remotefilelog: tell runbgcommand to not block on child process startup

These two invocations will always find a binary because they're
re-running hg. As a result, we can skip waiting for the subprocess to
start running and save a little bit of wall-time.

Differential Revision: https://phab.mercurial-scm.org/D6539

hgext/remotefilelog/repack.py

@@ -1,777 +1,778 b''
 from __future__ import absolute_import
 
 import os
 import time
 
 from mercurial.i18n import _
 from mercurial.node import (
     nullid,
     short,
 )
 from mercurial import (
     encoding,
     error,
     mdiff,
     policy,
     pycompat,
     scmutil,
     util,
     vfs,
 )
 from mercurial.utils import procutil
 from . import (
     constants,
     contentstore,
     datapack,
     extutil,
     historypack,
     metadatastore,
     shallowutil,
 )
 
 osutil = policy.importmod(r'osutil')
 
 class RepackAlreadyRunning(error.Abort):
     pass
 
 def backgroundrepack(repo, incremental=True, packsonly=False):
     cmd = [procutil.hgexecutable(), '-R', repo.origroot, 'repack']
     msg = _("(running background repack)\n")
     if incremental:
         cmd.append('--incremental')
         msg = _("(running background incremental repack)\n")
     if packsonly:
         cmd.append('--packsonly')
     repo.ui.warn(msg)
-    procutil.runbgcommand(cmd, encoding.environ)
+    # We know this command will find a binary, so don't block on it starting.
+    procutil.runbgcommand(cmd, encoding.environ, ensurestart=False)
 
 def fullrepack(repo, options=None):
     """If ``packsonly`` is True, stores creating only loose objects are skipped.
     """
     if util.safehasattr(repo, 'shareddatastores'):
         datasource = contentstore.unioncontentstore(
             *repo.shareddatastores)
         historysource = metadatastore.unionmetadatastore(
             *repo.sharedhistorystores,
             allowincomplete=True)
 
         packpath = shallowutil.getcachepackpath(
             repo,
             constants.FILEPACK_CATEGORY)
         _runrepack(repo, datasource, historysource, packpath,
                    constants.FILEPACK_CATEGORY, options=options)
 
     if util.safehasattr(repo.manifestlog, 'datastore'):
         localdata, shareddata = _getmanifeststores(repo)
         lpackpath, ldstores, lhstores = localdata
         spackpath, sdstores, shstores = shareddata
 
         # Repack the shared manifest store
         datasource = contentstore.unioncontentstore(*sdstores)
         historysource = metadatastore.unionmetadatastore(
                         *shstores,
                         allowincomplete=True)
         _runrepack(repo, datasource, historysource, spackpath,
                    constants.TREEPACK_CATEGORY, options=options)
 
         # Repack the local manifest store
         datasource = contentstore.unioncontentstore(
                         *ldstores,
                         allowincomplete=True)
         historysource = metadatastore.unionmetadatastore(
                         *lhstores,
                         allowincomplete=True)
         _runrepack(repo, datasource, historysource, lpackpath,
                    constants.TREEPACK_CATEGORY, options=options)
 
 def incrementalrepack(repo, options=None):
     """This repacks the repo by looking at the distribution of pack files in the
     repo and performing the most minimal repack to keep the repo in good shape.
     """
     if util.safehasattr(repo, 'shareddatastores'):
         packpath = shallowutil.getcachepackpath(
             repo,
             constants.FILEPACK_CATEGORY)
         _incrementalrepack(repo,
                            repo.shareddatastores,
                            repo.sharedhistorystores,
                            packpath,
                            constants.FILEPACK_CATEGORY,
                            options=options)
 
     if util.safehasattr(repo.manifestlog, 'datastore'):
         localdata, shareddata = _getmanifeststores(repo)
         lpackpath, ldstores, lhstores = localdata
         spackpath, sdstores, shstores = shareddata
 
         # Repack the shared manifest store
         _incrementalrepack(repo,
                            sdstores,
                            shstores,
                            spackpath,
                            constants.TREEPACK_CATEGORY,
                            options=options)
 
         # Repack the local manifest store
         _incrementalrepack(repo,
                            ldstores,
                            lhstores,
                            lpackpath,
                            constants.TREEPACK_CATEGORY,
                            allowincompletedata=True,
                            options=options)
 
 def _getmanifeststores(repo):
     shareddatastores = repo.manifestlog.shareddatastores
     localdatastores = repo.manifestlog.localdatastores
     sharedhistorystores = repo.manifestlog.sharedhistorystores
     localhistorystores = repo.manifestlog.localhistorystores
 
     sharedpackpath = shallowutil.getcachepackpath(repo,
                                             constants.TREEPACK_CATEGORY)
     localpackpath = shallowutil.getlocalpackpath(repo.svfs.vfs.base,
                                             constants.TREEPACK_CATEGORY)
 
     return ((localpackpath, localdatastores, localhistorystores),
             (sharedpackpath, shareddatastores, sharedhistorystores))
 
 def _topacks(packpath, files, constructor):
     paths = list(os.path.join(packpath, p) for p in files)
     packs = list(constructor(p) for p in paths)
     return packs
 
 def _deletebigpacks(repo, folder, files):
     """Deletes packfiles that are bigger than ``packs.maxpacksize``.
 
