largefiles: move basestore._openstore into new module to remove cycle

hgext/largefiles/basestore.py

@@ -1,226 +1,164
 # 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.
 
 '''base class for store implementations and store-related utility code'''
 
-import re
-
-from mercurial import util, node, hg, error
+from mercurial import util, node
 from mercurial.i18n import _
 
 import lfutil
 
 class StoreError(Exception):
     '''Raised when there is a problem getting files from or putting
     files to a central store.'''
     def __init__(self, filename, hash, url, detail):
         self.filename = filename
         self.hash = hash
         self.url = url
         self.detail = detail
 
     def longmessage(self):
         return (_("error getting id %s from url %s for file %s: %s\n") %
                  (self.hash, util.hidepassword(self.url), self.filename,
                   self.detail))
 
     def __str__(self):
         return "%s: %s" % (util.hidepassword(self.url), self.detail)
 
 class basestore(object):
     def __init__(self, ui, repo, url):
         self.ui = ui
         self.repo = repo
         self.url = url
 
     def put(self, source, hash):
         '''Put source file into the store so it can be retrieved by hash.'''
         raise NotImplementedError('abstract method')
 
     def exists(self, hashes):
         '''Check to see if the store contains the given hashes. Given an
         iterable of hashes it returns a mapping from hash to bool.'''
         raise NotImplementedError('abstract method')
 
     def get(self, files):
         '''Get the specified largefiles from the store and write to local
         files under repo.root.  files is a list of (filename, hash)
         tuples.  Return (success, missing), lists of files successfully
         downloaded and those not found in the store.  success is a list
         of (filename, hash) tuples; missing is a list of filenames that
         we could not get.  (The detailed error message will already have
         been presented to the user, so missing is just supplied as a
         summary.)'''
         success = []
         missing = []
         ui = self.ui
 
         at = 0
         available = self.exists(set(hash for (_filename, hash) in files))
         for filename, hash in files:
             ui.progress(_('getting largefiles'), at, unit=_('files'),
                 total=len(files))
             at += 1
             ui.note(_('getting %s:%s\n') % (filename, hash))
 
             if not available.get(hash):
                 ui.warn(_('%s: largefile %s not available from %s\n')
                         % (filename, hash, util.hidepassword(self.url)))
                 missing.append(filename)
                 continue
 
             if self._gethash(filename, hash):
                 success.append((filename, hash))
             else:
                 missing.append(filename)
 
         ui.progress(_('getting largefiles'), None)
         return (success, missing)
 
     def _gethash(self, filename, hash):
         """Get file with the provided hash and store it in the local repo's
         store and in the usercache.
         filename is for informational messages only.
         """
         util.makedirs(lfutil.storepath(self.repo, ''))
         storefilename = lfutil.storepath(self.repo, hash)
 
         tmpname = storefilename + '.tmp'
         tmpfile = util.atomictempfile(tmpname,
                                       createmode=self.repo.store.createmode)
 
         try:
             gothash = self._getfile(tmpfile, filename, hash)
         except StoreError as err:
             self.ui.warn(err.longmessage())
             gothash = ""
         tmpfile.close()
 
         if gothash != hash:
             if gothash != "":
                 self.ui.warn(_('%s: data corruption (expected %s, got %s)\n')
                              % (filename, hash, gothash))
             util.unlink(tmpname)
             return False
 
         util.rename(tmpname, storefilename)
         lfutil.linktousercache(self.repo, hash)
         return True
 
     def verify(self, revs, contents=False):
         '''Verify the existence (and, optionally, contents) of every big
         file revision referenced by every changeset in revs.
         Return 0 if all is well, non-zero on any errors.'''
 
         self.ui.status(_('searching %d changesets for largefiles\n') %
                        len(revs))
         verified = set()                # set of (filename, filenode) tuples
         filestocheck = []               # list of (cset, filename, expectedhash)
         for rev in revs:
             cctx = self.repo[rev]
             cset = "%d:%s" % (cctx.rev(), node.short(cctx.node()))
 
             for standin in cctx:
                 filename = lfutil.splitstandin(standin)
                 if filename:
                     fctx = cctx[standin]
                     key = (filename, fctx.filenode())
                     if key not in verified:
                         verified.add(key)
                         expectedhash = fctx.data()[0:40]
                         filestocheck.append((cset, filename, expectedhash))
 
         failed = self._verifyfiles(contents, filestocheck)
 
         numrevs = len(verified)
         numlfiles = len(set([fname for (fname, fnode) in verified]))
         if contents:
             self.ui.status(
                 _('verified contents of %d revisions of %d largefiles\n')
                 % (numrevs, numlfiles))
         else:
             self.ui.status(
                 _('verified existence of %d revisions of %d largefiles\n')
                 % (numrevs, numlfiles))
         return int(failed)
 
     def _getfile(self, tmpfile, filename, hash):
         '''Fetch one revision of one file from the store and write it
         to tmpfile.  Compute the hash of the file on-the-fly as it
         downloads and return the hash.  Close tmpfile.  Raise
         StoreError if unable to download the file (e.g. it does not
         exist in the store).'''
         raise NotImplementedError('abstract method')
 
     def _verifyfiles(self, contents, filestocheck):
         '''Perform the actual verification of files in the store.
         'contents' controls verification of content hash.
         'filestocheck' is list of files to check.
         Returns _true_ if any problems are found!
         '''
         raise NotImplementedError('abstract method')
-
-import localstore, wirestore
-
-_storeprovider = {
-    'file':  [localstore.localstore],
-    'http':  [wirestore.wirestore],
-    'https': [wirestore.wirestore],
-    'ssh': [wirestore.wirestore],
-    }
-
-_scheme_re = re.compile(r'^([a-zA-Z0-9+-.]+)://')
-
-# During clone this function is passed the src's ui object
-# but it needs the dest's ui object so it can read out of
-# the config file. Use repo.ui instead.
-def _openstore(repo, remote=None, put=False):
-    ui = repo.ui
-
-    if not remote:
-        lfpullsource = getattr(repo, 'lfpullsource', None)
-        if lfpullsource:
-            path = ui.expandpath(lfpullsource)
-        elif put:
-            path = ui.expandpath('default-push', 'default')
-        else:
-            path = ui.expandpath('default')
-
-        # ui.expandpath() leaves 'default-push' and 'default' alone if
-        # they cannot be expanded: fallback to the empty string,
-        # meaning the current directory.
-        if path == 'default-push' or path == 'default':
-            path = ''
-            remote = repo
-        else:
-            path, _branches = hg.parseurl(path)
-            remote = hg.peer(repo, {}, path)
-
-    # The path could be a scheme so use Mercurial's normal functionality
-    # to resolve the scheme to a repository and use its path
-    path = util.safehasattr(remote, 'url') and remote.url() or remote.path
-
-    match = _scheme_re.match(path)
-    if not match:                       # regular filesystem path
-        scheme = 'file'
-    else:
-        scheme = match.group(1)
-
-    try:
-        storeproviders = _storeprovider[scheme]
-    except KeyError:
-        raise error.Abort(_('unsupported URL scheme %r') % scheme)
-
-    for classobj in storeproviders:
-        try:
-            return classobj(ui, repo, remote)
-        except lfutil.storeprotonotcapable:
-            pass
-
-    raise error.Abort(_('%s does not appear to be a largefile store') %
-                     util.hidepassword(path))

