Mercurial > hg
view mercurial/branchmap.py @ 21811:789b69d597cc
dirstate.status: assign members one by one instead of unpacking the tuple
With this patch, hg status and hg diff regain their previous speed.
The following tests are run against a working copy with over 270,000 files.
Here, 'before' means without this or the previous patch applied.
Note that in this case `hg perfstatus` isn't representative since it doesn't
take dirstate parsing time into account.
$ time hg status # best of 5
before: 2.03s user 1.25s system 99% cpu 3.290 total
after: 2.01s user 1.25s system 99% cpu 3.261 total
$ time hg diff # best of 5
before: 1.32s user 0.78s system 99% cpu 2.105 total
after: 1.27s user 0.79s system 99% cpu 2.066 total
author | Siddharth Agarwal <sid0@fb.com> |
---|---|
date | Tue, 27 May 2014 21:02:16 -0700 |
parents | 15baed3f24ee |
children | d5cef58d8ec8 |
line wrap: on
line source
# branchmap.py - logic to computes, maintain and stores branchmap for local repo # # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> # # This software may be used and distributed according to the terms of the # GNU General Public License version 2 or any later version. from node import bin, hex, nullid, nullrev import encoding import util import time def _filename(repo): """name of a branchcache file for a given repo or repoview""" filename = "cache/branch2" if repo.filtername: filename = '%s-%s' % (filename, repo.filtername) return filename def read(repo): try: f = repo.opener(_filename(repo)) lines = f.read().split('\n') f.close() except (IOError, OSError): return None try: cachekey = lines.pop(0).split(" ", 2) last, lrev = cachekey[:2] last, lrev = bin(last), int(lrev) filteredhash = None if len(cachekey) > 2: filteredhash = bin(cachekey[2]) partial = branchcache(tipnode=last, tiprev=lrev, filteredhash=filteredhash) if not partial.validfor(repo): # invalidate the cache raise ValueError('tip differs') for l in lines: if not l: continue node, state, label = l.split(" ", 2) if state not in 'oc': raise ValueError('invalid branch state') label = encoding.tolocal(label.strip()) if not node in repo: raise ValueError('node %s does not exist' % node) node = bin(node) partial.setdefault(label, []).append(node) if state == 'c': partial._closednodes.add(node) except KeyboardInterrupt: raise except Exception, inst: if repo.ui.debugflag: msg = 'invalid branchheads cache' if repo.filtername is not None: msg += ' (%s)' % repo.filtername msg += ': %s\n' repo.ui.debug(msg % inst) partial = None return partial ### Nearest subset relation # Nearest subset of filter X is a filter Y so that: # * Y is included in X, # * X - Y is as small as possible. # This create and ordering used for branchmap purpose. # the ordering may be partial subsettable = {None: 'visible', 'visible': 'served', 'served': 'immutable', 'immutable': 'base'} def updatecache(repo): cl = repo.changelog filtername = repo.filtername partial = repo._branchcaches.get(filtername) revs = [] if partial is None or not partial.validfor(repo): partial = read(repo) if partial is None: subsetname = subsettable.get(filtername) if subsetname is None: partial = branchcache() else: subset = repo.filtered(subsetname) partial = subset.branchmap().copy() extrarevs = subset.changelog.filteredrevs - cl.filteredrevs revs.extend(r for r in extrarevs if r <= partial.tiprev) revs.extend(cl.revs(start=partial.tiprev + 1)) if revs: partial.update(repo, revs) partial.write(repo) assert partial.validfor(repo), filtername repo._branchcaches[repo.filtername] = partial class branchcache(dict): """A dict like object that hold branches heads cache. This cache is used to avoid costly computations to determine all the branch heads of a repo. The cache is serialized on disk in the following format: <tip hex node> <tip rev number> [optional filtered repo hex hash] <branch head hex node> <open/closed state> <branch name> <branch head hex node> <open/closed state> <branch name> ... The first line is used to check if the cache is still valid. If the branch cache is for a filtered repo view, an optional third hash is included that hashes the hashes of all filtered revisions. The open/closed state is represented by a single letter 'o' or 'c'. This field can be used to avoid changelog reads when determining if a branch head closes a branch or not. """ def __init__(self, entries=(), tipnode=nullid, tiprev=nullrev, filteredhash=None, closednodes=None): super(branchcache, self).