reduce memory used by util.opener when making a temp copy of a file.

mercurial/util.py

@@ -1,878 +1,880 b''
 """
 util.py - Mercurial utility functions and platform specfic implementations
 
  Copyright 2005 K. Thananchayan <[e-mail]>
 
 This software may be used and distributed according to the terms
 of the GNU General Public License, incorporated herein by reference.
 
 This contains helper routines that are independent of the SCM core and hide
 platform-specific details from the core.
 """
 
 import os, errno
 from i18n import gettext as _
 from demandload import *
 demandload(globals(), "cStringIO errno popen2 re shutil sys tempfile")
 demandload(globals(), "threading time")
 
 class SignalInterrupt(Exception):
     """Exception raised on SIGTERM and SIGHUP."""
 
 def pipefilter(s, cmd):
     '''filter string S through command CMD, returning its output'''
     (pout, pin) = popen2.popen2(cmd, -1, 'b')
     def writer():
         try:
             pin.write(s)
             pin.close()
         except IOError, inst:
             if inst.errno != errno.EPIPE:
                 raise
 
     # we should use select instead on UNIX, but this will work on most
     # systems, including Windows
     w = threading.Thread(target=writer)
     w.start()
     f = pout.read()
     pout.close()
     w.join()
     return f
 
 def tempfilter(s, cmd):
     '''filter string S through a pair of temporary files with CMD.
     CMD is used as a template to create the real command to be run,
     with the strings INFILE and OUTFILE replaced by the real names of
     the temporary files generated.'''
     inname, outname = None, None
     try:
         infd, inname = tempfile.mkstemp(prefix='hg-filter-in-')
         fp = os.fdopen(infd, 'wb')
         fp.write(s)
         fp.close()
         outfd, outname = tempfile.mkstemp(prefix='hg-filter-out-')
         os.close(outfd)
         cmd = cmd.replace('INFILE', inname)
         cmd = cmd.replace('OUTFILE', outname)
         code = os.system(cmd)
         if code: raise Abort(_("command '%s' failed: %s") %
                              (cmd, explain_exit(code)))
         return open(outname, 'rb').read()
     finally:
         try:
             if inname: os.unlink(inname)
         except: pass
         try:
             if outname: os.unlink(outname)
         except: pass
 
 filtertable = {
     'tempfile:': tempfilter,
     'pipe:': pipefilter,
     }
 
 def filter(s, cmd):
     "filter a string through a command that transforms its input to its output"
     for name, fn in filtertable.iteritems():
         if cmd.startswith(name):
             return fn(s, cmd[len(name):].lstrip())
     return pipefilter(s, cmd)
 
 def find_in_path(name, path, default=None):
     '''find name in search path. path can be string (will be split
     with os.pathsep), or iterable thing that returns strings.  if name
     found, return path to name. else return default.'''
     if isinstance(path, str):
         path = path.split(os.pathsep)
     for p in path:
         p_name = os.path.join(p, name)
         if os.path.exists(p_name):
             return p_name
     return default
 
 def patch(strip, patchname, ui):
     """apply the patch <patchname> to the working directory.
     a list of patched files is returned"""
     patcher = find_in_path('gpatch', os.environ.get('PATH', ''), 'patch')
     fp = os.popen('"%s" -p%d < "%s"' % (patcher, strip, patchname))
     files = {}
     for line in fp:
         line = line.rstrip()
         ui.status("%s\n" % line)
         if line.startswith('patching file '):
             pf = parse_patch_output(line)
             files.setdefault(pf, 1)
     code = fp.close()
     if code:
         raise Abort(_("patch command failed: %s") % explain_exit(code)[0])
     return files.keys()
 
 def binary(s):
     """return true if a string is binary data using diff's heuristic"""
     if s and '\0' in s[:4096]:
         return True
     return False
 
 def unique(g):
     """return the uniq elements of iterable g"""
     seen = {}
     for f in g:
         if f not in seen:
             seen[f] = 1
             yield f
 
 class Abort(Exception):
     """Raised if a command needs to print an error and exit."""
 
