largefiles: move "copyalltostore" invocation into "markcommitted"

Before this patch, while "hg convert", largefiles avoids copying
largefiles in the working directory into the store area by combination
of setting "repo._isconverting" in "mercurialsink{before|after}" and
checking it in "copytostoreabsolute".

This avoiding is needed while "hg convert", because converting doesn't
update largefiles in the working directory.

But this implementation is not efficient, because:

  - invocation in "markcommitted" can easily ensure updating
    largefiles in the working directory

    "markcommitted" is invoked only when new revision is committed via
    "commit" of "localrepository" (= with files in the working
    directory). On the other hand, "commitctx" may be invoked directly
    for in-memory committing.

  - committing without updating the working directory (e.g. "import
    --bypass") also needs this kind of avoiding

For efficiency of this kind of avoiding, this patch does:

  - move "copyalltostore" invocation into "markcommitted"
  - remove meaningless procedures below:
    - hooking "mercurialsink{before|after}" to (un)set "repo._isconverting"
    - checking "repo._isconverting" in "copytostoreabsolute"

This patch invokes "copyalltostore" also in "_commitcontext", because
"_commitcontext" expects that largefiles in the working directory are
copied into store area after "commitctx". In this case, the working
directory is used as a kind of temporary area to write largefiles out,
even though converted revisions are committed via "commitctx" (without
updating normal files).

hgext/largefiles/lfcommands.py

@@ -1,578 +1,579 b''
 # Copyright 2009-2010 Gregory P. Ward
 # Copyright 2009-2010 Intelerad Medical Systems Incorporated
 # Copyright 2010-2011 Fog Creek Software
 # Copyright 2010-2011 Unity Technologies
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 '''High-level command function for lfconvert, plus the cmdtable.'''
 
 import os, errno
 import shutil
 
 from mercurial import util, match as match_, hg, node, context, error, \
     cmdutil, scmutil, commands
 from mercurial.i18n import _
 from mercurial.lock import release
 
 import lfutil
 import basestore
 
 # -- Commands ----------------------------------------------------------
 
 cmdtable = {}
 command = cmdutil.command(cmdtable)
 
 @command('lfconvert',
     [('s', 'size', '',
       _('minimum size (MB) for files to be converted as largefiles'), 'SIZE'),
     ('', 'to-normal', False,
      _('convert from a largefiles repo to a normal repo')),
     ],
     _('hg lfconvert SOURCE DEST [FILE ...]'),
     norepo=True,
     inferrepo=True)
 def lfconvert(ui, src, dest, *pats, **opts):
     '''convert a normal repository to a largefiles repository
 
     Convert repository SOURCE to a new repository DEST, identical to
     SOURCE except that certain files will be converted as largefiles:
     specifically, any file that matches any PATTERN *or* whose size is
     above the minimum size threshold is converted as a largefile. The
     size used to determine whether or not to track a file as a
     largefile is the size of the first version of the file. The
     minimum size can be specified either with --size or in
     configuration as ``largefiles.size``.
 
     After running this command you will need to make sure that
     largefiles is enabled anywhere you intend to push the new
     repository.
 
     Use --to-normal to convert largefiles back to normal files; after
     this, the DEST repository can be used without largefiles at all.'''
 
     if opts['to_normal']:
         tolfile = False
     else:
         tolfile = True
         size = lfutil.getminsize(ui, True, opts.get('size'), default=None)
 
     if not hg.islocal(src):
         raise util.Abort(_('%s is not a local Mercurial repo') % src)
     if not hg.islocal(dest):
         raise util.Abort(_('%s is not a local Mercurial repo') % dest)
 
     rsrc = hg.repository(ui, src)
     ui.status(_('initializing destination %s\n') % dest)
     rdst = hg.repository(ui, dest, create=True)
 
     success = False
     dstwlock = dstlock = None
     try:
         # Lock destination to prevent modification while it is converted to.
         # Don't need to lock src because we are just reading from its history
         # which can't change.
         dstwlock = rdst.wlock()
         dstlock = rdst.lock()
 
         # Get a list of all changesets in the source.  The easy way to do this
         # is to simply walk the changelog, using changelog.nodesbetween().
         # Take a look at mercurial/revlog.py:639 for more details.
         # Use a generator instead of a list to decrease memory usage
         ctxs = (rsrc[ctx] for ctx in rsrc.changelog.nodesbetween(None,
             rsrc.heads())[0])
         revmap = {node.nullid: node.nullid}
         if tolfile:
             lfiles = set()
             normalfiles = set()
             if not pats:
                 pats = ui.configlist(lfutil.longname, 'patterns', default=[])
             if pats:
                 matcher = match_.match(rsrc.root, '', list(pats))
             else:
                 matcher = None
 
             lfiletohash = {}
             for ctx in ctxs:
                 ui.progress(_('converting revisions'), ctx.rev(),
                     unit=_('revision'), total=rsrc['tip'].rev())
                 _lfconvert_addchangeset(rsrc, rdst, ctx, revmap,
                     lfiles, normalfiles, matcher, size, lfiletohash)
             ui.progress(_('converting revisions'), None)
 
             if os.path.exists(rdst.wjoin(lfutil.shortname)):
                 shutil.rmtree(rdst.wjoin(lfutil.shortname))
 
             for f in lfiletohash.keys():
                 if os.path.isfile(rdst.wjoin(f)):
                     os.unlink(rdst.wjoin(f))
                 try:
                     os.removedirs(os.path.dirname(rdst.wjoin(f)))
                 except OSError:
                     pass
 
             # If there were any files converted to largefiles, add largefiles
             # to the destination repository's requirements.
             if lfiles:
                 rdst.requirements.add('largefiles')
                 rdst._writerequirements()
         else:
             for ctx in ctxs:
                 ui.progress(_('converting revisions'), ctx.rev(),
                     unit=_('revision'), total=rsrc['tip'].rev())
                 _addchangeset(ui, rsrc, rdst, ctx, revmap)
 
             ui.progress(_('converting revisions'), None)
         success = True
     finally:
         rdst.dirstate.clear()
         release(dstlock, dstwlock)
         if not success:
             # we failed, remove the new directory
             shutil.rmtree(rdst.root)
 
 def _addchangeset(ui, rsrc, rdst, ctx, revmap):
     # Convert src parents to dst parents
     parents = _convertparents(ctx, revmap)
 
     # Generate list of changed files
     files = _getchangedfiles(ctx, parents)
 
     def getfilectx(repo, memctx, f):
         if lfutil.standin(f) in files:
             # if the file isn't in the manifest then it was removed
             # or renamed, raise IOError to indicate this
             try:
                 fctx = ctx.filectx(lfutil.standin(f))
             except error.LookupError:
                 return None
             renamed = fctx.renamed()
             if renamed:
                 renamed = lfutil.splitstandin(renamed[0])
 
             hash = fctx.data().strip()
             path = lfutil.findfile(rsrc, hash)
 
             # If one file is missing, likely all files from this rev are
             if path is None:
                 cachelfiles(ui, rsrc, ctx.node())
                 path = lfutil.findfile(rsrc, hash)
 
                 if path is None:
                     raise util.Abort(
                         _("missing largefile \'%s\' from revision %s")
                          % (f, node.hex(ctx.node())))
 
             data = ''
             fd = None
             try:
                 fd = open(path, 'rb')
                 data = fd.read()
             finally:
                 if fd:
                     fd.close()
             return context.memfilectx(repo, f, data, 'l' in fctx.flags(),
                                       'x' in fctx.flags(), renamed)
         else:
             return _getnormalcontext(repo, ctx, f, revmap)
 
     dstfiles = []
     for file in files:
         if lfutil.isstandin(file):
             dstfiles.append(lfutil.splitstandin(file))
         else:
             dstfiles.append(file)
     # Commit
     _commitcontext(rdst, parents, ctx, dstfiles, getfilectx, revmap)
 
 def _lfconvert_addchangeset(rsrc, rdst, ctx, revmap, lfiles, normalfiles,
         matcher, size, lfiletohash):
     # Convert src parents to dst parents
     parents = _convertparents(ctx, revmap)
 
     # Generate list of changed files
     files = _getchangedfiles(ctx, parents)
 
     dstfiles = []
     for f in files:
         if f not in lfiles and f not in normalfiles:
             islfile = _islfile(f, ctx, matcher, size)
             # If this file was renamed or copied then copy
             # the largefile-ness of its predecessor
             if f in ctx.manifest():
                 fctx = ctx.filectx(f)
                 renamed = fctx.renamed()
                 renamedlfile = renamed and renamed[0] in lfiles
                 islfile |= renamedlfile
                 if 'l' in fctx.flags():
                     if renamedlfile:
                         raise util.Abort(
                             _('renamed/copied largefile %s becomes symlink')
                             % f)
                     islfile = False
             if islfile:
                 lfiles.add(f)
             else:
                 normalfiles.add(f)
 
         if f in lfiles:
             dstfiles.append(lfutil.standin(f))
             # largefile in manifest if it has not been removed/renamed
             if f in ctx.manifest():
                 fctx = ctx.filectx(f)
                 if 'l' in fctx.flags():
                     renamed = fctx.renamed()
                     if renamed and renamed[0] in lfiles:
                         raise util.Abort(_('largefile %s becomes symlink') % f)
 
                 # largefile was modified, update standins
                 m = util.sha1('')
                 m.update(ctx[f].data())
                 hash = m.hexdigest()
                 if f not in lfiletohash or lfiletohash[f] != hash:
                     rdst.wwrite(f, ctx[f].data(), ctx[f].flags())
                     executable = 'x' in ctx[f].flags()
                     lfutil.writestandin(rdst, lfutil.standin(f), hash,
                         executable)
                     lfiletohash[f] = hash
         else:
             # normal file
             dstfiles.append(f)
 
     def getfilectx(repo, memctx, f):
         if lfutil.isstandin(f):
             # if the file isn't in the manifest then it was removed
             # or renamed, raise IOError to indicate this
             srcfname = lfutil.splitstandin(f)
             try:
                 fctx = ctx.filectx(srcfname)
             except error.LookupError:
                 return None
             renamed = fctx.renamed()
             if renamed:
                 # standin is always a largefile because largefile-ness
                 # doesn't change after rename or copy
                 renamed = lfutil.standin(renamed[0])
 
             return context.memfilectx(repo, f, lfiletohash[srcfname] + '\n',
                                       'l' in fctx.flags(), 'x' in fctx.flags(),
                                       renamed)
         else:
             return _getnormalcontext(repo, ctx, f, revmap)
 
     # Commit
     _commitcontext(rdst, parents, ctx, dstfiles, getfilectx, revmap)
 
 def _commitcontext(rdst, parents, ctx, dstfiles, getfilectx, revmap):
     mctx = context.memctx(rdst, parents, ctx.description(), dstfiles,
                           getfilectx, ctx.user(), ctx.date(), ctx.extra())
     ret = rdst.commitctx(mctx)
+    lfutil.copyalltostore(rdst, ret)
     rdst.setparents(ret)
     revmap[ctx.node()] = rdst.changelog.tip()
 
 # Generate list of changed files
 def _getchangedfiles(ctx, parents):
     files = set(ctx.files())
     if node.nullid not in parents:
         mc = ctx.manifest()
         mp1 = ctx.parents()[0].manifest()
         mp2 = ctx.parents()[1].manifest()
         files |= (set(mp1) | set(mp2)) - set(mc)
         for f in mc:
             if mc[f] != mp1.get(f, None) or mc[f] != mp2.get(f, None):
                 files.add(f)
     return files
 
 # Convert src parents to dst parents
 def _convertparents(ctx, revmap):
     parents = []
     for p in ctx.parents():
         parents.append(revmap[p.node()])
     while len(parents) < 2:
         parents.append(node.nullid)
     return parents
 
 # Get memfilectx for a normal file
 def _getnormalcontext(repo, ctx, f, revmap):
     try:
         fctx = ctx.filectx(f)
     except error.LookupError:
         return None
     renamed = fctx.renamed()
     if renamed:
         renamed = renamed[0]
 
     data = fctx.data()
     if f == '.hgtags':
         data = _converttags (repo.ui, revmap, data)
     return context.memfilectx(repo, f, data, 'l' in fctx.flags(),
                               'x' in fctx.flags(), renamed)
 
 # Remap tag data using a revision map
 def _converttags(ui, revmap, data):
     newdata = []
     for line in data.splitlines():
         try:
             id, name = line.split(' ', 1)
         except ValueError:
             ui.warn(_('skipping incorrectly formatted tag %s\n')
                 % line)
             continue
         try:
             newid = node.bin(id)
         except TypeError:
             ui.warn(_('skipping incorrectly formatted id %s\n')
                 % id)
             continue
         try:
             newdata.append('%s %s\n' % (node.hex(revmap[newid]),
                 name))
         except KeyError:
             ui.warn(_('no mapping for id %s\n') % id)
             continue
     return ''.join(newdata)
 
 def _islfile(file, ctx, matcher, size):
     '''Return true if file should be considered a largefile, i.e.
     matcher matches it or it is larger than size.'''
     # never store special .hg* files as largefiles
     if file == '.hgtags' or file == '.hgignore' or file == '.hgsigs':
         return False
     if matcher and matcher(file):
         return True
     try:
         return ctx.filectx(file).size() >= size * 1024 * 1024
     except error.LookupError:
         return False
 
 def uploadlfiles(ui, rsrc, rdst, files):
     '''upload largefiles to the central store'''
 
     if not files:
         return
 
     store = basestore._openstore(rsrc, rdst, put=True)
 
     at = 0
     ui.debug("sending statlfile command for %d largefiles\n" % len(files))
     retval = store.exists(files)
     files = filter(lambda h: not retval[h], files)
     ui.debug("%d largefiles need to be uploaded\n" % len(files))
 
     for hash in files:
         ui.progress(_('uploading largefiles'), at, unit='largefile',
                     total=len(files))
         source = lfutil.findfile(rsrc, hash)
         if not source:
             raise util.Abort(_('largefile %s missing from store'
                                ' (needs to be uploaded)') % hash)
         # XXX check for errors here
         store.put(source, hash)
         at += 1
     ui.progress(_('uploading largefiles'), None)
 
 def verifylfiles(ui, repo, all=False, contents=False):
     '''Verify that every largefile revision in the current changeset
     exists in the central store.  With --contents, also verify that
     the contents of each local largefile file revision are correct (SHA-1 hash
     matches the revision ID).  With --all, check every changeset in
     this repository.'''
     if all:
         # Pass a list to the function rather than an iterator because we know a
         # list will work.
         revs = range(len(repo))
     else:
         revs = ['.']
 
     store = basestore._openstore(repo)
     return store.verify(revs, contents=contents)
 
 def cachelfiles(ui, repo, node, filelist=None):
     '''cachelfiles ensures that all largefiles needed by the specified revision
     are present in the repository's largefile cache.
 
     returns a tuple (cached, missing).  cached is the list of files downloaded
     by this operation; missing is the list of files that were needed but could
     not be found.'''
     lfiles = lfutil.listlfiles(repo, node)
     if filelist:
         lfiles = set(lfiles) & set(filelist)
     toget = []
 
     for lfile in lfiles:
         try:
             expectedhash = repo[node][lfutil.standin(lfile)].data().strip()
         except IOError, err:
             if err.errno == errno.ENOENT:
                 continue # node must be None and standin wasn't found in wctx
             raise
         if not lfutil.findfile(repo, expectedhash):
             toget.append((lfile, expectedhash))
 
     if toget:
         store = basestore._openstore(repo)
         ret = store.get(toget)
         return ret
 
     return ([], [])
 
 def downloadlfiles(ui, repo, rev=None):
     matchfn = scmutil.match(repo[None],
                             [repo.wjoin(lfutil.shortname)], {})
     def prepare(ctx, fns):
         pass
     totalsuccess = 0
     totalmissing = 0
     if rev != []: # walkchangerevs on empty list would return all revs
         for ctx in cmdutil.walkchangerevs(repo, matchfn, {'rev' : rev},
                                           prepare):
             success, missing = cachelfiles(ui, repo, ctx.node())
             totalsuccess += len(success)
             totalmissing += len(missing)
     ui.status(_("%d additional largefiles cached\n") % totalsuccess)
     if totalmissing > 0:
         ui.status(_("%d largefiles failed to download\n") % totalmissing)
     return totalsuccess, totalmissing
 
 def updatelfiles(ui, repo, filelist=None, printmessage=None,
                  normallookup=False):
     '''Update largefiles according to standins in the working directory
 
