largefiles: use progress helper

Differential Revision: https://phab.mercurial-scm.org/D3808

hgext/largefiles/basestore.py

@@ -1,164 +1,165 b''
 # Copyright 2009-2010 Gregory P. Ward
 # Copyright 2009-2010 Intelerad Medical Systems Incorporated
 # Copyright 2010-2011 Fog Creek Software
 # Copyright 2010-2011 Unity Technologies
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 '''base class for store implementations and store-related utility code'''
 from __future__ import absolute_import
 
 from mercurial.i18n import _
 
 from mercurial import node, util
 
 from . 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))
+        progress = ui.makeprogress(_('getting largefiles'), unit=_('files'),
+                                   total=len(files))
         for filename, hash in files:
-            ui.progress(_('getting largefiles'), at, unit=_('files'),
-                total=len(files))
+            progress.update(at)
             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)
+        progress.complete()
         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'
         with util.atomictempfile(tmpname,
                 createmode=self.repo.store.createmode) as tmpfile:
             try:
                 gothash = self._getfile(tmpfile, filename, hash)
             except StoreError as err:
                 self.ui.warn(err.longmessage())
                 gothash = ""
 
         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 = lfutil.readasstandin(fctx)
                         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')

hgext/largefiles/lfcommands.py

@@ -1,604 +1,607 b''
 # Copyright 2009-2010 Gregory P. Ward
 # Copyright 2009-2010 Intelerad Medical Systems Incorporated
 # Copyright 2010-2011 Fog Creek Software
 # Copyright 2010-2011 Unity Technologies
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 '''High-level command function for lfconvert, plus the cmdtable.'''
 from __future__ import absolute_import
 
 import errno
 import hashlib
 import os
 import shutil
 
 from mercurial.i18n import _
 
 from mercurial import (
     cmdutil,
     context,
     error,
     hg,
     lock,
     match as matchmod,
     node,
     pycompat,
     registrar,
     scmutil,
     util,
 )
 
 from ..convert import (
     convcmd,
     filemap,
 )
 
 from . import (
     lfutil,
     storefactory
 )
 
 release = lock.release
 
 # -- Commands ----------------------------------------------------------
 
 cmdtable = {}
 command = registrar.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.'''
 
     opts = pycompat.byteskwargs(opts)
     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')
             if pats:
                 matcher = matchmod.match(rsrc.root, '', list(pats))
             else:
                 matcher = None
 
             lfiletohash = {}
+            progress = ui.makeprogress(_('converting revisions'),
+                                       unit=_('revisions'),
+                                       total=rsrc['tip'].rev())
             for ctx in ctxs:
-                ui.progress(_('converting revisions'), ctx.rev(),
-                    unit=_('revisions'), total=rsrc['tip'].rev())
+                progress.update(ctx.rev())
                 _lfconvert_addchangeset(rsrc, rdst, ctx, revmap,
                     lfiles, normalfiles, matcher, size, lfiletohash)
-            ui.progress(_('converting revisions'), None)
+            progress.complete()
 
             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, source_type='hg', dest_type='hg')
             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:
             fstandin = lfutil.standin(f)
             dstfiles.append(fstandin)
             # 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 = hashlib.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, fstandin, hash,
                         executable)
                     lfiletohash[f] = hash
         else:
             # normal file
             dstfiles.append(f)
 
     def getfilectx(repo, memctx, f):
         srcfname = lfutil.splitstandin(f)
         if srcfname is not None:
             # if the file isn't in the manifest then it was removed
             # or renamed, return None to indicate this
             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, memctx, 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, ctx, 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 = storefactory.openstore(rsrc, rdst, put=True)
 
     at = 0
     ui.debug("sending statlfile command for %d largefiles\n" % len(files))
     retval = store.exists(files)
     files = [h for h in files if not retval[h]]
     ui.debug("%d largefiles need to be uploaded\n" % len(files))
 
+    progress = ui.makeprogress(_('uploading largefiles'), unit=_('files'),
+                               total=len(files))
     for hash in files:
-        ui.progress(_('uploading largefiles'), at, unit=_('files'),
-                    total=len(files))
+        progress.update(at)
         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)
+    progress.complete()
 
