largefiles: rename functions and methods to match desired behavior

The original intent was that the largefiles would primarily be in the
repository, with the global cache being only that--a cache.  The naming
conventions and actual intent have both strayed.  In this first patch, the
naming conventions are switched to match the actual intent, as are the
configuration options.

hgext/largefiles/basestore.py

@@ -1,202 +1,202
 # 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 os
 import tempfile
 import binascii
 import re
 
 from mercurial import util, node, hg
 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):
         if self.url:
             return ('%s: %s\n'
                     '(failed URL: %s)\n'
                     % (self.filename, self.detail, self.url))
         else:
             return ('%s: %s\n'
                     '(no default or default-push path set in hgrc)\n'
                     % (self.filename, self.detail))
 
     def __str__(self):
         return "%s: %s" % (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 under <filename>/<hash>.'''
         raise NotImplementedError('abstract method')
 
     def exists(self, hash):
         '''Check to see if the store contains the given hash.'''
         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 successfuly
         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
         for filename, hash in files:
             ui.progress(_('getting largefiles'), at, unit='lfile',
                 total=len(files))
             at += 1
             ui.note(_('getting %s:%s\n') % (filename, hash))
 
-            cachefilename = lfutil.cachepath(self.repo, hash)
-            cachedir = os.path.dirname(cachefilename)
+            storefilename = lfutil.storepath(self.repo, hash)
+            storedir = os.path.dirname(storefilename)
 
             # No need to pass mode='wb' to fdopen(), since mkstemp() already
             # opened the file in binary mode.
             (tmpfd, tmpfilename) = tempfile.mkstemp(
-                dir=cachedir, prefix=os.path.basename(filename))
+                dir=storedir, prefix=os.path.basename(filename))
             tmpfile = os.fdopen(tmpfd, 'w')
 
             try:
                 hhash = binascii.hexlify(self._getfile(tmpfile, filename, hash))
             except StoreError, err:
                 ui.warn(err.longmessage())
                 hhash = ""
 
             if hhash != hash:
                 if hhash != "":
                     ui.warn(_('%s: data corruption (expected %s, got %s)\n')
                             % (filename, hash, hhash))
                 tmpfile.close() # no-op if it's already closed
                 os.remove(tmpfilename)
                 missing.append(filename)
                 continue
 
-            if os.path.exists(cachefilename): # Windows
-                os.remove(cachefilename)
-            os.rename(tmpfilename, cachefilename)
-            lfutil.linktosystemcache(self.repo, hash)
+            if os.path.exists(storefilename): # Windows
+                os.remove(storefilename)
+            os.rename(tmpfilename, storefilename)
+            lfutil.linktousercache(self.repo, hash)
             success.append((filename, hhash))
 
         ui.progress(_('getting largefiles'), None)
         return (success, missing)
 
     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.'''
         write = self.ui.write
         failed = False
 
         write(_('searching %d changesets for largefiles\n') % len(revs))
         verified = set()                # set of (filename, filenode) tuples
 
         for rev in revs:
             cctx = self.repo[rev]
             cset = "%d:%s" % (cctx.rev(), node.short(cctx.node()))
 
             failed = lfutil.any_(self._verifyfile(
                 cctx, cset, contents, standin, verified) for standin in cctx)
 
         num_revs = len(verified)
         num_lfiles = len(set([fname for (fname, fnode) in verified]))
         if contents:
             write(_('verified contents of %d revisions of %d largefiles\n')
                   % (num_revs, num_lfiles))
         else:
             write(_('verified existence of %d revisions of %d largefiles\n')
                   % (num_revs, num_lfiles))
 
         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 binary 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 _verifyfile(self, cctx, cset, contents, standin, verified):
         '''Perform the actual verification of a file in the store.
         '''
         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:
         path = (getattr(repo, 'lfpullsource', None) or
                 ui.expandpath('default-push', '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:
             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 util.Abort(_('unsupported URL scheme %r') % scheme)
 
     for class_obj in storeproviders:
         try:
             return class_obj(ui, repo, remote)
         except lfutil.storeprotonotcapable:
             pass
 
     raise util.Abort(_('%s does not appear to be a largefile store') % path)

hgext/largefiles/lfutil.py

@@ -1,448 +1,448
 # Copyright 2009-2010 Gregory P. Ward
 # Copyright 2009-2010 Intelerad Medical Systems Incorporated
 # Copyright 2010-2011 Fog Creek Software
 # Copyright 2010-2011 Unity Technologies
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 '''largefiles utility code: must not import other modules in this package.'''
 
 import os
 import errno
 import shutil
 import stat
 import hashlib
 
 from mercurial import dirstate, httpconnection, match as match_, util, scmutil
 from mercurial.i18n import _
 
 shortname = '.hglf'
 longname = 'largefiles'
 