     If ``printmessage`` is other than ``None``, it means "print (or
     ignore, for false) message forcibly".
     '''
     statuswriter = lfutil.getstatuswriter(ui, repo, printmessage)
     wlock = repo.wlock()
     try:
         lfdirstate = lfutil.openlfdirstate(ui, repo)
         lfiles = set(lfutil.listlfiles(repo)) | set(lfdirstate)
 
         if filelist is not None:
             filelist = set(filelist)
             lfiles = [f for f in lfiles if f in filelist]
 
         update = {}
         updated, removed = 0, 0
         for lfile in lfiles:
             abslfile = repo.wjoin(lfile)
             absstandin = repo.wjoin(lfutil.standin(lfile))
             if os.path.exists(absstandin):
                 if (os.path.exists(absstandin + '.orig') and
                     os.path.exists(abslfile)):
                     shutil.copyfile(abslfile, abslfile + '.orig')
                     util.unlinkpath(absstandin + '.orig')
                 expecthash = lfutil.readstandin(repo, lfile)
                 if (expecthash != '' and
                     (not os.path.exists(abslfile) or
                      expecthash != lfutil.hashfile(abslfile))):
                     if lfile not in repo[None]: # not switched to normal file
                         util.unlinkpath(abslfile, ignoremissing=True)
                     # use normallookup() to allocate an entry in largefiles
                     # dirstate, because lack of it misleads
                     # lfilesrepo.status() into recognition that such cache
                     # missing files are removed.
                     lfdirstate.normallookup(lfile)
                     update[lfile] = expecthash
             else:
                 # Remove lfiles for which the standin is deleted, unless the
                 # lfile is added to the repository again. This happens when a
                 # largefile is converted back to a normal file: the standin
                 # disappears, but a new (normal) file appears as the lfile.
                 if (os.path.exists(abslfile) and
                     repo.dirstate.normalize(lfile) not in repo[None]):
                     util.unlinkpath(abslfile)
                     removed += 1
 
         # largefile processing might be slow and be interrupted - be prepared
         lfdirstate.write()
 
         if lfiles:
             statuswriter(_('getting changed largefiles\n'))
             cachelfiles(ui, repo, None, lfiles)
 
         for lfile in lfiles:
             update1 = 0
 
             expecthash = update.get(lfile)
             if expecthash:
                 if not lfutil.copyfromcache(repo, expecthash, lfile):
                     # failed ... but already removed and set to normallookup
                     continue
                 # Synchronize largefile dirstate to the last modified
                 # time of the file
                 lfdirstate.normal(lfile)
                 update1 = 1
 
             # copy the state of largefile standin from the repository's
             # dirstate to its state in the lfdirstate.
             abslfile = repo.wjoin(lfile)
             absstandin = repo.wjoin(lfutil.standin(lfile))
             if os.path.exists(absstandin):
                 mode = os.stat(absstandin).st_mode
                 if mode != os.stat(abslfile).st_mode:
                     os.chmod(abslfile, mode)
                     update1 = 1
 
             updated += update1
 
             lfutil.synclfdirstate(repo, lfdirstate, lfile, normallookup)
 
         if filelist is not None:
             # If "local largefile" is chosen at file merging, it is
             # not listed in "filelist" (= dirstate syncing is
             # omitted), because the standin file is not changed before and
             # after merging.
             # But the status of such files may have to be changed by
             # merging. For example, locally modified ("M") largefile
             # has to become re-added("A"), if it is "normal" file in
             # the target revision of linear-merging.
             for lfile in lfdirstate:
                 if lfile not in filelist:
                     lfutil.synclfdirstate(repo, lfdirstate, lfile, True)
 
         lfdirstate.write()
         if lfiles:
             statuswriter(_('%d largefiles updated, %d removed\n') % (updated,
                 removed))
     finally:
         wlock.release()
 
 @command('lfpull',
     [('r', 'rev', [], _('pull largefiles for these revisions'))
     ] + commands.remoteopts,
     _('-r REV... [-e CMD] [--remotecmd CMD] [SOURCE]'))
 def lfpull(ui, repo, source="default", **opts):
     """pull largefiles for the specified revisions from the specified source
 
     Pull largefiles that are referenced from local changesets but missing
     locally, pulling from a remote repository to the local cache.
 
     If SOURCE is omitted, the 'default' path will be used.
     See :hg:`help urls` for more information.
 
     .. container:: verbose
 
       Some examples:
 
       - pull largefiles for all branch heads::
 
           hg lfpull -r "head() and not closed()"
 
       - pull largefiles on the default branch::
 
           hg lfpull -r "branch(default)"
     """
     repo.lfpullsource = source
 
     revs = opts.get('rev', [])
     if not revs:
         raise util.Abort(_('no revisions specified'))
     revs = scmutil.revrange(repo, revs)
 
     numcached = 0
     for rev in revs:
         ui.note(_('pulling largefiles for revision %s\n') % rev)
         (cached, missing) = cachelfiles(ui, repo, rev)
         numcached += len(cached)
     ui.status(_("%d largefiles cached\n") % numcached)

hgext/largefiles/lfutil.py

@@ -1,573 +1,576 b''
 # Copyright 2009-2010 Gregory P. Ward
 # Copyright 2009-2010 Intelerad Medical Systems Incorporated
 # Copyright 2010-2011 Fog Creek Software
 # Copyright 2010-2011 Unity Technologies
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 '''largefiles utility code: must not import other modules in this package.'''
 
 import os
 import platform
 import shutil
 import stat
 import copy
 
 from mercurial import dirstate, httpconnection, match as match_, util, scmutil
 from mercurial.i18n import _
 from mercurial import node
 
 shortname = '.hglf'
 shortnameslash = shortname + '/'
 longname = 'largefiles'
 
 
 # -- Private worker functions ------------------------------------------
 
 def getminsize(ui, assumelfiles, opt, default=10):
     lfsize = opt
     if not lfsize and assumelfiles:
         lfsize = ui.config(longname, 'minsize', default=default)
     if lfsize:
         try:
             lfsize = float(lfsize)
         except ValueError:
             raise util.Abort(_('largefiles: size must be number (not %s)\n')
                              % lfsize)
     if lfsize is None:
         raise util.Abort(_('minimum size for largefiles must be specified'))
     return lfsize
 
 def link(src, dest):
     util.makedirs(os.path.dirname(dest))
     try:
         util.oslink(src, dest)
     except OSError:
         # if hardlinks fail, fallback on atomic copy
         dst = util.atomictempfile(dest)
         for chunk in util.filechunkiter(open(src, 'rb')):
             dst.write(chunk)
         dst.close()
         os.chmod(dest, os.stat(src).st_mode)
 
 def usercachepath(ui, hash):
     path = ui.configpath(longname, 'usercache', None)
     if path:
         path = os.path.join(path, hash)
     else:
         if os.name == 'nt':
             appdata = os.getenv('LOCALAPPDATA', os.getenv('APPDATA'))
             if appdata:
                 path = os.path.join(appdata, longname, hash)
         elif platform.system() == 'Darwin':
             home = os.getenv('HOME')
             if home:
                 path = os.path.join(home, 'Library', 'Caches',
                                     longname, hash)
         elif os.name == 'posix':
             path = os.getenv('XDG_CACHE_HOME')
             if path:
                 path = os.path.join(path, longname, hash)
             else:
                 home = os.getenv('HOME')
                 if home:
                     path = os.path.join(home, '.cache', longname, hash)
         else:
             raise util.Abort(_('unknown operating system: %s\n') % os.name)
     return path
 
 def inusercache(ui, hash):
     path = usercachepath(ui, hash)
     return path and os.path.exists(path)
 
 def findfile(repo, hash):
     if instore(repo, hash):
         repo.ui.note(_('found %s in store\n') % hash)
         return storepath(repo, hash)
     elif inusercache(repo.ui, hash):
         repo.ui.note(_('found %s in system cache\n') % hash)
         path = storepath(repo, hash)
         link(usercachepath(repo.ui, hash), path)
         return path
     return None
 
 class largefilesdirstate(dirstate.dirstate):
     def __getitem__(self, key):
         return super(largefilesdirstate, self).__getitem__(unixpath(key))
     def normal(self, f):
         return super(largefilesdirstate, self).normal(unixpath(f))
     def remove(self, f):
         return super(largefilesdirstate, self).remove(unixpath(f))
     def add(self, f):
         return super(largefilesdirstate, self).add(unixpath(f))
     def drop(self, f):
         return super(largefilesdirstate, self).drop(unixpath(f))
     def forget(self, f):
         return super(largefilesdirstate, self).forget(unixpath(f))
     def normallookup(self, f):
         return super(largefilesdirstate, self).normallookup(unixpath(f))
     def _ignore(self, f):
         return False
 
 def openlfdirstate(ui, repo, create=True):
     '''
     Return a dirstate object that tracks largefiles: i.e. its root is
     the repo root, but it is saved in .hg/largefiles/dirstate.
     '''
     lfstoredir = repo.join(longname)
     opener = scmutil.opener(lfstoredir)
     lfdirstate = largefilesdirstate(opener, ui, repo.root,
                                      repo.dirstate._validate)
 
     # If the largefiles dirstate does not exist, populate and create
     # it. This ensures that we create it on the first meaningful
     # largefiles operation in a new clone.
     if create and not os.path.exists(os.path.join(lfstoredir, 'dirstate')):
         matcher = getstandinmatcher(repo)
         standins = repo.dirstate.walk(matcher, [], False, False)
 
         if len(standins) > 0:
             util.makedirs(lfstoredir)
 
         for standin in standins:
             lfile = splitstandin(standin)
             lfdirstate.normallookup(lfile)
     return lfdirstate
 
 def lfdirstatestatus(lfdirstate, repo):
     wctx = repo['.']
     match = match_.always(repo.root, repo.getcwd())
     unsure, s = lfdirstate.status(match, [], False, False, False)
     modified, clean = s.modified, s.clean
     for lfile in unsure:
         try:
             fctx = wctx[standin(lfile)]
         except LookupError:
             fctx = None
         if not fctx or fctx.data().strip() != hashfile(repo.wjoin(lfile)):
             modified.append(lfile)
         else:
             clean.append(lfile)
             lfdirstate.normal(lfile)
     return s
 
 def listlfiles(repo, rev=None, matcher=None):
     '''return a list of largefiles in the working copy or the
     specified changeset'''
 
     if matcher is None:
         matcher = getstandinmatcher(repo)
 
     # ignore unknown files in working directory
     return [splitstandin(f)
             for f in repo[rev].walk(matcher)
             if rev is not None or repo.dirstate[f] != '?']
 
 def instore(repo, hash):
     return os.path.exists(storepath(repo, hash))
 
 def storepath(repo, hash):
     return repo.join(os.path.join(longname, hash))
 
 def copyfromcache(repo, hash, filename):
     '''Copy the specified largefile from the repo or system cache to
     filename in the repository. Return true on success or false if the
     file was not found in either cache (which should not happened:
     this is meant to be called only after ensuring that the needed
     largefile exists in the cache).'''
     path = findfile(repo, hash)
     if path is None:
         return False
     util.makedirs(os.path.dirname(repo.wjoin(filename)))
     # The write may fail before the file is fully written, but we
     # don't use atomic writes in the working copy.
     shutil.copy(path, repo.wjoin(filename))
     return True
 
 def copytostore(repo, rev, file, uploaded=False):
     hash = readstandin(repo, file, rev)
     if instore(repo, hash):
         return
     copytostoreabsolute(repo, repo.wjoin(file), hash)
 
 def copyalltostore(repo, node):
     '''Copy all largefiles in a given revision to the store'''
 
     ctx = repo[node]
     for filename in ctx.files():
         if isstandin(filename) and filename in ctx.manifest():
             realfile = splitstandin(filename)
             copytostore(repo, ctx.node(), realfile)
 
 
 def copytostoreabsolute(repo, file, hash):
     if inusercache(repo.ui, hash):
         link(usercachepath(repo.ui, hash), storepath(repo, hash))
-    elif not getattr(repo, "_isconverting", False):
+    else:
         util.makedirs(os.path.dirname(storepath(repo, hash)))
         dst = util.atomictempfile(storepath(repo, hash),
                                   createmode=repo.store.createmode)
         for chunk in util.filechunkiter(open(file, 'rb')):
             dst.write(chunk)
         dst.close()
         linktousercache(repo, hash)
 
 def linktousercache(repo, hash):
     path = usercachepath(repo.ui, hash)
     if path:
         link(storepath(repo, hash), path)
 
 def getstandinmatcher(repo, pats=[], opts={}):
     '''Return a match object that applies pats to the standin directory'''
     standindir = repo.wjoin(shortname)
     if pats:
         pats = [os.path.join(standindir, pat) for pat in pats]
     else:
         # no patterns: relative to repo root
         pats = [standindir]
     # no warnings about missing files or directories
     match = scmutil.match(repo[None], pats, opts)
     match.bad = lambda f, msg: None
     return match
 
 def composestandinmatcher(repo, rmatcher):
     '''Return a matcher that accepts standins corresponding to the
     files accepted by rmatcher. Pass the list of files in the matcher
     as the paths specified by the user.'''
     smatcher = getstandinmatcher(repo, rmatcher.files())
     isstandin = smatcher.matchfn
     def composedmatchfn(f):
         return isstandin(f) and rmatcher.matchfn(splitstandin(f))
     smatcher.matchfn = composedmatchfn
 
     return smatcher
 
 def standin(filename):
     '''Return the repo-relative path to the standin for the specified big
     file.'''
     # Notes:
     # 1) Some callers want an absolute path, but for instance addlargefiles
     #    needs it repo-relative so it can be passed to repo[None].add().  So
     #    leave it up to the caller to use repo.wjoin() to get an absolute path.
     # 2) Join with '/' because that's what dirstate always uses, even on
     #    Windows. Change existing separator to '/' first in case we are
     #    passed filenames from an external source (like the command line).
     return shortnameslash + util.pconvert(filename)
 
 def isstandin(filename):
     '''Return true if filename is a big file standin. filename must be
     in Mercurial's internal form (slash-separated).'''
     return filename.startswith(shortnameslash)
 
 def splitstandin(filename):
     # Split on / because that's what dirstate always uses, even on Windows.
     # Change local separator to / first just in case we are passed filenames
     # from an external source (like the command line).
     bits = util.pconvert(filename).split('/', 1)
     if len(bits) == 2 and bits[0] == shortname:
         return bits[1]
     else:
         return None
 
 def updatestandin(repo, standin):
     file = repo.wjoin(splitstandin(standin))
     if os.path.exists(file):
         hash = hashfile(file)
         executable = getexecutable(file)
         writestandin(repo, standin, hash, executable)
 
 def readstandin(repo, filename, node=None):
     '''read hex hash from standin for filename at given node, or working
     directory if no node is given'''
     return repo[node][standin(filename)].data().strip()
 
 def writestandin(repo, standin, hash, executable):
     '''write hash to <repo.root>/<standin>'''
     repo.wwrite(standin, hash + '\n', executable and 'x' or '')
 
 def copyandhash(instream, outfile):
     '''Read bytes from instream (iterable) and write them to outfile,
     computing the SHA-1 hash of the data along the way. Return the hash.'''
     hasher = util.sha1('')
     for data in instream:
         hasher.update(data)
         outfile.write(data)
     return hasher.hexdigest()
 
 def hashrepofile(repo, file):
     return hashfile(repo.wjoin(file))
 
 def hashfile(file):
     if not os.path.exists(file):
         return ''
     hasher = util.sha1('')
     fd = open(file, 'rb')
     for data in util.filechunkiter(fd, 128 * 1024):
         hasher.update(data)
     fd.close()
     return hasher.hexdigest()
 
 def getexecutable(filename):
     mode = os.stat(filename).st_mode
     return ((mode & stat.S_IXUSR) and
             (mode & stat.S_IXGRP) and
             (mode & stat.S_IXOTH))
 
 def urljoin(first, second, *arg):
     def join(left, right):
         if not left.endswith('/'):
             left += '/'
         if right.startswith('/'):
             right = right[1:]
         return left + right
 
     url = join(first, second)
     for a in arg:
         url = join(url, a)
     return url
 
 def hexsha1(data):
     """hexsha1 returns the hex-encoded sha1 sum of the data in the file-like
     object data"""
     h = util.sha1()
     for chunk in util.filechunkiter(data):
         h.update(chunk)
     return h.hexdigest()
 
 def httpsendfile(ui, filename):
     return httpconnection.httpsendfile(ui, filename, 'rb')
 
 def unixpath(path):
     '''Return a version of path normalized for use with the lfdirstate.'''
     return util.pconvert(os.path.normpath(path))
 
 def islfilesrepo(repo):
     if ('largefiles' in repo.requirements and
             util.any(shortnameslash in f[0] for f in repo.store.datafiles())):
         return True
 
     return util.any(openlfdirstate(repo.ui, repo, False))
 
 class storeprotonotcapable(Exception):
     def __init__(self, storetypes):
         self.storetypes = storetypes
 
 def getstandinsstate(repo):
     standins = []
     matcher = getstandinmatcher(repo)
     for standin in repo.dirstate.walk(matcher, [], False, False):
         lfile = splitstandin(standin)
         try:
             hash = readstandin(repo, lfile)
         except IOError:
             hash = None
         standins.append((lfile, hash))
     return standins
 
 def synclfdirstate(repo, lfdirstate, lfile, normallookup):
     lfstandin = standin(lfile)
     if lfstandin in repo.dirstate:
         stat = repo.dirstate._map[lfstandin]
         state, mtime = stat[0], stat[3]
     else:
         state, mtime = '?', -1
     if state == 'n':
         if normallookup or mtime < 0:
             # state 'n' doesn't ensure 'clean' in this case
             lfdirstate.normallookup(lfile)
         else:
             lfdirstate.normal(lfile)
     elif state == 'm':
         lfdirstate.normallookup(lfile)
     elif state == 'r':
         lfdirstate.remove(lfile)
     elif state == 'a':
         lfdirstate.add(lfile)
     elif state == '?':
         lfdirstate.drop(lfile)
 
 def markcommitted(orig, ctx, node):
     repo = ctx._repo
 
     orig(node)
 