hgext/largefiles/lfcommands.py

@@ -1,550 +1,550
 # 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
 
 from hgext.convert import convcmd
 from hgext.convert import filemap
 
 import lfutil
-import basestore
+import storefactory
 
 # -- 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 error.Abort(_('%s is not a local Mercurial repo') % src)
     if not hg.islocal(dest):
         raise error.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:
         # 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:
             # 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()
 
             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=_('revisions'), total=rsrc['tip'].rev())
                 _lfconvert_addchangeset(rsrc, rdst, ctx, revmap,
                     lfiles, normalfiles, matcher, size, lfiletohash)
             ui.progress(_('converting revisions'), None)
 
             if rdst.wvfs.exists(lfutil.shortname):
                 rdst.wvfs.rmtree(lfutil.shortname)
 
             for f in lfiletohash.keys():
                 if rdst.wvfs.isfile(f):
                     rdst.wvfs.unlink(f)
                 try:
                     rdst.wvfs.removedirs(rdst.wvfs.dirname(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:
             class lfsource(filemap.filemap_source):
                 def __init__(self, ui, source):
                     super(lfsource, self).__init__(ui, source, None)
                     self.filemapper.rename[lfutil.shortname] = '.'
 
                 def getfile(self, name, rev):
                     realname, realrev = rev
                     f = super(lfsource, self).getfile(name, rev)
 
                     if (not realname.startswith(lfutil.shortnameslash)
                             or f[0] is None):
                         return f
 
                     # Substitute in the largefile data for the hash
                     hash = f[0].strip()
                     path = lfutil.findfile(rsrc, hash)
 
                     if path is None:
                         raise error.Abort(_("missing largefile for '%s' in %s")
                                           % (realname, realrev))
                     return util.readfile(path), f[1]
 
             class converter(convcmd.converter):
                 def __init__(self, ui, source, dest, revmapfile, opts):
                     src = lfsource(ui, source)
 
                     super(converter, self).__init__(ui, src, dest, revmapfile,
                                                     opts)
 
             found, missing = downloadlfiles(ui, rsrc)
             if missing != 0:
                 raise error.Abort(_("all largefiles must be present locally"))
 
             orig = convcmd.converter
             convcmd.converter = converter
 
             try:
                 convcmd.convert(ui, src, dest)
             finally:
                 convcmd.converter = orig
         success = True
     finally:
         if tolfile:
             rdst.dirstate.clear()
             release(dstlock, dstwlock)
         if not success:
             # we failed, remove the new directory
             shutil.rmtree(rdst.root)
 
 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 error.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 error.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)
+    store = storefactory._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=_('files'),
                     total=len(files))
         source = lfutil.findfile(rsrc, hash)
         if not source:
             raise error.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:
         revs = repo.revs('all()')
     else:
         revs = ['.']
 
-    store = basestore._openstore(repo)
+    store = storefactory._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 as 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)
+        store = storefactory._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)
     with repo.wlock():
         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
         wvfs = repo.wvfs
         for lfile in lfiles:
             rellfile = lfile
             rellfileorig = os.path.relpath(
                 scmutil.origpath(ui, repo, wvfs.join(rellfile)),
                 start=repo.root)
             relstandin = lfutil.standin(lfile)
             relstandinorig = os.path.relpath(
                 scmutil.origpath(ui, repo, wvfs.join(relstandin)),
                 start=repo.root)
             if wvfs.exists(relstandin):
                 if (wvfs.exists(relstandinorig) and
                     wvfs.exists(rellfile)):
                     shutil.copyfile(wvfs.join(rellfile),
                                     wvfs.join(rellfileorig))
                     wvfs.unlinkpath(relstandinorig)
                 expecthash = lfutil.readstandin(repo, lfile)
                 if expecthash != '':
                     if lfile not in repo[None]: # not switched to normal file
                         wvfs.unlinkpath(rellfile, ignoremissing=True)
                     # use normallookup() to allocate an entry in largefiles
                     # dirstate to prevent lfilesrepo.status() from reporting
                     # missing files as 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 (wvfs.exists(rellfile) and
                     repo.dirstate.normalize(lfile) not in repo[None]):
                     wvfs.unlinkpath(rellfile)
                     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.
             rellfile = lfile
             relstandin = lfutil.standin(lfile)
             if wvfs.exists(relstandin):
                 mode = wvfs.stat(relstandin).st_mode
                 if mode != wvfs.stat(rellfile).st_mode:
                     wvfs.chmod(rellfile, mode)
                     update1 = 1
 
             updated += update1
 
             lfutil.synclfdirstate(repo, lfdirstate, lfile, normallookup)
 
         lfdirstate.write()
         if lfiles:
             statuswriter(_('%d largefiles updated, %d removed\n') % (updated,
                 removed))
 
 @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 error.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/overrides.py