 
 # -- Portability wrappers ----------------------------------------------
 
 def dirstate_walk(dirstate, matcher, unknown=False, ignored=False):
     return dirstate.walk(matcher, [], unknown, ignored)
 
 def repo_add(repo, list):
     add = repo[None].add
     return add(list)
 
 def repo_remove(repo, list, unlink=False):
     def remove(list, unlink):
         wlock = repo.wlock()
         try:
             if unlink:
                 for f in list:
                     try:
                         util.unlinkpath(repo.wjoin(f))
                     except OSError, inst:
                         if inst.errno != errno.ENOENT:
                             raise
             repo[None].forget(list)
         finally:
             wlock.release()
     return remove(list, unlink=unlink)
 
 def repo_forget(repo, list):
     forget = repo[None].forget
     return forget(list)
 
 def findoutgoing(repo, remote, force):
     from mercurial import discovery
     common, _anyinc, _heads = discovery.findcommonincoming(repo,
         remote, force=force)
     return repo.changelog.findmissing(common)
 
 # -- Private worker functions ------------------------------------------
 
 def getminsize(ui, assumelfiles, opt, default=10):
     lfsize = opt
     if not lfsize and assumelfiles:
         lfsize = ui.config(longname, 'minsize', default=default)
     if lfsize:
         try:
             lfsize = float(lfsize)
         except ValueError:
             raise util.Abort(_('largefiles: size must be number (not %s)\n')
                              % lfsize)
     if lfsize is None:
         raise util.Abort(_('minimum size for largefiles must be specified'))
     return lfsize
 
 def link(src, dest):
     try:
         util.oslink(src, dest)
     except OSError:
         # if hardlinks fail, fallback on copy
         shutil.copyfile(src, dest)
         os.chmod(dest, os.stat(src).st_mode)
 
-def systemcachepath(ui, hash):
-    path = ui.config(longname, 'systemcache', None)
+def usercachepath(ui, hash):
+    path = ui.config(longname, 'usercache', None)
     if path:
         path = os.path.join(path, hash)
     else:
         if os.name == 'nt':
             appdata = os.getenv('LOCALAPPDATA', os.getenv('APPDATA'))
             path = os.path.join(appdata, longname, hash)
         elif os.name == 'posix':
             path = os.path.join(os.getenv('HOME'), '.' + longname, hash)
         else:
             raise util.Abort(_('unknown operating system: %s\n') % os.name)
     return path
 
-def insystemcache(ui, hash):
-    return os.path.exists(systemcachepath(ui, hash))
+def inusercache(ui, hash):
+    return os.path.exists(usercachepath(ui, hash))
 
 def findfile(repo, hash):
-    if incache(repo, hash):
-        repo.ui.note(_('Found %s in cache\n') % hash)
-        return cachepath(repo, hash)
-    if insystemcache(repo.ui, hash):
+    if instore(repo, hash):
+        repo.ui.note(_('Found %s in store\n') % hash)
+        return storepath(repo, hash)
+    if inusercache(repo.ui, hash):
         repo.ui.note(_('Found %s in system cache\n') % hash)
-        return systemcachepath(repo.ui, hash)
+        return usercachepath(repo.ui, hash)
     return None
 
 class largefiles_dirstate(dirstate.dirstate):
     def __getitem__(self, key):
         return super(largefiles_dirstate, self).__getitem__(unixpath(key))
     def normal(self, f):
         return super(largefiles_dirstate, self).normal(unixpath(f))
     def remove(self, f):
         return super(largefiles_dirstate, self).remove(unixpath(f))
     def add(self, f):
         return super(largefiles_dirstate, self).add(unixpath(f))
     def drop(self, f):
         return super(largefiles_dirstate, self).drop(unixpath(f))
     def forget(self, f):
         return super(largefiles_dirstate, self).forget(unixpath(f))
 
 def openlfdirstate(ui, repo):
     '''
     Return a dirstate object that tracks largefiles: i.e. its root is
     the repo root, but it is saved in .hg/largefiles/dirstate.
     '''
     admin = repo.join(longname)
     opener = scmutil.opener(admin)
     if util.safehasattr(repo.dirstate, '_validate'):
         lfdirstate = largefiles_dirstate(opener, ui, repo.root,
             repo.dirstate._validate)
     else:
         lfdirstate = largefiles_dirstate(opener, ui, repo.root)
 
     # 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. It also gives us an easy
     # way to forcibly rebuild largefiles state:
     #   rm .hg/largefiles/dirstate && hg status
     # Or even, if things are really messed up:
     #   rm -rf .hg/largefiles && hg status
     if not os.path.exists(os.path.join(admin, 'dirstate')):
         util.makedirs(admin)
         matcher = getstandinmatcher(repo)
         for standin in dirstate_walk(repo.dirstate, matcher):
             lfile = splitstandin(standin)
             hash = readstandin(repo, lfile)
             lfdirstate.normallookup(lfile)
             try:
                 if hash == hashfile(lfile):
                     lfdirstate.normal(lfile)
             except IOError, err:
                 if err.errno != errno.ENOENT:
                     raise
 
         lfdirstate.write()
 
     return lfdirstate
 
 def lfdirstate_status(lfdirstate, repo, rev):
     wlock = repo.wlock()
     try:
         match = match_.always(repo.root, repo.getcwd())
         s = lfdirstate.status(match, [], False, False, False)
         unsure, modified, added, removed, missing, unknown, ignored, clean = s
         for lfile in unsure:
             if repo[rev][standin(lfile)].data().strip() != \
                     hashfile(repo.wjoin(lfile)):
                 modified.append(lfile)
             else:
                 clean.append(lfile)
                 lfdirstate.normal(lfile)
         lfdirstate.write()
     finally:
         wlock.release()
     return (modified, added, removed, missing, unknown, ignored, clean)
 