     # ATTENTION: "ctx.files()" may differ from "repo[node].files()"
     # because files coming from the 2nd parent are omitted in the latter.
     #
     # The former should be used to get targets of "synclfdirstate",
     # because such files:
     # - are marked as "a" by "patch.patch()" (e.g. via transplant), and
     # - have to be marked as "n" after commit, but
     # - aren't listed in "repo[node].files()"
 
     lfdirstate = openlfdirstate(repo.ui, repo)
     for f in ctx.files():
         if isstandin(f):
             lfile = splitstandin(f)
             synclfdirstate(repo, lfdirstate, lfile, False)
     lfdirstate.write()
 
+    # As part of committing, copy all of the largefiles into the cache.
+    copyalltostore(repo, node)
+
 def getlfilestoupdate(oldstandins, newstandins):
     changedstandins = set(oldstandins).symmetric_difference(set(newstandins))
     filelist = []
     for f in changedstandins:
         if f[0] not in filelist:
             filelist.append(f[0])
     return filelist
 
 def getlfilestoupload(repo, missing, addfunc):
     for n in missing:
         parents = [p for p in repo.changelog.parents(n) if p != node.nullid]
         ctx = repo[n]
         files = set(ctx.files())
         if len(parents) == 2:
             mc = ctx.manifest()
             mp1 = ctx.parents()[0].manifest()
             mp2 = ctx.parents()[1].manifest()
             for f in mp1:
                 if f not in mc:
                     files.add(f)
             for f in mp2:
                 if f not in mc:
                     files.add(f)
             for f in mc:
                 if mc[f] != mp1.get(f, None) or mc[f] != mp2.get(f, None):
                     files.add(f)
         for fn in files:
             if isstandin(fn) and fn in ctx:
                 addfunc(fn, ctx[fn].data().strip())
 
 def updatestandinsbymatch(repo, match):
     '''Update standins in the working directory according to specified match
 
     This returns (possibly modified) ``match`` object to be used for
     subsequent commit process.
     '''
 
     ui = repo.ui
 
     # Case 1: user calls commit with no specific files or
     # include/exclude patterns: refresh and commit all files that
     # are "dirty".
     if match is None or match.always():
         # Spend a bit of time here to get a list of files we know
         # are modified so we can compare only against those.
         # It can cost a lot of time (several seconds)
         # otherwise to update all standins if the largefiles are
         # large.
         lfdirstate = openlfdirstate(ui, repo)
         dirtymatch = match_.always(repo.root, repo.getcwd())
         unsure, s = lfdirstate.status(dirtymatch, [], False, False,
                                       False)
         modifiedfiles = unsure + s.modified + s.added + s.removed
         lfiles = listlfiles(repo)
         # this only loops through largefiles that exist (not
         # removed/renamed)
         for lfile in lfiles:
             if lfile in modifiedfiles:
                 if os.path.exists(
                         repo.wjoin(standin(lfile))):
                     # this handles the case where a rebase is being
                     # performed and the working copy is not updated
                     # yet.
                     if os.path.exists(repo.wjoin(lfile)):
                         updatestandin(repo,
                             standin(lfile))
 
         return match
 
     lfiles = listlfiles(repo)
     match._files = repo._subdirlfs(match.files(), lfiles)
 
     # Case 2: user calls commit with specified patterns: refresh
     # any matching big files.
     smatcher = composestandinmatcher(repo, match)
     standins = repo.dirstate.walk(smatcher, [], False, False)
 
     # No matching big files: get out of the way and pass control to
     # the usual commit() method.
     if not standins:
         return match
 
     # Refresh all matching big files.  It's possible that the
     # commit will end up failing, in which case the big files will
     # stay refreshed.  No harm done: the user modified them and
     # asked to commit them, so sooner or later we're going to
     # refresh the standins.  Might as well leave them refreshed.
     lfdirstate = openlfdirstate(ui, repo)
     for fstandin in standins:
         lfile = splitstandin(fstandin)
         if lfdirstate[lfile] != 'r':
             updatestandin(repo, fstandin)
 
     # Cook up a new matcher that only matches regular files or
     # standins corresponding to the big files requested by the
     # user.  Have to modify _files to prevent commit() from
     # complaining "not tracked" for big files.
     match = copy.copy(match)
     origmatchfn = match.matchfn
 
     # Check both the list of largefiles and the list of
     # standins because if a largefile was removed, it
     # won't be in the list of largefiles at this point
     match._files += sorted(standins)
 
     actualfiles = []
     for f in match._files:
         fstandin = standin(f)
 
         # ignore known largefiles and standins
         if f in lfiles or fstandin in standins:
             continue
 
         actualfiles.append(f)
     match._files = actualfiles
 
     def matchfn(f):
         if origmatchfn(f):
             return f not in lfiles
         else:
             return f in standins
 
     match.matchfn = matchfn
 
     return match
 
 class automatedcommithook(object):
     '''Statefull hook to update standins at the 1st commit of resuming
 
     For efficiency, updating standins in the working directory should
     be avoided while automated committing (like rebase, transplant and
     so on), because they should be updated before committing.
 
     But the 1st commit of resuming automated committing (e.g. ``rebase
     --continue``) should update them, because largefiles may be
     modified manually.
     '''
     def __init__(self, resuming):
         self.resuming = resuming
 
     def __call__(self, repo, match):
         if self.resuming:
             self.resuming = False # avoids updating at subsequent commits
             return updatestandinsbymatch(repo, match)
         else:
             return match
 
 def getstatuswriter(ui, repo, forcibly=None):
     '''Return the function to write largefiles specific status out
 
     If ``forcibly`` is ``None``, this returns the last element of
     ``repo._lfupdatereporters`` as "default" writer function.
 
     Otherwise, this returns the function to always write out (or
     ignore if ``not forcibly``) status.
     '''
     if forcibly is None:
         return repo._lfstatuswriters[-1]
     else:
         if forcibly:
             return ui.status # forcibly WRITE OUT
         else:
             return lambda *msg, **opts: None # forcibly IGNORE

hgext/largefiles/overrides.py

@@ -1,1295 +1,1287 b''
 # Copyright 2009-2010 Gregory P. Ward
 # Copyright 2009-2010 Intelerad Medical Systems Incorporated
 # Copyright 2010-2011 Fog Creek Software
 # Copyright 2010-2011 Unity Technologies
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 '''Overridden Mercurial commands and functions for the largefiles extension'''
 
 import os
 import copy
 
 from mercurial import hg, util, cmdutil, scmutil, match as match_, \
         archival, pathutil, revset
 from mercurial.i18n import _
 from mercurial.node import hex
 
 import lfutil
 import lfcommands
 import basestore
 
 # -- Utility functions: commonly/repeatedly needed functionality ---------------
 
 def installnormalfilesmatchfn(manifest):
     '''installmatchfn with a matchfn that ignores all largefiles'''
     def overridematch(ctx, pats=[], opts={}, globbed=False,
             default='relpath'):
         match = oldmatch(ctx, pats, opts, globbed, default)
         m = copy.copy(match)
         notlfile = lambda f: not (lfutil.isstandin(f) or lfutil.standin(f) in
                 manifest)
         m._files = filter(notlfile, m._files)
         m._fmap = set(m._files)
         m._always = False
         origmatchfn = m.matchfn
         m.matchfn = lambda f: notlfile(f) and origmatchfn(f)
         return m
     oldmatch = installmatchfn(overridematch)
 
 def installmatchfn(f):
     '''monkey patch the scmutil module with a custom match function.
     Warning: it is monkey patching the _module_ on runtime! Not thread safe!'''
     oldmatch = scmutil.match
     setattr(f, 'oldmatch', oldmatch)
     scmutil.match = f
     return oldmatch
 
 def restorematchfn():
     '''restores scmutil.match to what it was before installmatchfn
     was called.  no-op if scmutil.match is its original function.
 
     Note that n calls to installmatchfn will require n calls to
     restore matchfn to reverse'''
     scmutil.match = getattr(scmutil.match, 'oldmatch')
 
 def installmatchandpatsfn(f):
     oldmatchandpats = scmutil.matchandpats
     setattr(f, 'oldmatchandpats', oldmatchandpats)
     scmutil.matchandpats = f
     return oldmatchandpats
 
 def restorematchandpatsfn():
     '''restores scmutil.matchandpats to what it was before
     installmatchandpatsfn was called. No-op if scmutil.matchandpats
     is its original function.
 
     Note that n calls to installmatchandpatsfn will require n calls
     to restore matchfn to reverse'''
     scmutil.matchandpats = getattr(scmutil.matchandpats, 'oldmatchandpats',
             scmutil.matchandpats)
 
 def addlargefiles(ui, repo, *pats, **opts):
     large = opts.pop('large', None)
     lfsize = lfutil.getminsize(
         ui, lfutil.islfilesrepo(repo), opts.pop('lfsize', None))
 
     lfmatcher = None
     if lfutil.islfilesrepo(repo):
         lfpats = ui.configlist(lfutil.longname, 'patterns', default=[])
         if lfpats:
             lfmatcher = match_.match(repo.root, '', list(lfpats))
 
     lfnames = []
     m = scmutil.match(repo[None], pats, opts)
     m.bad = lambda x, y: None
     wctx = repo[None]
     for f in repo.walk(m):
         exact = m.exact(f)
         lfile = lfutil.standin(f) in wctx
         nfile = f in wctx
         exists = lfile or nfile
 
         # Don't warn the user when they attempt to add a normal tracked file.
         # The normal add code will do that for us.
         if exact and exists:
             if lfile:
                 ui.warn(_('%s already a largefile\n') % f)
             continue
 
         if (exact or not exists) and not lfutil.isstandin(f):
             wfile = repo.wjoin(f)
 
             # In case the file was removed previously, but not committed
             # (issue3507)
             if not os.path.exists(wfile):
                 continue
 
             abovemin = (lfsize and
                         os.lstat(wfile).st_size >= lfsize * 1024 * 1024)
             if large or abovemin or (lfmatcher and lfmatcher(f)):
                 lfnames.append(f)
                 if ui.verbose or not exact:
                     ui.status(_('adding %s as a largefile\n') % m.rel(f))
 
     bad = []
 
     # Need to lock, otherwise there could be a race condition between
     # when standins are created and added to the repo.
     wlock = repo.wlock()
     try:
         if not opts.get('dry_run'):
             standins = []
             lfdirstate = lfutil.openlfdirstate(ui, repo)
             for f in lfnames:
                 standinname = lfutil.standin(f)
                 lfutil.writestandin(repo, standinname, hash='',
                     executable=lfutil.getexecutable(repo.wjoin(f)))
                 standins.append(standinname)
                 if lfdirstate[f] == 'r':
                     lfdirstate.normallookup(f)
                 else:
                     lfdirstate.add(f)
             lfdirstate.write()
             bad += [lfutil.splitstandin(f)
                     for f in repo[None].add(standins)
                     if f in m.files()]
     finally:
         wlock.release()
     return bad
 
 def removelargefiles(ui, repo, isaddremove, *pats, **opts):
     after = opts.get('after')
     if not pats and not after:
         raise util.Abort(_('no files specified'))
     m = scmutil.match(repo[None], pats, opts)
     try:
         repo.lfstatus = True
         s = repo.status(match=m, clean=True)
     finally:
         repo.lfstatus = False
     manifest = repo[None].manifest()
     modified, added, deleted, clean = [[f for f in list
                                         if lfutil.standin(f) in manifest]
                                        for list in (s.modified, s.added,
                                                     s.deleted, s.clean)]
 
     def warn(files, msg):
         for f in files:
             ui.warn(msg % m.rel(f))
         return int(len(files) > 0)
 
     result = 0
 
     if after:
         remove = deleted
         result = warn(modified + added + clean,
                       _('not removing %s: file still exists\n'))
     else:
         remove = deleted + clean
         result = warn(modified, _('not removing %s: file is modified (use -f'
                                   ' to force removal)\n'))
         result = warn(added, _('not removing %s: file has been marked for add'
                                ' (use forget to undo)\n')) or result
 
     for f in sorted(remove):
         if ui.verbose or not m.exact(f):
             ui.status(_('removing %s\n') % m.rel(f))
 
     # Need to lock because standin files are deleted then removed from the
     # repository and we could race in-between.
     wlock = repo.wlock()
     try:
         lfdirstate = lfutil.openlfdirstate(ui, repo)
         for f in remove:
             if not after:
                 # If this is being called by addremove, notify the user that we
                 # are removing the file.
                 if isaddremove:
                     ui.status(_('removing %s\n') % f)
                 util.unlinkpath(repo.wjoin(f), ignoremissing=True)
             lfdirstate.remove(f)
         lfdirstate.write()
         remove = [lfutil.standin(f) for f in remove]
         # If this is being called by addremove, let the original addremove
         # function handle this.
         if not isaddremove:
             for f in remove:
                 util.unlinkpath(repo.wjoin(f), ignoremissing=True)
         repo[None].forget(remove)
     finally:
         wlock.release()
 
     return result
 
 # For overriding mercurial.hgweb.webcommands so that largefiles will
 # appear at their right place in the manifests.
 def decodepath(orig, path):
     return lfutil.splitstandin(path) or path
 