@@ -1,1419 +1,1419
 # 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, registrar, revset, error
 from mercurial.i18n import _
 
 import lfutil
 import lfcommands
-import basestore
+import storefactory
 
 # -- Utility functions: commonly/repeatedly needed functionality ---------------
 
 def composelargefilematcher(match, manifest):
     '''create a matcher that matches only the largefiles in the original
     matcher'''
     m = copy.copy(match)
     lfile = lambda f: lfutil.standin(f) in manifest
     m._files = filter(lfile, m._files)
     m._fileroots = set(m._files)
     m._always = False
     origmatchfn = m.matchfn
     m.matchfn = lambda f: lfile(f) and origmatchfn(f)
     return m
 
 def composenormalfilematcher(match, manifest, exclude=None):
     excluded = set()
     if exclude is not None:
         excluded.update(exclude)
 
     m = copy.copy(match)
     notlfile = lambda f: not (lfutil.isstandin(f) or lfutil.standin(f) in
             manifest or f in excluded)
     m._files = filter(notlfile, m._files)
     m._fileroots = set(m._files)
     m._always = False
     origmatchfn = m.matchfn
     m.matchfn = lambda f: notlfile(f) and origmatchfn(f)
     return m
 
 def installnormalfilesmatchfn(manifest):
     '''installmatchfn with a matchfn that ignores all largefiles'''
     def overridematch(ctx, pats=(), opts=None, globbed=False,
             default='relpath', badfn=None):
         if opts is None:
             opts = {}
         match = oldmatch(ctx, pats, opts, globbed, default, badfn=badfn)
         return composenormalfilematcher(match, manifest)
     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 the original matchfn.'''
     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 the original matchfn.'''
     scmutil.matchandpats = getattr(scmutil.matchandpats, 'oldmatchandpats',
             scmutil.matchandpats)
 
 def addlargefiles(ui, repo, isaddremove, matcher, **opts):
     large = opts.get('large')
     lfsize = lfutil.getminsize(
         ui, lfutil.islfilesrepo(repo), opts.get('lfsize'))
 
     lfmatcher = None
     if lfutil.islfilesrepo(repo):
         lfpats = ui.configlist(lfutil.longname, 'patterns', default=[])
         if lfpats:
             lfmatcher = match_.match(repo.root, '', list(lfpats))
 
     lfnames = []
     m = matcher
 
     wctx = repo[None]
     for f in repo.walk(match_.badmatch(m, lambda x, y: None)):
         exact = m.exact(f)
         lfile = lfutil.standin(f) in wctx
         nfile = f in wctx
         exists = lfile or nfile
 
         # addremove in core gets fancy with the name, add doesn't
         if isaddremove:
             name = m.uipath(f)
         else:
             name = m.rel(f)
 
         # 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') % name)
             continue
 
         if (exact or not exists) and not lfutil.isstandin(f):
             # In case the file was removed previously, but not committed
             # (issue3507)
             if not repo.wvfs.exists(f):
                 continue
 
             abovemin = (lfsize and
                         repo.wvfs.lstat(f).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') % name)
 
     bad = []
 
     # Need to lock, otherwise there could be a race condition between
     # when standins are created and added to the repo.
     with repo.wlock():
         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()]
 
         added = [f for f in lfnames if f not in bad]
     return added, bad
 
 def removelargefiles(ui, repo, isaddremove, matcher, **opts):
     after = opts.get('after')
     m = composelargefilematcher(matcher, repo[None].manifest())
     try:
         repo.lfstatus = True
         s = repo.status(match=m, clean=not isaddremove)
     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
 
     # Need to lock because standin files are deleted then removed from the
     # repository and we could race in-between.
     with repo.wlock():
         lfdirstate = lfutil.openlfdirstate(ui, repo)
         for f in sorted(remove):
             if ui.verbose or not m.exact(f):
                 # addremove in core gets fancy with the name, remove doesn't
                 if isaddremove:
                     name = m.uipath(f)
                 else:
                     name = m.rel(f)
                 ui.status(_('removing %s\n') % name)
 
             if not opts.get('dry_run'):
                 if not after:
                     util.unlinkpath(repo.wjoin(f), ignoremissing=True)
 
         if opts.get('dry_run'):
             return result
 
         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)
 
         for f in remove:
             lfutil.synclfdirstate(repo, lfdirstate, lfutil.splitstandin(f),
                                   False)
 
         lfdirstate.write()
 
     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 --------------------------------
 
 def overrideadd(orig, ui, repo, *pats, **opts):
     if opts.get('normal') and opts.get('large'):
         raise error.Abort(_('--normal cannot be used with --large'))
     return orig(ui, repo, *pats, **opts)
 
 def cmdutiladd(orig, ui, repo, matcher, prefix, explicitonly, **opts):
     # The --normal flag short circuits this override
     if opts.get('normal'):
         return orig(ui, repo, matcher, prefix, explicitonly, **opts)
 
     ladded, lbad = addlargefiles(ui, repo, False, matcher, **opts)
     normalmatcher = composenormalfilematcher(matcher, repo[None].manifest(),
                                              ladded)
     bad = orig(ui, repo, normalmatcher, prefix, explicitonly, **opts)
 
     bad.extend(f for f in lbad)
     return bad
 
 def cmdutilremove(orig, ui, repo, matcher, prefix, after, force, subrepos):
     normalmatcher = composenormalfilematcher(matcher, repo[None].manifest())
     result = orig(ui, repo, normalmatcher, prefix, after, force, subrepos)
     return removelargefiles(ui, repo, False, matcher, after=after,
                             force=force) 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=None, globbed=False,
             default='relpath', badfn=None):
         """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.
         """
         if opts is None:
             opts = {}
         matchandpats = oldmatchandpats(ctx, pats, opts, globbed, default,
                                        badfn=badfn)
         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)
 
         def fixpats(pat, tostandin=lfutil.standin):
             if pat.startswith('set:'):
                 return pat
 
             kindpat = match_._patsplit(pat, None)
 
             if kindpat[0] is not None:
                 return kindpat[0] + ':' + tostandin(kindpat[1])
             return tostandin(kindpat[1])
 
         if m._cwd:
             hglf = lfutil.shortname
             back = util.pconvert(m.rel(hglf)[:-len(hglf)])
 
             def tostandin(f):
                 # The file may already be a standin, so truncate the back
                 # prefix and test before mangling it.  This avoids turning
                 # 'glob:../.hglf/foo*' into 'glob:../.hglf/../.hglf/foo*'.
                 if f.startswith(back) and lfutil.splitstandin(f[len(back):]):
                     return f
 