 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 = 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 = []
 
     ctx = repo[node]
     for lfile in lfiles:
         try:
             expectedhash = lfutil.readasstandin(ctx[lfutil.standin(lfile)])
         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 = storefactory.openstore(repo)
         ret = store.get(toget)
         return ret
 
     return ([], [])
 
 def downloadlfiles(ui, repo, rev=None):
     match = 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, match, {'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 = {}
         dropped = set()
         updated, removed = 0, 0
         wvfs = repo.wvfs
         wctx = repo[None]
         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.readasstandin(wctx[relstandin])
                 if expecthash != '':
                     if lfile not in wctx: # not switched to normal file
                         if repo.dirstate[relstandin] != '?':
                             wvfs.unlinkpath(rellfile, ignoremissing=True)
                         else:
                             dropped.add(rellfile)
 
                     # 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 wctx):
                     wvfs.unlinkpath(rellfile)
                     removed += 1
 
         # largefile processing might be slow and be interrupted - be prepared
         lfdirstate.write()
 
         if lfiles:
             lfiles = [f for f in lfiles if f not in dropped]
 
             for f in dropped:
                 repo.wvfs.unlinkpath(lfutil.standin(f))
 
                 # This needs to happen for dropped files, otherwise they stay in
                 # the M state.
                 lfutil.synclfdirstate(repo, lfdirstate, f, normallookup)
 
             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 exec mode of largefile standin from the repository's
             # dirstate to its state in the lfdirstate.
             rellfile = lfile
             relstandin = lfutil.standin(lfile)
             if wvfs.exists(relstandin):
                 # exec is decided by the users permissions using mask 0o100
                 standinexec = wvfs.stat(relstandin).st_mode & 0o100
                 st = wvfs.stat(rellfile)
                 mode = st.st_mode
                 if standinexec != mode & 0o100:
                     # first remove all X bits, then shift all R bits to X
                     mode &= ~0o111
                     if standinexec:
                         mode |= (mode >> 2) & 0o111 & ~util.umask
                     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'))
     ] + cmdutil.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(r'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 %d\n') % rev)
         (cached, missing) = cachelfiles(ui, repo, rev)
         numcached += len(cached)
     ui.status(_("%d largefiles cached\n") % numcached)
 
 @command('debuglfput',
     [] + cmdutil.remoteopts,
     _('FILE'))
 def debuglfput(ui, repo, filepath, **kwargs):
     hash = lfutil.hashfile(filepath)
     storefactory.openstore(repo).put(filepath, hash)
     ui.write('%s\n' % hash)
     return 0

hgext/largefiles/lfutil.py

@@ -1,674 +1,675 b''
 # Copyright 2009-2010 Gregory P. Ward
 # Copyright 2009-2010 Intelerad Medical Systems Incorporated
 # Copyright 2010-2011 Fog Creek Software
 # Copyright 2010-2011 Unity Technologies
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 '''largefiles utility code: must not import other modules in this package.'''
 from __future__ import absolute_import
 
 import copy
 import hashlib
 import os
 import stat
 
 from mercurial.i18n import _
 from mercurial.node import hex
 
 from mercurial import (
     dirstate,
     encoding,
     error,
     httpconnection,
     match as matchmod,
     node,
     pycompat,
     scmutil,
     sparse,
     util,
     vfs as vfsmod,
 )
 
 shortname = '.hglf'
 shortnameslash = shortname + '/'
 longname = 'largefiles'
 
 # -- Private worker functions ------------------------------------------
 
 def getminsize(ui, assumelfiles, opt, default=10):
     lfsize = opt
     if not lfsize and assumelfiles:
         lfsize = ui.config(longname, 'minsize', default=default)
     if lfsize:
         try:
             lfsize = float(lfsize)
         except ValueError:
             raise error.Abort(_('largefiles: size must be number (not %s)\n')
                              % lfsize)
     if lfsize is None:
         raise error.Abort(_('minimum size for largefiles must be specified'))
     return lfsize
 
 def link(src, dest):
     """Try to create hardlink - if that fails, efficiently make a copy."""
     util.makedirs(os.path.dirname(dest))
     try:
         util.oslink(src, dest)
     except OSError:
         # if hardlinks fail, fallback on atomic copy
         with open(src, 'rb') as srcf, util.atomictempfile(dest) as dstf:
             for chunk in util.filechunkiter(srcf):
                 dstf.write(chunk)
         os.chmod(dest, os.stat(src).st_mode)
 