 # -- Wrappers: modify existing commands --------------------------------
 
 # Add works by going through the files that the user wanted to add and
 # checking if they should be added as largefiles. Then it makes a new
 # matcher which matches only the normal files and runs the original
 # version of add.
 def overrideadd(orig, ui, repo, *pats, **opts):
     normal = opts.pop('normal')
     if normal:
         if opts.get('large'):
             raise util.Abort(_('--normal cannot be used with --large'))
         return orig(ui, repo, *pats, **opts)
     bad = addlargefiles(ui, repo, *pats, **opts)
     installnormalfilesmatchfn(repo[None].manifest())
     result = orig(ui, repo, *pats, **opts)
     restorematchfn()
 
     return (result == 1 or bad) and 1 or 0
 
 def overrideremove(orig, ui, repo, *pats, **opts):
     installnormalfilesmatchfn(repo[None].manifest())
     result = orig(ui, repo, *pats, **opts)
     restorematchfn()
     return removelargefiles(ui, repo, False, *pats, **opts) or result
 
 def overridestatusfn(orig, repo, rev2, **opts):
     try:
         repo._repo.lfstatus = True
         return orig(repo, rev2, **opts)
     finally:
         repo._repo.lfstatus = False
 
 def overridestatus(orig, ui, repo, *pats, **opts):
     try:
         repo.lfstatus = True
         return orig(ui, repo, *pats, **opts)
     finally:
         repo.lfstatus = False
 
 def overridedirty(orig, repo, ignoreupdate=False):
     try:
         repo._repo.lfstatus = True
         return orig(repo, ignoreupdate)
     finally:
         repo._repo.lfstatus = False
 
 def overridelog(orig, ui, repo, *pats, **opts):
     def overridematchandpats(ctx, pats=[], opts={}, globbed=False,
             default='relpath'):
         """Matcher that merges root directory with .hglf, suitable for log.
         It is still possible to match .hglf directly.
         For any listed files run log on the standin too.
         matchfn tries both the given filename and with .hglf stripped.
         """
         matchandpats = oldmatchandpats(ctx, pats, opts, globbed, default)
         m, p = copy.copy(matchandpats)
 
         if m.always():
             # We want to match everything anyway, so there's no benefit trying
             # to add standins.
             return matchandpats
 
         pats = set(p)
         # TODO: handling of patterns in both cases below
         if m._cwd:
             if os.path.isabs(m._cwd):
                 # TODO: handle largefile magic when invoked from other cwd
                 return matchandpats
             back = (m._cwd.count('/') + 1) * '../'
             pats.update(back + lfutil.standin(m._cwd + '/' + f) for f in p)
         else:
             pats.update(lfutil.standin(f) for f in p)
 
         for i in range(0, len(m._files)):
             standin = lfutil.standin(m._files[i])
             if standin in repo[ctx.node()]:
                 m._files[i] = standin
             elif m._files[i] not in repo[ctx.node()]:
                 m._files.append(standin)
             pats.add(standin)
 
         m._fmap = set(m._files)
         m._always = False
         origmatchfn = m.matchfn
         def lfmatchfn(f):
             lf = lfutil.splitstandin(f)
             if lf is not None and origmatchfn(lf):
                 return True
             r = origmatchfn(f)
             return r
         m.matchfn = lfmatchfn
 
         return m, pats
 
     # For hg log --patch, the match object is used in two different senses:
     # (1) to determine what revisions should be printed out, and
     # (2) to determine what files to print out diffs for.
     # The magic matchandpats override should be used for case (1) but not for
     # case (2).
     def overridemakelogfilematcher(repo, pats, opts):
         pctx = repo[None]
         match, pats = oldmatchandpats(pctx, pats, opts)
         return lambda rev: match
 
     oldmatchandpats = installmatchandpatsfn(overridematchandpats)
     oldmakelogfilematcher = cmdutil._makenofollowlogfilematcher
     setattr(cmdutil, '_makenofollowlogfilematcher', overridemakelogfilematcher)
 
     try:
         return orig(ui, repo, *pats, **opts)
     finally:
         restorematchandpatsfn()
         setattr(cmdutil, '_makenofollowlogfilematcher', oldmakelogfilematcher)
 
 def overrideverify(orig, ui, repo, *pats, **opts):
     large = opts.pop('large', False)
     all = opts.pop('lfa', False)
     contents = opts.pop('lfc', False)
 
     result = orig(ui, repo, *pats, **opts)
     if large or all or contents:
         result = result or lfcommands.verifylfiles(ui, repo, all, contents)
     return result
 
 def overridedebugstate(orig, ui, repo, *pats, **opts):
     large = opts.pop('large', False)
     if large:
         class fakerepo(object):
             dirstate = lfutil.openlfdirstate(ui, repo)
         orig(ui, fakerepo, *pats, **opts)
     else:
         orig(ui, repo, *pats, **opts)
 
 # Override needs to refresh standins so that update's normal merge
 # will go through properly. Then the other update hook (overriding repo.update)
 # will get the new files. Filemerge is also overridden so that the merge
 # will merge standins correctly.
 def overrideupdate(orig, ui, repo, *pats, **opts):
     # Need to lock between the standins getting updated and their
     # largefiles getting updated
     wlock = repo.wlock()
     try:
         if opts['check']:
             lfdirstate = lfutil.openlfdirstate(ui, repo)
             unsure, s = lfdirstate.status(
                 match_.always(repo.root, repo.getcwd()),
                 [], False, False, False)
 
             mod = len(s.modified) > 0
             for lfile in unsure:
                 standin = lfutil.standin(lfile)
                 if repo['.'][standin].data().strip() != \
                         lfutil.hashfile(repo.wjoin(lfile)):
                     mod = True
                 else:
                     lfdirstate.normal(lfile)
             lfdirstate.write()
             if mod:
                 raise util.Abort(_('uncommitted changes'))
         return orig(ui, repo, *pats, **opts)
     finally:
         wlock.release()
 
 # Before starting the manifest merge, merge.updates will call
 # _checkunknown to check if there are any files in the merged-in
 # changeset that collide with unknown files in the working copy.
 #
 # The largefiles are seen as unknown, so this prevents us from merging
 # in a file 'foo' if we already have a largefile with the same name.
 #
 # The overridden function filters the unknown files by removing any
 # largefiles. This makes the merge proceed and we can then handle this
 # case further in the overridden manifestmerge function below.
 def overridecheckunknownfile(origfn, repo, wctx, mctx, f):
     if lfutil.standin(repo.dirstate.normalize(f)) in wctx:
         return False
     return origfn(repo, wctx, mctx, f)
 
 # The manifest merge handles conflicts on the manifest level. We want
 # to handle changes in largefile-ness of files at this level too.
 #
 # The strategy is to run the original manifestmerge and then process
 # the action list it outputs. There are two cases we need to deal with:
 #
 # 1. Normal file in p1, largefile in p2. Here the largefile is
 #    detected via its standin file, which will enter the working copy
 #    with a "get" action. It is not "merge" since the standin is all
 #    Mercurial is concerned with at this level -- the link to the
 #    existing normal file is not relevant here.
 #
 # 2. Largefile in p1, normal file in p2. Here we get a "merge" action
 #    since the largefile will be present in the working copy and
 #    different from the normal file in p2. Mercurial therefore
 #    triggers a merge action.
 #
 # In both cases, we prompt the user and emit new actions to either
 # remove the standin (if the normal file was kept) or to remove the
 # normal file and get the standin (if the largefile was kept). The
 # default prompt answer is to use the largefile version since it was
 # presumably changed on purpose.
 #
 # Finally, the merge.applyupdates function will then take care of
 # writing the files into the working copy and lfcommands.updatelfiles
 # will update the largefiles.
 def overridecalculateupdates(origfn, repo, p1, p2, pas, branchmerge, force,
                              partial, acceptremote, followcopies):
     overwrite = force and not branchmerge
     actions = origfn(repo, p1, p2, pas, branchmerge, force, partial,
                      acceptremote, followcopies)
 
     if overwrite:
         return actions
 
     removes = set(a[0] for a in actions['r'])
 
     newglist = []
     lfmr = [] # LargeFiles: Mark as Removed
     for action in actions['g']:
         f, args, msg = action
         splitstandin = f and lfutil.splitstandin(f)
         if (splitstandin is not None and
             splitstandin in p1 and splitstandin not in removes):
             # Case 1: normal file in the working copy, largefile in
             # the second parent
             lfile = splitstandin
             standin = f
             msg = _('remote turned local normal file %s into a largefile\n'
                     'use (l)argefile or keep (n)ormal file?'
                     '$$ &Largefile $$ &Normal file') % lfile
             if repo.ui.promptchoice(msg, 0) == 0:
                 actions['r'].append((lfile, None, msg))
                 newglist.append((standin, (p2.flags(standin),), msg))
             else:
                 actions['r'].append((standin, None, msg))
         elif lfutil.standin(f) in p1 and lfutil.standin(f) not in removes:
             # Case 2: largefile in the working copy, normal file in
             # the second parent
             standin = lfutil.standin(f)
             lfile = f
             msg = _('remote turned local largefile %s into a normal file\n'
                     'keep (l)argefile or use (n)ormal file?'
                     '$$ &Largefile $$ &Normal file') % lfile
             if repo.ui.promptchoice(msg, 0) == 0:
                 if branchmerge:
                     # largefile can be restored from standin safely
                     actions['r'].append((lfile, None, msg))
                 else:
                     # "lfile" should be marked as "removed" without
                     # removal of itself
                     lfmr.append((lfile, None, msg))
 
                     # linear-merge should treat this largefile as 're-added'
                     actions['a'].append((standin, None, msg))
             else:
                 actions['r'].append((standin, None, msg))
                 newglist.append((lfile, (p2.flags(lfile),), msg))
         else:
             newglist.append(action)
 
     newglist.sort()
     actions['g'] = newglist
     if lfmr:
         lfmr.sort()
         actions['lfmr'] = lfmr
 
     return actions
 
 def mergerecordupdates(orig, repo, actions, branchmerge):
     if 'lfmr' in actions:
         # this should be executed before 'orig', to execute 'remove'
         # before all other actions
         for lfile, args, msg in actions['lfmr']:
             repo.dirstate.remove(lfile)
 
     return orig(repo, actions, branchmerge)
 
 
 # Override filemerge to prompt the user about how they wish to merge
 # largefiles. This will handle identical edits without prompting the user.
 def overridefilemerge(origfn, repo, mynode, orig, fcd, fco, fca, labels=None):
     if not lfutil.isstandin(orig):
         return origfn(repo, mynode, orig, fcd, fco, fca, labels=labels)
 
     ahash = fca.data().strip().lower()
     dhash = fcd.data().strip().lower()
     ohash = fco.data().strip().lower()
     if (ohash != ahash and
         ohash != dhash and
         (dhash == ahash or
          repo.ui.promptchoice(
              _('largefile %s has a merge conflict\nancestor was %s\n'
                'keep (l)ocal %s or\ntake (o)ther %s?'
                '$$ &Local $$ &Other') %
                (lfutil.splitstandin(orig), ahash, dhash, ohash),
              0) == 1)):
         repo.wwrite(fcd.path(), fco.data(), fco.flags())
     return 0
 
 # Copy first changes the matchers to match standins instead of
 # largefiles.  Then it overrides util.copyfile in that function it
 # checks if the destination largefile already exists. It also keeps a
 # list of copied files so that the largefiles can be copied and the
 # dirstate updated.
 def overridecopy(orig, ui, repo, pats, opts, rename=False):
     # doesn't remove largefile on rename
     if len(pats) < 2:
         # this isn't legal, let the original function deal with it
         return orig(ui, repo, pats, opts, rename)
 
     def makestandin(relpath):
         path = pathutil.canonpath(repo.root, repo.getcwd(), relpath)
         return os.path.join(repo.wjoin(lfutil.standin(path)))
 
     fullpats = scmutil.expandpats(pats)
     dest = fullpats[-1]
 
     if os.path.isdir(dest):
         if not os.path.isdir(makestandin(dest)):
             os.makedirs(makestandin(dest))
     # This could copy both lfiles and normal files in one command,
     # but we don't want to do that. First replace their matcher to
     # only match normal files and run it, then replace it to just
     # match largefiles and run it again.
     nonormalfiles = False
     nolfiles = False
     installnormalfilesmatchfn(repo[None].manifest())
     try:
         try:
             result = orig(ui, repo, pats, opts, rename)
         except util.Abort, e:
             if str(e) != _('no files to copy'):
                 raise e
             else:
                 nonormalfiles = True
             result = 0
     finally:
         restorematchfn()
 
     # The first rename can cause our current working directory to be removed.
     # In that case there is nothing left to copy/rename so just quit.
     try:
         repo.getcwd()
     except OSError:
         return result
 
     try:
         try:
             # When we call orig below it creates the standins but we don't add
             # them to the dir state until later so lock during that time.
             wlock = repo.wlock()
 
             manifest = repo[None].manifest()
             def overridematch(ctx, pats=[], opts={}, globbed=False,
                     default='relpath'):
                 newpats = []
                 # The patterns were previously mangled to add the standin
                 # directory; we need to remove that now
                 for pat in pats:
                     if match_.patkind(pat) is None and lfutil.shortname in pat:
                         newpats.append(pat.replace(lfutil.shortname, ''))
                     else:
                         newpats.append(pat)
                 match = oldmatch(ctx, newpats, opts, globbed, default)
                 m = copy.copy(match)
                 lfile = lambda f: lfutil.standin(f) in manifest
                 m._files = [lfutil.standin(f) for f in m._files if lfile(f)]
                 m._fmap = set(m._files)
                 origmatchfn = m.matchfn
                 m.matchfn = lambda f: (lfutil.isstandin(f) and
                                     (f in manifest) and
                                     origmatchfn(lfutil.splitstandin(f)) or
                                     None)
                 return m
             oldmatch = installmatchfn(overridematch)
             listpats = []
             for pat in pats:
                 if match_.patkind(pat) is not None:
                     listpats.append(pat)
                 else:
                     listpats.append(makestandin(pat))
 
             try:
                 origcopyfile = util.copyfile
                 copiedfiles = []
                 def overridecopyfile(src, dest):
                     if (lfutil.shortname in src and
                         dest.startswith(repo.wjoin(lfutil.shortname))):
                         destlfile = dest.replace(lfutil.shortname, '')
                         if not opts['force'] and os.path.exists(destlfile):
                             raise IOError('',
                                 _('destination largefile already exists'))
                     copiedfiles.append((src, dest))
                     origcopyfile(src, dest)
 
                 util.copyfile = overridecopyfile
                 result += orig(ui, repo, listpats, opts, rename)
             finally:
                 util.copyfile = origcopyfile
 
             lfdirstate = lfutil.openlfdirstate(ui, repo)
             for (src, dest) in copiedfiles:
                 if (lfutil.shortname in src and
                     dest.startswith(repo.wjoin(lfutil.shortname))):
                     srclfile = src.replace(repo.wjoin(lfutil.standin('')), '')
                     destlfile = dest.replace(repo.wjoin(lfutil.standin('')), '')
                     destlfiledir = os.path.dirname(repo.wjoin(destlfile)) or '.'
                     if not os.path.isdir(destlfiledir):
                         os.makedirs(destlfiledir)
                     if rename:
                         os.rename(repo.wjoin(srclfile), repo.wjoin(destlfile))
 
                         # The file is gone, but this deletes any empty parent
                         # directories as a side-effect.
                         util.unlinkpath(repo.wjoin(srclfile), True)
                         lfdirstate.remove(srclfile)
                     else:
                         util.copyfile(repo.wjoin(srclfile),
                                       repo.wjoin(destlfile))
 
                     lfdirstate.add(destlfile)
             lfdirstate.write()
         except util.Abort, e:
             if str(e) != _('no files to copy'):
                 raise e
             else:
                 nolfiles = True
     finally:
         restorematchfn()
         wlock.release()
 
     if nolfiles and nonormalfiles:
         raise util.Abort(_('no files to copy'))
 
     return result
 
 # When the user calls revert, we have to be careful to not revert any
 # changes to other largefiles accidentally. This means we have to keep
 # track of the largefiles that are being reverted so we only pull down
 # the necessary largefiles.
 #
 # Standins are only updated (to match the hash of largefiles) before
 # commits. Update the standins then run the original revert, changing
 # the matcher to hit standins instead of largefiles. Based on the
 # resulting standins update the largefiles.
 def overriderevert(orig, ui, repo, *pats, **opts):
     # Because we put the standins in a bad state (by updating them)
     # and then return them to a correct state we need to lock to
     # prevent others from changing them in their incorrect state.
     wlock = repo.wlock()
     try:
         lfdirstate = lfutil.openlfdirstate(ui, repo)
         s = lfutil.lfdirstatestatus(lfdirstate, repo)
         lfdirstate.write()
         for lfile in s.modified:
             lfutil.updatestandin(repo, lfutil.standin(lfile))
         for lfile in s.deleted:
             if (os.path.exists(repo.wjoin(lfutil.standin(lfile)))):
                 os.unlink(repo.wjoin(lfutil.standin(lfile)))
 
         oldstandins = lfutil.getstandinsstate(repo)
 
         def overridematch(ctx, pats=[], opts={}, globbed=False,
                 default='relpath'):
             match = oldmatch(ctx, pats, opts, globbed, default)
             m = copy.copy(match)
             def tostandin(f):
                 if lfutil.standin(f) in ctx:
                     return lfutil.standin(f)
                 elif lfutil.standin(f) in repo[None]:
                     return None
                 return f
             m._files = [tostandin(f) for f in m._files]
             m._files = [f for f in m._files if f is not None]
             m._fmap = set(m._files)
             origmatchfn = m.matchfn
             def matchfn(f):
                 if lfutil.isstandin(f):
                     return (origmatchfn(lfutil.splitstandin(f)) and
                             (f in repo[None] or f in ctx))
                 return origmatchfn(f)
             m.matchfn = matchfn
             return m
         oldmatch = installmatchfn(overridematch)
         try:
             orig(ui, repo, *pats, **opts)
         finally:
             restorematchfn()
 
         newstandins = lfutil.getstandinsstate(repo)
         filelist = lfutil.getlfilestoupdate(oldstandins, newstandins)
         # lfdirstate should be 'normallookup'-ed for updated files,
         # because reverting doesn't touch dirstate for 'normal' files
         # when target revision is explicitly specified: in such case,
         # 'n' and valid timestamp in dirstate doesn't ensure 'clean'
         # of target (standin) file.
         lfcommands.updatelfiles(ui, repo, filelist, printmessage=False,
                                 normallookup=True)
 
     finally:
         wlock.release()
 