 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 incache(repo, hash):
-    return os.path.exists(cachepath(repo, hash))
+def instore(repo, hash):
+    return os.path.exists(storepath(repo, hash))
 
 def createdir(dir):
     if not os.path.exists(dir):
         os.makedirs(dir)
 
-def cachepath(repo, hash):
+def storepath(repo, hash):
     return repo.join(os.path.join(longname, hash))
 
 def copyfromcache(repo, hash, filename):
     '''Copy the specified largefile from the repo or system cache to
     filename in the repository. Return true on success or false if the
     file was not found in either cache (which should not happened:
     this is meant to be called only after ensuring that the needed
     largefile exists in the cache).'''
     path = findfile(repo, hash)
     if path is None:
         return False
     util.makedirs(os.path.dirname(repo.wjoin(filename)))
     shutil.copy(path, repo.wjoin(filename))
     return True
 
-def copytocache(repo, rev, file, uploaded=False):
+def copytostore(repo, rev, file, uploaded=False):
     hash = readstandin(repo, file)
-    if incache(repo, hash):
+    if instore(repo, hash):
         return
-    copytocacheabsolute(repo, repo.wjoin(file), hash)
+    copytostoreabsolute(repo, repo.wjoin(file), hash)
 
-def copytocacheabsolute(repo, file, hash):
-    createdir(os.path.dirname(cachepath(repo, hash)))
-    if insystemcache(repo.ui, hash):
-        link(systemcachepath(repo.ui, hash), cachepath(repo, hash))
+def copytostoreabsolute(repo, file, hash):
+    createdir(os.path.dirname(storepath(repo, hash)))
+    if inusercache(repo.ui, hash):
+        link(usercachepath(repo.ui, hash), storepath(repo, hash))
     else:
-        shutil.copyfile(file, cachepath(repo, hash))
-        os.chmod(cachepath(repo, hash), os.stat(file).st_mode)
-        linktosystemcache(repo, hash)
+        shutil.copyfile(file, storepath(repo, hash))
+        os.chmod(storepath(repo, hash), os.stat(file).st_mode)
+        linktousercache(repo, hash)
 
-def linktosystemcache(repo, hash):
-    createdir(os.path.dirname(systemcachepath(repo.ui, hash)))
-    link(cachepath(repo, hash), systemcachepath(repo.ui, hash))
+def linktousercache(repo, hash):
+    createdir(os.path.dirname(usercachepath(repo.ui, hash)))
+    link(storepath(repo, hash), usercachepath(repo.ui, hash))
 
 def getstandinmatcher(repo, pats=[], opts={}):
     '''Return a match object that applies pats to the standin directory'''
     standindir = repo.pathto(shortname)
     if pats:
         # patterns supplied: search standin directory relative to current dir
         cwd = repo.getcwd()
         if os.path.isabs(cwd):
             # cwd is an absolute path for hg -R <reponame>
             # work relative to the repository root in this case
             cwd = ''
         pats = [os.path.join(standindir, cwd, pat) for pat in pats]
     elif os.path.isdir(standindir):
         # no patterns: relative to repo root
         pats = [standindir]
     else:
         # no patterns and no standin dir: return matcher that matches nothing
         match = match_.match(repo.root, None, [], exact=True)
         match.matchfn = lambda f: False
         return match
     return getmatcher(repo, pats, opts, showbad=False)
 
 def getmatcher(repo, pats=[], opts={}, showbad=True):
     '''Wrapper around scmutil.match() that adds showbad: if false,
     neuter the match object's bad() method so it does not print any
     warnings about missing files or directories.'''
     match = scmutil.match(repo[None], pats, opts)
 
     if not showbad:
         match.bad = lambda f, msg: None
     return match
 
 def composestandinmatcher(repo, rmatcher):
     '''Return a matcher that accepts standins corresponding to the
     files accepted by rmatcher. Pass the list of files in the matcher
     as the paths specified by the user.'''
     smatcher = getstandinmatcher(repo, rmatcher.files())
     isstandin = smatcher.matchfn
     def composed_matchfn(f):
         return isstandin(f) and rmatcher.matchfn(splitstandin(f))
     smatcher.matchfn = composed_matchfn
 
     return smatcher
 
 def standin(filename):
     '''Return the repo-relative path to the standin for the specified big
     file.'''
     # Notes:
     # 1) Most callers want an absolute path, but _create_standin() needs
     #    it repo-relative so lfadd() can pass it to repo_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 shortname + '/' + filename.replace(os.sep, '/')
 
 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(shortname + '/')
 