     Returns ``files` with the removed files omitted."""
     maxsize = repo.ui.configbytes("packs", "maxpacksize")
     if maxsize <= 0:
         return files
 
     # This only considers datapacks today, but we could broaden it to include
     # historypacks.
     VALIDEXTS = [".datapack", ".dataidx"]
 
     # Either an oversize index or datapack will trigger cleanup of the whole
     # pack:
     oversized = {os.path.splitext(path)[0] for path, ftype, stat in files
         if (stat.st_size > maxsize and (os.path.splitext(path)[1]
                                         in VALIDEXTS))}
 
     for rootfname in oversized:
         rootpath = os.path.join(folder, rootfname)
         for ext in VALIDEXTS:
             path = rootpath + ext
             repo.ui.debug('removing oversize packfile %s (%s)\n' %
                           (path, util.bytecount(os.stat(path).st_size)))
             os.unlink(path)
     return [row for row in files if os.path.basename(row[0]) not in oversized]
 
 def _incrementalrepack(repo, datastore, historystore, packpath, category,
         allowincompletedata=False, options=None):
     shallowutil.mkstickygroupdir(repo.ui, packpath)
 
     files = osutil.listdir(packpath, stat=True)
     files = _deletebigpacks(repo, packpath, files)
     datapacks = _topacks(packpath,
         _computeincrementaldatapack(repo.ui, files),
         datapack.datapack)
     datapacks.extend(s for s in datastore
                      if not isinstance(s, datapack.datapackstore))
 
     historypacks = _topacks(packpath,
         _computeincrementalhistorypack(repo.ui, files),
         historypack.historypack)
     historypacks.extend(s for s in historystore
                         if not isinstance(s, historypack.historypackstore))
 
     # ``allhistory{files,packs}`` contains all known history packs, even ones we
     # don't plan to repack. They are used during the datapack repack to ensure
     # good ordering of nodes.
     allhistoryfiles = _allpackfileswithsuffix(files, historypack.PACKSUFFIX,
                             historypack.INDEXSUFFIX)
     allhistorypacks = _topacks(packpath,
         (f for f, mode, stat in allhistoryfiles),
         historypack.historypack)
     allhistorypacks.extend(s for s in historystore
                         if not isinstance(s, historypack.historypackstore))
     _runrepack(repo,
                contentstore.unioncontentstore(
                    *datapacks,
                    allowincomplete=allowincompletedata),
                metadatastore.unionmetadatastore(
                    *historypacks,
                    allowincomplete=True),
                packpath, category,
                fullhistory=metadatastore.unionmetadatastore(
                    *allhistorypacks,
                    allowincomplete=True),
                 options=options)
 
 def _computeincrementaldatapack(ui, files):
     opts = {
         'gencountlimit' : ui.configint(
             'remotefilelog', 'data.gencountlimit'),
         'generations' : ui.configlist(
             'remotefilelog', 'data.generations'),
         'maxrepackpacks' : ui.configint(
             'remotefilelog', 'data.maxrepackpacks'),
         'repackmaxpacksize' : ui.configbytes(
             'remotefilelog', 'data.repackmaxpacksize'),
         'repacksizelimit' : ui.configbytes(
             'remotefilelog', 'data.repacksizelimit'),
     }
 
     packfiles = _allpackfileswithsuffix(
         files, datapack.PACKSUFFIX, datapack.INDEXSUFFIX)
     return _computeincrementalpack(packfiles, opts)
 
 def _computeincrementalhistorypack(ui, files):
     opts = {
         'gencountlimit' : ui.configint(
             'remotefilelog', 'history.gencountlimit'),
         'generations' : ui.configlist(
             'remotefilelog', 'history.generations', ['100MB']),
         'maxrepackpacks' : ui.configint(
             'remotefilelog', 'history.maxrepackpacks'),
         'repackmaxpacksize' : ui.configbytes(
             'remotefilelog', 'history.repackmaxpacksize', '400MB'),
         'repacksizelimit' : ui.configbytes(
             'remotefilelog', 'history.repacksizelimit'),
     }
 
     packfiles = _allpackfileswithsuffix(
         files, historypack.PACKSUFFIX, historypack.INDEXSUFFIX)
     return _computeincrementalpack(packfiles, opts)
 
 def _allpackfileswithsuffix(files, packsuffix, indexsuffix):
     result = []
     fileset = set(fn for fn, mode, stat in files)
     for filename, mode, stat in files:
         if not filename.endswith(packsuffix):
             continue
 
         prefix = filename[:-len(packsuffix)]
 
         # Don't process a pack if it doesn't have an index.
         if (prefix + indexsuffix) not in fileset:
             continue
         result.append((prefix, mode, stat))
 
     return result
 
 def _computeincrementalpack(files, opts):
     """Given a set of pack files along with the configuration options, this
     function computes the list of files that should be packed as part of an
     incremental repack.
 