 # after pulling changesets, we need to take some extra care to get
 # largefiles updated remotely
 def overridepull(orig, ui, repo, source=None, **opts):
     revsprepull = len(repo)
     if not source:
         source = 'default'
     repo.lfpullsource = source
     result = orig(ui, repo, source, **opts)
     revspostpull = len(repo)
     lfrevs = opts.get('lfrev', [])
     if opts.get('all_largefiles'):
         lfrevs.append('pulled()')
     if lfrevs and revspostpull > revsprepull:
         numcached = 0
         repo.firstpulled = revsprepull # for pulled() revset expression
         try:
             for rev in scmutil.revrange(repo, lfrevs):
                 ui.note(_('pulling largefiles for revision %s\n') % rev)
                 (cached, missing) = lfcommands.cachelfiles(ui, repo, rev)
                 numcached += len(cached)
         finally:
             del repo.firstpulled
         ui.status(_("%d largefiles cached\n") % numcached)
     return result
 
 def pulledrevsetsymbol(repo, subset, x):
     """``pulled()``
     Changesets that just has been pulled.
 
     Only available with largefiles from pull --lfrev expressions.
 
     .. container:: verbose
 
       Some examples:
 
       - pull largefiles for all new changesets::
 
           hg pull -lfrev "pulled()"
 
       - pull largefiles for all new branch heads::
 
           hg pull -lfrev "head(pulled()) and not closed()"
 
     """
 
     try:
         firstpulled = repo.firstpulled
     except AttributeError:
         raise util.Abort(_("pulled() only available in --lfrev"))
     return revset.baseset([r for r in subset if r >= firstpulled])
 
 def overrideclone(orig, ui, source, dest=None, **opts):
     d = dest
     if d is None:
         d = hg.defaultdest(source)
     if opts.get('all_largefiles') and not hg.islocal(d):
             raise util.Abort(_(
             '--all-largefiles is incompatible with non-local destination %s') %
             d)
 
     return orig(ui, source, dest, **opts)
 
 def hgclone(orig, ui, opts, *args, **kwargs):
     result = orig(ui, opts, *args, **kwargs)
 
     if result is not None:
         sourcerepo, destrepo = result
         repo = destrepo.local()
 
         # Caching is implicitly limited to 'rev' option, since the dest repo was
         # truncated at that point.  The user may expect a download count with
         # this option, so attempt whether or not this is a largefile repo.
         if opts.get('all_largefiles'):
             success, missing = lfcommands.downloadlfiles(ui, repo, None)
 
             if missing != 0:
                 return None
 
     return result
 
 def overriderebase(orig, ui, repo, **opts):
     resuming = opts.get('continue')
     repo._lfcommithooks.append(lfutil.automatedcommithook(resuming))
     repo._lfstatuswriters.append(lambda *msg, **opts: None)
     try:
         return orig(ui, repo, **opts)
     finally:
         repo._lfstatuswriters.pop()
         repo._lfcommithooks.pop()
 
 def overridearchive(orig, repo, dest, node, kind, decode=True, matchfn=None,
             prefix=None, mtime=None, subrepos=None):
     # No need to lock because we are only reading history and
     # largefile caches, neither of which are modified.
     lfcommands.cachelfiles(repo.ui, repo, node)
 
     if kind not in archival.archivers:
         raise util.Abort(_("unknown archive type '%s'") % kind)
 
     ctx = repo[node]
 
     if kind == 'files':
         if prefix:
             raise util.Abort(
                 _('cannot give prefix when archiving to files'))
     else:
         prefix = archival.tidyprefix(dest, kind, prefix)
 
     def write(name, mode, islink, getdata):
         if matchfn and not matchfn(name):
             return
         data = getdata()
         if decode:
             data = repo.wwritedata(name, data)
         archiver.addfile(prefix + name, mode, islink, data)
 
     archiver = archival.archivers[kind](dest, mtime or ctx.date()[0])
 
     if repo.ui.configbool("ui", "archivemeta", True):
         def metadata():
             base = 'repo: %s\nnode: %s\nbranch: %s\n' % (
                 hex(repo.changelog.node(0)), hex(node), ctx.branch())
 
             tags = ''.join('tag: %s\n' % t for t in ctx.tags()
                            if repo.tagtype(t) == 'global')
             if not tags:
                 repo.ui.pushbuffer()
                 opts = {'template': '{latesttag}\n{latesttagdistance}',
                         'style': '', 'patch': None, 'git': None}
                 cmdutil.show_changeset(repo.ui, repo, opts).show(ctx)
                 ltags, dist = repo.ui.popbuffer().split('\n')
                 tags = ''.join('latesttag: %s\n' % t for t in ltags.split(':'))
                 tags += 'latesttagdistance: %s\n' % dist
 
             return base + tags
 
         write('.hg_archival.txt', 0644, False, metadata)
 
     for f in ctx:
         ff = ctx.flags(f)
         getdata = ctx[f].data
         if lfutil.isstandin(f):
             path = lfutil.findfile(repo, getdata().strip())
             if path is None:
                 raise util.Abort(
                     _('largefile %s not found in repo store or system cache')
                     % lfutil.splitstandin(f))
             f = lfutil.splitstandin(f)
 
             def getdatafn():
                 fd = None
                 try:
                     fd = open(path, 'rb')
                     return fd.read()
                 finally:
                     if fd:
                         fd.close()
 
             getdata = getdatafn
         write(f, 'x' in ff and 0755 or 0644, 'l' in ff, getdata)
 
     if subrepos:
         for subpath in sorted(ctx.substate):
             sub = ctx.sub(subpath)
             submatch = match_.narrowmatcher(subpath, matchfn)
             sub.archive(repo.ui, archiver, prefix, submatch)
 
     archiver.done()
 
 def hgsubrepoarchive(orig, repo, ui, archiver, prefix, match=None):
     repo._get(repo._state + ('hg',))
     rev = repo._state[1]
     ctx = repo._repo[rev]
 
     lfcommands.cachelfiles(ui, repo._repo, ctx.node())
 
     def write(name, mode, islink, getdata):
         # At this point, the standin has been replaced with the largefile name,
         # so the normal matcher works here without the lfutil variants.
         if match and not match(f):
             return
         data = getdata()
 
         archiver.addfile(prefix + repo._path + '/' + name, mode, islink, data)
 
     for f in ctx:
         ff = ctx.flags(f)
         getdata = ctx[f].data
         if lfutil.isstandin(f):
             path = lfutil.findfile(repo._repo, getdata().strip())
             if path is None:
                 raise util.Abort(
                     _('largefile %s not found in repo store or system cache')
                     % lfutil.splitstandin(f))
             f = lfutil.splitstandin(f)
 
             def getdatafn():
                 fd = None
                 try:
                     fd = open(os.path.join(prefix, path), 'rb')
                     return fd.read()
                 finally:
                     if fd:
                         fd.close()
 
             getdata = getdatafn
 
         write(f, 'x' in ff and 0755 or 0644, 'l' in ff, getdata)
 
     for subpath in sorted(ctx.substate):
         sub = ctx.sub(subpath)
         submatch = match_.narrowmatcher(subpath, match)
         sub.archive(ui, archiver, os.path.join(prefix, repo._path) + '/',
                     submatch)
 
 # If a largefile is modified, the change is not reflected in its
 # standin until a commit. cmdutil.bailifchanged() raises an exception
 # if the repo has uncommitted changes. Wrap it to also check if
 # largefiles were changed. This is used by bisect and backout.
 def overridebailifchanged(orig, repo):
     orig(repo)
     repo.lfstatus = True
     s = repo.status()
     repo.lfstatus = False
     if s.modified or s.added or s.removed or s.deleted:
         raise util.Abort(_('uncommitted changes'))
 
 # Fetch doesn't use cmdutil.bailifchanged so override it to add the check
 def overridefetch(orig, ui, repo, *pats, **opts):
     repo.lfstatus = True
     s = repo.status()
     repo.lfstatus = False
     if s.modified or s.added or s.removed or s.deleted:
         raise util.Abort(_('uncommitted changes'))
     return orig(ui, repo, *pats, **opts)
 
 def overrideforget(orig, ui, repo, *pats, **opts):
     installnormalfilesmatchfn(repo[None].manifest())
     result = orig(ui, repo, *pats, **opts)
     restorematchfn()
     m = scmutil.match(repo[None], pats, opts)
 
     try:
         repo.lfstatus = True
         s = repo.status(match=m, clean=True)
     finally:
         repo.lfstatus = False
     forget = sorted(s.modified + s.added + s.deleted + s.clean)
     forget = [f for f in forget if lfutil.standin(f) in repo[None].manifest()]
 
     for f in forget:
         if lfutil.standin(f) not in repo.dirstate and not \
                 os.path.isdir(m.rel(lfutil.standin(f))):
             ui.warn(_('not removing %s: file is already untracked\n')
                     % m.rel(f))
             result = 1
 
     for f in forget:
         if ui.verbose or not m.exact(f):
             ui.status(_('removing %s\n') % m.rel(f))
 
     # Need to lock because standin files are deleted then removed from the
     # repository and we could race in-between.
     wlock = repo.wlock()
     try:
         lfdirstate = lfutil.openlfdirstate(ui, repo)
         for f in forget:
             if lfdirstate[f] == 'a':
                 lfdirstate.drop(f)
             else:
                 lfdirstate.remove(f)
         lfdirstate.write()
         standins = [lfutil.standin(f) for f in forget]
         for f in standins:
             util.unlinkpath(repo.wjoin(f), ignoremissing=True)
         repo[None].forget(standins)
     finally:
         wlock.release()
 
     return result
 
 def _getoutgoings(repo, other, missing, addfunc):
     """get pairs of filename and largefile hash in outgoing revisions
     in 'missing'.
 
     largefiles already existing on 'other' repository are ignored.
 
     'addfunc' is invoked with each unique pairs of filename and
     largefile hash value.
     """
     knowns = set()
     lfhashes = set()
     def dedup(fn, lfhash):
         k = (fn, lfhash)
         if k not in knowns:
             knowns.add(k)
             lfhashes.add(lfhash)
     lfutil.getlfilestoupload(repo, missing, dedup)
     if lfhashes:
         lfexists = basestore._openstore(repo, other).exists(lfhashes)
         for fn, lfhash in knowns:
             if not lfexists[lfhash]: # lfhash doesn't exist on "other"
                 addfunc(fn, lfhash)
 
 def outgoinghook(ui, repo, other, opts, missing):
     if opts.pop('large', None):
         lfhashes = set()
         if ui.debugflag:
             toupload = {}
             def addfunc(fn, lfhash):
                 if fn not in toupload:
                     toupload[fn] = []
                 toupload[fn].append(lfhash)
                 lfhashes.add(lfhash)
             def showhashes(fn):
                 for lfhash in sorted(toupload[fn]):
                     ui.debug('    %s\n' % (lfhash))
         else:
             toupload = set()
             def addfunc(fn, lfhash):
                 toupload.add(fn)
                 lfhashes.add(lfhash)
             def showhashes(fn):
                 pass
         _getoutgoings(repo, other, missing, addfunc)
 
         if not toupload:
             ui.status(_('largefiles: no files to upload\n'))
         else:
             ui.status(_('largefiles to upload (%d entities):\n')
                       % (len(lfhashes)))
             for file in sorted(toupload):
                 ui.status(lfutil.splitstandin(file) + '\n')
                 showhashes(file)
             ui.status('\n')
 
 def summaryremotehook(ui, repo, opts, changes):
     largeopt = opts.get('large', False)
     if changes is None:
         if largeopt:
             return (False, True) # only outgoing check is needed
         else:
             return (False, False)
     elif largeopt:
         url, branch, peer, outgoing = changes[1]
         if peer is None:
             # i18n: column positioning for "hg summary"
             ui.status(_('largefiles: (no remote repo)\n'))
             return
 
         toupload = set()
         lfhashes = set()
         def addfunc(fn, lfhash):
             toupload.add(fn)
             lfhashes.add(lfhash)
         _getoutgoings(repo, peer, outgoing.missing, addfunc)
 
         if not toupload:
             # i18n: column positioning for "hg summary"
             ui.status(_('largefiles: (no files to upload)\n'))
         else:
             # i18n: column positioning for "hg summary"
             ui.status(_('largefiles: %d entities for %d files to upload\n')
                       % (len(lfhashes), len(toupload)))
 
 def overridesummary(orig, ui, repo, *pats, **opts):
     try:
         repo.lfstatus = True
         orig(ui, repo, *pats, **opts)
     finally:
         repo.lfstatus = False
 
 def scmutiladdremove(orig, repo, pats=[], opts={}, dry_run=None,
                      similarity=None):
     if not lfutil.islfilesrepo(repo):
         return orig(repo, pats, opts, dry_run, similarity)
     # Get the list of missing largefiles so we can remove them
     lfdirstate = lfutil.openlfdirstate(repo.ui, repo)
     unsure, s = lfdirstate.status(match_.always(repo.root, repo.getcwd()), [],
                                   False, False, False)
 
     # Call into the normal remove code, but the removing of the standin, we want
     # to have handled by original addremove.  Monkey patching here makes sure
     # we don't remove the standin in the largefiles code, preventing a very
     # confused state later.
     if s.deleted:
         m = [repo.wjoin(f) for f in s.deleted]
         removelargefiles(repo.ui, repo, True, *m, **opts)
     # Call into the normal add code, and any files that *should* be added as
     # largefiles will be
     addlargefiles(repo.ui, repo, *pats, **opts)
     # Now that we've handled largefiles, hand off to the original addremove
     # function to take care of the rest.  Make sure it doesn't do anything with
     # largefiles by installing a matcher that will ignore them.
     installnormalfilesmatchfn(repo[None].manifest())
     result = orig(repo, pats, opts, dry_run, similarity)
     restorematchfn()
     return result
 
