view mercurial/util.py @ 7913:1b1b3dd630a5

windows: hoist expand_glob() back into util.py

The windows version of expand_glob() requires patkind(). To
avoid a circular dependency, move function back into util.py.

"""
util.py - Mercurial utility functions and platform specfic implementations

 Copyright 2005 K. Thananchayan <[e-mail]>
 Copyright 2005-2007 Matt Mackall <[e-mail]>
 Copyright 2006 Vadim Gelfer <[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.
"""

from i18n import _
import cStringIO, errno, re, shutil, sys, tempfile, traceback, error
import os, stat, threading, time, calendar, ConfigParser, locale, glob, osutil
import imp, unicodedata

# Python compatibility

try:
    set = set
    frozenset = frozenset
except NameError:
    from sets import Set as set, ImmutableSet as frozenset

_md5 = None
def md5(s):
    global _md5
    if _md5 is None:
        try:
            import hashlib
            _md5 = hashlib.md5
        except ImportError:
            import md5
            _md5 = md5.md5
    return _md5(s)

_sha1 = None
def sha1(s):
    global _sha1
    if _sha1 is None:
        try:
            import hashlib
            _sha1 = hashlib.sha1
        except ImportError:
            import sha
            _sha1 = sha.sha
    return _sha1(s)

try:
    import subprocess
    subprocess.Popen  # trigger ImportError early
    closefds = os.name == 'posix'
    def popen2(cmd, mode='t', bufsize=-1):
        p = subprocess.Popen(cmd, shell=True, bufsize=bufsize,
                             close_fds=closefds,
                             stdin=subprocess.PIPE, stdout=subprocess.PIPE)
        return p.stdin, p.stdout
    def popen3(cmd, mode='t', bufsize=-1):
        p = subprocess.Popen(cmd, shell=True, bufsize=bufsize,
                             close_fds=closefds,
                             stdin=subprocess.PIPE, stdout=subprocess.PIPE,
                             stderr=subprocess.PIPE)
        return p.stdin, p.stdout, p.stderr
    def Popen3(cmd, capturestderr=False, bufsize=-1):
        stderr = capturestderr and subprocess.PIPE or None
        p = subprocess.Popen(cmd, shell=True, bufsize=bufsize,
                             close_fds=closefds,
                             stdin=subprocess.PIPE, stdout=subprocess.PIPE,
                             stderr=stderr)
        p.fromchild = p.stdout
        p.tochild = p.stdin
        p.childerr = p.stderr
        return p
except ImportError:
    subprocess = None
    from popen2 import Popen3
    popen2 = os.popen2
    popen3 = os.popen3


_encodingfixup = {'646': 'ascii', 'ANSI_X3.4-1968': 'ascii'}

try:
    _encoding = os.environ.get("HGENCODING")
    if sys.platform == 'darwin' and not _encoding:
        # On darwin, getpreferredencoding ignores the locale environment and
        # always returns mac-roman. We override this if the environment is
        # not C (has been customized by the user).
        locale.setlocale(locale.LC_CTYPE, '')
        _encoding = locale.getlocale()[1]
    if not _encoding:
        _encoding = locale.getpreferredencoding() or 'ascii'
        _encoding = _encodingfixup.get(_encoding, _encoding)
except locale.Error:
    _encoding = 'ascii'
_encodingmode = os.environ.get("HGENCODINGMODE", "strict")
_fallbackencoding = 'ISO-8859-1'

def tolocal(s):
    """
    Convert a string from internal UTF-8 to local encoding

    All internal strings should be UTF-8 but some repos before the
    implementation of locale support may contain latin1 or possibly
    other character sets. We attempt to decode everything strictly
    using UTF-8, then Latin-1, and failing that, we use UTF-8 and
    replace unknown characters.
    """
    for e in ('UTF-8', _fallbackencoding):
        try:
            u = s.decode(e) # attempt strict decoding
            return u.encode(_encoding, "replace")
        except LookupError, k:
            raise Abort(_("%s, please check your locale settings") % k)
        except UnicodeDecodeError:
            pass
    u = s.decode("utf-8", "replace") # last ditch
    return u.encode(_encoding, "replace")

def fromlocal(s):
    """
    Convert a string from the local character encoding to UTF-8