 def always(fn): return True
 def never(fn): return False
 
 def patkind(name, dflt_pat='glob'):
     """Split a string into an optional pattern kind prefix and the
     actual pattern."""
     for prefix in 're', 'glob', 'path', 'relglob', 'relpath', 'relre':
         if name.startswith(prefix + ':'): return name.split(':', 1)
     return dflt_pat, name
 
 def globre(pat, head='^', tail='$'):
     "convert a glob pattern into a regexp"
     i, n = 0, len(pat)
     res = ''
     group = False
     def peek(): return i < n and pat[i]
     while i < n:
         c = pat[i]
         i = i+1
         if c == '*':
             if peek() == '*':
                 i += 1
                 res += '.*'
             else:
                 res += '[^/]*'
         elif c == '?':
             res += '.'
         elif c == '[':
             j = i
             if j < n and pat[j] in '!]':
                 j += 1
             while j < n and pat[j] != ']':
                 j += 1
             if j >= n:
                 res += '\\['
             else:
                 stuff = pat[i:j].replace('\\','\\\\')
                 i = j + 1
                 if stuff[0] == '!':
                     stuff = '^' + stuff[1:]
                 elif stuff[0] == '^':
                     stuff = '\\' + stuff
                 res = '%s[%s]' % (res, stuff)
         elif c == '{':
             group = True
             res += '(?:'
         elif c == '}' and group:
             res += ')'
             group = False
         elif c == ',' and group:
             res += '|'
         elif c == '\\':
             p = peek()
             if p:
                 i += 1
                 res += re.escape(p)
             else:
                 res += re.escape(c)
         else:
             res += re.escape(c)
     return head + res + tail
 
 _globchars = {'[': 1, '{': 1, '*': 1, '?': 1}
 
 def pathto(n1, n2):
     '''return the relative path from one place to another.
     this returns a path in the form used by the local filesystem, not hg.'''
     if not n1: return localpath(n2)
     a, b = n1.split('/'), n2.split('/')
     a.reverse()
     b.reverse()
     while a and b and a[-1] == b[-1]:
         a.pop()
         b.pop()
     b.reverse()
     return os.sep.join((['..'] * len(a)) + b)
 
 def canonpath(root, cwd, myname):
     """return the canonical path of myname, given cwd and root"""
     if root == os.sep:
         rootsep = os.sep
     else:
         rootsep = root + os.sep
     name = myname
     if not os.path.isabs(name):
         name = os.path.join(root, cwd, name)
     name = os.path.normpath(name)
     if name.startswith(rootsep):
         name = name[len(rootsep):]
         audit_path(name)
         return pconvert(name)
     elif name == root:
         return ''
     else:
         # Determine whether `name' is in the hierarchy at or beneath `root',
         # by iterating name=dirname(name) until that causes no change (can't
         # check name == '/', because that doesn't work on windows).  For each
         # `name', compare dev/inode numbers.  If they match, the list `rel'
         # holds the reversed list of components making up the relative file
         # name we want.
         root_st = os.stat(root)
         rel = []
         while True:
             try:
                 name_st = os.stat(name)
             except OSError:
                 break
             if samestat(name_st, root_st):
                 rel.reverse()
                 name = os.path.join(*rel)
                 audit_path(name)
                 return pconvert(name)
             dirname, basename = os.path.split(name)
             rel.append(basename)
             if dirname == name:
                 break
             name = dirname
 
         raise Abort('%s not under root' % myname)
 
 def matcher(canonroot, cwd='', names=['.'], inc=[], exc=[], head='', src=None):
     return _matcher(canonroot, cwd, names, inc, exc, head, 'glob', src)
 
 def cmdmatcher(canonroot, cwd='', names=['.'], inc=[], exc=[], head='', src=None):
     if os.name == 'nt':
         dflt_pat = 'glob'
     else:
         dflt_pat = 'relpath'
     return _matcher(canonroot, cwd, names, inc, exc, head, dflt_pat, src)
 
 def _matcher(canonroot, cwd, names, inc, exc, head, dflt_pat, src):
     """build a function to match a set of file patterns
 
     arguments:
     canonroot - the canonical root of the tree you're matching against
     cwd - the current working directory, if relevant
     names - patterns to find
     inc - patterns to include
     exc - patterns to exclude
     head - a regex to prepend to patterns to control whether a match is rooted
 
     a pattern is one of:
     'glob:<rooted glob>'
     're:<rooted regexp>'
     'path:<rooted path>'
     'relglob:<relative glob>'
     'relpath:<relative path>'
     'relre:<relative regexp>'
     '<rooted path or regexp>'
 
     returns:
     a 3-tuple containing
     - list of explicit non-pattern names passed in
     - a bool match(filename) function
     - a bool indicating if any patterns were passed in
 
     todo:
     make head regex a rooted bool
     """
 
     def contains_glob(name):
         for c in name:
             if c in _globchars: return True
         return False
 