 # Calling purge with --all will cause the largefiles to be deleted.
 # Override repo.status to prevent this from happening.
 def overridepurge(orig, ui, repo, *dirs, **opts):
     # XXX large file status is buggy when used on repo proxy.
     # XXX this needs to be investigate.
     repo = repo.unfiltered()
     oldstatus = repo.status
     def overridestatus(node1='.', node2=None, match=None, ignored=False,
                         clean=False, unknown=False, listsubrepos=False):
         r = oldstatus(node1, node2, match, ignored, clean, unknown,
                       listsubrepos)
         lfdirstate = lfutil.openlfdirstate(ui, repo)
         unknown = [f for f in r.unknown if lfdirstate[f] == '?']
         ignored = [f for f in r.ignored if lfdirstate[f] == '?']
         return scmutil.status(r.modified, r.added, r.removed, r.deleted,
                               unknown, ignored, r.clean)
     repo.status = overridestatus
     orig(ui, repo, *dirs, **opts)
     repo.status = oldstatus
 def overriderollback(orig, ui, repo, **opts):
     wlock = repo.wlock()
     try:
         before = repo.dirstate.parents()
         orphans = set(f for f in repo.dirstate
                       if lfutil.isstandin(f) and repo.dirstate[f] != 'r')
         result = orig(ui, repo, **opts)
         after = repo.dirstate.parents()
         if before == after:
             return result # no need to restore standins
 
         pctx = repo['.']
         for f in repo.dirstate:
             if lfutil.isstandin(f):
                 orphans.discard(f)
                 if repo.dirstate[f] == 'r':
                     repo.wvfs.unlinkpath(f, ignoremissing=True)
                 elif f in pctx:
                     fctx = pctx[f]
                     repo.wwrite(f, fctx.data(), fctx.flags())
                 else:
                     # content of standin is not so important in 'a',
                     # 'm' or 'n' (coming from the 2nd parent) cases
                     lfutil.writestandin(repo, f, '', False)
         for standin in orphans:
             repo.wvfs.unlinkpath(standin, ignoremissing=True)
 
         lfdirstate = lfutil.openlfdirstate(ui, repo)
         orphans = set(lfdirstate)
         lfiles = lfutil.listlfiles(repo)
         for file in lfiles:
             lfutil.synclfdirstate(repo, lfdirstate, file, True)
             orphans.discard(file)
         for lfile in orphans:
             lfdirstate.drop(lfile)
         lfdirstate.write()
     finally:
         wlock.release()
     return result
 
 def overridetransplant(orig, ui, repo, *revs, **opts):
     resuming = opts.get('continue')
     repo._lfcommithooks.append(lfutil.automatedcommithook(resuming))
     repo._lfstatuswriters.append(lambda *msg, **opts: None)
     try:
         result = orig(ui, repo, *revs, **opts)
     finally:
         repo._lfstatuswriters.pop()
         repo._lfcommithooks.pop()
     return result
 
 def overridecat(orig, ui, repo, file1, *pats, **opts):
     ctx = scmutil.revsingle(repo, opts.get('rev'))
     err = 1
     notbad = set()
     m = scmutil.match(ctx, (file1,) + pats, opts)
     origmatchfn = m.matchfn
     def lfmatchfn(f):
         if origmatchfn(f):
             return True
         lf = lfutil.splitstandin(f)
         if lf is None:
             return False
         notbad.add(lf)
         return origmatchfn(lf)
     m.matchfn = lfmatchfn
     origbadfn = m.bad
     def lfbadfn(f, msg):
         if not f in notbad:
             origbadfn(f, msg)
     m.bad = lfbadfn
     for f in ctx.walk(m):
         fp = cmdutil.makefileobj(repo, opts.get('output'), ctx.node(),
                                  pathname=f)
         lf = lfutil.splitstandin(f)
         if lf is None or origmatchfn(f):
             # duplicating unreachable code from commands.cat
             data = ctx[f].data()
             if opts.get('decode'):
                 data = repo.wwritedata(f, data)
             fp.write(data)
         else:
             hash = lfutil.readstandin(repo, lf, ctx.rev())
             if not lfutil.inusercache(repo.ui, hash):
                 store = basestore._openstore(repo)
                 success, missing = store.get([(lf, hash)])
                 if len(success) != 1:
                     raise util.Abort(
                         _('largefile %s is not in cache and could not be '
                           'downloaded')  % lf)
             path = lfutil.usercachepath(repo.ui, hash)
             fpin = open(path, "rb")
             for chunk in util.filechunkiter(fpin, 128 * 1024):
                 fp.write(chunk)
             fpin.close()
         fp.close()
         err = 0
     return err
 
-def mercurialsinkbefore(orig, sink):
-    sink.repo._isconverting = True
-    orig(sink)
-
-def mercurialsinkafter(orig, sink):
-    sink.repo._isconverting = False
-    orig(sink)
-
 def mergeupdate(orig, repo, node, branchmerge, force, partial,
                 *args, **kwargs):
     wlock = repo.wlock()
     try:
         # branch |       |         |
         #  merge | force | partial | action
         # -------+-------+---------+--------------
         #    x   |   x   |    x    | linear-merge
         #    o   |   x   |    x    | branch-merge
         #    x   |   o   |    x    | overwrite (as clean update)
         #    o   |   o   |    x    | force-branch-merge (*1)
         #    x   |   x   |    o    |   (*)
         #    o   |   x   |    o    |   (*)
         #    x   |   o   |    o    | overwrite (as revert)
         #    o   |   o   |    o    |   (*)
         #
         # (*) don't care
         # (*1) deprecated, but used internally (e.g: "rebase --collapse")
 
         linearmerge = not branchmerge and not force and not partial
 
         if linearmerge or (branchmerge and force and not partial):
             # update standins for linear-merge or force-branch-merge,
             # because largefiles in the working directory may be modified
             lfdirstate = lfutil.openlfdirstate(repo.ui, repo)
             unsure, s = lfdirstate.status(match_.always(repo.root,
                                                         repo.getcwd()),
                                           [], False, False, False)
             for lfile in unsure + s.modified + s.added:
                 lfutil.updatestandin(repo, lfutil.standin(lfile))
 
         if linearmerge:
             # Only call updatelfiles on the standins that have changed
             # to save time
             oldstandins = lfutil.getstandinsstate(repo)
 
         result = orig(repo, node, branchmerge, force, partial, *args, **kwargs)
 
         filelist = None
         if linearmerge:
             newstandins = lfutil.getstandinsstate(repo)
             filelist = lfutil.getlfilestoupdate(oldstandins, newstandins)
 
         lfcommands.updatelfiles(repo.ui, repo, filelist=filelist,
                                 normallookup=partial)
 
         return result
     finally:
         wlock.release()
 
 def scmutilmarktouched(orig, repo, files, *args, **kwargs):
     result = orig(repo, files, *args, **kwargs)
 
     filelist = [lfutil.splitstandin(f) for f in files if lfutil.isstandin(f)]
     if filelist:
         lfcommands.updatelfiles(repo.ui, repo, filelist=filelist,
                                 printmessage=False, normallookup=True)
 
     return result

hgext/largefiles/reposetup.py

@@ -1,367 +1,364 b''
 # Copyright 2009-2010 Gregory P. Ward
 # Copyright 2009-2010 Intelerad Medical Systems Incorporated
 # Copyright 2010-2011 Fog Creek Software
 # Copyright 2010-2011 Unity Technologies
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 '''setup for largefiles repositories: reposetup'''
 import copy
 import os
 
 from mercurial import error, manifest, match as match_, util
 from mercurial.i18n import _
 from mercurial import localrepo, scmutil
 
 import lfcommands
 import lfutil
 
 def reposetup(ui, repo):
     # wire repositories should be given new wireproto functions
     # by "proto.wirereposetup()" via "hg.wirepeersetupfuncs"
     if not repo.local():
         return
 
     class lfilesrepo(repo.__class__):
         lfstatus = False
         def status_nolfiles(self, *args, **kwargs):
             return super(lfilesrepo, self).status(*args, **kwargs)
 
         # When lfstatus is set, return a context that gives the names
         # of largefiles instead of their corresponding standins and
         # identifies the largefiles as always binary, regardless of
         # their actual contents.
         def __getitem__(self, changeid):
             ctx = super(lfilesrepo, self).__getitem__(changeid)
             if self.lfstatus:
                 class lfilesmanifestdict(manifest.manifestdict):
                     def __contains__(self, filename):
                         orig = super(lfilesmanifestdict, self).__contains__
                         return orig(filename) or orig(lfutil.standin(filename))
                 class lfilesctx(ctx.__class__):
                     def files(self):
                         filenames = super(lfilesctx, self).files()
                         return [lfutil.splitstandin(f) or f for f in filenames]
                     def manifest(self):
                         man1 = super(lfilesctx, self).manifest()
                         man1.__class__ = lfilesmanifestdict
                         return man1
                     def filectx(self, path, fileid=None, filelog=None):
                         orig = super(lfilesctx, self).filectx
                         try:
                             if filelog is not None:
                                 result = orig(path, fileid, filelog)
                             else:
                                 result = orig(path, fileid)
                         except error.LookupError:
                             # Adding a null character will cause Mercurial to
                             # identify this as a binary file.
                             if filelog is not None:
                                 result = orig(lfutil.standin(path), fileid,
                                               filelog)
                             else:
                                 result = orig(lfutil.standin(path), fileid)
                             olddata = result.data
                             result.data = lambda: olddata() + '\0'
                         return result
                 ctx.__class__ = lfilesctx
             return ctx
 
         # Figure out the status of big files and insert them into the
         # appropriate list in the result. Also removes standin files
         # from the listing. Revert to the original status if
         # self.lfstatus is False.
         # XXX large file status is buggy when used on repo proxy.
         # XXX this needs to be investigated.
         @localrepo.unfilteredmethod
         def status(self, node1='.', node2=None, match=None, ignored=False,
                 clean=False, unknown=False, listsubrepos=False):
             listignored, listclean, listunknown = ignored, clean, unknown
             orig = super(lfilesrepo, self).status
             if not self.lfstatus:
                 return orig(node1, node2, match, listignored, listclean,
                             listunknown, listsubrepos)
 
             # some calls in this function rely on the old version of status
             self.lfstatus = False
             ctx1 = self[node1]
             ctx2 = self[node2]
             working = ctx2.rev() is None
             parentworking = working and ctx1 == self['.']
 
             if match is None:
                 match = match_.always(self.root, self.getcwd())
 
             wlock = None
             try:
                 try:
                     # updating the dirstate is optional
                     # so we don't wait on the lock
                     wlock = self.wlock(False)
                 except error.LockError:
                     pass
 
                 # First check if paths or patterns were specified on the
                 # command line.  If there were, and they don't match any
                 # largefiles, we should just bail here and let super
                 # handle it -- thus gaining a big performance boost.
                 lfdirstate = lfutil.openlfdirstate(ui, self)
                 if not match.always():
                     for f in lfdirstate:
                         if match(f):
                             break
                     else:
                         return orig(node1, node2, match, listignored, listclean,
                                     listunknown, listsubrepos)
 
                 # Create a copy of match that matches standins instead
                 # of largefiles.
                 def tostandins(files):
                     if not working:
                         return files
                     newfiles = []
                     dirstate = self.dirstate
                     for f in files:
                         sf = lfutil.standin(f)
                         if sf in dirstate:
                             newfiles.append(sf)
                         elif sf in dirstate.dirs():
                             # Directory entries could be regular or
                             # standin, check both
                             newfiles.extend((f, sf))
                         else:
                             newfiles.append(f)
                     return newfiles
 
                 m = copy.copy(match)
                 m._files = tostandins(m._files)
 
                 result = orig(node1, node2, m, ignored, clean, unknown,
                               listsubrepos)
                 if working:
 
                     def sfindirstate(f):
                         sf = lfutil.standin(f)
                         dirstate = self.dirstate
                         return sf in dirstate or sf in dirstate.dirs()
 
                     match._files = [f for f in match._files
                                     if sfindirstate(f)]
                     # Don't waste time getting the ignored and unknown
                     # files from lfdirstate
                     unsure, s = lfdirstate.status(match, [], False, listclean,
                                                   False)
                     (modified, added, removed, clean) = (s.modified, s.added,
                                                          s.removed, s.clean)
                     if parentworking:
                         for lfile in unsure:
                             standin = lfutil.standin(lfile)
                             if standin not in ctx1:
                                 # from second parent
                                 modified.append(lfile)
                             elif ctx1[standin].data().strip() \
                                     != lfutil.hashfile(self.wjoin(lfile)):
                                 modified.append(lfile)
                             else:
                                 if listclean:
                                     clean.append(lfile)
                                 lfdirstate.normal(lfile)
                     else:
                         tocheck = unsure + modified + added + clean
                         modified, added, clean = [], [], []
 
                         for lfile in tocheck:
                             standin = lfutil.standin(lfile)
                             if standin in ctx1:
                                 abslfile = self.wjoin(lfile)
                                 if ((ctx1[standin].data().strip() !=
                                      lfutil.hashfile(abslfile)) or
                                     (('x' in ctx1.flags(standin)) !=
                                      bool(lfutil.getexecutable(abslfile)))):
                                     modified.append(lfile)
                                 elif listclean:
                                     clean.append(lfile)
                             else:
                                 added.append(lfile)
 
                         # at this point, 'removed' contains largefiles
                         # marked as 'R' in the working context.
                         # then, largefiles not managed also in the target
                         # context should be excluded from 'removed'.
                         removed = [lfile for lfile in removed
                                    if lfutil.standin(lfile) in ctx1]
 
                     # Standins no longer found in lfdirstate has been
                     # removed
                     for standin in ctx1.walk(lfutil.getstandinmatcher(self)):
                         lfile = lfutil.splitstandin(standin)
                         if not match(lfile):
                             continue
                         if lfile not in lfdirstate:
                             removed.append(lfile)
 
                     # Filter result lists
                     result = list(result)
 
                     # Largefiles are not really removed when they're
                     # still in the normal dirstate. Likewise, normal
                     # files are not really removed if they are still in
                     # lfdirstate. This happens in merges where files
                     # change type.
                     removed = [f for f in removed
                                if f not in self.dirstate]
                     result[2] = [f for f in result[2]
                                  if f not in lfdirstate]
 
                     lfiles = set(lfdirstate._map)
                     # Unknown files
                     result[4] = set(result[4]).difference(lfiles)
                     # Ignored files
                     result[5] = set(result[5]).difference(lfiles)
                     # combine normal files and largefiles
                     normals = [[fn for fn in filelist
                                 if not lfutil.isstandin(fn)]
                                for filelist in result]
                     lfstatus = (modified, added, removed, s.deleted, [], [],
                                 clean)
                     result = [sorted(list1 + list2)
                               for (list1, list2) in zip(normals, lfstatus)]
                 else: # not against working directory
                     result = [[lfutil.splitstandin(f) or f for f in items]
                               for items in result]
 
                 if wlock:
                     lfdirstate.write()
 
             finally:
                 if wlock:
                     wlock.release()
 
             self.lfstatus = True
             return scmutil.status(*result)
 
-        # As part of committing, copy all of the largefiles into the
-        # cache.
         def commitctx(self, ctx, *args, **kwargs):
             node = super(lfilesrepo, self).commitctx(ctx, *args, **kwargs)
-            lfutil.copyalltostore(self, node)
             class lfilesctx(ctx.__class__):
                 def markcommitted(self, node):
                     orig = super(lfilesctx, self).markcommitted
                     return lfutil.markcommitted(orig, self, node)
             ctx.__class__ = lfilesctx
             return node
 
         # Before commit, largefile standins have not had their
         # contents updated to reflect the hash of their largefile.
         # Do that here.
         def commit(self, text="", user=None, date=None, match=None,
                 force=False, editor=False, extra={}):
             orig = super(lfilesrepo, self).commit
 
             wlock = self.wlock()
             try:
                 lfcommithook = self._lfcommithooks[-1]
                 match = lfcommithook(self, match)
                 result = orig(text=text, user=user, date=date, match=match,
                                 force=force, editor=editor, extra=extra)
                 return result
             finally:
                 wlock.release()
 
         def push(self, remote, force=False, revs=None, newbranch=False):
             if remote.local():
                 missing = set(self.requirements) - remote.local().supported
                 if missing:
                     msg = _("required features are not"
                             " supported in the destination:"
                             " %s") % (', '.join(sorted(missing)))
                     raise util.Abort(msg)
             return super(lfilesrepo, self).push(remote, force=force, revs=revs,
                 newbranch=newbranch)
 