    We attempt to decode strings using the encoding mode set by
    HGENCODINGMODE, which defaults to 'strict'. In this mode, unknown
    characters will cause an error message. Other modes include
    'replace', which replaces unknown characters with a special
    Unicode character, and 'ignore', which drops the character.
    """
    try:
        return s.decode(_encoding, _encodingmode).encode("utf-8")
    except UnicodeDecodeError, inst:
        sub = s[max(0, inst.start-10):inst.start+10]
        raise Abort("decoding near '%s': %s!" % (sub, inst))
    except LookupError, k:
        raise Abort(_("%s, please check your locale settings") % k)

def colwidth(s):
    "Find the column width of a UTF-8 string for display"
    d = s.decode(_encoding, 'replace')
    if hasattr(unicodedata, 'east_asian_width'):
        w = unicodedata.east_asian_width
        return sum([w(c) in 'WF' and 2 or 1 for c in d])
    return len(d)

def version():
    """Return version information if available."""
    try:
        import __version__
        return __version__.version
    except ImportError:
        return 'unknown'

# used by parsedate
defaultdateformats = (
    '%Y-%m-%d %H:%M:%S',
    '%Y-%m-%d %I:%M:%S%p',
    '%Y-%m-%d %H:%M',
    '%Y-%m-%d %I:%M%p',
    '%Y-%m-%d',
    '%m-%d',
    '%m/%d',
    '%m/%d/%y',
    '%m/%d/%Y',
    '%a %b %d %H:%M:%S %Y',
    '%a %b %d %I:%M:%S%p %Y',
    '%a, %d %b %Y %H:%M:%S',        #  GNU coreutils "/bin/date --rfc-2822"
    '%b %d %H:%M:%S %Y',
    '%b %d %I:%M:%S%p %Y',
    '%b %d %H:%M:%S',
    '%b %d %I:%M:%S%p',
    '%b %d %H:%M',
    '%b %d %I:%M%p',
    '%b %d %Y',
    '%b %d',
    '%H:%M:%S',
    '%I:%M:%SP',
    '%H:%M',
    '%I:%M%p',
)

extendeddateformats = defaultdateformats + (
    "%Y",
    "%Y-%m",
    "%b",
    "%b %Y",
    )

# differences from SafeConfigParser:
# - case-sensitive keys
# - allows values that are not strings (this means that you may not
#   be able to save the configuration to a file)
class configparser(ConfigParser.SafeConfigParser):
    def optionxform(self, optionstr):
        return optionstr

    def set(self, section, option, value):
        return ConfigParser.ConfigParser.set(self, section, option, value)

    def _interpolate(self, section, option, rawval, vars):
        if not isinstance(rawval, basestring):
            return rawval
        return ConfigParser.SafeConfigParser._interpolate(self, section,
                                                          option, rawval, vars)

def cachefunc(func):
    '''cache the result of function calls'''
    # XXX doesn't handle keywords args
    cache = {}
    if func.func_code.co_argcount == 1:
        # we gain a small amount of time because
        # we don't need to pack/unpack the list
        def f(arg):
            if arg not in cache:
                cache[arg] = func(arg)
            return cache[arg]
    else:
        def f(*args):
            if args not in cache:
                cache[args] = func(*args)
            return cache[args]

    return f

def pipefilter(s, cmd):
    '''filter string S through command CMD, returning its output'''
    (pin, pout) = popen2(cmd, '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 sys.platform == 'OpenVMS' and code & 1:
            code = 0
        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 binary(s):
    """return true if a string is binary data"""
    if s and '\0' in s:
        return True
    return False

def unique(g):
    """return the uniq elements of iterable g"""
    return dict.fromkeys(g).keys()

def sort(l):
    if not isinstance(l, list):
        l = list(l)
    l.sort()
    return l

def increasingchunks(source, min=1024, max=65536):
    '''return no less than min bytes per chunk while data remains,
    doubling min after each chunk until it reaches max'''
    def log2(x):
        if not x:
            return 0
        i = 0
        while x:
            x >>= 1
            i += 1
        return i - 1

    buf = []
    blen = 0
    for chunk in source:
        buf.append(chunk)
        blen += len(chunk)
        if blen >= min:
            if min < max:
                min = min << 1
                nmin = 1 << log2(blen)
                if nmin > min:
                    min = nmin
                if min > max:
                    min = max
            yield ''.join(buf)
            blen = 0
            buf = []
    if buf:
        yield ''.join(buf)