                 # An absolute path is from outside the repo, so truncate the
                 # path to the root before building the standin.  Otherwise cwd
                 # is somewhere in the repo, relative to root, and needs to be
                 # prepended before building the standin.
                 if os.path.isabs(m._cwd):
                     f = f[len(back):]
                 else:
                     f = m._cwd + '/' + f
                 return back + lfutil.standin(f)
 
             pats.update(fixpats(f, tostandin) for f in p)
         else:
             def tostandin(f):
                 if lfutil.splitstandin(f):
                     return f
                 return lfutil.standin(f)
             pats.update(fixpats(f, tostandin) for f in p)
 
         for i in range(0, len(m._files)):
             # Don't add '.hglf' to m.files, since that is already covered by '.'
             if m._files[i] == '.':
                 continue
             standin = lfutil.standin(m._files[i])
             # If the "standin" is a directory, append instead of replace to
             # support naming a directory on the command line with only
             # largefiles.  The original directory is kept to support normal
             # files.
             if standin in repo[ctx.node()]:
                 m._files[i] = standin
             elif m._files[i] not in repo[ctx.node()] \
                     and repo.wvfs.isdir(standin):
                 m._files.append(standin)
 
         m._fileroots = 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
 
         ui.debug('updated patterns: %s\n' % sorted(pats))
         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, badfn=None):
         wctx = repo[None]
         match, pats = oldmatchandpats(wctx, pats, opts, badfn=badfn)
         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)
 
 # Before starting the manifest merge, merge.updates will call
 # _checkunknownfile 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 calculateupdates function below.
 def overridecheckunknownfile(origfn, repo, wctx, mctx, f, f2=None):
     if lfutil.standin(repo.dirstate.normalize(f)) in wctx:
         return False
     return origfn(repo, wctx, mctx, f, f2)
 
 # 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 calculateupdates 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,
                              acceptremote, *args, **kwargs):
     overwrite = force and not branchmerge
     actions, diverge, renamedelete = origfn(
         repo, p1, p2, pas, branchmerge, force, acceptremote, *args, **kwargs)
 
     if overwrite:
         return actions, diverge, renamedelete
 
     # Convert to dictionary with filename as key and action as value.
     lfiles = set()
     for f in actions:
         splitstandin = lfutil.splitstandin(f)
         if splitstandin in p1:
             lfiles.add(splitstandin)
         elif lfutil.standin(f) in p1:
             lfiles.add(f)
 
     for lfile in sorted(lfiles):
         standin = lfutil.standin(lfile)
         (lm, largs, lmsg) = actions.get(lfile, (None, None, None))
         (sm, sargs, smsg) = actions.get(standin, (None, None, None))
         if sm in ('g', 'dc') and lm != 'r':
             if sm == 'dc':
                 f1, f2, fa, move, anc = sargs
                 sargs = (p2[f2].flags(), False)
             # Case 1: normal file in the working copy, largefile in
             # the second parent
             usermsg = _('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(usermsg, 0) == 0: # pick remote largefile
                 actions[lfile] = ('r', None, 'replaced by standin')
                 actions[standin] = ('g', sargs, 'replaces standin')
             else: # keep local normal file
                 actions[lfile] = ('k', None, 'replaces standin')
                 if branchmerge:
                     actions[standin] = ('k', None, 'replaced by non-standin')
                 else:
                     actions[standin] = ('r', None, 'replaced by non-standin')
         elif lm in ('g', 'dc') and sm != 'r':
             if lm == 'dc':
                 f1, f2, fa, move, anc = largs
                 largs = (p2[f2].flags(), False)
             # Case 2: largefile in the working copy, normal file in
             # the second parent
             usermsg = _('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(usermsg, 0) == 0: # keep local largefile
                 if branchmerge:
                     # largefile can be restored from standin safely
                     actions[lfile] = ('k', None, 'replaced by standin')
                     actions[standin] = ('k', None, 'replaces standin')
                 else:
                     # "lfile" should be marked as "removed" without
                     # removal of itself
                     actions[lfile] = ('lfmr', None,
                                       'forget non-standin largefile')
 
                     # linear-merge should treat this largefile as 're-added'
                     actions[standin] = ('a', None, 'keep standin')
             else: # pick remote normal file
                 actions[lfile] = ('g', largs, 'replaces standin')
                 actions[standin] = ('r', None, 'replaced by non-standin')
 
     return actions, diverge, renamedelete
 
 def mergerecordupdates(orig, repo, actions, branchmerge):
     if 'lfmr' in actions:
         lfdirstate = lfutil.openlfdirstate(repo.ui, repo)
         for lfile, args, msg in actions['lfmr']:
             # this should be executed before 'orig', to execute 'remove'
             # before all other actions
             repo.dirstate.remove(lfile)
             # make sure lfile doesn't get synclfdirstate'd as normal
             lfdirstate.add(lfile)
         lfdirstate.write()
 
     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, premerge, repo, mynode, orig, fcd, fco, fca,
                       labels=None):
     if not lfutil.isstandin(orig) or fcd.isabsent() or fco.isabsent():
         return origfn(premerge, 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 True, 0, False
 
 def copiespathcopies(orig, ctx1, ctx2, match=None):
     copies = orig(ctx1, ctx2, match=match)
     updated = {}
 
     for k, v in copies.iteritems():
         updated[lfutil.splitstandin(k) or k] = lfutil.splitstandin(v) or v
 
     return updated
 
 # 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)
 
     # 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:
         result = orig(ui, repo, pats, opts, rename)
     except error.Abort as 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
 
     def makestandin(relpath):
         path = pathutil.canonpath(repo.root, repo.getcwd(), relpath)
         return repo.wvfs.join(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))
 
     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=None, globbed=False,
                 default='relpath', badfn=None):
             if opts is None:
                 opts = {}
             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, badfn=badfn)
             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._fileroots = 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 = repo.wvfs.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 error.Abort as e:
         if str(e) != _('no files to copy'):
             raise e
         else:
             nolfiles = True
     finally:
         restorematchfn()
         wlock.release()
 
     if nolfiles and nonormalfiles:
         raise error.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, ctx, parents, *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.
     with repo.wlock():
         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 (repo.wvfs.exists(lfutil.standin(lfile))):
                 repo.wvfs.unlink(lfutil.standin(lfile))
 
         oldstandins = lfutil.getstandinsstate(repo)
 
         def overridematch(mctx, pats=(), opts=None, globbed=False,
                 default='relpath', badfn=None):
             if opts is None:
                 opts = {}
             match = oldmatch(mctx, pats, opts, globbed, default, badfn=badfn)
             m = copy.copy(match)
 