     It tries to strike a balance between keeping incremental repacks cheap (i.e.
     packing small things when possible, and rolling the packs up to the big ones
     over time).
     """
 
     limits = list(sorted((util.sizetoint(s) for s in opts['generations']),
                                 reverse=True))
     limits.append(0)
 
     # Group the packs by generation (i.e. by size)
     generations = []
     for i in pycompat.xrange(len(limits)):
         generations.append([])
 
     sizes = {}
     for prefix, mode, stat in files:
         size = stat.st_size
         if size > opts['repackmaxpacksize']:
             continue
 
         sizes[prefix] = size
         for i, limit in enumerate(limits):
             if size > limit:
                 generations[i].append(prefix)
                 break
 
     # Steps for picking what packs to repack:
     # 1. Pick the largest generation with > gencountlimit pack files.
     # 2. Take the smallest three packs.
     # 3. While total-size-of-packs < repacksizelimit: add another pack
 
     # Find the largest generation with more than gencountlimit packs
     genpacks = []
     for i, limit in enumerate(limits):
         if len(generations[i]) > opts['gencountlimit']:
             # Sort to be smallest last, for easy popping later
             genpacks.extend(sorted(generations[i], reverse=True,
                                    key=lambda x: sizes[x]))
             break
 
     # Take as many packs from the generation as we can
     chosenpacks = genpacks[-3:]
     genpacks = genpacks[:-3]
     repacksize = sum(sizes[n] for n in chosenpacks)
     while (repacksize < opts['repacksizelimit'] and genpacks and
            len(chosenpacks) < opts['maxrepackpacks']):
         chosenpacks.append(genpacks.pop())
         repacksize += sizes[chosenpacks[-1]]
 
     return chosenpacks
 
 def _runrepack(repo, data, history, packpath, category, fullhistory=None,
                options=None):
     shallowutil.mkstickygroupdir(repo.ui, packpath)
 
     def isold(repo, filename, node):
         """Check if the file node is older than a limit.
         Unless a limit is specified in the config the default limit is taken.
         """
         filectx = repo.filectx(filename, fileid=node)
         filetime = repo[filectx.linkrev()].date()
 
         ttl = repo.ui.configint('remotefilelog', 'nodettl')
 
         limit = time.time() - ttl
         return filetime[0] < limit
 
     garbagecollect = repo.ui.configbool('remotefilelog', 'gcrepack')
     if not fullhistory:
         fullhistory = history
     packer = repacker(repo, data, history, fullhistory, category,
                       gc=garbagecollect, isold=isold, options=options)
 
     with datapack.mutabledatapack(repo.ui, packpath) as dpack:
         with historypack.mutablehistorypack(repo.ui, packpath) as hpack:
             try:
                 packer.run(dpack, hpack)
             except error.LockHeld:
                 raise RepackAlreadyRunning(_("skipping repack - another repack "
                                              "is already running"))
 
 def keepset(repo, keyfn, lastkeepkeys=None):
     """Computes a keepset which is not garbage collected.
     'keyfn' is a function that maps filename, node to a unique key.
     'lastkeepkeys' is an optional argument and if provided the keepset
     function updates lastkeepkeys with more keys and returns the result.
     """
     if not lastkeepkeys:
         keepkeys = set()
     else:
         keepkeys = lastkeepkeys
 
     # We want to keep:
     # 1. Working copy parent
     # 2. Draft commits
     # 3. Parents of draft commits
     # 4. Pullprefetch and bgprefetchrevs revsets if specified
     revs = ['.', 'draft()', 'parents(draft())']
     prefetchrevs = repo.ui.config('remotefilelog', 'pullprefetch', None)
     if prefetchrevs:
         revs.append('(%s)' % prefetchrevs)
     prefetchrevs = repo.ui.config('remotefilelog', 'bgprefetchrevs', None)
     if prefetchrevs:
         revs.append('(%s)' % prefetchrevs)
     revs = '+'.join(revs)
 
     revs = ['sort((%s), "topo")' % revs]
     keep = scmutil.revrange(repo, revs)
 
     processed = set()
     lastmanifest = None
 
     # process the commits in toposorted order starting from the oldest
     for r in reversed(keep._list):
         if repo[r].p1().rev() in processed:
             # if the direct parent has already been processed
             # then we only need to process the delta
             m = repo[r].manifestctx().readdelta()
         else:
             # otherwise take the manifest and diff it
             # with the previous manifest if one exists
             if lastmanifest:
                 m = repo[r].manifest().diff(lastmanifest)
             else:
                 m = repo[r].manifest()
         lastmanifest = repo[r].manifest()
         processed.add(r)
 
