dirstate: make dir collision logic faster

- shortcircuit decpath if we haven't built the _dirs map
- increment only for leafnodes of directory tree
  (this should make construction more like O(nlog n) than O(n^2))

mercurial/dirstate.py

@@ -1,547 +1,558
 """
 dirstate.py - working directory tracking for mercurial
 
 Copyright 2005-2007 Matt Mackall <[e-mail]>
 
 This software may be used and distributed according to the terms
 of the GNU General Public License, incorporated herein by reference.
 """
 
 from node import *
 from i18n import _
 import struct, os, time, bisect, stat, strutil, util, re, errno, ignore
 import cStringIO
 
 _unknown = ('?', 0, 0, 0)
 _format = ">cllll"
 
 class dirstate(object):
 
     def __init__(self, opener, ui, root):
         self._opener = opener
         self._root = root
         self._dirty = False
         self._dirtypl = False
         self._ui = ui
 
     def __getattr__(self, name):
         if name == '_map':
             self._read()
             return self._map
         elif name == '_copymap':
             self._read()
             return self._copymap
         elif name == '_branch':
             try:
                 self._branch = (self._opener("branch").read().strip()
                                 or "default")
             except IOError:
                 self._branch = "default"
             return self._branch
         elif name == '_pl':
             self._pl = [nullid, nullid]
             try:
                 st = self._opener("dirstate").read(40)
                 if len(st) == 40:
                     self._pl = st[:20], st[20:40]
             except IOError, err:
                 if err.errno != errno.ENOENT: raise
             return self._pl
         elif name == '_dirs':
             self._dirs = {}
             for f in self._map:
                 self._incpath(f)
             return self._dirs
         elif name == '_ignore':
             files = [self._join('.hgignore')]
             for name, path in self._ui.configitems("ui"):
                 if name == 'ignore' or name.startswith('ignore.'):
                     files.append(os.path.expanduser(path))
             self._ignore = ignore.ignore(self._root, files, self._ui.warn)
             return self._ignore
         elif name == '_slash':
             self._slash = self._ui.configbool('ui', 'slash') and os.sep != '/'
             return self._slash
         else:
             raise AttributeError, name
 
     def _join(self, f):
         return os.path.join(self._root, f)
 
     def getcwd(self):
         cwd = os.getcwd()
         if cwd == self._root: return ''
         # self._root ends with a path separator if self._root is '/' or 'C:\'
         rootsep = self._root
         if not rootsep.endswith(os.sep):
             rootsep += os.sep
         if cwd.startswith(rootsep):
             return cwd[len(rootsep):]
         else:
             # we're outside the repo. return an absolute path.
             return cwd
 
     def pathto(self, f, cwd=None):
         if cwd is None:
             cwd = self.getcwd()
         path = util.pathto(self._root, cwd, f)
         if self._slash:
             return path.replace(os.sep, '/')
         return path
 
     def __getitem__(self, key):
         ''' current states:
         n  normal
         m  needs merging
         r  marked for removal
         a  marked for addition
         ?  not tracked'''
         return self._map.get(key, ("?",))[0]
 
     def __contains__(self, key):
         return key in self._map
 
     def __iter__(self):
         a = self._map.keys()
         a.sort()
         for x in a:
             yield x
 
     def parents(self):
         return self._pl
 
     def branch(self):
         return self._branch
 
     def setparents(self, p1, p2=nullid):
         self._dirty = self._dirtypl = True
         self._pl = p1, p2
 
     def setbranch(self, branch):
         self._branch = branch
         self._opener("branch", "w").write(branch + '\n')
 
     def _read(self):
         self._map = {}
         self._copymap = {}
         if not self._dirtypl:
             self._pl = [nullid, nullid]
         try:
             st = self._opener("dirstate").read()
         except IOError, err:
             if err.errno != errno.ENOENT: raise
             return
         if not st:
             return
 
         if not self._dirtypl:
             self._pl = [st[:20], st[20: 40]]
 