             # revert supports recursing into subrepos, and though largefiles
             # currently doesn't work correctly in that case, this match is
             # called, so the lfdirstate above may not be the correct one for
             # this invocation of match.
             lfdirstate = lfutil.openlfdirstate(mctx.repo().ui, mctx.repo(),
                                                False)
 
             def tostandin(f):
                 standin = lfutil.standin(f)
                 if standin in ctx or standin in mctx:
                     return standin
                 elif standin in repo[None] or lfdirstate[f] == 'r':
                     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._fileroots = set(m._files)
             origmatchfn = m.matchfn
             def matchfn(f):
                 if lfutil.isstandin(f):
                     return (origmatchfn(lfutil.splitstandin(f)) and
                             (f in ctx or f in mctx))
                 return origmatchfn(f)
             m.matchfn = matchfn
             return m
         oldmatch = installmatchfn(overridematch)
         try:
             orig(ui, repo, ctx, parents, *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)
 
 # 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 overridepush(orig, ui, repo, *args, **kwargs):
     """Override push command and store --lfrev parameters in opargs"""
     lfrevs = kwargs.pop('lfrev', None)
     if lfrevs:
         opargs = kwargs.setdefault('opargs', {})
         opargs['lfrevs'] = scmutil.revrange(repo, lfrevs)
     return orig(ui, repo, *args, **kwargs)
 
 def exchangepushoperation(orig, *args, **kwargs):
     """Override pushoperation constructor and store lfrevs parameter"""
     lfrevs = kwargs.pop('lfrevs', None)
     pushop = orig(*args, **kwargs)
     pushop.lfrevs = lfrevs
     return pushop
 
 revsetpredicate = registrar.revsetpredicate()
 
 @revsetpredicate('pulled()')
 def pulledrevsetsymbol(repo, subset, x):
     """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 error.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 error.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()
 
         # When cloning to a remote repo (like through SSH), no repo is available
         # from the peer.   Therefore the largefiles can't be downloaded and the
         # hgrc can't be updated.
         if not repo:
             return result
 
         # If largefiles is required for this repo, permanently enable it locally
         if 'largefiles' in repo.requirements:
             fp = repo.vfs('hgrc', 'a', text=True)
             try:
                 fp.write('\n[extensions]\nlargefiles=\n')
             finally:
                 fp.close()
 
         # 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):
     if not util.safehasattr(repo, '_largefilesenabled'):
         return 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 overridearchivecmd(orig, ui, repo, dest, **opts):
     repo.unfiltered().lfstatus = True
 
     try:
         return orig(ui, repo.unfiltered(), dest, **opts)
     finally:
         repo.unfiltered().lfstatus = False
 
 def hgwebarchive(orig, web, req, tmpl):
     web.repo.lfstatus = True
 
     try:
         return orig(web, req, tmpl)
     finally:
         web.repo.lfstatus = False
 
 def overridearchive(orig, repo, dest, node, kind, decode=True, matchfn=None,
             prefix='', mtime=None, subrepos=None):
     # For some reason setting repo.lfstatus in hgwebarchive only changes the
     # unfiltered repo's attr, so check that as well.
     if not repo.lfstatus and not repo.unfiltered().lfstatus:
         return orig(repo, dest, node, kind, decode, matchfn, prefix, mtime,
                     subrepos)
 
     # No need to lock because we are only reading history and
     # largefile caches, neither of which are modified.
     if node is not None:
         lfcommands.cachelfiles(repo.ui, repo, node)
 
     if kind not in archival.archivers:
         raise error.Abort(_("unknown archive type '%s'") % kind)
 
     ctx = repo[node]
 
     if kind == 'files':
         if prefix:
             raise error.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):
         write('.hg_archival.txt', 0o644, False,
               lambda: archival.buildmetadata(ctx))
 
     for f in ctx:
         ff = ctx.flags(f)
         getdata = ctx[f].data
         if lfutil.isstandin(f):
             if node is not None:
                 path = lfutil.findfile(repo, getdata().strip())
 
                 if path is None:
                     raise error.Abort(
                        _('largefile %s not found in repo store or system cache')
                        % lfutil.splitstandin(f))
             else:
                 path = lfutil.splitstandin(f)
 
             f = lfutil.splitstandin(f)
 
             getdata = lambda: util.readfile(path)
         write(f, 'x' in ff and 0o755 or 0o644, 'l' in ff, getdata)
 
     if subrepos:
         for subpath in sorted(ctx.substate):
             sub = ctx.workingsub(subpath)
             submatch = match_.subdirmatcher(subpath, matchfn)
             sub._repo.lfstatus = True
             sub.archive(archiver, prefix, submatch)
 
     archiver.done()
 
 def hgsubrepoarchive(orig, repo, archiver, prefix, match=None):
     if not repo._repo.lfstatus:
         return orig(repo, archiver, prefix, match)
 
     repo._get(repo._state + ('hg',))
     rev = repo._state[1]
     ctx = repo._repo[rev]
 
     if ctx.node() is not None:
         lfcommands.cachelfiles(repo.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):
             if ctx.node() is not None:
                 path = lfutil.findfile(repo._repo, getdata().strip())
 
                 if path is None:
                     raise error.Abort(
                        _('largefile %s not found in repo store or system cache')
                        % lfutil.splitstandin(f))
             else:
                 path = lfutil.splitstandin(f)
 
             f = lfutil.splitstandin(f)
 
             getdata = lambda: util.readfile(os.path.join(prefix, path))
 
         write(f, 'x' in ff and 0o755 or 0o644, 'l' in ff, getdata)
 
     for subpath in sorted(ctx.substate):
         sub = ctx.workingsub(subpath)
         submatch = match_.subdirmatcher(subpath, match)
         sub._repo.lfstatus = True
         sub.archive(archiver, 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, backout and fetch.
 def overridebailifchanged(orig, repo, *args, **kwargs):
     orig(repo, *args, **kwargs)
     repo.lfstatus = True
     s = repo.status()
     repo.lfstatus = False
     if s.modified or s.added or s.removed or s.deleted:
         raise error.Abort(_('uncommitted changes'))
 
 def postcommitstatus(orig, repo, *args, **kwargs):
     repo.lfstatus = True
     try:
         return orig(repo, *args, **kwargs)
     finally:
         repo.lfstatus = False
 
 def cmdutilforget(orig, ui, repo, match, prefix, explicitonly):
     normalmatcher = composenormalfilematcher(match, repo[None].manifest())
     bad, forgot = orig(ui, repo, normalmatcher, prefix, explicitonly)
     m = composelargefilematcher(match, repo[None].manifest())
 