         # TODO: _subdirlfs should be moved into "lfutil.py", because
         # it is referred only from "lfutil.updatestandinsbymatch"
         def _subdirlfs(self, files, lfiles):
             '''
             Adjust matched file list
             If we pass a directory to commit whose only commitable files
             are largefiles, the core commit code aborts before finding
             the largefiles.
             So we do the following:
             For directories that only have largefiles as matches,
             we explicitly add the largefiles to the match list and remove
             the directory.
             In other cases, we leave the match list unmodified.
             '''
             actualfiles = []
             dirs = []
             regulars = []
 
             for f in files:
                 if lfutil.isstandin(f + '/'):
                     raise util.Abort(
                         _('file "%s" is a largefile standin') % f,
                         hint=('commit the largefile itself instead'))
                 # Scan directories
                 if os.path.isdir(self.wjoin(f)):
                     dirs.append(f)
                 else:
                     regulars.append(f)
 
             for f in dirs:
                 matcheddir = False
                 d = self.dirstate.normalize(f) + '/'
                 # Check for matched normal files
                 for mf in regulars:
                     if self.dirstate.normalize(mf).startswith(d):
                         actualfiles.append(f)
                         matcheddir = True
                         break
                 if not matcheddir:
                     # If no normal match, manually append
                     # any matching largefiles
                     for lf in lfiles:
                         if self.dirstate.normalize(lf).startswith(d):
                             actualfiles.append(lf)
                             if not matcheddir:
                                 actualfiles.append(lfutil.standin(f))
                                 matcheddir = True
                 # Nothing in dir, so readd it
                 # and let commit reject it
                 if not matcheddir:
                     actualfiles.append(f)
 
             # Always add normal files
             actualfiles += regulars
             return actualfiles
 
     repo.__class__ = lfilesrepo
 
     # stack of hooks being executed before committing.
     # only last element ("_lfcommithooks[-1]") is used for each committing.
     repo._lfcommithooks = [lfutil.updatestandinsbymatch]
 
     # Stack of status writer functions taking "*msg, **opts" arguments
     # like "ui.status()". Only last element ("_lfupdatereporters[-1]")
     # is used to write status out.
     repo._lfstatuswriters = [ui.status]
 
     def prepushoutgoinghook(local, remote, outgoing):
         if outgoing.missing:
             toupload = set()
             addfunc = lambda fn, lfhash: toupload.add(lfhash)
             lfutil.getlfilestoupload(local, outgoing.missing, addfunc)
             lfcommands.uploadlfiles(ui, local, remote, toupload)
     repo.prepushoutgoinghooks.add("largefiles", prepushoutgoinghook)
 
     def checkrequireslfiles(ui, repo, **kwargs):
         if 'largefiles' not in repo.requirements and util.any(
                 lfutil.shortname+'/' in f[0] for f in repo.store.datafiles()):
             repo.requirements.add('largefiles')
             repo._writerequirements()
 
     ui.setconfig('hooks', 'changegroup.lfiles', checkrequireslfiles,
                  'largefiles')
     ui.setconfig('hooks', 'commit.lfiles', checkrequireslfiles, 'largefiles')

hgext/largefiles/uisetup.py

@@ -1,183 +1,176 b''
 # Copyright 2009-2010 Gregory P. Ward
 # Copyright 2009-2010 Intelerad Medical Systems Incorporated
 # Copyright 2010-2011 Fog Creek Software
 # Copyright 2010-2011 Unity Technologies
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 '''setup for largefiles extension: uisetup'''
 
 from mercurial import archival, cmdutil, commands, extensions, filemerge, hg, \
     httppeer, merge, scmutil, sshpeer, wireproto, revset, subrepo
 from mercurial.i18n import _
 from mercurial.hgweb import hgweb_mod, webcommands
 
 import overrides
 import proto
 
 def uisetup(ui):
     # Disable auto-status for some commands which assume that all
     # files in the result are under Mercurial's control
 
     entry = extensions.wrapcommand(commands.table, 'add',
                                    overrides.overrideadd)
     addopt = [('', 'large', None, _('add as largefile')),
               ('', 'normal', None, _('add as normal file')),
               ('', 'lfsize', '', _('add all files above this size '
                                    '(in megabytes) as largefiles '
                                    '(default: 10)'))]
     entry[1].extend(addopt)
 
     # The scmutil function is called both by the (trivial) addremove command,
     # and in the process of handling commit -A (issue3542)
     entry = extensions.wrapfunction(scmutil, 'addremove',
                                     overrides.scmutiladdremove)
     entry = extensions.wrapcommand(commands.table, 'remove',
                                    overrides.overrideremove)
     entry = extensions.wrapcommand(commands.table, 'forget',
                                    overrides.overrideforget)
 
     # Subrepos call status function
     entry = extensions.wrapcommand(commands.table, 'status',
                                    overrides.overridestatus)
     entry = extensions.wrapfunction(subrepo.hgsubrepo, 'status',
                                     overrides.overridestatusfn)
 
     entry = extensions.wrapcommand(commands.table, 'log',
                                    overrides.overridelog)
     entry = extensions.wrapcommand(commands.table, 'rollback',
                                    overrides.overriderollback)
     entry = extensions.wrapcommand(commands.table, 'verify',
                                    overrides.overrideverify)
 
     verifyopt = [('', 'large', None,
                   _('verify that all largefiles in current revision exists')),
                  ('', 'lfa', None,
                   _('verify largefiles in all revisions, not just current')),
                  ('', 'lfc', None,
                   _('verify local largefile contents, not just existence'))]
     entry[1].extend(verifyopt)
 
     entry = extensions.wrapcommand(commands.table, 'debugstate',
                                    overrides.overridedebugstate)
     debugstateopt = [('', 'large', None, _('display largefiles dirstate'))]
     entry[1].extend(debugstateopt)
 
     outgoing = lambda orgfunc, *arg, **kwargs: orgfunc(*arg, **kwargs)
     entry = extensions.wrapcommand(commands.table, 'outgoing', outgoing)
     outgoingopt = [('', 'large', None, _('display outgoing largefiles'))]
     entry[1].extend(outgoingopt)
     cmdutil.outgoinghooks.add('largefiles', overrides.outgoinghook)
     entry = extensions.wrapcommand(commands.table, 'summary',
                                    overrides.overridesummary)
     summaryopt = [('', 'large', None, _('display outgoing largefiles'))]
     entry[1].extend(summaryopt)
     cmdutil.summaryremotehooks.add('largefiles', overrides.summaryremotehook)
 
     entry = extensions.wrapcommand(commands.table, 'update',
                                    overrides.overrideupdate)
     entry = extensions.wrapcommand(commands.table, 'pull',
                                    overrides.overridepull)
     pullopt = [('', 'all-largefiles', None,
                  _('download all pulled versions of largefiles (DEPRECATED)')),
                ('', 'lfrev', [],
                 _('download largefiles for these revisions'), _('REV'))]
     entry[1].extend(pullopt)
     revset.symbols['pulled'] = overrides.pulledrevsetsymbol
 
     entry = extensions.wrapcommand(commands.table, 'clone',
                                    overrides.overrideclone)
     cloneopt = [('', 'all-largefiles', None,
                  _('download all versions of all largefiles'))]
     entry[1].extend(cloneopt)
     entry = extensions.wrapfunction(hg, 'clone', overrides.hgclone)
 
     entry = extensions.wrapcommand(commands.table, 'cat',
                                    overrides.overridecat)
     entry = extensions.wrapfunction(merge, '_checkunknownfile',
                                     overrides.overridecheckunknownfile)
     entry = extensions.wrapfunction(merge, 'calculateupdates',
                                     overrides.overridecalculateupdates)
     entry = extensions.wrapfunction(merge, 'recordupdates',
                                     overrides.mergerecordupdates)
     entry = extensions.wrapfunction(merge, 'update',
                                     overrides.mergeupdate)
     entry = extensions.wrapfunction(filemerge, 'filemerge',
                                     overrides.overridefilemerge)
     entry = extensions.wrapfunction(cmdutil, 'copy',
                                     overrides.overridecopy)
 
     # Summary calls dirty on the subrepos
     entry = extensions.wrapfunction(subrepo.hgsubrepo, 'dirty',
                                     overrides.overridedirty)
 
     # Backout calls revert so we need to override both the command and the
     # function
     entry = extensions.wrapcommand(commands.table, 'revert',
                                    overrides.overriderevert)
     entry = extensions.wrapfunction(commands, 'revert',
                                     overrides.overriderevert)
 
     extensions.wrapfunction(archival, 'archive', overrides.overridearchive)
     extensions.wrapfunction(subrepo.hgsubrepo, 'archive',
                             overrides.hgsubrepoarchive)
     extensions.wrapfunction(cmdutil, 'bailifchanged',
                             overrides.overridebailifchanged)
 
     extensions.wrapfunction(scmutil, 'marktouched',
                             overrides.scmutilmarktouched)
 
     # create the new wireproto commands ...
     wireproto.commands['putlfile'] = (proto.putlfile, 'sha')
     wireproto.commands['getlfile'] = (proto.getlfile, 'sha')
     wireproto.commands['statlfile'] = (proto.statlfile, 'sha')
 
     # ... and wrap some existing ones
     wireproto.commands['capabilities'] = (proto.capabilities, '')
     wireproto.commands['heads'] = (proto.heads, '')
     wireproto.commands['lheads'] = (wireproto.heads, '')
 
     # make putlfile behave the same as push and {get,stat}lfile behave
     # the same as pull w.r.t. permissions checks
     hgweb_mod.perms['putlfile'] = 'push'
     hgweb_mod.perms['getlfile'] = 'pull'
     hgweb_mod.perms['statlfile'] = 'pull'
 
     extensions.wrapfunction(webcommands, 'decodepath', overrides.decodepath)
 
     # the hello wireproto command uses wireproto.capabilities, so it won't see
     # our largefiles capability unless we replace the actual function as well.
     proto.capabilitiesorig = wireproto.capabilities
     wireproto.capabilities = proto.capabilities
 
     # can't do this in reposetup because it needs to have happened before
     # wirerepo.__init__ is called
     proto.ssholdcallstream = sshpeer.sshpeer._callstream
     proto.httpoldcallstream = httppeer.httppeer._callstream
     sshpeer.sshpeer._callstream = proto.sshrepocallstream
     httppeer.httppeer._callstream = proto.httprepocallstream
 
     # override some extensions' stuff as well
     for name, module in extensions.extensions():
         if name == 'fetch':
             extensions.wrapcommand(getattr(module, 'cmdtable'), 'fetch',
                 overrides.overridefetch)
         if name == 'purge':
             extensions.wrapcommand(getattr(module, 'cmdtable'), 'purge',
                 overrides.overridepurge)
         if name == 'rebase':
             extensions.wrapcommand(getattr(module, 'cmdtable'), 'rebase',
                 overrides.overriderebase)
             extensions.wrapfunction(module, 'rebase',
                                     overrides.overriderebase)
         if name == 'transplant':
             extensions.wrapcommand(getattr(module, 'cmdtable'), 'transplant',
                 overrides.overridetransplant)
-        if name == 'convert':
-            convcmd = getattr(module, 'convcmd')
-            hgsink = getattr(convcmd, 'mercurial_sink')
-            extensions.wrapfunction(hgsink, 'before',
-                                    overrides.mercurialsinkbefore)
-            extensions.wrapfunction(hgsink, 'after',
-                                    overrides.mercurialsinkafter)

tests/test-largefiles-update.t

@@ -1,613 +1,626 b''
 This file focuses mainly on updating largefiles in the working
 directory (and ".hg/largefiles/dirstate")
 
   $ cat >> $HGRCPATH <<EOF
   > [ui]
   > merge = internal:fail
   > [extensions]
   > largefiles =
   > EOF
 
   $ hg init repo
   $ cd repo
 
   $ echo large1 > large1
   $ echo large2 > large2
   $ hg add --large large1 large2
   $ echo normal1 > normal1
   $ hg add normal1
   $ hg commit -m '#0'
   $ echo 'large1 in #1' > large1
   $ echo 'normal1 in #1' > normal1
   $ hg commit -m '#1'
   $ hg update -q -C 0
   $ echo 'large2 in #2' > large2
   $ hg commit -m '#2'
   created new head
 
 Test that "hg merge" updates largefiles from "other" correctly
 
 (getting largefiles from "other" normally)
 
   $ hg status -A large1
   C large1
   $ cat large1
   large1
   $ cat .hglf/large1
   4669e532d5b2c093a78eca010077e708a071bb64
   $ hg merge --config debug.dirstate.delaywrite=2
   getting changed largefiles
   1 largefiles updated, 0 removed
   2 files updated, 0 files merged, 0 files removed, 0 files unresolved
   (branch merge, don't forget to commit)
   $ hg status -A large1
   M large1
   $ cat large1
   large1 in #1
   $ cat .hglf/large1
   58e24f733a964da346e2407a2bee99d9001184f5
   $ hg diff -c 1 --nodates .hglf/large1 | grep '^[+-][0-9a-z]'
   -4669e532d5b2c093a78eca010077e708a071bb64
   +58e24f733a964da346e2407a2bee99d9001184f5
 
 (getting largefiles from "other" via conflict prompt)
 
   $ hg update -q -C 2
   $ echo 'large1 in #3' > large1
   $ echo 'normal1 in #3' > normal1
   $ hg commit -m '#3'
   $ cat .hglf/large1
   e5bb990443d6a92aaf7223813720f7566c9dd05b
   $ hg merge --config debug.dirstate.delaywrite=2 --config ui.interactive=True <<EOF
   > o
   > EOF
   largefile large1 has a merge conflict
   ancestor was 4669e532d5b2c093a78eca010077e708a071bb64
   keep (l)ocal e5bb990443d6a92aaf7223813720f7566c9dd05b or
   take (o)ther 58e24f733a964da346e2407a2bee99d9001184f5? o
   merging normal1
   warning: conflicts during merge.
   merging normal1 incomplete! (edit conflicts, then use 'hg resolve --mark')
   getting changed largefiles
   1 largefiles updated, 0 removed
   0 files updated, 1 files merged, 0 files removed, 1 files unresolved
   use 'hg resolve' to retry unresolved file merges or 'hg update -C .' to abandon
   [1]
   $ hg status -A large1
   M large1
   $ cat large1
   large1 in #1
   $ cat .hglf/large1
   58e24f733a964da346e2407a2bee99d9001184f5
 
 Test that "hg revert -r REV" updates largefiles from "REV" correctly
 
   $ hg update -q -C 3
   $ hg status -A large1
   C large1
   $ cat large1
   large1 in #3
   $ cat .hglf/large1
   e5bb990443d6a92aaf7223813720f7566c9dd05b
   $ hg diff -c 1 --nodates .hglf/large1 | grep '^[+-][0-9a-z]'
   -4669e532d5b2c093a78eca010077e708a071bb64
   +58e24f733a964da346e2407a2bee99d9001184f5
   $ hg revert --no-backup -r 1 --config debug.dirstate.delaywrite=2 large1
   $ hg status -A large1
   M large1
   $ cat large1
   large1 in #1
   $ cat .hglf/large1
   58e24f733a964da346e2407a2bee99d9001184f5
 
 Test that "hg rollback" restores status of largefiles correctly
 
   $ hg update -C -q
   $ hg remove large1
   $ test -f .hglf/large1
   [1]
   $ hg forget large2
   $ test -f .hglf/large2
   [1]
   $ echo largeX > largeX
   $ hg add --large largeX
   $ cat .hglf/largeX
   
   $ hg commit -m 'will be rollback-ed soon'
   $ echo largeY > largeY
   $ hg add --large largeY
 #if windows
   $ hg status -A large1
   large1: * (glob)
 #else
   $ hg status -A large1
   large1: No such file or directory
 #endif
   $ hg status -A large2
   ? large2
   $ hg status -A largeX
   C largeX
   $ hg status -A largeY
   A largeY
   $ hg rollback
   repository tip rolled back to revision 3 (undo commit)
   working directory now based on revision 3
   $ hg status -A large1
   R large1
   $ test -f .hglf/large1
   [1]
   $ hg status -A large2
   R large2
   $ test -f .hglf/large2
   [1]
   $ hg status -A largeX
   A largeX
   $ cat .hglf/largeX
   
   $ hg status -A largeY
   ? largeY
   $ test -f .hglf/largeY
   [1]
 
 Test that "hg rollback" restores standins correctly
 
   $ hg commit -m 'will be rollback-ed soon'
   $ hg update -q -C 2
   $ cat large1
   large1
   $ cat .hglf/large1
   4669e532d5b2c093a78eca010077e708a071bb64
   $ cat large2
   large2 in #2
   $ cat .hglf/large2
   3cfce6277e7668985707b6887ce56f9f62f6ccd9
 
   $ hg rollback -q -f
   $ cat large1
   large1
   $ cat .hglf/large1
   4669e532d5b2c093a78eca010077e708a071bb64
   $ cat large2
   large2 in #2
   $ cat .hglf/large2
   3cfce6277e7668985707b6887ce56f9f62f6ccd9
 