         # deref fields so they will be local in loop
         dmap = self._map
         copymap = self._copymap
         unpack = struct.unpack
 
         pos = 40
         e_size = struct.calcsize(_format)
 
         while pos < len(st):
             newpos = pos + e_size
             e = unpack(_format, st[pos:newpos])
             l = e[4]
             pos = newpos
             newpos = pos + l
             f = st[pos:newpos]
             if '\0' in f:
                 f, c = f.split('\0')
                 copymap[f] = c
             dmap[f] = e[:4]
             pos = newpos
 
     def invalidate(self):
         for a in "_map _copymap _branch _pl _dirs _ignore".split():
             if a in self.__dict__:
                 delattr(self, a)
         self._dirty = False
 
     def copy(self, source, dest):
         self._dirty = True
         self._copymap[dest] = source
 
     def copied(self, file):
         return self._copymap.get(file, None)
 
     def copies(self):
         return self._copymap
 
     def _incpath(self, path):
-        for c in strutil.findall(path, '/'):
-            pc = path[:c]
-            self._dirs.setdefault(pc, 0)
-            self._dirs[pc] += 1
+        c = path.rfind('/')
+        if c >= 0:
+            dirs = self._dirs
+            base = path[:c]
+            if base not in dirs:
+                self._incpath(base)
+                dirs[base] = 1
+            else:
+                dirs[base] += 1
 
     def _decpath(self, path):
-        for c in strutil.findall(path, '/'):
-            pc = path[:c]
-            self._dirs.setdefault(pc, 0)
-            self._dirs[pc] -= 1
+        if "_dirs" in self.__dict__:
+            c = path.rfind('/')
+            if c >= 0:
+                base = path[:c]
+                dirs = self._dirs
+                if dirs[base] == 1:
+                    del dirs[base]
+                    self._decpath(base)
+                else:
+                    dirs[base] -= 1
 
     def _incpathcheck(self, f):
         if '\r' in f or '\n' in f:
             raise util.Abort(_("'\\n' and '\\r' disallowed in filenames"))
         # shadows
         if f in self._dirs:
             raise util.Abort(_('directory %r already in dirstate') % f)
         for c in strutil.rfindall(f, '/'):
             d = f[:c]
             if d in self._dirs:
                 break
             if d in self._map:
                 raise util.Abort(_('file %r in dirstate clashes with %r') %
                                  (d, f))
         self._incpath(f)
 
     def normal(self, f):
         'mark a file normal and clean'
         self._dirty = True
         s = os.lstat(self._join(f))
         self._map[f] = ('n', s.st_mode, s.st_size, s.st_mtime)
         if self._copymap.has_key(f):
             del self._copymap[f]
 
     def normallookup(self, f):
         'mark a file normal, but possibly dirty'
         self._dirty = True
         self._map[f] = ('n', 0, -1, -1)
         if f in self._copymap:
             del self._copymap[f]
 
     def normaldirty(self, f):
         'mark a file normal, but dirty'
         self._dirty = True
         self._map[f] = ('n', 0, -2, -1)
         if f in self._copymap:
             del self._copymap[f]
 
     def add(self, f):
         'mark a file added'
         self._dirty = True
         self._incpathcheck(f)
         self._map[f] = ('a', 0, -1, -1)
         if f in self._copymap:
             del self._copymap[f]
 
     def remove(self, f):
         'mark a file removed'
         self._dirty = True
         self._map[f] = ('r', 0, 0, 0)
         self._decpath(f)
         if f in self._copymap:
             del self._copymap[f]
 
     def merge(self, f):
         'mark a file merged'
         self._dirty = True
         s = os.lstat(self._join(f))
         self._map[f] = ('m', s.st_mode, s.st_size, s.st_mtime)
         if f in self._copymap:
             del self._copymap[f]
 
     def forget(self, f):
         'forget a file'
         self._dirty = True
         try:
             del self._map[f]
             self._decpath(f)
         except KeyError:
             self._ui.warn(_("not in dirstate: %s!\n") % f)
 
     def clear(self):
         self._map = {}
         self._copymap = {}
         self._pl = [nullid, nullid]
         self._dirty = True
 