     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 \
                 repo.wvfs.isdir(lfutil.standin(f)):
             ui.warn(_('not removing %s: file is already untracked\n')
                     % m.rel(f))
             bad.append(f)
 
     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.
     with repo.wlock():
         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)
         rejected = repo[None].forget(standins)
 
     bad.extend(f for f in rejected if f in m.files())
     forgot.extend(f for f in forget if f not in rejected)
     return bad, forgot
 
 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)
+        lfexists = storefactory._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, matcher, prefix, opts=None, dry_run=None,
                      similarity=None):
     if opts is None:
         opts = {}
     if not lfutil.islfilesrepo(repo):
         return orig(repo, matcher, prefix, 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 = copy.copy(matcher)
 
         # The m._files and m._map attributes are not changed to the deleted list
         # because that affects the m.exact() test, which in turn governs whether
         # or not the file name is printed, and how.  Simply limit the original
         # matches to those in the deleted status list.
         matchfn = m.matchfn
         m.matchfn = lambda f: f in s.deleted and matchfn(f)
 
         removelargefiles(repo.ui, repo, True, m, **opts)
     # Call into the normal add code, and any files that *should* be added as
     # largefiles will be
     added, bad = addlargefiles(repo.ui, repo, True, matcher, **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 passing a matcher that will ignore them.
     matcher = composenormalfilematcher(matcher, repo[None].manifest(), added)
     return orig(repo, matcher, prefix, opts, dry_run, similarity)
 
 # 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 Monkey patching a repoview will not work. The assigned attribute will
     # be set on the unfiltered repo, but we will only lookup attributes in the
     # unfiltered repo if the lookup in the repoview object itself fails. As the
     # monkey patched method exists on the repoview class the lookup will not
     # fail. As a result, the original version will shadow the monkey patched
     # one, defeating the monkey patch.
     #
     # As a work around we use an unfiltered repo here. We should do something
     # cleaner instead.
     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):
     with repo.wlock():
         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()
     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
 
     origvisitdirfn = m.visitdir
     def lfvisitdirfn(dir):
         if dir == lfutil.shortname:
             return True
         ret = origvisitdirfn(dir)
         if ret:
             return ret
         lf = lfutil.splitstandin(dir)
         if lf is None:
             return False
         return origvisitdirfn(lf)
     m.visitdir = lfvisitdirfn
 
     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)
+                store = storefactory._openstore(repo)
                 success, missing = store.get([(lf, hash)])
                 if len(success) != 1:
                     raise error.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 mergeupdate(orig, repo, node, branchmerge, force,
                 *args, **kwargs):
     matcher = kwargs.get('matcher', None)
     # note if this is a partial update
     partial = matcher and not matcher.always()
     with repo.wlock():
         # 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")
 
         lfdirstate = lfutil.openlfdirstate(repo.ui, repo)
         unsure, s = lfdirstate.status(match_.always(repo.root,
                                                     repo.getcwd()),
                                       [], False, False, False)
         pctx = repo['.']
         for lfile in unsure + s.modified:
             lfileabs = repo.wvfs.join(lfile)
             if not repo.wvfs.exists(lfileabs):
                 continue
             lfhash = lfutil.hashrepofile(repo, lfile)
             standin = lfutil.standin(lfile)
             lfutil.writestandin(repo, standin, lfhash,
                                 lfutil.getexecutable(lfileabs))
             if (standin in pctx and
                 lfhash == lfutil.readstandin(repo, lfile, '.')):
                 lfdirstate.normal(lfile)
         for lfile in s.added:
             lfutil.updatestandin(repo, lfutil.standin(lfile))
         lfdirstate.write()
 
         oldstandins = lfutil.getstandinsstate(repo)
 
         result = orig(repo, node, branchmerge, force, *args, **kwargs)
 
         newstandins = lfutil.getstandinsstate(repo)
         filelist = lfutil.getlfilestoupdate(oldstandins, newstandins)
         if branchmerge or force or partial:
             filelist.extend(s.deleted + s.removed)
 
         lfcommands.updatelfiles(repo.ui, repo, filelist=filelist,
                                 normallookup=partial)
 
         return result
 
 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/storefactory.py

@@ -1,226 +1,78
-# 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.
 
-'''base class for store implementations and store-related utility code'''
+from __future__ import absolute_import
 
 import re
 
-from mercurial import util, node, hg, error
 from mercurial.i18n import _
 
-import lfutil
-
-class StoreError(Exception):
-    '''Raised when there is a problem getting files from or putting
-    files to a central store.'''
-    def __init__(self, filename, hash, url, detail):
-        self.filename = filename
-        self.hash = hash
-        self.url = url
-        self.detail = detail
-
-    def longmessage(self):
-        return (_("error getting id %s from url %s for file %s: %s\n") %
-                 (self.hash, util.hidepassword(self.url), self.filename,
-                  self.detail))
-
-    def __str__(self):
-        return "%s: %s" % (util.hidepassword(self.url), self.detail)
-
-class basestore(object):
-    def __init__(self, ui, repo, url):
-        self.ui = ui
-        self.repo = repo
-        self.url = url
-
-    def put(self, source, hash):
-        '''Put source file into the store so it can be retrieved by hash.'''
-        raise NotImplementedError('abstract method')
-
-    def exists(self, hashes):
-        '''Check to see if the store contains the given hashes. Given an
-        iterable of hashes it returns a mapping from hash to bool.'''
-        raise NotImplementedError('abstract method')
-
-    def get(self, files):
-        '''Get the specified largefiles from the store and write to local
-        files under repo.root.  files is a list of (filename, hash)
-        tuples.  Return (success, missing), lists of files successfully
-        downloaded and those not found in the store.  success is a list
-        of (filename, hash) tuples; missing is a list of filenames that
-        we could not get.  (The detailed error message will already have
-        been presented to the user, so missing is just supplied as a
-        summary.)'''
-        success = []
-        missing = []
-        ui = self.ui
-
-        at = 0
-        available = self.exists(set(hash for (_filename, hash) in files))
-        for filename, hash in files:
-            ui.progress(_('getting largefiles'), at, unit=_('files'),
-                total=len(files))
-            at += 1
-            ui.note(_('getting %s:%s\n') % (filename, hash))
-
-            if not available.get(hash):
-                ui.warn(_('%s: largefile %s not available from %s\n')
-                        % (filename, hash, util.hidepassword(self.url)))
-                missing.append(filename)
-                continue
-
-            if self._gethash(filename, hash):
-                success.append((filename, hash))
-            else:
-                missing.append(filename)
-
-        ui.progress(_('getting largefiles'), None)
-        return (success, missing)
-
-    def _gethash(self, filename, hash):
-        """Get file with the provided hash and store it in the local repo's
-        store and in the usercache.
-        filename is for informational messages only.
-        """
-        util.makedirs(lfutil.storepath(self.repo, ''))
-        storefilename = lfutil.storepath(self.repo, hash)
-
-        tmpname = storefilename + '.tmp'
-        tmpfile = util.atomictempfile(tmpname,
-                                      createmode=self.repo.store.createmode)
+from mercurial import (
+    error,
+    hg,
+    util,
+)
 