 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 = filename.replace(os.sep, '/').split('/', 1)
     if len(bits) == 2 and bits[0] == shortname:
         return bits[1]
     else:
         return None
 
 def updatestandin(repo, standin):
     file = repo.wjoin(splitstandin(standin))
     if os.path.exists(file):
         hash = hashfile(file)
         executable = getexecutable(file)
         writestandin(repo, standin, hash, executable)
 
 def readstandin(repo, filename, node=None):
     '''read hex hash from standin for filename at given node, or working
     directory if no node is given'''
     return repo[node][standin(filename)].data().strip()
 
 def writestandin(repo, standin, hash, executable):
     '''write hash to <repo.root>/<standin>'''
     writehash(hash, repo.wjoin(standin), executable)
 
 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.  Close outfile
     when done and return the binary hash.'''
     hasher = util.sha1('')
     for data in instream:
         hasher.update(data)
         outfile.write(data)
 
     # Blecch: closing a file that somebody else opened is rude and
     # wrong. But it's so darn convenient and practical! After all,
     # outfile was opened just to copy and hash.
     outfile.close()
 
     return hasher.digest()
 
 def hashrepofile(repo, file):
     return hashfile(repo.wjoin(file))
 
 def hashfile(file):
     if not os.path.exists(file):
         return ''
     hasher = util.sha1('')
     fd = open(file, 'rb')
     for data in blockstream(fd):
         hasher.update(data)
     fd.close()
     return hasher.hexdigest()
 
 class limitreader(object):
     def __init__(self, f, limit):
         self.f = f
         self.limit = limit
 
     def read(self, length):
         if self.limit == 0:
             return ''
         length = length > self.limit and self.limit or length
         self.limit -= length
         return self.f.read(length)
 
     def close(self):
         pass
 
 def blockstream(infile, blocksize=128 * 1024):
     """Generator that yields blocks of data from infile and closes infile."""
     while True:
         data = infile.read(blocksize)
         if not data:
             break
         yield data
     # same blecch as copyandhash() above
     infile.close()
 
 def readhash(filename):
     rfile = open(filename, 'rb')
     hash = rfile.read(40)
     rfile.close()
     if len(hash) < 40:
         raise util.Abort(_('bad hash in \'%s\' (only %d bytes long)')
                          % (filename, len(hash)))
     return hash
 
 def writehash(hash, filename, executable):
     util.makedirs(os.path.dirname(filename))
     if os.path.exists(filename):
         os.unlink(filename)
     wfile = open(filename, 'wb')
 
     try:
         wfile.write(hash)
         wfile.write('\n')
     finally:
         wfile.close()
     if os.path.exists(filename):
         os.chmod(filename, getmode(executable))
 
 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 getmode(executable):
     if executable:
         return 0755
     else:
         return 0644
 
 def urljoin(first, second, *arg):
     def join(left, right):
         if not left.endswith('/'):
             left += '/'
         if right.startswith('/'):
             right = right[1:]
         return left + right
 
     url = join(first, second)
     for a in arg:
         url = join(url, a)
     return url
 
 def hexsha1(data):
     """hexsha1 returns the hex-encoded sha1 sum of the data in the file-like
     object data"""
     h = hashlib.sha1()
     for chunk in util.filechunkiter(data):
         h.update(chunk)
     return h.hexdigest()
 
 def httpsendfile(ui, filename):
     return httpconnection.httpsendfile(ui, filename, 'rb')
 
 def unixpath(path):
     '''Return a version of path normalized for use with the lfdirstate.'''
     return os.path.normpath(path).replace(os.sep, '/')
 
 def islfilesrepo(repo):
     return ('largefiles' in repo.requirements and
             any_(shortname + '/' in f[0] for f in repo.store.datafiles()))
 
 def any_(gen):
     for x in gen:
         if x:
             return True
     return False
 
 class storeprotonotcapable(BaseException):
     def __init__(self, storetypes):
         self.storetypes = storetypes

hgext/largefiles/localstore.py

@@ -1,71 +1,71
 # 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.
 
 '''store class for local filesystem'''
 
 import os
 
 from mercurial import util
 from mercurial.i18n import _
 
 import lfutil
 import basestore
 
 class localstore(basestore.basestore):
     '''Because there is a system-wide cache, the local store always
     uses that cache. Since the cache is updated elsewhere, we can
     just read from it here as if it were the store.'''
 
     def __init__(self, ui, repo, remote):
         url = os.path.join(remote.path, '.hg', lfutil.longname)
         super(localstore, self).__init__(ui, repo, util.expandpath(url))
 
     def put(self, source, filename, hash):
         '''Any file that is put must already be in the system-wide
         cache so do nothing.'''
         return
 
     def exists(self, hash):
-        return lfutil.insystemcache(self.repo.ui, hash)
+        return lfutil.inusercache(self.repo.ui, hash)
 
     def _getfile(self, tmpfile, filename, hash):
-        if lfutil.insystemcache(self.ui, hash):
-            return lfutil.systemcachepath(self.ui, hash)
+        if lfutil.inusercache(self.ui, hash):
+            return lfutil.usercachepath(self.ui, hash)
         raise basestore.StoreError(filename, hash, '',
             _("Can't get file locally"))
 
     def _verifyfile(self, cctx, cset, contents, standin, verified):
         filename = lfutil.splitstandin(standin)
         if not filename:
             return False
         fctx = cctx[standin]
         key = (filename, fctx.filenode())
         if key in verified:
             return False
 