     def regex(kind, name, tail):
         '''convert a pattern into a regular expression'''
         if kind == 're':
             return name
         elif kind == 'path':
             return '^' + re.escape(name) + '(?:/|$)'
         elif kind == 'relglob':
             return head + globre(name, '(?:|.*/)', tail)
         elif kind == 'relpath':
             return head + re.escape(name) + tail
         elif kind == 'relre':
             if name.startswith('^'):
                 return name
             return '.*' + name
         return head + globre(name, '', tail)
 
     def matchfn(pats, tail):
         """build a matching function from a set of patterns"""
         if not pats:
             return
         matches = []
         for k, p in pats:
             try:
                 pat = '(?:%s)' % regex(k, p, tail)
                 matches.append(re.compile(pat).match)
             except re.error:
                 if src: raise Abort("%s: invalid pattern (%s): %s" % (src, k, p))
                 else: raise Abort("invalid pattern (%s): %s" % (k, p))
 
         def buildfn(text):
             for m in matches:
                 r = m(text)
                 if r:
                     return r
 
         return buildfn
 
     def globprefix(pat):
         '''return the non-glob prefix of a path, e.g. foo/* -> foo'''
         root = []
         for p in pat.split(os.sep):
             if contains_glob(p): break
             root.append(p)
         return '/'.join(root)
 
     pats = []
     files = []
     roots = []
     for kind, name in [patkind(p, dflt_pat) for p in names]:
         if kind in ('glob', 'relpath'):
             name = canonpath(canonroot, cwd, name)
             if name == '':
                 kind, name = 'glob', '**'
         if kind in ('glob', 'path', 're'):
             pats.append((kind, name))
         if kind == 'glob':
             root = globprefix(name)
             if root: roots.append(root)
         elif kind == 'relpath':
             files.append((kind, name))
             roots.append(name)
 
     patmatch = matchfn(pats, '$') or always
     filematch = matchfn(files, '(?:/|$)') or always
     incmatch = always
     if inc:
         incmatch = matchfn(map(patkind, inc), '(?:/|$)')
     excmatch = lambda fn: False
     if exc:
         excmatch = matchfn(map(patkind, exc), '(?:/|$)')
 
     return (roots,
             lambda fn: (incmatch(fn) and not excmatch(fn) and
                         (fn.endswith('/') or
                          (not pats and not files) or
                          (pats and patmatch(fn)) or
                          (files and filematch(fn)))),
             (inc or exc or (pats and pats != [('glob', '**')])) and True)
 
 def system(cmd, environ={}, cwd=None, onerr=None, errprefix=None):
     '''enhanced shell command execution.
     run with environment maybe modified, maybe in different dir.
 
     if command fails and onerr is None, return status.  if ui object,
     print error message and return status, else raise onerr object as
     exception.'''
     oldenv = {}
     for k in environ:
         oldenv[k] = os.environ.get(k)
     if cwd is not None:
         oldcwd = os.getcwd()
     try:
         for k, v in environ.iteritems():
             os.environ[k] = str(v)
         if cwd is not None and oldcwd != cwd:
             os.chdir(cwd)
         rc = os.system(cmd)
         if rc and onerr:
             errmsg = '%s %s' % (os.path.basename(cmd.split(None, 1)[0]),
                                 explain_exit(rc)[0])
             if errprefix:
                 errmsg = '%s: %s' % (errprefix, errmsg)
             try:
                 onerr.warn(errmsg + '\n')
             except AttributeError:
                 raise onerr(errmsg)
         return rc
     finally:
         for k, v in oldenv.iteritems():
             if v is None:
                 del os.environ[k]
             else:
                 os.environ[k] = v
         if cwd is not None and oldcwd != cwd:
             os.chdir(oldcwd)
 
 def rename(src, dst):
     """forcibly rename a file"""
     try:
         os.rename(src, dst)
     except OSError, err:
         # on windows, rename to existing file is not allowed, so we
         # must delete destination first. but if file is open, unlink
         # schedules it for delete but does not delete it. rename
         # happens immediately even for open files, so we create
         # temporary file, delete it, rename destination to that name,
         # then delete that. then rename is safe to do.
         fd, temp = tempfile.mkstemp(dir=os.path.dirname(dst) or '.')
         os.close(fd)
         os.unlink(temp)
         os.rename(dst, temp)
         os.unlink(temp)
         os.rename(src, dst)
 