         # populate keepkeys with keys from the current manifest
         if type(m) is dict:
             # m is a result of diff of two manifests and is a dictionary that
             # maps filename to ((newnode, newflag), (oldnode, oldflag)) tuple
             for filename, diff in m.iteritems():
                 if diff[0][0] is not None:
                     keepkeys.add(keyfn(filename, diff[0][0]))
         else:
             # m is a manifest object
             for filename, filenode in m.iteritems():
                 keepkeys.add(keyfn(filename, filenode))
 
     return keepkeys
 
 class repacker(object):
     """Class for orchestrating the repack of data and history information into a
     new format.
     """
     def __init__(self, repo, data, history, fullhistory, category, gc=False,
                  isold=None, options=None):
         self.repo = repo
         self.data = data
         self.history = history
         self.fullhistory = fullhistory
         self.unit = constants.getunits(category)
         self.garbagecollect = gc
         self.options = options
         if self.garbagecollect:
             if not isold:
                 raise ValueError("Function 'isold' is not properly specified")
             # use (filename, node) tuple as a keepset key
             self.keepkeys = keepset(repo, lambda f, n : (f, n))
             self.isold = isold
 
     def run(self, targetdata, targethistory):
         ledger = repackledger()
 
         with extutil.flock(repacklockvfs(self.repo).join("repacklock"),
                            _('repacking %s') % self.repo.origroot, timeout=0):
             self.repo.hook('prerepack')
 
             # Populate ledger from source
             self.data.markledger(ledger, options=self.options)
             self.history.markledger(ledger, options=self.options)
 
             # Run repack
             self.repackdata(ledger, targetdata)
             self.repackhistory(ledger, targethistory)
 
             # Call cleanup on each source
             for source in ledger.sources:
                 source.cleanup(ledger)
 
     def _chainorphans(self, ui, filename, nodes, orphans, deltabases):
         """Reorderes ``orphans`` into a single chain inside ``nodes`` and
         ``deltabases``.
 
         We often have orphan entries (nodes without a base that aren't
         referenced by other nodes -- i.e., part of a chain) due to gaps in
         history. Rather than store them as individual fulltexts, we prefer to
         insert them as one chain sorted by size.
         """
         if not orphans:
             return nodes
 
         def getsize(node, default=0):
             meta = self.data.getmeta(filename, node)
             if constants.METAKEYSIZE in meta:
                 return meta[constants.METAKEYSIZE]
             else:
                 return default
 
         # Sort orphans by size; biggest first is preferred, since it's more
         # likely to be the newest version assuming files grow over time.
         # (Sort by node first to ensure the sort is stable.)
         orphans = sorted(orphans)
         orphans = list(sorted(orphans, key=getsize, reverse=True))
         if ui.debugflag:
             ui.debug("%s: orphan chain: %s\n" % (filename,
                 ", ".join([short(s) for s in orphans])))
 
         # Create one contiguous chain and reassign deltabases.
         for i, node in enumerate(orphans):
             if i == 0:
                 deltabases[node] = (nullid, 0)
             else:
                 parent = orphans[i - 1]
                 deltabases[node] = (parent, deltabases[parent][1] + 1)
         nodes = [n for n in nodes if n not in orphans]
         nodes += orphans
         return nodes
 
     def repackdata(self, ledger, target):
         ui = self.repo.ui
         maxchainlen = ui.configint('packs', 'maxchainlen', 1000)
 
         byfile = {}
         for entry in ledger.entries.itervalues():
             if entry.datasource:
                 byfile.setdefault(entry.filename, {})[entry.node] = entry
 
         count = 0
         repackprogress = ui.makeprogress(_("repacking data"), unit=self.unit,
                                             total=len(byfile))
         for filename, entries in sorted(byfile.iteritems()):
             repackprogress.update(count)
 
             ancestors = {}
             nodes = list(node for node in entries)
             nohistory = []
             buildprogress = ui.makeprogress(_("building history"), unit='nodes',
                                             total=len(nodes))
             for i, node in enumerate(nodes):
                 if node in ancestors:
                     continue
                 buildprogress.update(i)
                 try:
                     ancestors.update(self.fullhistory.getancestors(filename,
                         node, known=ancestors))
                 except KeyError:
                     # Since we're packing data entries, we may not have the
                     # corresponding history entries for them. It's not a big
                     # deal, but the entries won't be delta'd perfectly.
                     nohistory.append(node)
             buildprogress.complete()
 