         expecthash = fctx.data()[0:40]
         verified.add(key)
-        if not lfutil.insystemcache(self.ui, expecthash):
+        if not lfutil.inusercache(self.ui, expecthash):
             self.ui.warn(
                 _('changeset %s: %s missing\n'
                   '  (looked for hash %s)\n')
                 % (cset, filename, expecthash))
             return True                 # failed
 
         if contents:
-            storepath = lfutil.systemcachepath(self.ui, expecthash)
+            storepath = lfutil.usercachepath(self.ui, expecthash)
             actualhash = lfutil.hashfile(storepath)
             if actualhash != expecthash:
                 self.ui.warn(
                     _('changeset %s: %s: contents differ\n'
                       '  (%s:\n'
                       '  expected hash %s,\n'
                       '  but got %s)\n')
                     % (cset, filename, storepath, expecthash, actualhash))
                 return True             # failed
         return False

hgext/largefiles/proto.py

@@ -1,160 +1,160
 # Copyright 2011 Fog Creek Software
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 import os
 import tempfile
 import urllib2
 
 from mercurial import error, httprepo, util, wireproto
 from mercurial.i18n import _
 
 import lfutil
 
 LARGEFILES_REQUIRED_MSG = ('\nThis repository uses the largefiles extension.'
                            '\n\nPlease enable it in your Mercurial config '
                            'file.\n')
 
 def putlfile(repo, proto, sha):
     '''Put a largefile into a repository's local cache and into the
     system cache.'''
     f = None
     proto.redirect()
     try:
         try:
             f = tempfile.NamedTemporaryFile(mode='wb+', prefix='hg-putlfile-')
             proto.getfile(f)
             f.seek(0)
             if sha != lfutil.hexsha1(f):
                 return wireproto.pushres(1)
-            lfutil.copytocacheabsolute(repo, f.name, sha)
+            lfutil.copytostoreabsolute(repo, f.name, sha)
         except IOError:
             repo.ui.warn(
                 _('error: could not put received data into largefile store'))
             return wireproto.pushres(1)
     finally:
         if f:
             f.close()
 
     return wireproto.pushres(0)
 
 def getlfile(repo, proto, sha):
     '''Retrieve a largefile from the repository-local cache or system
     cache.'''
     filename = lfutil.findfile(repo, sha)
     if not filename:
         raise util.Abort(_('requested largefile %s not present in cache') % sha)
     f = open(filename, 'rb')
     length = os.fstat(f.fileno())[6]
 
     # Since we can't set an HTTP content-length header here, and
     # Mercurial core provides no way to give the length of a streamres
     # (and reading the entire file into RAM would be ill-advised), we
     # just send the length on the first line of the response, like the
     # ssh proto does for string responses.
     def generator():
         yield '%d\n' % length
         for chunk in f:
             yield chunk
     return wireproto.streamres(generator())
 
 def statlfile(repo, proto, sha):
     '''Return '2\n' if the largefile is missing, '1\n' if it has a
     mismatched checksum, or '0\n' if it is in good condition'''
     filename = lfutil.findfile(repo, sha)
     if not filename:
         return '2\n'
     fd = None
     try:
         fd = open(filename, 'rb')
         return lfutil.hexsha1(fd) == sha and '0\n' or '1\n'
     finally:
         if fd:
             fd.close()
 
 def wirereposetup(ui, repo):
     class lfileswirerepository(repo.__class__):
         def putlfile(self, sha, fd):
             # unfortunately, httprepository._callpush tries to convert its
             # input file-like into a bundle before sending it, so we can't use
             # it ...
             if issubclass(self.__class__, httprepo.httprepository):
                 try:
                     return int(self._call('putlfile', data=fd, sha=sha,
                         headers={'content-type':'application/mercurial-0.1'}))
                 except (ValueError, urllib2.HTTPError):
                     return 1
             # ... but we can't use sshrepository._call because the data=
             # argument won't get sent, and _callpush does exactly what we want
             # in this case: send the data straight through
             else:
                 try:
                     ret, output = self._callpush("putlfile", fd, sha=sha)
                     if ret == "":
                         raise error.ResponseError(_('putlfile failed:'),
                                 output)
                     return int(ret)
                 except IOError:
                     return 1
                 except ValueError:
                     raise error.ResponseError(
                         _('putlfile failed (unexpected response):'), ret)
 
         def getlfile(self, sha):
             stream = self._callstream("getlfile", sha=sha)
             length = stream.readline()
             try:
                 length = int(length)
             except ValueError:
                 self._abort(error.ResponseError(_("unexpected response:"),
                                                 length))
             return (length, stream)
 
         def statlfile(self, sha):
             try:
                 return int(self._call("statlfile", sha=sha))
             except (ValueError, urllib2.HTTPError):
                 # If the server returns anything but an integer followed by a
                 # newline, newline, it's not speaking our language; if we get
                 # an HTTP error, we can't be sure the largefile is present;
                 # either way, consider it missing.
                 return 2
 
     repo.__class__ = lfileswirerepository
 
 # advertise the largefiles=serve capability
 def capabilities(repo, proto):
     return capabilities_orig(repo, proto) + ' largefiles=serve'
 