 def unlink(f):
     """unlink and remove the directory if it is empty"""
     os.unlink(f)
     # try removing directories that might now be empty
     try:
         os.removedirs(os.path.dirname(f))
     except OSError:
         pass
 
 def copyfiles(src, dst, hardlink=None):
     """Copy a directory tree using hardlinks if possible"""
 
     if hardlink is None:
         hardlink = (os.stat(src).st_dev ==
                     os.stat(os.path.dirname(dst)).st_dev)
 
     if os.path.isdir(src):
         os.mkdir(dst)
         for name in os.listdir(src):
             srcname = os.path.join(src, name)
             dstname = os.path.join(dst, name)
             copyfiles(srcname, dstname, hardlink)
     else:
         if hardlink:
             try:
                 os_link(src, dst)
             except (IOError, OSError):
                 hardlink = False
                 shutil.copy(src, dst)
         else:
             shutil.copy(src, dst)
 
 def audit_path(path):
     """Abort if path contains dangerous components"""
     parts = os.path.normcase(path).split(os.sep)
     if (os.path.splitdrive(path)[0] or parts[0] in ('.hg', '')
         or os.pardir in parts):
         raise Abort(_("path contains illegal component: %s\n") % path)
 
 def _makelock_file(info, pathname):
     ld = os.open(pathname, os.O_CREAT | os.O_WRONLY | os.O_EXCL)
     os.write(ld, info)
     os.close(ld)
 
 def _readlock_file(pathname):
     return posixfile(pathname).read()
 
 def nlinks(pathname):
     """Return number of hardlinks for the given file."""
     return os.stat(pathname).st_nlink
 
 if hasattr(os, 'link'):
     os_link = os.link
 else:
     def os_link(src, dst):
         raise OSError(0, _("Hardlinks not supported"))
 
 def fstat(fp):
     '''stat file object that may not have fileno method.'''
     try:
         return os.fstat(fp.fileno())
     except AttributeError:
         return os.stat(fp.name)
 
 posixfile = file
 
 # Platform specific variants
 if os.name == 'nt':
     demandload(globals(), "msvcrt")
     nulldev = 'NUL:'
 
     class winstdout:
         '''stdout on windows misbehaves if sent through a pipe'''
 
         def __init__(self, fp):
             self.fp = fp
 
         def __getattr__(self, key):
             return getattr(self.fp, key)
 
         def close(self):
             try:
                 self.fp.close()
             except: pass
 
         def write(self, s):
             try:
                 return self.fp.write(s)
             except IOError, inst:
                 if inst.errno != 0: raise
                 self.close()
                 raise IOError(errno.EPIPE, 'Broken pipe')
 
     sys.stdout = winstdout(sys.stdout)
 
     def system_rcpath():
         try:
             return system_rcpath_win32()
         except:
             return [r'c:\mercurial\mercurial.ini']
 
     def os_rcpath():
         '''return default os-specific hgrc search path'''
         return system_rcpath() + [os.path.join(os.path.expanduser('~'),
                                                'mercurial.ini')]
 
     def parse_patch_output(output_line):
         """parses the output produced by patch and returns the file name"""
         pf = output_line[14:]
         if pf[0] == '`':
             pf = pf[1:-1] # Remove the quotes
         return pf
 
     def testpid(pid):
         '''return False if pid dead, True if running or not known'''
         return True
 
     def is_exec(f, last):
         return last
 
     def set_exec(f, mode):
         pass
 
     def set_binary(fd):
         msvcrt.setmode(fd.fileno(), os.O_BINARY)
 
     def pconvert(path):
         return path.replace("\\", "/")
 
     def localpath(path):
         return path.replace('/', '\\')
 
     def normpath(path):
         return pconvert(os.path.normpath(path))
 
     makelock = _makelock_file
     readlock = _readlock_file
 
     def samestat(s1, s2):
         return False
 
     def explain_exit(code):
         return _("exited with status %d") % code, code
 
     try:
         # override functions with win32 versions if possible
         from util_win32 import *
     except ImportError:
         pass
 
 else:
     nulldev = '/dev/null'
 
     def rcfiles(path):
         rcs = [os.path.join(path, 'hgrc')]
         rcdir = os.path.join(path, 'hgrc.d')
         try:
             rcs.extend([os.path.join(rcdir, f) for f in os.listdir(rcdir)
                         if f.endswith(".rc")])
         except OSError, inst: pass
         return rcs
 