__init__(entries) self.tipnode = tipnode self.tiprev = tiprev self.filteredhash = filteredhash # closednodes is a set of nodes that close their branch. If the branch # cache has been updated, it may contain nodes that are no longer # heads. if closednodes is None: self._closednodes = set() else: self._closednodes = closednodes def _hashfiltered(self, repo): """build hash of revision filtered in the current cache Tracking tipnode and tiprev is not enough to ensure validity of the cache as they do not help to distinct cache that ignored various revision bellow tiprev. To detect such difference, we build a cache of all ignored revisions. """ cl = repo.changelog if not cl.filteredrevs: return None key = None revs = sorted(r for r in cl.filteredrevs if r <= self.tiprev) if revs: s = util.sha1() for rev in revs: s.update('%s;' % rev) key = s.digest() return key def validfor(self, repo): """Is the cache content valid regarding a repo - False when cached tipnode is unknown or if we detect a strip. - True when cache is up to date or a subset of current repo.""" try: return ((self.tipnode == repo.changelog.node(self.tiprev)) and (self.filteredhash == self._hashfiltered(repo))) except IndexError: return False def _branchtip(self, heads): '''Return tuple with last open head in heads and false, otherwise return last closed head and true.''' tip = heads[-1] closed = True for h in reversed(heads): if h not in self._closednodes: tip = h closed = False break return tip, closed def branchtip(self, branch): '''Return the tipmost open head on branch head, otherwise return the tipmost closed head on branch. Raise KeyError for unknown branch.''' return self._branchtip(self[branch])[0] def branchheads(self, branch, closed=False): heads = self[branch] if not closed: heads = [h for h in heads if h not in self._closednodes] return heads def iterbranches(self): for bn, heads in self.iteritems(): yield (bn, heads) + self._branchtip(heads) def copy(self): """return an deep copy of the branchcache object""" return branchcache(self, self.tipnode, self.tiprev, self.filteredhash, self._closednodes) def write(self, repo): try: f = repo.opener(_filename(repo), "w", atomictemp=True) cachekey = [hex(self.tipnode), str(self.tiprev)] if self.filteredhash is not None: cachekey.append(hex(self.filteredhash)) f.write(" ".join(cachekey) + '\n') nodecount = 0 for label, nodes in sorted(self.iteritems()): for node in nodes: nodecount += 1 if node in self._closednodes: state = 'c' else: state = 'o' f.write("%s %s %s\n" % (hex(node), state, encoding.fromlocal(label))) f.close() repo.ui.log('branchcache', 'wrote %s branch cache with %d labels and %d nodes\n', repo.filtername, len(self), nodecount) except (IOError, OSError, util.Abort), inst: repo.ui.debug("couldn't write branch cache: %s\n" % inst) # Abort may be raise by read only opener pass def update(self, repo, revgen): """Given a branchhead cache, self, that may have extra nodes or be missing heads, and a generator of nodes that are strictly a superset of heads missing, this function updates self to be correct. """ starttime = time.time() cl = repo.changelog # collect new branch entries newbranches = {} getbranchinfo = cl.branchinfo for r in revgen: branch, closesbranch = getbranchinfo(r) newbranches.setdefault(branch, []).append(r) if closesbranch: self._closednodes.add(cl.node(r)) # if older branchheads are reachable from new ones, they aren't # really branchheads. Note checking parents is insufficient: # 1 (branch a) -> 2 (branch b) -> 3 (branch a) for branch, newheadrevs in newbranches.iteritems(): bheads = self.setdefault(branch, []) bheadset = set(cl.rev(node) for node in bheads) # This have been tested True on all internal usage of this function. # run it again in case of doubt # assert not (set(bheadrevs) & set(newheadrevs)) newheadrevs.sort() bheadset.update(newheadrevs) # This loop prunes out two kinds of heads - heads that are # superseded by a head in newheadrevs, and newheadrevs that are not # heads because an existing head is their descendant. while newheadrevs: latest = newheadrevs.pop() if latest not in bheadset: continue ancestors = set(cl.ancestors([latest], min(bheadset))) bheadset -= ancestors bheadrevs = sorted(bheadset) self[branch] = [cl.node(rev) for rev in bheadrevs] tiprev = bheadrevs[-1] if tiprev > self.tiprev: self.tipnode = cl.node(tiprev) self.tiprev = tiprev if not self.validfor(repo): # cache key are not valid anymore self.tipnode = nullid self.tiprev = nullrev for heads in self.values(): tiprev = max(cl.rev(node) for node in heads) if tiprev > self.tiprev: self.tipnode = cl.node(tiprev) self.tiprev = tiprev self.filteredhash = self._hashfiltered(repo) duration = time.time() - starttime repo.ui.log('branchcache', 'updated %s branch cache in %.4f seconds\n', repo.filtername, duration)