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, default):
    """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 default, name

def globre(pat, head='^', tail='$'):
    "convert a glob pattern into a regexp"
    i, n = 0, len(pat)
    res = ''
    group = 0
    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 += 1
            res += '(?:'
        elif c == '}' and group:
            res += ')'
            group -= 1
        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(root, n1, n2):
    '''return the relative path from one place to another.
    root should use os.sep to separate directories
    n1 should use os.sep to separate directories
    n2 should use "/" to separate directories
    returns an os.sep-separated path.

    If n1 is a relative path, it's assumed it's
    relative to root.
    n2 should always be relative to root.
    '''
    if not n1: return localpath(n2)
    if os.path.isabs(n1):
        if os.path.splitdrive(root)[0] != os.path.splitdrive(n1)[0]:
            return os.path.join(root, localpath(n2))
        n2 = '/'.join((pconvert(root), n2))
    a, b = splitpath(n1), 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) or '.'

def canonpath(root, cwd, myname):
    """return the canonical path of myname, given cwd and root"""
    if root == os.sep:
        rootsep = os.sep
    elif endswithsep(root):
        rootsep = root
    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)
    audit_path = path_auditor(root)
    if name != rootsep and 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):
                if not rel:
                    # name was actually the same as root (maybe a symlink)
                    return ''
                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=[], src=None, dflt_pat='glob'):
    """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
    dflt_pat - if a pattern in names has no explicit type, assume this one
    src - where these patterns came from (e.g. .hgignore)

    a pattern is one of:
    'glob:<glob>' - a glob relative to cwd
    're:<regexp>' - a regular expression
    'path:<path>' - a path relative to canonroot
    'relglob:<glob>' - an unrooted glob (*.c matches C files in all dirs)
    'relpath:<path>' - a path relative to cwd
    'relre:<regexp>' - a regexp that doesn't have to match the start of a name
    '<something>' - one of the cases above, selected by the dflt_pat argument

    returns:
    a 3-tuple containing
    - list of roots (places where one should start a recursive walk of the fs);
      this often matches the explicit non-pattern names passed in, but also
      includes the initial part of glob: patterns that has no glob characters
    - a bool match(filename) function
    - a bool indicating if any patterns were passed in
    """

    # a common case: no patterns at all
    if not names and not inc and not exc:
        return [], always, False

    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 not name:
            return ''
        if kind == 're':
            return name
        elif kind == 'path':
            return '^' + re.escape(name) + '(?:/|$)'
        elif kind == 'relglob':
            return globre(name, '(?:|.*/)', tail)
        elif kind == 'relpath':
            return re.escape(name) + '(?:/|$)'
        elif kind == 'relre':
            if name.startswith('^'):
                return name
            return '.*' + name
        return globre(name, '', tail)

    def matchfn(pats, tail):
        """build a matching function from a set of patterns"""
        if not pats:
            return
        try:
            pat = '(?:%s)' % '|'.join([regex(k, p, tail) for (k, p) in pats])
            if len(pat) > 20000:
                raise OverflowError()
            return re.compile(pat).match
        except OverflowError:
            # We're using a Python with a tiny regex engine and we
            # made it explode, so we'll divide the pattern list in two
            # until it works
            l = len(pats)
            if l < 2:
                raise
            a, b = matchfn(pats[:l//2], tail), matchfn(pats[l//2:], tail)
            return lambda s: a(s) or b(s)
        except re.error:
            for k, p in pats:
                try:
                    re.compile('(?:%s)' % regex(k, p, tail))
                except re.error:
                    if src:
                        raise Abort("%s: invalid pattern (%s): %s" %
                                    (src, k, p))
                    else:
                        raise Abort("invalid pattern (%s): %s" % (k, p))
            raise Abort("invalid pattern")

    def globprefix(pat):
        '''return the non-glob prefix of a path, e.g. foo/* -> foo'''
        root = []
        for p in pat.split('/'):
            if contains_glob(p): break
            root.append(p)
        return '/'.join(root) or '.'