-        try:
-            gothash = self._getfile(tmpfile, filename, hash)
-        except StoreError as err:
-            self.ui.warn(err.longmessage())
-            gothash = ""
-        tmpfile.close()
-
-        if gothash != hash:
-            if gothash != "":
-                self.ui.warn(_('%s: data corruption (expected %s, got %s)\n')
-                             % (filename, hash, gothash))
-            util.unlink(tmpname)
-            return False
-
-        util.rename(tmpname, storefilename)
-        lfutil.linktousercache(self.repo, hash)
-        return True
-
-    def verify(self, revs, contents=False):
-        '''Verify the existence (and, optionally, contents) of every big
-        file revision referenced by every changeset in revs.
-        Return 0 if all is well, non-zero on any errors.'''
-
-        self.ui.status(_('searching %d changesets for largefiles\n') %
-                       len(revs))
-        verified = set()                # set of (filename, filenode) tuples
-        filestocheck = []               # list of (cset, filename, expectedhash)
-        for rev in revs:
-            cctx = self.repo[rev]
-            cset = "%d:%s" % (cctx.rev(), node.short(cctx.node()))
-
-            for standin in cctx:
-                filename = lfutil.splitstandin(standin)
-                if filename:
-                    fctx = cctx[standin]
-                    key = (filename, fctx.filenode())
-                    if key not in verified:
-                        verified.add(key)
-                        expectedhash = fctx.data()[0:40]
-                        filestocheck.append((cset, filename, expectedhash))
-
-        failed = self._verifyfiles(contents, filestocheck)
-
-        numrevs = len(verified)
-        numlfiles = len(set([fname for (fname, fnode) in verified]))
-        if contents:
-            self.ui.status(
-                _('verified contents of %d revisions of %d largefiles\n')
-                % (numrevs, numlfiles))
-        else:
-            self.ui.status(
-                _('verified existence of %d revisions of %d largefiles\n')
-                % (numrevs, numlfiles))
-        return int(failed)
-
-    def _getfile(self, tmpfile, filename, hash):
-        '''Fetch one revision of one file from the store and write it
-        to tmpfile.  Compute the hash of the file on-the-fly as it
-        downloads and return the hash.  Close tmpfile.  Raise
-        StoreError if unable to download the file (e.g. it does not
-        exist in the store).'''
-        raise NotImplementedError('abstract method')
-
-    def _verifyfiles(self, contents, filestocheck):
-        '''Perform the actual verification of files in the store.
-        'contents' controls verification of content hash.
-        'filestocheck' is list of files to check.
-        Returns _true_ if any problems are found!
-        '''
-        raise NotImplementedError('abstract method')
-
-import localstore, wirestore
-
-_storeprovider = {
-    'file':  [localstore.localstore],
-    'http':  [wirestore.wirestore],
-    'https': [wirestore.wirestore],
-    'ssh': [wirestore.wirestore],
-    }
-
-_scheme_re = re.compile(r'^([a-zA-Z0-9+-.]+)://')
+from . import (
+    lfutil,
+    localstore,
+    wirestore,
+)
 
 # During clone this function is passed the src's ui object
 # but it needs the dest's ui object so it can read out of
 # the config file. Use repo.ui instead.
 def _openstore(repo, remote=None, put=False):
     ui = repo.ui
 
     if not remote:
         lfpullsource = getattr(repo, 'lfpullsource', None)
         if lfpullsource:
             path = ui.expandpath(lfpullsource)
         elif put:
             path = ui.expandpath('default-push', 'default')
         else:
             path = ui.expandpath('default')
 
         # ui.expandpath() leaves 'default-push' and 'default' alone if
         # they cannot be expanded: fallback to the empty string,
         # meaning the current directory.
         if path == 'default-push' or path == 'default':
             path = ''
             remote = repo
         else:
             path, _branches = hg.parseurl(path)
             remote = hg.peer(repo, {}, path)
 
     # The path could be a scheme so use Mercurial's normal functionality
     # to resolve the scheme to a repository and use its path
     path = util.safehasattr(remote, 'url') and remote.url() or remote.path
 
     match = _scheme_re.match(path)
     if not match:                       # regular filesystem path
         scheme = 'file'
     else:
         scheme = match.group(1)
 
     try:
         storeproviders = _storeprovider[scheme]
     except KeyError:
         raise error.Abort(_('unsupported URL scheme %r') % scheme)
 
     for classobj in storeproviders:
         try:
             return classobj(ui, repo, remote)
         except lfutil.storeprotonotcapable:
             pass
 
     raise error.Abort(_('%s does not appear to be a largefile store') %
                      util.hidepassword(path))
+
+_storeprovider = {
+    'file':  [localstore.localstore],
+    'http':  [wirestore.wirestore],
+    'https': [wirestore.wirestore],
+    'ssh': [wirestore.wirestore],
+    }
+
+_scheme_re = re.compile(r'^([a-zA-Z0-9+-.]+)://')

tests/test-check-module-imports.t

@@ -1,183 +1,181
 #require test-repo
 
   $ . "$TESTDIR/helpers-testrepo.sh"
   $ import_checker="$TESTDIR"/../contrib/import-checker.py
 
 Run the doctests from the import checker, and make sure
 it's working correctly.
   $ TERM=dumb
   $ export TERM
   $ python -m doctest $import_checker
 
 Run additional tests for the import checker
 
   $ mkdir testpackage
   $ touch testpackage/__init__.py
 
   $ cat > testpackage/multiple.py << EOF
   > from __future__ import absolute_import
   > import os, sys
   > EOF
 