 def usercachepath(ui, hash):
     '''Return the correct location in the "global" largefiles cache for a file
     with the given hash.
     This cache is used for sharing of largefiles across repositories - both
     to preserve download bandwidth and storage space.'''
     return os.path.join(_usercachedir(ui), hash)
 
 def _usercachedir(ui, name=longname):
     '''Return the location of the "global" largefiles cache.'''
     path = ui.configpath(name, 'usercache')
     if path:
         return path
     if pycompat.iswindows:
         appdata = encoding.environ.get('LOCALAPPDATA',\
                         encoding.environ.get('APPDATA'))
         if appdata:
             return os.path.join(appdata, name)
     elif pycompat.isdarwin:
         home = encoding.environ.get('HOME')
         if home:
             return os.path.join(home, 'Library', 'Caches', name)
     elif pycompat.isposix:
         path = encoding.environ.get('XDG_CACHE_HOME')
         if path:
             return os.path.join(path, name)
         home = encoding.environ.get('HOME')
         if home:
             return os.path.join(home, '.cache', name)
     else:
         raise error.Abort(_('unknown operating system: %s\n')
                           % pycompat.osname)
     raise error.Abort(_('unknown %s usercache location') % name)
 
 def inusercache(ui, hash):
     path = usercachepath(ui, hash)
     return os.path.exists(path)
 
 def findfile(repo, hash):
     '''Return store path of the largefile with the specified hash.
     As a side effect, the file might be linked from user cache.
     Return None if the file can't be found locally.'''
     path, exists = findstorepath(repo, hash)
     if exists:
         repo.ui.note(_('found %s in store\n') % hash)
         return path
     elif inusercache(repo.ui, hash):
         repo.ui.note(_('found %s in system cache\n') % hash)
         path = storepath(repo, hash)
         link(usercachepath(repo.ui, hash), path)
         return path
     return None
 
 class largefilesdirstate(dirstate.dirstate):
     def __getitem__(self, key):
         return super(largefilesdirstate, self).__getitem__(unixpath(key))
     def normal(self, f):
         return super(largefilesdirstate, self).normal(unixpath(f))
     def remove(self, f):
         return super(largefilesdirstate, self).remove(unixpath(f))
     def add(self, f):
         return super(largefilesdirstate, self).add(unixpath(f))
     def drop(self, f):
         return super(largefilesdirstate, self).drop(unixpath(f))
     def forget(self, f):
         return super(largefilesdirstate, self).forget(unixpath(f))
     def normallookup(self, f):
         return super(largefilesdirstate, self).normallookup(unixpath(f))
     def _ignore(self, f):
         return False
     def write(self, tr=False):
         # (1) disable PENDING mode always
         #     (lfdirstate isn't yet managed as a part of the transaction)
         # (2) avoid develwarn 'use dirstate.write with ....'
         super(largefilesdirstate, self).write(None)
 
 def openlfdirstate(ui, repo, create=True):
     '''
     Return a dirstate object that tracks largefiles: i.e. its root is
     the repo root, but it is saved in .hg/largefiles/dirstate.
     '''
     vfs = repo.vfs
     lfstoredir = longname
     opener = vfsmod.vfs(vfs.join(lfstoredir))
     lfdirstate = largefilesdirstate(opener, ui, repo.root,
                                     repo.dirstate._validate,
                                     lambda: sparse.matcher(repo))
 
     # If the largefiles dirstate does not exist, populate and create
     # it. This ensures that we create it on the first meaningful
     # largefiles operation in a new clone.
     if create and not vfs.exists(vfs.join(lfstoredir, 'dirstate')):
         matcher = getstandinmatcher(repo)
         standins = repo.dirstate.walk(matcher, subrepos=[], unknown=False,
                                       ignored=False)
 
         if len(standins) > 0:
             vfs.makedirs(lfstoredir)
 
         for standin in standins:
             lfile = splitstandin(standin)
             lfdirstate.normallookup(lfile)
     return lfdirstate
 
 def lfdirstatestatus(lfdirstate, repo):
     pctx = repo['.']
     match = matchmod.always(repo.root, repo.getcwd())
     unsure, s = lfdirstate.status(match, subrepos=[], ignored=False,
                                   clean=False, unknown=False)
     modified, clean = s.modified, s.clean
     for lfile in unsure:
         try:
             fctx = pctx[standin(lfile)]
         except LookupError:
             fctx = None
         if not fctx or readasstandin(fctx) != hashfile(repo.wjoin(lfile)):
             modified.append(lfile)
         else:
             clean.append(lfile)
             lfdirstate.normal(lfile)
     return s
 
 def listlfiles(repo, rev=None, matcher=None):
     '''return a list of largefiles in the working copy or the
     specified changeset'''
 
     if matcher is None:
         matcher = getstandinmatcher(repo)
 