             # Order the nodes children first, so we can produce reverse deltas
             orderednodes = list(reversed(self._toposort(ancestors)))
             if len(nohistory) > 0:
                 ui.debug('repackdata: %d nodes without history\n' %
                          len(nohistory))
             orderednodes.extend(sorted(nohistory))
 
             # Filter orderednodes to just the nodes we want to serialize (it
             # currently also has the edge nodes' ancestors).
             orderednodes = list(filter(lambda node: node in nodes,
                                 orderednodes))
 
             # Garbage collect old nodes:
             if self.garbagecollect:
                 neworderednodes = []
                 for node in orderednodes:
                     # If the node is old and is not in the keepset, we skip it,
                     # and mark as garbage collected
                     if ((filename, node) not in self.keepkeys and
                         self.isold(self.repo, filename, node)):
                         entries[node].gced = True
                         continue
                     neworderednodes.append(node)
                 orderednodes = neworderednodes
 
             # Compute delta bases for nodes:
             deltabases = {}
             nobase = set()
             referenced = set()
             nodes = set(nodes)
             processprogress = ui.makeprogress(_("processing nodes"),
                                               unit='nodes',
                                               total=len(orderednodes))
             for i, node in enumerate(orderednodes):
                 processprogress.update(i)
                 # Find delta base
                 # TODO: allow delta'ing against most recent descendant instead
                 # of immediate child
                 deltatuple = deltabases.get(node, None)
                 if deltatuple is None:
                     deltabase, chainlen = nullid, 0
                     deltabases[node] = (nullid, 0)
                     nobase.add(node)
                 else:
                     deltabase, chainlen = deltatuple
                     referenced.add(deltabase)
 
                 # Use available ancestor information to inform our delta choices
                 ancestorinfo = ancestors.get(node)
                 if ancestorinfo:
                     p1, p2, linknode, copyfrom = ancestorinfo
 
                     # The presence of copyfrom means we're at a point where the
                     # file was copied from elsewhere. So don't attempt to do any
                     # deltas with the other file.
                     if copyfrom:
                         p1 = nullid
 
                     if chainlen < maxchainlen:
                         # Record this child as the delta base for its parents.
                         # This may be non optimal, since the parents may have
                         # many children, and this will only choose the last one.
                         # TODO: record all children and try all deltas to find
                         # best
                         if p1 != nullid:
                             deltabases[p1] = (node, chainlen + 1)
                         if p2 != nullid:
                             deltabases[p2] = (node, chainlen + 1)
 
             # experimental config: repack.chainorphansbysize
             if ui.configbool('repack', 'chainorphansbysize'):
                 orphans = nobase - referenced
                 orderednodes = self._chainorphans(ui, filename, orderednodes,
                     orphans, deltabases)
 
             # Compute deltas and write to the pack
             for i, node in enumerate(orderednodes):
                 deltabase, chainlen = deltabases[node]
                 # Compute delta
                 # TODO: Optimize the deltachain fetching. Since we're
                 # iterating over the different version of the file, we may
                 # be fetching the same deltachain over and over again.
                 if deltabase != nullid:
                     deltaentry = self.data.getdelta(filename, node)
                     delta, deltabasename, origdeltabase, meta = deltaentry
                     size = meta.get(constants.METAKEYSIZE)
                     if (deltabasename != filename or origdeltabase != deltabase
                         or size is None):
                         deltabasetext = self.data.get(filename, deltabase)
                         original = self.data.get(filename, node)
                         size = len(original)
                         delta = mdiff.textdiff(deltabasetext, original)
                 else:
                     delta = self.data.get(filename, node)
                     size = len(delta)
                     meta = self.data.getmeta(filename, node)
 
                 # TODO: don't use the delta if it's larger than the fulltext
                 if constants.METAKEYSIZE not in meta:
                     meta[constants.METAKEYSIZE] = size
                 target.add(filename, node, deltabase, delta, meta)
 
                 entries[node].datarepacked = True
 
             processprogress.complete()
             count += 1
 
         repackprogress.complete()
         target.close(ledger=ledger)
 
     def repackhistory(self, ledger, target):
         ui = self.repo.ui
 
         byfile = {}
         for entry in ledger.entries.itervalues():
             if entry.historysource:
                 byfile.setdefault(entry.filename, {})[entry.node] = entry
 
         progress = ui.makeprogress(_("repacking history"), unit=self.unit,
                                    total=len(byfile))
         for filename, entries in sorted(byfile.iteritems()):
             ancestors = {}
             nodes = list(node for node in entries)
 
             for node in nodes:
                 if node in ancestors:
                     continue
                 ancestors.update(self.history.getancestors(filename, node,
                                                            known=ancestors))
 