 # duplicate what Mercurial's new out-of-band errors mechanism does, because
 # clients old and new alike both handle it well
 def webproto_refuseclient(self, message):
     self.req.header([('Content-Type', 'application/hg-error')])
     return message
 
 def sshproto_refuseclient(self, message):
     self.ui.write_err('%s\n-\n' % message)
     self.fout.write('\n')
     self.fout.flush()
 
     return ''
 
 def heads(repo, proto):
     if lfutil.islfilesrepo(repo):
         return wireproto.ooberror(LARGEFILES_REQUIRED_MSG)
     return wireproto.heads(repo, proto)
 
 def sshrepo_callstream(self, cmd, **args):
     if cmd == 'heads' and self.capable('largefiles'):
         cmd = 'lheads'
     if cmd == 'batch' and self.capable('largefiles'):
         args['cmds'] = args['cmds'].replace('heads ', 'lheads ')
     return ssh_oldcallstream(self, cmd, **args)
 
 def httprepo_callstream(self, cmd, **args):
     if cmd == 'heads' and self.capable('largefiles'):
         cmd = 'lheads'
     if cmd == 'batch' and self.capable('largefiles'):
         args['cmds'] = args['cmds'].replace('heads ', 'lheads ')
     return http_oldcallstream(self, cmd, **args)

hgext/largefiles/reposetup.py

@@ -1,411 +1,411
 # Copyright 2009-2010 Gregory P. Ward
 # Copyright 2009-2010 Intelerad Medical Systems Incorporated
 # Copyright 2010-2011 Fog Creek Software
 # Copyright 2010-2011 Unity Technologies
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 '''setup for largefiles repositories: reposetup'''
 import copy
 import types
 import os
 import re
 
 from mercurial import context, error, manifest, match as match_, node, util
 from mercurial.i18n import _
 
 import lfcommands
 import proto
 import lfutil
 
 def reposetup(ui, repo):
     # wire repositories should be given new wireproto functions but not the
     # other largefiles modifications
     if not repo.local():
         return proto.wirereposetup(ui, repo)
 
     for name in ('status', 'commitctx', 'commit', 'push'):
         method = getattr(repo, name)
         #if not (isinstance(method, types.MethodType) and
         #        method.im_func is repo.__class__.commitctx.im_func):
         if (isinstance(method, types.FunctionType) and
             method.func_name == 'wrap'):
             ui.warn(_('largefiles: repo method %r appears to have already been'
                     ' wrapped by another extension: '
                     'largefiles may behave incorrectly\n')
                     % name)
 
     class lfiles_repo(repo.__class__):
         lfstatus = False
         def status_nolfiles(self, *args, **kwargs):
             return super(lfiles_repo, self).status(*args, **kwargs)
 
         # When lfstatus is set, return a context that gives the names
         # of largefiles instead of their corresponding standins and
         # identifies the largefiles as always binary, regardless of
         # their actual contents.
         def __getitem__(self, changeid):
             ctx = super(lfiles_repo, self).__getitem__(changeid)
             if self.lfstatus:
                 class lfiles_manifestdict(manifest.manifestdict):
                     def __contains__(self, filename):
                         if super(lfiles_manifestdict,
                                 self).__contains__(filename):
                             return True
                         return super(lfiles_manifestdict,
                             self).__contains__(lfutil.shortname+'/' + filename)
                 class lfiles_ctx(ctx.__class__):
                     def files(self):
                         filenames = super(lfiles_ctx, self).files()
                         return [re.sub('^\\'+lfutil.shortname+'/', '',
                                        filename) for filename in filenames]
                     def manifest(self):
                         man1 = super(lfiles_ctx, self).manifest()
                         man1.__class__ = lfiles_manifestdict
                         return man1
                     def filectx(self, path, fileid=None, filelog=None):
                         try:
                             result = super(lfiles_ctx, self).filectx(path,
                                 fileid, filelog)
                         except error.LookupError:
                             # Adding a null character will cause Mercurial to
                             # identify this as a binary file.
                             result = super(lfiles_ctx, self).filectx(
                                 lfutil.shortname + '/' + path, fileid,
                                 filelog)
                             olddata = result.data
                             result.data = lambda: olddata() + '\0'
                         return result
                 ctx.__class__ = lfiles_ctx
             return ctx
 
         # Figure out the status of big files and insert them into the
         # appropriate list in the result. Also removes standin files
         # from the listing. Revert to the original status if
         # self.lfstatus is False.
         def status(self, node1='.', node2=None, match=None, ignored=False,
                 clean=False, unknown=False, listsubrepos=False):
             listignored, listclean, listunknown = ignored, clean, unknown
             if not self.lfstatus:
                 try:
                     return super(lfiles_repo, self).status(node1, node2, match,
                         listignored, listclean, listunknown, listsubrepos)
                 except TypeError:
                     return super(lfiles_repo, self).status(node1, node2, match,
                         listignored, listclean, listunknown)
             else:
                 # some calls in this function rely on the old version of status
                 self.lfstatus = False
                 if isinstance(node1, context.changectx):
                     ctx1 = node1
                 else:
                     ctx1 = repo[node1]
                 if isinstance(node2, context.changectx):
                     ctx2 = node2
                 else:
                     ctx2 = repo[node2]
                 working = ctx2.rev() is None
                 parentworking = working and ctx1 == self['.']
 