     # ignore unknown files in working directory
     return [splitstandin(f)
             for f in repo[rev].walk(matcher)
             if rev is not None or repo.dirstate[f] != '?']
 
 def instore(repo, hash, forcelocal=False):
     '''Return true if a largefile with the given hash exists in the store'''
     return os.path.exists(storepath(repo, hash, forcelocal))
 
 def storepath(repo, hash, forcelocal=False):
     '''Return the correct location in the repository largefiles store for a
     file with the given hash.'''
     if not forcelocal and repo.shared():
         return repo.vfs.reljoin(repo.sharedpath, longname, hash)
     return repo.vfs.join(longname, hash)
 
 def findstorepath(repo, hash):
     '''Search through the local store path(s) to find the file for the given
     hash.  If the file is not found, its path in the primary store is returned.
     The return value is a tuple of (path, exists(path)).
     '''
     # For shared repos, the primary store is in the share source.  But for
     # backward compatibility, force a lookup in the local store if it wasn't
     # found in the share source.
     path = storepath(repo, hash, False)
 
     if instore(repo, hash):
         return (path, True)
     elif repo.shared() and instore(repo, hash, True):
         return storepath(repo, hash, True), True
 
     return (path, False)
 
 def copyfromcache(repo, hash, filename):
     '''Copy the specified largefile from the repo or system cache to
     filename in the repository. Return true on success or false if the
     file was not found in either cache (which should not happened:
     this is meant to be called only after ensuring that the needed
     largefile exists in the cache).'''
     wvfs = repo.wvfs
     path = findfile(repo, hash)
     if path is None:
         return False
     wvfs.makedirs(wvfs.dirname(wvfs.join(filename)))
     # The write may fail before the file is fully written, but we
     # don't use atomic writes in the working copy.
     with open(path, 'rb') as srcfd, wvfs(filename, 'wb') as destfd:
         gothash = copyandhash(
             util.filechunkiter(srcfd), destfd)
     if gothash != hash:
         repo.ui.warn(_('%s: data corruption in %s with hash %s\n')
                      % (filename, path, gothash))
         wvfs.unlink(filename)
         return False
     return True
 
 def copytostore(repo, ctx, file, fstandin):
     wvfs = repo.wvfs
     hash = readasstandin(ctx[fstandin])
     if instore(repo, hash):
         return
     if wvfs.exists(file):
         copytostoreabsolute(repo, wvfs.join(file), hash)
     else:
         repo.ui.warn(_("%s: largefile %s not available from local store\n") %
                      (file, hash))
 
 def copyalltostore(repo, node):
     '''Copy all largefiles in a given revision to the store'''
 
     ctx = repo[node]
     for filename in ctx.files():
         realfile = splitstandin(filename)
         if realfile is not None and filename in ctx.manifest():
             copytostore(repo, ctx, realfile, filename)
 
 def copytostoreabsolute(repo, file, hash):
     if inusercache(repo.ui, hash):
         link(usercachepath(repo.ui, hash), storepath(repo, hash))
     else:
         util.makedirs(os.path.dirname(storepath(repo, hash)))
         with open(file, 'rb') as srcf:
             with util.atomictempfile(storepath(repo, hash),
                                      createmode=repo.store.createmode) as dstf:
                 for chunk in util.filechunkiter(srcf):
                     dstf.write(chunk)
         linktousercache(repo, hash)
 
 def linktousercache(repo, hash):
     '''Link / copy the largefile with the specified hash from the store
     to the cache.'''
     path = usercachepath(repo.ui, hash)
     link(storepath(repo, hash), path)
 
 def getstandinmatcher(repo, rmatcher=None):
     '''Return a match object that applies rmatcher to the standin directory'''
     wvfs = repo.wvfs
     standindir = shortname
 