             # Order the nodes children first
             orderednodes = reversed(self._toposort(ancestors))
 
             # Write to the pack
             dontprocess = set()
             for node in orderednodes:
                 p1, p2, linknode, copyfrom = ancestors[node]
 
                 # If the node is marked dontprocess, but it's also in the
                 # explicit entries set, that means the node exists both in this
                 # file and in another file that was copied to this file.
                 # Usually this happens if the file was copied to another file,
                 # then the copy was deleted, then reintroduced without copy
                 # metadata. The original add and the new add have the same hash
                 # since the content is identical and the parents are null.
                 if node in dontprocess and node not in entries:
                     # If copyfrom == filename, it means the copy history
                     # went to come other file, then came back to this one, so we
                     # should continue processing it.
                     if p1 != nullid and copyfrom != filename:
                         dontprocess.add(p1)
                     if p2 != nullid:
                         dontprocess.add(p2)
                     continue
 
                 if copyfrom:
                     dontprocess.add(p1)
 
                 target.add(filename, node, p1, p2, linknode, copyfrom)
 
                 if node in entries:
                     entries[node].historyrepacked = True
 
             progress.increment()
 
         progress.complete()
         target.close(ledger=ledger)
 
     def _toposort(self, ancestors):
         def parentfunc(node):
             p1, p2, linknode, copyfrom = ancestors[node]
             parents = []
             if p1 != nullid:
                 parents.append(p1)
             if p2 != nullid:
                 parents.append(p2)
             return parents
 
         sortednodes = shallowutil.sortnodes(ancestors.keys(), parentfunc)
         return sortednodes
 
 class repackledger(object):
     """Storage for all the bookkeeping that happens during a repack. It contains
     the list of revisions being repacked, what happened to each revision, and
     which source store contained which revision originally (for later cleanup).
     """
     def __init__(self):
         self.entries = {}
         self.sources = {}
         self.created = set()
 
     def markdataentry(self, source, filename, node):
         """Mark the given filename+node revision as having a data rev in the
         given source.
         """
         entry = self._getorcreateentry(filename, node)
         entry.datasource = True
         entries = self.sources.get(source)
         if not entries:
             entries = set()
             self.sources[source] = entries
         entries.add(entry)
 
     def markhistoryentry(self, source, filename, node):
         """Mark the given filename+node revision as having a history rev in the
         given source.
         """
         entry = self._getorcreateentry(filename, node)
         entry.historysource = True
         entries = self.sources.get(source)
         if not entries:
             entries = set()
             self.sources[source] = entries
         entries.add(entry)
 
     def _getorcreateentry(self, filename, node):
         key = (filename, node)
         value = self.entries.get(key)
         if not value:
             value = repackentry(filename, node)
             self.entries[key] = value
 
         return value
 
     def addcreated(self, value):
         self.created.add(value)
 
 class repackentry(object):
     """Simple class representing a single revision entry in the repackledger.
     """
     __slots__ = (r'filename', r'node', r'datasource', r'historysource',
                  r'datarepacked', r'historyrepacked', r'gced')
     def __init__(self, filename, node):
         self.filename = filename
         self.node = node
         # If the revision has a data entry in the source
         self.datasource = False
         # If the revision has a history entry in the source
         self.historysource = False
         # If the revision's data entry was repacked into the repack target
         self.datarepacked = False
         # If the revision's history entry was repacked into the repack target
         self.historyrepacked = False
         # If garbage collected
         self.gced = False
 
 def repacklockvfs(repo):
     if util.safehasattr(repo, 'name'):
         # Lock in the shared cache so repacks across multiple copies of the same
         # repo are coordinated.
         sharedcachepath = shallowutil.getcachepackpath(
             repo,
             constants.FILEPACK_CATEGORY)
         return vfs.vfs(sharedcachepath)
     else:
         return repo.svfs

hgext/remotefilelog/shallowrepo.py

@@ -1,298 +1,300 b''
 # shallowrepo.py - shallow repository that uses remote filelogs
 #
 # Copyright 2013 Facebook, Inc.
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 from __future__ import absolute_import
 
 import os
 
 from mercurial.i18n import _
 from mercurial.node import hex, nullid, nullrev
 from mercurial import (
     encoding,
     error,
     localrepo,
     match,
     scmutil,
     sparse,
     util,
 )
 from mercurial.utils import procutil
 from . import (
     connectionpool,
     constants,
     contentstore,
     datapack,
     fileserverclient,
     historypack,
     metadatastore,
     remotefilectx,
     remotefilelog,
     shallowutil,
 )
 
 # These make*stores functions are global so that other extensions can replace
 # them.
 def makelocalstores(repo):
     """In-repo stores, like .hg/store/data; can not be discarded."""
     localpath = os.path.join(repo.svfs.vfs.base, 'data')
     if not os.path.exists(localpath):
         os.makedirs(localpath)
 