     def rebuild(self, parent, files):
         self.clear()
         for f in files:
             if files.execf(f):
                 self._map[f] = ('n', 0777, -1, 0)
             else:
                 self._map[f] = ('n', 0666, -1, 0)
         self._pl = (parent, nullid)
         self._dirty = True
 
     def write(self):
         if not self._dirty:
             return
         cs = cStringIO.StringIO()
         cs.write("".join(self._pl))
         for f, e in self._map.iteritems():
             c = self.copied(f)
             if c:
                 f = f + "\0" + c
             e = struct.pack(_format, e[0], e[1], e[2], e[3], len(f))
             cs.write(e)
             cs.write(f)
         st = self._opener("dirstate", "w", atomictemp=True)
         st.write(cs.getvalue())
         st.rename()
         self._dirty = self._dirtypl = False
 
     def _filter(self, files):
         ret = {}
         unknown = []
 
         for x in files:
             if x == '.':
                 return self._map.copy()
             if x not in self._map:
                 unknown.append(x)
             else:
                 ret[x] = self._map[x]
 
         if not unknown:
             return ret
 
         b = self._map.keys()
         b.sort()
         blen = len(b)
 
         for x in unknown:
             bs = bisect.bisect(b, "%s%s" % (x, '/'))
             while bs < blen:
                 s = b[bs]
                 if len(s) > len(x) and s.startswith(x):
                     ret[s] = self._map[s]
                 else:
                     break
                 bs += 1
         return ret
 
     def _supported(self, f, mode, verbose=False):
         if stat.S_ISREG(mode) or stat.S_ISLNK(mode):
             return True
         if verbose:
             kind = 'unknown'
             if stat.S_ISCHR(mode): kind = _('character device')
             elif stat.S_ISBLK(mode): kind = _('block device')
             elif stat.S_ISFIFO(mode): kind = _('fifo')
             elif stat.S_ISSOCK(mode): kind = _('socket')
             elif stat.S_ISDIR(mode): kind = _('directory')
             self._ui.warn(_('%s: unsupported file type (type is %s)\n')
                           % (self.pathto(f), kind))
         return False
 
     def walk(self, files=None, match=util.always, badmatch=None):
         # filter out the stat
         for src, f, st in self.statwalk(files, match, badmatch=badmatch):
             yield src, f
 
     def statwalk(self, files=None, match=util.always, ignored=False,
                  badmatch=None, directories=False):
         '''
         walk recursively through the directory tree, finding all files
         matched by the match function
 
         results are yielded in a tuple (src, filename, st), where src
         is one of:
         'f' the file was found in the directory tree
         'd' the file is a directory of the tree
         'm' the file was only in the dirstate and not in the tree
         'b' file was not found and matched badmatch
 
         and st is the stat result if the file was found in the directory.
         '''
 
         # walk all files by default
         if not files:
             files = ['.']
             dc = self._map.copy()
         else:
             files = util.unique(files)
             dc = self._filter(files)
 
         def imatch(file_):
             if file_ not in dc and self._ignore(file_):
                 return False
             return match(file_)
 
         ignore = self._ignore
         if ignored:
             imatch = match
             ignore = util.never
 
         # self._root may end with a path separator when self._root == '/'
         common_prefix_len = len(self._root)
         if not self._root.endswith(os.sep):
             common_prefix_len += 1
 
         normpath = util.normpath
         listdir = os.listdir
         lstat = os.lstat
         bisect_left = bisect.bisect_left
         isdir = os.path.isdir
         pconvert = util.pconvert
         join = os.path.join
         s_isdir = stat.S_ISDIR
         supported = self._supported
         _join = self._join
         known = {'.hg': 1}
 