     # no warnings about missing files or directories
     badfn = lambda f, msg: None
 
     if rmatcher and not rmatcher.always():
         pats = [wvfs.join(standindir, pat) for pat in rmatcher.files()]
         if not pats:
             pats = [wvfs.join(standindir)]
         match = scmutil.match(repo[None], pats, badfn=badfn)
     else:
         # no patterns: relative to repo root
         match = scmutil.match(repo[None], [wvfs.join(standindir)], badfn=badfn)
     return match
 
 def composestandinmatcher(repo, rmatcher):
     '''Return a matcher that accepts standins corresponding to the
     files accepted by rmatcher. Pass the list of files in the matcher
     as the paths specified by the user.'''
     smatcher = getstandinmatcher(repo, rmatcher)
     isstandin = smatcher.matchfn
     def composedmatchfn(f):
         return isstandin(f) and rmatcher.matchfn(splitstandin(f))
     smatcher.matchfn = composedmatchfn
 
     return smatcher
 
 def standin(filename):
     '''Return the repo-relative path to the standin for the specified big
     file.'''
     # Notes:
     # 1) Some callers want an absolute path, but for instance addlargefiles
     #    needs it repo-relative so it can be passed to repo[None].add().  So
     #    leave it up to the caller to use repo.wjoin() to get an absolute path.
     # 2) Join with '/' because that's what dirstate always uses, even on
     #    Windows. Change existing separator to '/' first in case we are
     #    passed filenames from an external source (like the command line).
     return shortnameslash + util.pconvert(filename)
 
 def isstandin(filename):
     '''Return true if filename is a big file standin. filename must be
     in Mercurial's internal form (slash-separated).'''
     return filename.startswith(shortnameslash)
 
 def splitstandin(filename):
     # Split on / because that's what dirstate always uses, even on Windows.
     # Change local separator to / first just in case we are passed filenames
     # from an external source (like the command line).
     bits = util.pconvert(filename).split('/', 1)
     if len(bits) == 2 and bits[0] == shortname:
         return bits[1]
     else:
         return None
 
 def updatestandin(repo, lfile, standin):
     """Re-calculate hash value of lfile and write it into standin
 
     This assumes that "lfutil.standin(lfile) == standin", for efficiency.
     """
     file = repo.wjoin(lfile)
     if repo.wvfs.exists(lfile):
         hash = hashfile(file)
         executable = getexecutable(file)
         writestandin(repo, standin, hash, executable)
     else:
         raise error.Abort(_('%s: file not found!') % lfile)
 
 def readasstandin(fctx):
     '''read hex hash from given filectx of standin file
 
     This encapsulates how "standin" data is stored into storage layer.'''
     return fctx.data().strip()
 
 def writestandin(repo, standin, hash, executable):
     '''write hash to <repo.root>/<standin>'''
     repo.wwrite(standin, hash + '\n', executable and 'x' or '')
 
 def copyandhash(instream, outfile):
     '''Read bytes from instream (iterable) and write them to outfile,
     computing the SHA-1 hash of the data along the way. Return the hash.'''
     hasher = hashlib.sha1('')
     for data in instream:
         hasher.update(data)
         outfile.write(data)
     return hex(hasher.digest())
 
 def hashfile(file):
     if not os.path.exists(file):
         return ''
     with open(file, 'rb') as fd:
         return hexsha1(fd)
 
 def getexecutable(filename):
     mode = os.stat(filename).st_mode
     return ((mode & stat.S_IXUSR) and
             (mode & stat.S_IXGRP) and
             (mode & stat.S_IXOTH))
 
 def urljoin(first, second, *arg):
     def join(left, right):
         if not left.endswith('/'):
             left += '/'
         if right.startswith('/'):
             right = right[1:]
         return left + right
 
     url = join(first, second)
     for a in arg:
         url = join(url, a)
     return url
 
 def hexsha1(fileobj):
     """hexsha1 returns the hex-encoded sha1 sum of the data in the file-like
     object data"""
     h = hashlib.sha1()
     for chunk in util.filechunkiter(fileobj):
         h.update(chunk)
     return hex(h.digest())
 
 def httpsendfile(ui, filename):
     return httpconnection.httpsendfile(ui, filename, 'rb')
 
 def unixpath(path):
     '''Return a version of path normalized for use with the lfdirstate.'''
     return util.pconvert(os.path.normpath(path))
 
 def islfilesrepo(repo):
     '''Return true if the repo is a largefile repo.'''
     if ('largefiles' in repo.requirements and
             any(shortnameslash in f[0] for f in repo.store.datafiles())):
         return True
 
     return any(openlfdirstate(repo.ui, repo, False))
 