     # Instantiate local data stores
     localcontent = contentstore.remotefilelogcontentstore(
         repo, localpath, repo.name, shared=False)
     localmetadata = metadatastore.remotefilelogmetadatastore(
         repo, localpath, repo.name, shared=False)
     return localcontent, localmetadata
 
 def makecachestores(repo):
     """Typically machine-wide, cache of remote data; can be discarded."""
     # Instantiate shared cache stores
     cachepath = shallowutil.getcachepath(repo.ui)
     cachecontent = contentstore.remotefilelogcontentstore(
         repo, cachepath, repo.name, shared=True)
     cachemetadata = metadatastore.remotefilelogmetadatastore(
         repo, cachepath, repo.name, shared=True)
 
     repo.sharedstore = cachecontent
     repo.shareddatastores.append(cachecontent)
     repo.sharedhistorystores.append(cachemetadata)
 
     return cachecontent, cachemetadata
 
 def makeremotestores(repo, cachecontent, cachemetadata):
     """These stores fetch data from a remote server."""
     # Instantiate remote stores
     repo.fileservice = fileserverclient.fileserverclient(repo)
     remotecontent = contentstore.remotecontentstore(
         repo.ui, repo.fileservice, cachecontent)
     remotemetadata = metadatastore.remotemetadatastore(
         repo.ui, repo.fileservice, cachemetadata)
     return remotecontent, remotemetadata
 
 def makepackstores(repo):
     """Packs are more efficient (to read from) cache stores."""
     # Instantiate pack stores
     packpath = shallowutil.getcachepackpath(repo,
                                             constants.FILEPACK_CATEGORY)
     packcontentstore = datapack.datapackstore(repo.ui, packpath)
     packmetadatastore = historypack.historypackstore(repo.ui, packpath)
 
     repo.shareddatastores.append(packcontentstore)
     repo.sharedhistorystores.append(packmetadatastore)
     shallowutil.reportpackmetrics(repo.ui, 'filestore', packcontentstore,
         packmetadatastore)
     return packcontentstore, packmetadatastore
 
 def makeunionstores(repo):
     """Union stores iterate the other stores and return the first result."""
     repo.shareddatastores = []
     repo.sharedhistorystores = []
 
     packcontentstore, packmetadatastore = makepackstores(repo)
     cachecontent, cachemetadata = makecachestores(repo)
     localcontent, localmetadata = makelocalstores(repo)
     remotecontent, remotemetadata = makeremotestores(repo, cachecontent,
                                                      cachemetadata)
 
     # Instantiate union stores
     repo.contentstore = contentstore.unioncontentstore(
         packcontentstore, cachecontent,
         localcontent, remotecontent, writestore=localcontent)
     repo.metadatastore = metadatastore.unionmetadatastore(
         packmetadatastore, cachemetadata, localmetadata, remotemetadata,
         writestore=localmetadata)
 
     fileservicedatawrite = cachecontent
     fileservicehistorywrite = cachemetadata
     repo.fileservice.setstore(repo.contentstore, repo.metadatastore,
                               fileservicedatawrite, fileservicehistorywrite)
     shallowutil.reportpackmetrics(repo.ui, 'filestore',
         packcontentstore, packmetadatastore)
 
 def wraprepo(repo):
     class shallowrepository(repo.__class__):
         @util.propertycache
         def name(self):
             return self.ui.config('remotefilelog', 'reponame')
 
         @util.propertycache
         def fallbackpath(self):
             path = repo.ui.config("remotefilelog", "fallbackpath",
                                   repo.ui.config('paths', 'default'))
             if not path:
                 raise error.Abort("no remotefilelog server "
                     "configured - is your .hg/hgrc trusted?")
 
             return path
 
         def maybesparsematch(self, *revs, **kwargs):
             '''
             A wrapper that allows the remotefilelog to invoke sparsematch() if
             this is a sparse repository, or returns None if this is not a
             sparse repository.
             '''
             if revs:
                 ret = sparse.matcher(repo, revs=revs)
             else:
                 ret = sparse.matcher(repo)
 
             if ret.always():
                 return None
             return ret
 
         def file(self, f):
             if f[0] == '/':
                 f = f[1:]
 
             if self.shallowmatch(f):
                 return remotefilelog.remotefilelog(self.svfs, f, self)
             else:
                 return super(shallowrepository, self).file(f)
 
         def filectx(self, path, *args, **kwargs):
             if self.shallowmatch(path):
                 return remotefilectx.remotefilectx(self, path, *args, **kwargs)
             else:
                 return super(shallowrepository, self).filectx(path, *args,
                                                               **kwargs)
 
         @localrepo.unfilteredmethod
         def commitctx(self, ctx, error=False):
             """Add a new revision to current repository.
             Revision information is passed via the context argument.
             """
 