                 def inctx(file, ctx):
                     try:
                         if ctx.rev() is None:
                             return file in ctx.manifest()
                         ctx[file]
                         return True
                     except KeyError:
                         return False
 
                 if match is None:
                     match = match_.always(self.root, self.getcwd())
 
                 # Create a copy of match that matches standins instead
                 # of largefiles.
                 def tostandin(file):
                     if inctx(lfutil.standin(file), ctx2):
                         return lfutil.standin(file)
                     return file
 
                 m = copy.copy(match)
                 m._files = [tostandin(f) for f in m._files]
 
                 # get ignored, clean, and unknown but remove them
                 # later if they were not asked for
                 try:
                     result = super(lfiles_repo, self).status(node1, node2, m,
                         True, True, True, listsubrepos)
                 except TypeError:
                     result = super(lfiles_repo, self).status(node1, node2, m,
                         True, True, True)
                 if working:
                     # hold the wlock while we read largefiles and
                     # update the lfdirstate
                     wlock = repo.wlock()
                     try:
                         # Any non-largefiles that were explicitly listed must be
                         # taken out or lfdirstate.status will report an error.
                         # The status of these files was already computed using
                         # super's status.
                         lfdirstate = lfutil.openlfdirstate(ui, self)
                         match._files = [f for f in match._files if f in
                             lfdirstate]
                         s = lfdirstate.status(match, [], listignored,
                                 listclean, listunknown)
                         (unsure, modified, added, removed, missing, unknown,
                                 ignored, clean) = s
                         if parentworking:
                             for lfile in unsure:
                                 if ctx1[lfutil.standin(lfile)].data().strip() \
                                         != lfutil.hashfile(self.wjoin(lfile)):
                                     modified.append(lfile)
                                 else:
                                     clean.append(lfile)
                                     lfdirstate.normal(lfile)
                             lfdirstate.write()
                         else:
                             tocheck = unsure + modified + added + clean
                             modified, added, clean = [], [], []
 
                             for lfile in tocheck:
                                 standin = lfutil.standin(lfile)
                                 if inctx(standin, ctx1):
                                     if ctx1[standin].data().strip() != \
                                             lfutil.hashfile(self.wjoin(lfile)):
                                         modified.append(lfile)
                                     else:
                                         clean.append(lfile)
                                 else:
                                     added.append(lfile)
                     finally:
                         wlock.release()
 
                     for standin in ctx1.manifest():
                         if not lfutil.isstandin(standin):
                             continue
                         lfile = lfutil.splitstandin(standin)
                         if not match(lfile):
                             continue
                         if lfile not in lfdirstate:
                             removed.append(lfile)
                     # Handle unknown and ignored differently
                     lfiles = (modified, added, removed, missing, [], [], clean)
                     result = list(result)
                     # Unknown files
                     result[4] = [f for f in unknown
                                  if (repo.dirstate[f] == '?' and
                                      not lfutil.isstandin(f))]
                     # Ignored files must be ignored by both the dirstate and
                     # lfdirstate
                     result[5] = set(ignored).intersection(set(result[5]))
                     # combine normal files and largefiles
                     normals = [[fn for fn in filelist
                                 if not lfutil.isstandin(fn)]
                                for filelist in result]
                     result = [sorted(list1 + list2)
                               for (list1, list2) in zip(normals, lfiles)]
                 else:
                     def toname(f):
                         if lfutil.isstandin(f):
                             return lfutil.splitstandin(f)
                         return f
                     result = [[toname(f) for f in items] for items in result]
 
                 if not listunknown:
                     result[4] = []
                 if not listignored:
                     result[5] = []
                 if not listclean:
                     result[6] = []
                 self.lfstatus = True
                 return result
 
         # As part of committing, copy all of the largefiles into the
         # cache.
         def commitctx(self, *args, **kwargs):
             node = super(lfiles_repo, self).commitctx(*args, **kwargs)
             ctx = self[node]
             for filename in ctx.files():
                 if lfutil.isstandin(filename) and filename in ctx.manifest():
                     realfile = lfutil.splitstandin(filename)
-                    lfutil.copytocache(self, ctx.node(), realfile)
+                    lfutil.copytostore(self, ctx.node(), realfile)
 
             return node
 
         # Before commit, largefile standins have not had their
         # contents updated to reflect the hash of their largefile.
         # Do that here.
         def commit(self, text="", user=None, date=None, match=None,
                 force=False, editor=False, extra={}):
             orig = super(lfiles_repo, self).commit
 