   $ cat > testpackage/unsorted.py << EOF
   > from __future__ import absolute_import
   > import sys
   > import os
   > EOF
 
   $ cat > testpackage/stdafterlocal.py << EOF
   > from __future__ import absolute_import
   > from . import unsorted
   > import os
   > EOF
 
   $ cat > testpackage/requirerelative.py << EOF
   > from __future__ import absolute_import
   > import testpackage.unsorted
   > EOF
 
   $ cat > testpackage/importalias.py << EOF
   > from __future__ import absolute_import
   > import ui
   > EOF
 
   $ cat > testpackage/relativestdlib.py << EOF
   > from __future__ import absolute_import
   > from .. import os
   > EOF
 
   $ cat > testpackage/symbolimport.py << EOF
   > from __future__ import absolute_import
   > from .unsorted import foo
   > EOF
 
   $ cat > testpackage/latesymbolimport.py << EOF
   > from __future__ import absolute_import
   > from . import unsorted
   > from mercurial.node import hex
   > EOF
 
   $ cat > testpackage/multiplegroups.py << EOF
   > from __future__ import absolute_import
   > from . import unsorted
   > from . import more
   > EOF
 
   $ mkdir testpackage/subpackage
   $ cat > testpackage/subpackage/levelpriority.py << EOF
   > from __future__ import absolute_import
   > from . import foo
   > from .. import parent
   > EOF
 
   $ touch testpackage/subpackage/foo.py
   $ cat > testpackage/subpackage/__init__.py << EOF
   > from __future__ import absolute_import
   > from . import levelpriority  # should not cause cycle
   > EOF
 
   $ cat > testpackage/subpackage/localimport.py << EOF
   > from __future__ import absolute_import
   > from . import foo
   > def bar():
   >     # should not cause "higher-level import should come first"
   >     from .. import unsorted
   >     # but other errors should be detected
   >     from .. import more
   >     import testpackage.subpackage.levelpriority
   > EOF
 
   $ cat > testpackage/importmodulefromsub.py << EOF
   > from __future__ import absolute_import
   > from .subpackage import foo  # not a "direct symbol import"
   > EOF
 
   $ cat > testpackage/importsymbolfromsub.py << EOF
   > from __future__ import absolute_import
   > from .subpackage import foo, nonmodule
   > EOF
 
   $ cat > testpackage/sortedentries.py << EOF
   > from __future__ import absolute_import
   > from . import (
   >     foo,
   >     bar,
   > )
   > EOF
 
   $ cat > testpackage/importfromalias.py << EOF
   > from __future__ import absolute_import
   > from . import ui
   > EOF
 
   $ cat > testpackage/importfromrelative.py << EOF
   > from __future__ import absolute_import
   > from testpackage.unsorted import foo
   > EOF
 
   $ mkdir testpackage2
   $ touch testpackage2/__init__.py
 
   $ cat > testpackage2/latesymbolimport.py << EOF
   > from __future__ import absolute_import
   > from testpackage import unsorted
   > from mercurial.node import hex
   > EOF
 
   $ python "$import_checker" testpackage*/*.py testpackage/subpackage/*.py
   testpackage/importalias.py:2: ui module must be "as" aliased to uimod
   testpackage/importfromalias.py:2: ui from testpackage must be "as" aliased to uimod
   testpackage/importfromrelative.py:2: import should be relative: testpackage.unsorted
   testpackage/importfromrelative.py:2: direct symbol import foo from testpackage.unsorted
   testpackage/importsymbolfromsub.py:2: direct symbol import nonmodule from testpackage.subpackage
   testpackage/latesymbolimport.py:3: symbol import follows non-symbol import: mercurial.node
   testpackage/multiple.py:2: multiple imported names: os, sys
   testpackage/multiplegroups.py:3: multiple "from . import" statements
   testpackage/relativestdlib.py:2: relative import of stdlib module
   testpackage/requirerelative.py:2: import should be relative: testpackage.unsorted
   testpackage/sortedentries.py:2: imports from testpackage not lexically sorted: bar < foo
   testpackage/stdafterlocal.py:3: stdlib import "os" follows local import: testpackage
   testpackage/subpackage/levelpriority.py:3: higher-level import should come first: testpackage
   testpackage/subpackage/localimport.py:7: multiple "from .. import" statements
   testpackage/subpackage/localimport.py:8: import should be relative: testpackage.subpackage.levelpriority
   testpackage/symbolimport.py:2: direct symbol import foo from testpackage.unsorted
   testpackage/unsorted.py:3: imports not lexically sorted: os < sys
   testpackage2/latesymbolimport.py:3: symbol import follows non-symbol import: mercurial.node
   [1]
 
   $ cd "$TESTDIR"/..
 
 There are a handful of cases here that require renaming a module so it
 doesn't overlap with a stdlib module name. There are also some cycles
 here that we should still endeavor to fix, and some cycles will be
 hidden by deduplication algorithm in the cycle detector, so fixing
 these may expose other cycles.
 
 Known-bad files are excluded by -X as some of them would produce unstable
 outputs, which should be fixed later.
 
   $ hg locate 'set:**.py or grep(r"^#!.*?python")' \
   > 'tests/**.t' \
   > -X contrib/debugshell.py \
   > -X contrib/win32/hgwebdir_wsgi.py \
   > -X doc/gendoc.py \
   > -X doc/hgmanpage.py \
   > -X i18n/posplit \
   > -X tests/test-hgweb-auth.py \
   > -X tests/hypothesishelpers.py \
   > -X tests/test-ctxmanager.py \
   > -X tests/test-lock.py \
   > -X tests/test-verify-repo-operations.py \
   > -X tests/test-hook.t \
   > -X tests/test-import.t \
   > -X tests/test-check-module-imports.t \
   > -X tests/test-commit-interactive.t \
   > -X tests/test-contrib-check-code.t \
   > -X tests/test-extension.t \
   > -X tests/test-hghave.t \
   > -X tests/test-hgweb-no-path-info.t \
   > -X tests/test-hgweb-no-request-uri.t \
   > -X tests/test-hgweb-non-interactive.t \
   > | sed 's-\\-/-g' | python "$import_checker" -
-  Import cycle: hgext.largefiles.basestore -> hgext.largefiles.localstore -> hgext.largefiles.basestore
-  [1]