             # some contexts already have manifest nodes, they don't need any
             # prefetching (for example if we're just editing a commit message
             # we can reuse manifest
             if not ctx.manifestnode():
                 # prefetch files that will likely be compared
                 m1 = ctx.p1().manifest()
                 files = []
                 for f in ctx.modified() + ctx.added():
                     fparent1 = m1.get(f, nullid)
                     if fparent1 != nullid:
                         files.append((f, hex(fparent1)))
                 self.fileservice.prefetch(files)
             return super(shallowrepository, self).commitctx(ctx,
                                                             error=error)
 
         def backgroundprefetch(self, revs, base=None, repack=False, pats=None,
                                opts=None):
             """Runs prefetch in background with optional repack
             """
             cmd = [procutil.hgexecutable(), '-R', repo.origroot, 'prefetch']
             if repack:
                 cmd.append('--repack')
             if revs:
                 cmd += ['-r', revs]
-            procutil.runbgcommand(cmd, encoding.environ)
+            # We know this command will find a binary, so don't block
+            # on it starting.
+            procutil.runbgcommand(cmd, encoding.environ, ensurestart=False)
 
         def prefetch(self, revs, base=None, pats=None, opts=None):
             """Prefetches all the necessary file revisions for the given revs
             Optionally runs repack in background
             """
             with repo._lock(repo.svfs, 'prefetchlock', True, None, None,
                             _('prefetching in %s') % repo.origroot):
                 self._prefetch(revs, base, pats, opts)
 
         def _prefetch(self, revs, base=None, pats=None, opts=None):
             fallbackpath = self.fallbackpath
             if fallbackpath:
                 # If we know a rev is on the server, we should fetch the server
                 # version of those files, since our local file versions might
                 # become obsolete if the local commits are stripped.
                 localrevs = repo.revs('outgoing(%s)', fallbackpath)
                 if base is not None and base != nullrev:
                     serverbase = list(repo.revs('first(reverse(::%s) - %ld)',
                                                 base, localrevs))
                     if serverbase:
                         base = serverbase[0]
             else:
                 localrevs = repo
 
             mfl = repo.manifestlog
             mfrevlog = mfl.getstorage('')
             if base is not None:
                 mfdict = mfl[repo[base].manifestnode()].read()
                 skip = set(mfdict.iteritems())
             else:
                 skip = set()
 
             # Copy the skip set to start large and avoid constant resizing,
             # and since it's likely to be very similar to the prefetch set.
             files = skip.copy()
             serverfiles = skip.copy()
             visited = set()
             visited.add(nullrev)
             revcount = len(revs)
             progress = self.ui.makeprogress(_('prefetching'), total=revcount)
             progress.update(0)
             for rev in sorted(revs):
                 ctx = repo[rev]
                 if pats:
                     m = scmutil.match(ctx, pats, opts)
                 sparsematch = repo.maybesparsematch(rev)
 
                 mfnode = ctx.manifestnode()
                 mfrev = mfrevlog.rev(mfnode)
 
                 # Decompressing manifests is expensive.
                 # When possible, only read the deltas.
                 p1, p2 = mfrevlog.parentrevs(mfrev)
                 if p1 in visited and p2 in visited:
                     mfdict = mfl[mfnode].readfast()
                 else:
                     mfdict = mfl[mfnode].read()
 
                 diff = mfdict.iteritems()
                 if pats:
                     diff = (pf for pf in diff if m(pf[0]))
                 if sparsematch:
                     diff = (pf for pf in diff if sparsematch(pf[0]))
                 if rev not in localrevs:
                     serverfiles.update(diff)
                 else:
                     files.update(diff)
 
                 visited.add(mfrev)
                 progress.increment()
 
             files.difference_update(skip)
             serverfiles.difference_update(skip)
             progress.complete()
 
             # Fetch files known to be on the server
             if serverfiles:
                 results = [(path, hex(fnode)) for (path, fnode) in serverfiles]
                 repo.fileservice.prefetch(results, force=True)
 
             # Fetch files that may or may not be on the server
             if files:
                 results = [(path, hex(fnode)) for (path, fnode) in files]
                 repo.fileservice.prefetch(results)
 
         def close(self):
             super(shallowrepository, self).close()
             self.connectionpool.close()
 
     repo.__class__ = shallowrepository
 
     repo.shallowmatch = match.always()
 
     makeunionstores(repo)
 
     repo.includepattern = repo.ui.configlist("remotefilelog", "includepattern",
                                              None)
     repo.excludepattern = repo.ui.configlist("remotefilelog", "excludepattern",
                                              None)
     if not util.safehasattr(repo, 'connectionpool'):
         repo.connectionpool = connectionpool.connectionpool(repo)
 
     if repo.includepattern or repo.excludepattern:
         repo.shallowmatch = match.match(repo.root, '', None,
             repo.includepattern, repo.excludepattern)