         # recursion free walker, faster than os.walk.
         def findfiles(s):
             work = [s]
             wadd = work.append
             found = []
             add = found.append
             if directories:
                 add((normpath(s[common_prefix_len:]), 'd', lstat(s)))
             while work:
                 top = work.pop()
                 names = listdir(top)
                 names.sort()
                 # nd is the top of the repository dir tree
                 nd = normpath(top[common_prefix_len:])
                 if nd == '.':
                     nd = ''
                 else:
                     # do not recurse into a repo contained in this
                     # one. use bisect to find .hg directory so speed
                     # is good on big directory.
                     hg = bisect_left(names, '.hg')
                     if hg < len(names) and names[hg] == '.hg':
                         if isdir(join(top, '.hg')):
                             continue
                 for f in names:
                     np = pconvert(join(nd, f))
                     if np in known:
                         continue
                     known[np] = 1
                     p = join(top, f)
                     # don't trip over symlinks
                     st = lstat(p)
                     if s_isdir(st.st_mode):
                         if not ignore(np):
                             wadd(p)
                             if directories:
                                 add((np, 'd', st))
                         if np in dc and match(np):
                             add((np, 'm', st))
                     elif imatch(np):
                         if supported(np, st.st_mode):
                             add((np, 'f', st))
                         elif np in dc:
                             add((np, 'm', st))
             found.sort()
             return found
 
         # step one, find all files that match our criteria
         files.sort()
         for ff in files:
             nf = normpath(ff)
             f = _join(ff)
             try:
                 st = lstat(f)
             except OSError, inst:
                 found = False
                 for fn in dc:
                     if nf == fn or (fn.startswith(nf) and fn[len(nf)] == '/'):
                         found = True
                         break
                 if not found:
                     if inst.errno != errno.ENOENT or not badmatch:
                         self._ui.warn('%s: %s\n' %
                                       (self.pathto(ff), inst.strerror))
                     elif badmatch and badmatch(ff) and imatch(nf):
                         yield 'b', ff, None
                 continue
             if s_isdir(st.st_mode):
                 for f, src, st in findfiles(f):
                     yield src, f, st
             else:
                 if nf in known:
                     continue
                 known[nf] = 1
                 if match(nf):
                     if supported(ff, st.st_mode, verbose=True):
                         yield 'f', nf, st
                     elif ff in dc:
                         yield 'm', nf, st
 
         # step two run through anything left in the dc hash and yield
         # if we haven't already seen it
         ks = dc.keys()
         ks.sort()
         for k in ks:
             if k in known:
                 continue
             known[k] = 1
             if imatch(k):
                 yield 'm', k, None
 
     def status(self, files, match, list_ignored, list_clean):
         lookup, modified, added, unknown, ignored = [], [], [], [], []
         removed, deleted, clean = [], [], []
 
         _join = self._join
         lstat = os.lstat
         cmap = self._copymap
         dmap = self._map
         ladd = lookup.append
         madd = modified.append
         aadd = added.append
         uadd = unknown.append
         iadd = ignored.append
         radd = removed.append
         dadd = deleted.append
         cadd = clean.append
 
         for src, fn, st in self.statwalk(files, match, ignored=list_ignored):
             if fn in dmap:
                 type_, mode, size, time = dmap[fn]
             else:
                 if list_ignored and self._ignore(fn):
                     iadd(fn)
                 else:
                     uadd(fn)
                 continue
             if src == 'm':
                 nonexistent = True
                 if not st:
                     try:
                         st = lstat(_join(fn))
                     except OSError, inst:
                         if inst.errno != errno.ENOENT:
                             raise
                         st = None
                     # We need to re-check that it is a valid file
                     if st and self._supported(fn, st.st_mode):
                         nonexistent = False
                 # XXX: what to do with file no longer present in the fs
                 # who are not removed in the dirstate ?
                 if nonexistent and type_ in "nm":
                     dadd(fn)
                     continue
             # check the common case first
             if type_ == 'n':
                 if not st:
                     st = lstat(_join(fn))
                 if (size >= 0 and (size != st.st_size
                                    or (mode ^ st.st_mode) & 0100)
                     or size == -2
                     or fn in self._copymap):
                     madd(fn)
                 elif time != int(st.st_mtime):
                     ladd(fn)
                 elif list_clean:
                     cadd(fn)
             elif type_ == 'm':
                 madd(fn)
             elif type_ == 'a':
                 aadd(fn)
             elif type_ == 'r':
                 radd(fn)
 
         return (lookup, modified, added, removed, deleted, unknown, ignored,
                 clean)