     def os_rcpath():
         '''return default os-specific hgrc search path'''
         path = []
         if len(sys.argv) > 0:
             path.extend(rcfiles(os.path.dirname(sys.argv[0]) +
                                   '/../etc/mercurial'))
         path.extend(rcfiles('/etc/mercurial'))
         path.append(os.path.expanduser('~/.hgrc'))
         path = [os.path.normpath(f) for f in path]
         return path
 
     def parse_patch_output(output_line):
         """parses the output produced by patch and returns the file name"""
         pf = output_line[14:]
         if pf.startswith("'") and pf.endswith("'") and pf.find(" ") >= 0:
             pf = pf[1:-1] # Remove the quotes
         return pf
 
     def is_exec(f, last):
         """check whether a file is executable"""
         return (os.stat(f).st_mode & 0100 != 0)
 
     def set_exec(f, mode):
         s = os.stat(f).st_mode
         if (s & 0100 != 0) == mode:
             return
         if mode:
             # Turn on +x for every +r bit when making a file executable
             # and obey umask.
             umask = os.umask(0)
             os.umask(umask)
             os.chmod(f, s | (s & 0444) >> 2 & ~umask)
         else:
             os.chmod(f, s & 0666)
 
     def set_binary(fd):
         pass
 
     def pconvert(path):
         return path
 
     def localpath(path):
         return path
 
     normpath = os.path.normpath
     samestat = os.path.samestat
 
     def makelock(info, pathname):
         try:
             os.symlink(info, pathname)
         except OSError, why:
             if why.errno == errno.EEXIST:
                 raise
             else:
                 _makelock_file(info, pathname)
 
     def readlock(pathname):
         try:
             return os.readlink(pathname)
         except OSError, why:
             if why.errno == errno.EINVAL:
                 return _readlock_file(pathname)
             else:
                 raise
 
     def testpid(pid):
         '''return False if pid dead, True if running or not sure'''
         try:
             os.kill(pid, 0)
             return True
         except OSError, inst:
             return inst.errno != errno.ESRCH
 
     def explain_exit(code):
         """return a 2-tuple (desc, code) describing a process's status"""
         if os.WIFEXITED(code):
             val = os.WEXITSTATUS(code)
             return _("exited with status %d") % val, val
         elif os.WIFSIGNALED(code):
             val = os.WTERMSIG(code)
             return _("killed by signal %d") % val, val
         elif os.WIFSTOPPED(code):
             val = os.WSTOPSIG(code)
             return _("stopped by signal %d") % val, val
         raise ValueError(_("invalid exit code"))
 
 def opener(base, audit=True):
     """
     return a function that opens files relative to base
 
     this function is used to hide the details of COW semantics and
     remote file access from higher level code.
     """
     p = base
     audit_p = audit
 
     def mktempcopy(name):
         d, fn = os.path.split(name)
         fd, temp = tempfile.mkstemp(prefix='.%s-' % fn, dir=d)
         os.close(fd)
-        fp = posixfile(temp, "wb")
+        ofp = posixfile(temp, "wb")
         try:
             try:
-                s = posixfile(name, "rb").read()
+                ifp = posixfile(name, "rb")
             except IOError, inst:
                 if not getattr(inst, 'filename', None):
                     inst.filename = name
                 raise
-            fp.write(s)
+            for chunk in filechunkiter(ifp):
+                ofp.write(chunk)
+            ifp.close()
+            ofp.close()
         except:
             try: os.unlink(temp)
             except: pass
             raise
-        fp.close()
         st = os.lstat(name)
         os.chmod(temp, st.st_mode)
         return temp
 
     class atomictempfile(posixfile):
         """the file will only be copied when rename is called"""
         def __init__(self, name, mode):
             self.__name = name
             self.temp = mktempcopy(name)
             posixfile.__init__(self, self.temp, mode)
         def rename(self):
             if not self.closed:
                 posixfile.close(self)
                 rename(self.temp, self.__name)
         def __del__(self):
             if not self.closed:
                 try:
                     os.unlink(self.temp)
                 except: pass
                 posixfile.close(self)
 
     class atomicfile(atomictempfile):
         """the file will only be copied on close"""
         def __init__(self, name, mode):
             atomictempfile.__init__(self, name, mode)
         def close(self):
             self.rename()
         def __del__(self):
             self.rename()
 
     def o(path, mode="r", text=False, atomic=False, atomictemp=False):
         if audit_p:
             audit_path(path)
         f = os.path.join(p, path)
 
         if not text:
             mode += "b" # for that other OS
 
         if mode[0] != "r":
             try:
                 nlink = nlinks(f)
             except OSError:
                 d = os.path.dirname(f)
                 if not os.path.isdir(d):
                     os.makedirs(d)
             else:
                 if atomic:
                     return atomicfile(f, mode)
                 elif atomictemp:
                     return atomictempfile(f, mode)
                 if nlink > 1:
                     rename(mktempcopy(f), f)
         return posixfile(f, mode)
 
     return o
 
 class chunkbuffer(object):
     """Allow arbitrary sized chunks of data to be efficiently read from an
     iterator over chunks of arbitrary size."""
 