 (rollback the parent of the working directory, when the parent of it
 is not branch-tip)
 
   $ hg update -q -C 1
   $ cat .hglf/large1
   58e24f733a964da346e2407a2bee99d9001184f5
   $ cat .hglf/large2
   1deebade43c8c498a3c8daddac0244dc55d1331d
 
   $ echo normalX > normalX
   $ hg add normalX
   $ hg commit -m 'will be rollback-ed soon'
   $ hg rollback -q
 
   $ cat .hglf/large1
   58e24f733a964da346e2407a2bee99d9001184f5
   $ cat .hglf/large2
   1deebade43c8c498a3c8daddac0244dc55d1331d
 
 Test that "hg status" shows status of largefiles correctly just after
 automated commit like rebase/transplant
 
   $ cat >> .hg/hgrc <<EOF
   > [extensions]
   > rebase =
   > strip =
   > transplant =
   > EOF
   $ hg update -q -C 1
   $ hg remove large1
   $ echo largeX > largeX
   $ hg add --large largeX
   $ hg commit -m '#4'
 
   $ hg rebase -s 1 -d 2 --keep
 #if windows
   $ hg status -A large1
   large1: * (glob)
 #else
   $ hg status -A large1
   large1: No such file or directory
 #endif
   $ hg status -A largeX
   C largeX
   $ hg strip -q 5
 
   $ hg update -q -C 2
   $ hg transplant -q 1 4
 #if windows
   $ hg status -A large1
   large1: * (glob)
 #else
   $ hg status -A large1
   large1: No such file or directory
 #endif
   $ hg status -A largeX
   C largeX
   $ hg strip -q 5
 
   $ hg update -q -C 2
   $ hg transplant -q --merge 1 --merge 4
 #if windows
   $ hg status -A large1
   large1: * (glob)
 #else
   $ hg status -A large1
   large1: No such file or directory
 #endif
   $ hg status -A largeX
   C largeX
   $ hg strip -q 5
 
 Test that linear merge can detect modification (and conflict) correctly
 
 (linear merge without conflict)
 
   $ echo 'large2 for linear merge (no conflict)' > large2
   $ hg update 3 --config debug.dirstate.delaywrite=2
   getting changed largefiles
   1 largefiles updated, 0 removed
   2 files updated, 0 files merged, 0 files removed, 0 files unresolved
   $ hg status -A large2
   M large2
   $ cat large2
   large2 for linear merge (no conflict)
   $ cat .hglf/large2
   9c4bf8f1b33536d6e5f89447e10620cfe52ea710
 
 (linear merge with conflict, choosing "other")
 
   $ hg update -q -C 2
   $ echo 'large1 for linear merge (conflict)' > large1
   $ hg update 3 --config ui.interactive=True <<EOF
   > o
   > EOF
   largefile large1 has a merge conflict
   ancestor was 4669e532d5b2c093a78eca010077e708a071bb64
   keep (l)ocal ba94c2efe5b7c5e0af8d189295ce00553b0612b7 or
   take (o)ther e5bb990443d6a92aaf7223813720f7566c9dd05b? o
   getting changed largefiles
   1 largefiles updated, 0 removed
   1 files updated, 1 files merged, 0 files removed, 0 files unresolved
   $ hg status -A large1
   C large1
   $ cat large1
   large1 in #3
   $ cat .hglf/large1
   e5bb990443d6a92aaf7223813720f7566c9dd05b
 
 (linear merge with conflict, choosing "local")
 
   $ hg update -q -C 2
   $ echo 'large1 for linear merge (conflict)' > large1
   $ hg update 3 --config debug.dirstate.delaywrite=2
   largefile large1 has a merge conflict
   ancestor was 4669e532d5b2c093a78eca010077e708a071bb64
   keep (l)ocal ba94c2efe5b7c5e0af8d189295ce00553b0612b7 or
   take (o)ther e5bb990443d6a92aaf7223813720f7566c9dd05b? l
   1 files updated, 1 files merged, 0 files removed, 0 files unresolved
   $ hg status -A large1
   M large1
   $ cat large1
   large1 for linear merge (conflict)
   $ cat .hglf/large1
   ba94c2efe5b7c5e0af8d189295ce00553b0612b7
 
 Test a linear merge to a revision containing same-name normal file
 
   $ hg update -q -C 3
   $ hg remove large2
   $ echo 'large2 as normal file' > large2
   $ hg add large2
   $ echo 'large3 as normal file' > large3
   $ hg add large3
   $ hg commit -m '#5'
   $ hg manifest
   .hglf/large1
   large2
   large3
   normal1
 
 (modified largefile is already switched to normal)
 
   $ hg update -q -C 2
   $ echo 'modified large2 for linear merge' > large2
   $ hg update -q 5
   local changed .hglf/large2 which remote deleted
   use (c)hanged version or (d)elete? c
   remote turned local largefile large2 into a normal file
   keep (l)argefile or use (n)ormal file? l
   $ hg debugdirstate --nodates | grep large2
   a   0         -1 .hglf/large2
   r   0          0 large2
   $ hg status -A large2
   A large2
   $ cat large2
   modified large2 for linear merge
 
 (added largefile is already committed as normal)
 
   $ hg update -q -C 2
   $ echo 'large3 as large file for linear merge' > large3
   $ hg add --large large3
   $ hg update -q 5
   remote turned local largefile large3 into a normal file
   keep (l)argefile or use (n)ormal file? l
   $ hg debugdirstate --nodates | grep large3
   a   0         -1 .hglf/large3
   r   0          0 large3
   $ hg status -A large3
   A large3
   $ cat large3
   large3 as large file for linear merge
   $ rm -f large3 .hglf/large3
 
 Test that the internal linear merging works correctly
 (both heads are stripped to keep pairing of revision number and commit log)
 
   $ hg update -q -C 2
   $ hg strip 3 4
   saved backup bundle to $TESTTMP/repo/.hg/strip-backup/9530e27857f7-backup.hg (glob)
   $ mv .hg/strip-backup/9530e27857f7-backup.hg $TESTTMP
 
 (internal linear merging at "hg pull --update")
 
   $ echo 'large1 for linear merge (conflict)' > large1
   $ echo 'large2 for linear merge (conflict with normal file)' > large2
   $ hg pull --update --config debug.dirstate.delaywrite=2 $TESTTMP/9530e27857f7-backup.hg
   pulling from $TESTTMP/9530e27857f7-backup.hg (glob)
   searching for changes
   adding changesets
   adding manifests
   adding file changes
   added 3 changesets with 5 changes to 5 files
   local changed .hglf/large2 which remote deleted
   use (c)hanged version or (d)elete? c
   remote turned local largefile large2 into a normal file
   keep (l)argefile or use (n)ormal file? l
   largefile large1 has a merge conflict
   ancestor was 4669e532d5b2c093a78eca010077e708a071bb64
   keep (l)ocal ba94c2efe5b7c5e0af8d189295ce00553b0612b7 or
   take (o)ther e5bb990443d6a92aaf7223813720f7566c9dd05b? l
   2 files updated, 1 files merged, 0 files removed, 0 files unresolved
 
   $ hg status -A large1
   M large1
   $ cat large1
   large1 for linear merge (conflict)
   $ cat .hglf/large1
   ba94c2efe5b7c5e0af8d189295ce00553b0612b7
   $ hg status -A large2
   A large2
   $ cat large2
   large2 for linear merge (conflict with normal file)
   $ cat .hglf/large2
   d7591fe9be0f6227d90bddf3e4f52ff41fc1f544
 
 (internal linear merging at "hg unbundle --update")
 
   $ hg update -q -C 2
   $ hg rollback -q
 
   $ echo 'large1 for linear merge (conflict)' > large1
   $ echo 'large2 for linear merge (conflict with normal file)' > large2
   $ hg unbundle --update --config debug.dirstate.delaywrite=2 $TESTTMP/9530e27857f7-backup.hg
   adding changesets
   adding manifests
   adding file changes
   added 3 changesets with 5 changes to 5 files
   local changed .hglf/large2 which remote deleted
   use (c)hanged version or (d)elete? c
   remote turned local largefile large2 into a normal file
   keep (l)argefile or use (n)ormal file? l
   largefile large1 has a merge conflict
   ancestor was 4669e532d5b2c093a78eca010077e708a071bb64
   keep (l)ocal ba94c2efe5b7c5e0af8d189295ce00553b0612b7 or
   take (o)ther e5bb990443d6a92aaf7223813720f7566c9dd05b? l
   2 files updated, 1 files merged, 0 files removed, 0 files unresolved
 
   $ hg status -A large1
   M large1
   $ cat large1
   large1 for linear merge (conflict)
   $ cat .hglf/large1
   ba94c2efe5b7c5e0af8d189295ce00553b0612b7
   $ hg status -A large2
   A large2
   $ cat large2
   large2 for linear merge (conflict with normal file)
   $ cat .hglf/large2
   d7591fe9be0f6227d90bddf3e4f52ff41fc1f544
 
 (internal linear merging in subrepo at "hg update")
 
   $ cd ..
   $ hg init subparent
   $ cd subparent
 
   $ hg clone -q -u 2 ../repo sub
   $ cat > .hgsub <<EOF
   > sub = sub
   > EOF
   $ hg add .hgsub
   $ hg commit -m '#0@parent'
   $ cat .hgsubstate
   f74e50bd9e5594b7cf1e6c5cbab86ddd25f3ca2f sub
   $ hg -R sub update -q
   $ hg commit -m '#1@parent'
   $ cat .hgsubstate
   d65e59e952a9638e2ce863b41a420ca723dd3e8d sub
   $ hg update -q 0
 
   $ echo 'large1 for linear merge (conflict)' > sub/large1
   $ echo 'large2 for linear merge (conflict with normal file)' > sub/large2
   $ hg update --config ui.interactive=True --config debug.dirstate.delaywrite=2 <<EOF
   > m
   > r
   > c
   > l
   > l
   > EOF
    subrepository sub diverged (local revision: f74e50bd9e55, remote revision: d65e59e952a9)
   (M)erge, keep (l)ocal or keep (r)emote? m
    subrepository sources for sub differ (in checked out version)
   use (l)ocal source (f74e50bd9e55) or (r)emote source (d65e59e952a9)? r
   local changed .hglf/large2 which remote deleted
   use (c)hanged version or (d)elete? c
   remote turned local largefile large2 into a normal file
   keep (l)argefile or use (n)ormal file? l
   largefile large1 has a merge conflict
   ancestor was 4669e532d5b2c093a78eca010077e708a071bb64
   keep (l)ocal ba94c2efe5b7c5e0af8d189295ce00553b0612b7 or
   take (o)ther e5bb990443d6a92aaf7223813720f7566c9dd05b? l
   2 files updated, 1 files merged, 0 files removed, 0 files unresolved
   0 files updated, 0 files merged, 0 files removed, 0 files unresolved
 
   $ hg -R sub status -A sub/large1
   M sub/large1
   $ cat sub/large1
   large1 for linear merge (conflict)
   $ cat sub/.hglf/large1
   ba94c2efe5b7c5e0af8d189295ce00553b0612b7
   $ hg -R sub status -A sub/large2
   A sub/large2
   $ cat sub/large2
   large2 for linear merge (conflict with normal file)
   $ cat sub/.hglf/large2
   d7591fe9be0f6227d90bddf3e4f52ff41fc1f544
 
   $ cd ..
   $ cd repo
 
 Test that rebase updates largefiles in the working directory even if
 it is aborted by conflict.
 
   $ hg update -q -C 3
   $ cat .hglf/large1
   e5bb990443d6a92aaf7223813720f7566c9dd05b
   $ cat large1
   large1 in #3
   $ hg rebase -s 1 -d 3 --keep --config ui.interactive=True <<EOF
   > o
   > EOF
   largefile large1 has a merge conflict
   ancestor was 4669e532d5b2c093a78eca010077e708a071bb64
   keep (l)ocal e5bb990443d6a92aaf7223813720f7566c9dd05b or
   take (o)ther 58e24f733a964da346e2407a2bee99d9001184f5? o
   merging normal1
   warning: conflicts during merge.
   merging normal1 incomplete! (edit conflicts, then use 'hg resolve --mark')
   unresolved conflicts (see hg resolve, then hg rebase --continue)
   [1]
   $ cat .hglf/large1
   58e24f733a964da346e2407a2bee99d9001184f5
   $ cat large1
   large1 in #1
 
 Test that rebase updates standins for manually modified largefiles at
 the 1st commit of resuming.
 
   $ echo "manually modified before 'hg rebase --continue'" > large1
   $ hg resolve -m normal1
   (no more unresolved files)
   $ hg rebase --continue --config ui.interactive=True <<EOF
   > c
   > EOF
   local changed .hglf/large1 which remote deleted
   use (c)hanged version or (d)elete? c
 
   $ hg diff -c "tip~1" --nodates .hglf/large1 | grep '^[+-][0-9a-z]'
   -e5bb990443d6a92aaf7223813720f7566c9dd05b
   +8a4f783556e7dea21139ca0466eafce954c75c13
   $ rm -f large1
   $ hg update -q -C tip
   $ cat large1
   manually modified before 'hg rebase --continue'
 
 Test that transplant updates largefiles, of which standins are safely
 changed, even if it is aborted by conflict of other.
 
   $ hg update -q -C 5
   $ cat .hglf/large1
   e5bb990443d6a92aaf7223813720f7566c9dd05b
   $ cat large1
   large1 in #3
   $ hg diff -c 4 .hglf/largeX | grep '^[+-][0-9a-z]'
   +fa44618ea25181aff4f48b70428294790cec9f61
   $ hg transplant 4
   applying 07d6153b5c04
   patching file .hglf/large1
   Hunk #1 FAILED at 0
   1 out of 1 hunks FAILED -- saving rejects to file .hglf/large1.rej
   patch failed to apply
   abort: fix up the merge and run hg transplant --continue
   [255]
   $ hg status -A large1
   C large1
   $ cat .hglf/large1
   e5bb990443d6a92aaf7223813720f7566c9dd05b
   $ cat large1
   large1 in #3
   $ hg status -A largeX
   A largeX
   $ cat .hglf/largeX
   fa44618ea25181aff4f48b70428294790cec9f61
   $ cat largeX
   largeX
 
 Test that transplant updates standins for manually modified largefiles
 at the 1st commit of resuming.
 
   $ echo "manually modified before 'hg transplant --continue'" > large1
   $ hg transplant --continue
   07d6153b5c04 transplanted as f1bf30eb88cc
   $ hg diff -c tip .hglf/large1 | grep '^[+-][0-9a-z]'
   -e5bb990443d6a92aaf7223813720f7566c9dd05b
   +6a4f36d4075fbe0f30ec1d26ca44e63c05903671
   $ rm -f large1
   $ hg update -q -C tip
   $ cat large1
   manually modified before 'hg transplant --continue'
 
 Test that "hg status" doesn't show removal of largefiles not managed
 in the target context.
 
   $ hg update -q -C 4
   $ hg remove largeX
   $ hg status -A largeX
   R largeX
   $ hg status -A --rev '.^1' largeX
 
 #if execbit
 
 Test that "hg status" against revisions other than parent notices exec
 bit changes of largefiles.
 
   $ hg update -q -C 4
 
 (the case that large2 doesn't have exec bit in the target context but
 in the working context)
 
   $ chmod +x large2
   $ hg status -A --rev 0 large2
   M large2
   $ hg commit -m 'chmod +x large2'
 
 (the case that large2 has exec bit in the target context but not in
 the working context)
 
   $ echo dummy > dummy
   $ hg add dummy
   $ hg commit -m 'revision for separation'
   $ chmod -x large2
   $ hg status -A --rev '.^1' large2
   M large2
 
 #endif
 
   $ cd ..
+
+Test that "hg convert" avoids copying largefiles from the working
+directory into store, because "hg convert" doesn't update largefiles
+in the working directory (removing files under ".cache/largefiles"
+forces "hg convert" to copy corresponding largefiles)
+
+  $ cat >> $HGRCPATH <<EOF
+  > [extensions]
+  > convert =
+  > EOF
+
+  $ rm $TESTTMP/.cache/largefiles/6a4f36d4075fbe0f30ec1d26ca44e63c05903671
+  $ hg convert -q repo repo.converted