 class storeprotonotcapable(Exception):
     def __init__(self, storetypes):
         self.storetypes = storetypes
 
 def getstandinsstate(repo):
     standins = []
     matcher = getstandinmatcher(repo)
     wctx = repo[None]
     for standin in repo.dirstate.walk(matcher, subrepos=[], unknown=False,
                                       ignored=False):
         lfile = splitstandin(standin)
         try:
             hash = readasstandin(wctx[standin])
         except IOError:
             hash = None
         standins.append((lfile, hash))
     return standins
 
 def synclfdirstate(repo, lfdirstate, lfile, normallookup):
     lfstandin = standin(lfile)
     if lfstandin in repo.dirstate:
         stat = repo.dirstate._map[lfstandin]
         state, mtime = stat[0], stat[3]
     else:
         state, mtime = '?', -1
     if state == 'n':
         if (normallookup or mtime < 0 or
             not repo.wvfs.exists(lfile)):
             # state 'n' doesn't ensure 'clean' in this case
             lfdirstate.normallookup(lfile)
         else:
             lfdirstate.normal(lfile)
     elif state == 'm':
         lfdirstate.normallookup(lfile)
     elif state == 'r':
         lfdirstate.remove(lfile)
     elif state == 'a':
         lfdirstate.add(lfile)
     elif state == '?':
         lfdirstate.drop(lfile)
 
 def markcommitted(orig, ctx, node):
     repo = ctx.repo()
 
     orig(node)
 
     # ATTENTION: "ctx.files()" may differ from "repo[node].files()"
     # because files coming from the 2nd parent are omitted in the latter.
     #
     # The former should be used to get targets of "synclfdirstate",
     # because such files:
     # - are marked as "a" by "patch.patch()" (e.g. via transplant), and
     # - have to be marked as "n" after commit, but
     # - aren't listed in "repo[node].files()"
 
     lfdirstate = openlfdirstate(repo.ui, repo)
     for f in ctx.files():
         lfile = splitstandin(f)
         if lfile is not None:
             synclfdirstate(repo, lfdirstate, lfile, False)
     lfdirstate.write()
 
     # As part of committing, copy all of the largefiles into the cache.
     #
     # Using "node" instead of "ctx" implies additional "repo[node]"
     # lookup while copyalltostore(), but can omit redundant check for
     # files comming from the 2nd parent, which should exist in store
     # at merging.
     copyalltostore(repo, node)
 
 def getlfilestoupdate(oldstandins, newstandins):
     changedstandins = set(oldstandins).symmetric_difference(set(newstandins))
     filelist = []
     for f in changedstandins:
         if f[0] not in filelist:
             filelist.append(f[0])
     return filelist
 
 def getlfilestoupload(repo, missing, addfunc):
+    progress = repo.ui.makeprogress(_('finding outgoing largefiles'),
+                                    unit=_('revisions'), total=len(missing))
     for i, n in enumerate(missing):
-        repo.ui.progress(_('finding outgoing largefiles'), i,
-            unit=_('revisions'), total=len(missing))
+        progress.update(i)
         parents = [p for p in repo[n].parents() if p != node.nullid]
 
         oldlfstatus = repo.lfstatus
         repo.lfstatus = False
         try:
             ctx = repo[n]
         finally:
             repo.lfstatus = oldlfstatus
 
         files = set(ctx.files())
         if len(parents) == 2:
             mc = ctx.manifest()
             mp1 = ctx.parents()[0].manifest()
             mp2 = ctx.parents()[1].manifest()
             for f in mp1:
                 if f not in mc:
                     files.add(f)
             for f in mp2:
                 if f not in mc:
                     files.add(f)
             for f in mc:
                 if mc[f] != mp1.get(f, None) or mc[f] != mp2.get(f, None):
                     files.add(f)
         for fn in files:
             if isstandin(fn) and fn in ctx:
                 addfunc(fn, readasstandin(ctx[fn]))
-    repo.ui.progress(_('finding outgoing largefiles'), None)
+    progress.complete()
 
 def updatestandinsbymatch(repo, match):
     '''Update standins in the working directory according to specified match
 