     def __init__(self, in_iter, targetsize = 2**16):
         """in_iter is the iterator that's iterating over the input chunks.
         targetsize is how big a buffer to try to maintain."""
         self.in_iter = iter(in_iter)
         self.buf = ''
         self.targetsize = int(targetsize)
         if self.targetsize <= 0:
             raise ValueError(_("targetsize must be greater than 0, was %d") %
                              targetsize)
         self.iterempty = False
 
     def fillbuf(self):
         """Ignore target size; read every chunk from iterator until empty."""
         if not self.iterempty:
             collector = cStringIO.StringIO()
             collector.write(self.buf)
             for ch in self.in_iter:
                 collector.write(ch)
             self.buf = collector.getvalue()
             self.iterempty = True
 
     def read(self, l):
         """Read L bytes of data from the iterator of chunks of data.
         Returns less than L bytes if the iterator runs dry."""
         if l > len(self.buf) and not self.iterempty:
             # Clamp to a multiple of self.targetsize
             targetsize = self.targetsize * ((l // self.targetsize) + 1)
             collector = cStringIO.StringIO()
             collector.write(self.buf)
             collected = len(self.buf)
             for chunk in self.in_iter:
                 collector.write(chunk)
                 collected += len(chunk)
                 if collected >= targetsize:
                     break
             if collected < targetsize:
                 self.iterempty = True
             self.buf = collector.getvalue()
         s, self.buf = self.buf[:l], buffer(self.buf, l)
         return s
 
 def filechunkiter(f, size = 65536):
     """Create a generator that produces all the data in the file size
     (default 65536) bytes at a time.  Chunks may be less than size
     bytes if the chunk is the last chunk in the file, or the file is a
     socket or some other type of file that sometimes reads less data
     than is requested."""
     s = f.read(size)
     while len(s) > 0:
         yield s
         s = f.read(size)
 
 def makedate():
     lt = time.localtime()
     if lt[8] == 1 and time.daylight:
         tz = time.altzone
     else:
         tz = time.timezone
     return time.mktime(lt), tz
 
 def datestr(date=None, format='%a %b %d %H:%M:%S %Y', timezone=True):
     """represent a (unixtime, offset) tuple as a localized time.
     unixtime is seconds since the epoch, and offset is the time zone's
     number of seconds away from UTC. if timezone is false, do not
     append time zone to string."""
     t, tz = date or makedate()
     s = time.strftime(format, time.gmtime(float(t) - tz))
     if timezone:
         s += " %+03d%02d" % (-tz / 3600, ((-tz % 3600) / 60))
     return s
 
 def shortuser(user):
     """Return a short representation of a user name or email address."""
     f = user.find('@')
     if f >= 0:
         user = user[:f]
     f = user.find('<')
     if f >= 0:
         user = user[f+1:]
     return user
 
 def walkrepos(path):
     '''yield every hg repository under path, recursively.'''
     def errhandler(err):
         if err.filename == path:
             raise err
 
     for root, dirs, files in os.walk(path, onerror=errhandler):
         for d in dirs:
             if d == '.hg':
                 yield root
                 dirs[:] = []
                 break
 
 _rcpath = None
 
 def rcpath():
     '''return hgrc search path. if env var HGRCPATH is set, use it.
     for each item in path, if directory, use files ending in .rc,
     else use item.
     make HGRCPATH empty to only look in .hg/hgrc of current repo.
     if no HGRCPATH, use default os-specific path.'''
     global _rcpath
     if _rcpath is None:
         if 'HGRCPATH' in os.environ:
             _rcpath = []
             for p in os.environ['HGRCPATH'].split(os.pathsep):
                 if not p: continue
                 if os.path.isdir(p):
                     for f in os.listdir(p):
                         if f.endswith('.rc'):
                             _rcpath.append(os.path.join(p, f))
                 else:
                     _rcpath.append(p)
         else:
             _rcpath = os_rcpath()
     return _rcpath