.append((rev, False, p))
+ self._discard(*parents)
+ return self.parentmap[rev]
+
+ # We want this revision.
+ # Rewrite the parents of the commit object
+ self.commits[rev].parents = mparents
+ self.mark_wanted(rev, parents)
+ self.convertedorder.append((rev, True, None))
+ self._discard(*parents)
+ return self.base.getchanges(rev)
+
+ def getfile(self, name, rev):
+ return self.base.getfile(name, rev)
+
+ def gettags(self):
+ return self.base.gettags()
+
+ def hasnativeorder(self):
+ return self.base.hasnativeorder()
+
+ def lookuprev(self, rev):
+ return self.base.lookuprev(rev)
+
+ def getbookmarks(self):
+ return self.base.getbookmarks()
+
+# This class does two additional things compared to a regular source:
+#
+# - Filter and rename files. This is mostly wrapped by the filemapper
+# class above. We hide the original filename in the revision that is
+# returned by getchanges to be able to find things later in getfile.
+#
+# - Return only revisions that matter for the files we're interested in.
+# This involves rewriting the parents of the original revision to
+# create a graph that is restricted to those revisions.
+#
+# This set of revisions includes not only revisions that directly
+# touch files we're interested in, but also merges that merge two
+# or more interesting revisions.
+
+class filtermap_source(filtermap_base):
+ def __init__(self, ui, base,
+ inrevs = None,
+ outrevs = None,
+ filemap = None,
+ modify_file = None):
+ super(filtermap_source, self).__init__(ui, base)
+ self.inrevs = self.read_revset(inrevs) if inrevs else None
+ self.outrevs = self.read_revset(outrevs) if outrevs else None
+ self.filemapper = filemapper(ui, filemap) if filemap else None
+ self.modify_file = modify_file
+
+ def read_revset(self, fname):
+ def parse(line):
+ line = line.strip()
+ ret = self.base.lookuprev(line)
+ if ret is None:
+ self.ui.warn(_("Can't find revision %s." % line))
+ return ret
+ with open(fname, 'r') as fp:
+ return set([parse(line) for line in fp])
+
+ def want_rev(self, rev):
+ if self.inrevs is not None and rev not in self.inrevs:
+ return False
+ if self.outrevs is not None and rev in self.outrevs:
+ return False
+ return True
+
+ def want_files(self, rev, i):
+ if self.filemapper is None:
+ return True
+ try:
+ files = self.base.getchangedfiles(rev, i)
+ except NotImplementedError:
+ raise util.Abort(_("source repository doesn't support --filemap"))
+ for f in files:
+ if self.filemapper(f):
+ return True
+ return False
+
+ def wanted(self, rev, i):
+ return self.want_rev(rev) and self.want_files(rev, i)
+
+ def filemap_changes(self, changes):
+ if self.filemapper is None:
+ return changes
+
+ # Get the real changes and do the filtering/mapping. To be
+ # able to get the files later on in getfile, we hide the
+ # original filename in the rev part of the return value.
+
+ files, copies = changes
+
+ nfiles = []
+ for f, r in files:
+ newf = self.filemapper(f)
+ if newf:
+ nfiles.append((newf, (f, r)))
+
+ ncopies = {}
+ for c in copies:
+ newc = self.filemapper(c)
+ if newc:
+ newsource = self.filemapper(copies[c])
+ if newsource:
+ ncopies[newc] = newsource
+
+ return nfiles, ncopies
+
+ def getchanges(self, rev):
+ changes = super(filtermap_source, self).getchanges(rev)
+ if isinstance(changes, basestring):
+ return changes
+ return self.filemap_changes(changes)
+
+ def getfile(self, name, rev):
+ if self.filemapper is not None:
+ name, rev = rev
+ file = super(filtermap_source, self).getfile(name, rev)
+ if self.modify_file is not None:
+ return self.modify_file(name, file)
+ return file
+
More information about the Mercurial-devel
mailing list