             wlock = repo.wlock()
             try:
                 if getattr(repo, "_isrebasing", False):
                     # We have to take the time to pull down the new
                     # largefiles now. Otherwise if we are rebasing,
                     # any largefiles that were modified in the
                     # destination changesets get overwritten, either
                     # by the rebase or in the first commit after the
                     # rebase.
                     lfcommands.updatelfiles(repo.ui, repo)
                 # 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
                     (not match.anypats() and not match.files())):
                     # 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 = lfutil.openlfdirstate(ui, self)
                     dirtymatch = match_.always(repo.root, repo.getcwd())
                     s = lfdirstate.status(dirtymatch, [], False, False, False)
                     modifiedfiles = []
                     for i in s:
                         modifiedfiles.extend(i)
                     lfiles = lfutil.listlfiles(self)
                     # this only loops through largefiles that exist (not
                     # removed/renamed)
                     for lfile in lfiles:
                         if lfile in modifiedfiles:
                             if os.path.exists(self.wjoin(lfutil.standin(lfile))):
                                 # this handles the case where a rebase is being
                                 # performed and the working copy is not updated
                                 # yet.
                                 if os.path.exists(self.wjoin(lfile)):
                                     lfutil.updatestandin(self,
                                         lfutil.standin(lfile))
                                     lfdirstate.normal(lfile)
                     for lfile in lfdirstate:
                         if lfile in modifiedfiles:
                             if not os.path.exists(
                                     repo.wjoin(lfutil.standin(lfile))):
                                 lfdirstate.drop(lfile)
                     lfdirstate.write()
 
                     return orig(text=text, user=user, date=date, match=match,
                                     force=force, editor=editor, extra=extra)
 
                 for f in match.files():
                     if lfutil.isstandin(f):
                         raise util.Abort(
                             _('file "%s" is a largefile standin') % f,
                             hint=('commit the largefile itself instead'))
 
                 # Case 2: user calls commit with specified patterns: refresh
                 # any matching big files.
                 smatcher = lfutil.composestandinmatcher(self, match)
                 standins = lfutil.dirstate_walk(self.dirstate, smatcher)
 
                 # No matching big files: get out of the way and pass control to
                 # the usual commit() method.
                 if not standins:
                     return orig(text=text, user=user, date=date, match=match,
                                     force=force, editor=editor, extra=extra)
 
                 # 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 = lfutil.openlfdirstate(ui, self)
                 for standin in standins:
                     lfile = lfutil.splitstandin(standin)
                     if lfdirstate[lfile] <> 'r':
                         lfutil.updatestandin(self, standin)
                         lfdirstate.normal(lfile)
                     else:
                         lfdirstate.drop(lfile)
                 lfdirstate.write()
 
                 # 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.
                 lfiles = lfutil.listlfiles(repo)
                 match = copy.copy(match)
                 orig_matchfn = 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 = lfutil.standin(f)
 
                     # ignore known largefiles and standins
                     if f in lfiles or fstandin in standins:
                         continue
 
                     # append directory separator to avoid collisions
                     if not fstandin.endswith(os.sep):
                         fstandin += os.sep
 
                     # prevalidate matching standin directories
                     if lfutil.any_(st for st in match._files
                                    if st.startswith(fstandin)):
                         continue
                     actualfiles.append(f)
                 match._files = actualfiles
 
                 def matchfn(f):
                     if orig_matchfn(f):
                         return f not in lfiles
                     else:
                         return f in standins
 
                 match.matchfn = matchfn
                 return orig(text=text, user=user, date=date, match=match,
                                 force=force, editor=editor, extra=extra)
             finally:
                 wlock.release()
 
         def push(self, remote, force=False, revs=None, newbranch=False):
             o = lfutil.findoutgoing(repo, remote, force)
             if o:
                 toupload = set()
                 o = repo.changelog.nodesbetween(o, revs)[0]
                 for n in o:
                     parents = [p for p in repo.changelog.parents(n)
                                if p != node.nullid]
                     ctx = repo[n]
                     files = set(ctx.files())
                     if len(parents) == 2:
                         mc = ctx.manifest()
                         mp1 = ctx.parents()[0].manifest()
                         mp2 = ctx.parents()[1].manifest()
                         for f in mp1:
                             if f not in mc:
                                 files.add(f)
                         for f in mp2:
                             if f not in mc:
                                 files.add(f)
                         for f in mc:
                             if mc[f] != mp1.get(f, None) or mc[f] != mp2.get(f,
                                     None):
                                 files.add(f)
 
                     toupload = toupload.union(
                         set([ctx[f].data().strip()
                              for f in files
                              if lfutil.isstandin(f) and f in ctx]))
                 lfcommands.uploadlfiles(ui, self, remote, toupload)
             return super(lfiles_repo, self).push(remote, force, revs,
                 newbranch)
 
     repo.__class__ = lfiles_repo
 
     def checkrequireslfiles(ui, repo, **kwargs):
         if 'largefiles' not in repo.requirements and lfutil.any_(
                 lfutil.shortname+'/' in f[0] for f in repo.store.datafiles()):
             repo.requirements.add('largefiles')
             repo._writerequirements()
 
     checkrequireslfiles(ui, repo)
 
     ui.setconfig('hooks', 'changegroup.lfiles', checkrequireslfiles)
     ui.setconfig('hooks', 'commit.lfiles', checkrequireslfiles)