     This returns (possibly modified) ``match`` object to be used for
     subsequent commit process.
     '''
 
     ui = repo.ui
 
     # Case 1: user calls commit with no specific files or
     # include/exclude patterns: refresh and commit all files that
     # are "dirty".
     if match is None or match.always():
         # Spend a bit of time here to get a list of files we know
         # are modified so we can compare only against those.
         # It can cost a lot of time (several seconds)
         # otherwise to update all standins if the largefiles are
         # large.
         lfdirstate = openlfdirstate(ui, repo)
         dirtymatch = matchmod.always(repo.root, repo.getcwd())
         unsure, s = lfdirstate.status(dirtymatch, subrepos=[], ignored=False,
                                       clean=False, unknown=False)
         modifiedfiles = unsure + s.modified + s.added + s.removed
         lfiles = listlfiles(repo)
         # this only loops through largefiles that exist (not
         # removed/renamed)
         for lfile in lfiles:
             if lfile in modifiedfiles:
                 fstandin = standin(lfile)
                 if repo.wvfs.exists(fstandin):
                     # this handles the case where a rebase is being
                     # performed and the working copy is not updated
                     # yet.
                     if repo.wvfs.exists(lfile):
                         updatestandin(repo, lfile, fstandin)
 
         return match
 
     lfiles = listlfiles(repo)
     match._files = repo._subdirlfs(match.files(), lfiles)
 
     # Case 2: user calls commit with specified patterns: refresh
     # any matching big files.
     smatcher = composestandinmatcher(repo, match)
     standins = repo.dirstate.walk(smatcher, subrepos=[], unknown=False,
                                   ignored=False)
 
     # No matching big files: get out of the way and pass control to
     # the usual commit() method.
     if not standins:
         return match
 
     # Refresh all matching big files.  It's possible that the
     # commit will end up failing, in which case the big files will
     # stay refreshed.  No harm done: the user modified them and
     # asked to commit them, so sooner or later we're going to
     # refresh the standins.  Might as well leave them refreshed.
     lfdirstate = openlfdirstate(ui, repo)
     for fstandin in standins:
         lfile = splitstandin(fstandin)
         if lfdirstate[lfile] != 'r':
             updatestandin(repo, lfile, fstandin)
 
     # Cook up a new matcher that only matches regular files or
     # standins corresponding to the big files requested by the
     # user.  Have to modify _files to prevent commit() from
     # complaining "not tracked" for big files.
     match = copy.copy(match)
     origmatchfn = match.matchfn
 
     # Check both the list of largefiles and the list of
     # standins because if a largefile was removed, it
     # won't be in the list of largefiles at this point
     match._files += sorted(standins)
 
     actualfiles = []
     for f in match._files:
         fstandin = standin(f)
 
         # For largefiles, only one of the normal and standin should be
         # committed (except if one of them is a remove).  In the case of a
         # standin removal, drop the normal file if it is unknown to dirstate.
         # Thus, skip plain largefile names but keep the standin.
         if f in lfiles or fstandin in standins:
             if repo.dirstate[fstandin] != 'r':
                 if repo.dirstate[f] != 'r':
                     continue
             elif repo.dirstate[f] == '?':
                 continue
 
         actualfiles.append(f)
     match._files = actualfiles
 
     def matchfn(f):
         if origmatchfn(f):
             return f not in lfiles
         else:
             return f in standins
 
     match.matchfn = matchfn
 
     return match
 
 class automatedcommithook(object):
     '''Stateful hook to update standins at the 1st commit of resuming
 
     For efficiency, updating standins in the working directory should
     be avoided while automated committing (like rebase, transplant and
     so on), because they should be updated before committing.
 
     But the 1st commit of resuming automated committing (e.g. ``rebase
     --continue``) should update them, because largefiles may be
     modified manually.
     '''
     def __init__(self, resuming):
         self.resuming = resuming
 
     def __call__(self, repo, match):
         if self.resuming:
             self.resuming = False # avoids updating at subsequent commits
             return updatestandinsbymatch(repo, match)
         else:
             return match
 
 def getstatuswriter(ui, repo, forcibly=None):
     '''Return the function to write largefiles specific status out
 
     If ``forcibly`` is ``None``, this returns the last element of
     ``repo._lfstatuswriters`` as "default" writer function.
 
     Otherwise, this returns the function to always write out (or
     ignore if ``not forcibly``) status.
     '''
     if forcibly is None and util.safehasattr(repo, '_largefilesenabled'):
         return repo._lfstatuswriters[-1]
     else:
         if forcibly:
             return ui.status # forcibly WRITE OUT
         else:
             return lambda *msg, **opts: None # forcibly IGNORE