    def normalizepats(names, default):
        pats = []
        roots = []
        anypats = False
        for kind, name in [patkind(p, default) for p in names]:
            if kind in ('glob', 'relpath'):
                name = canonpath(canonroot, cwd, name)
            elif kind in ('relglob', 'path'):
                name = normpath(name)

            pats.append((kind, name))

            if kind in ('glob', 're', 'relglob', 'relre'):
                anypats = True

            if kind == 'glob':
                root = globprefix(name)
                roots.append(root)
            elif kind in ('relpath', 'path'):
                roots.append(name or '.')
            elif kind == 'relglob':
                roots.append('.')
        return roots, pats, anypats

    roots, pats, anypats = normalizepats(names, dflt_pat)

    patmatch = matchfn(pats, '$') or always
    incmatch = always
    if inc:
        dummy, inckinds, dummy = normalizepats(inc, 'glob')
        incmatch = matchfn(inckinds, '(?:/|$)')
    excmatch = never
    if exc:
        dummy, exckinds, dummy = normalizepats(exc, 'glob')
        excmatch = matchfn(exckinds, '(?:/|$)')

    if not names and inc and not exc:
        # common case: hgignore patterns
        match = incmatch
    else:
        match = lambda fn: incmatch(fn) and not excmatch(fn) and patmatch(fn)

    return (roots, match, (inc or exc or anypats) and True)

_hgexecutable = None

def main_is_frozen():
    """return True if we are a frozen executable.

    The code supports py2exe (most common, Windows only) and tools/freeze
    (portable, not much used).
    """
    return (hasattr(sys, "frozen") or # new py2exe
            hasattr(sys, "importers") or # old py2exe
            imp.is_frozen("__main__")) # tools/freeze

def hgexecutable():
    """return location of the 'hg' executable.

    Defaults to $HG or 'hg' in the search path.
    """
    if _hgexecutable is None:
        hg = os.environ.get('HG')
        if hg:
            set_hgexecutable(hg)
        elif main_is_frozen():
            set_hgexecutable(sys.executable)
        else:
            set_hgexecutable(find_exe('hg') or 'hg')
    return _hgexecutable

def set_hgexecutable(path):
    """set location of the 'hg' executable"""
    global _hgexecutable
    _hgexecutable = path

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.'''
    def py2shell(val):
        'convert python object into string that is useful to shell'
        if val in (None, False):
            return '0'
        if val == True:
            return '1'
        return str(val)
    oldenv = {}
    for k in environ:
        oldenv[k] = os.environ.get(k)
    if cwd is not None:
        oldcwd = os.getcwd()
    origcmd = cmd
    if os.name == 'nt':
        cmd = '"%s"' % cmd
    try:
        for k, v in environ.iteritems():
            os.environ[k] = py2shell(v)
        os.environ['HG'] = hgexecutable()
        if cwd is not None and oldcwd != cwd:
            os.chdir(cwd)
        rc = os.system(cmd)
        if sys.platform == 'OpenVMS' and rc & 1:
            rc = 0
        if rc and onerr:
            errmsg = '%s %s' % (os.path.basename(origcmd.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 checksignature(func):
    '''wrap a function with code to check for calling errors'''
    def check(*args, **kwargs):
        try:
            return func(*args, **kwargs)
        except TypeError:
            if len(traceback.extract_tb(sys.exc_info()[2])) == 1:
                raise error.SignatureError
            raise

    return check

# os.path.lexists is not available on python2.3
def lexists(filename):
    "test whether a file with this name exists. does not follow symlinks"
    try:
        os.lstat(filename)
    except:
        return False
    return True

def rename(src, dst):
    """forcibly rename a file"""
    try:
        os.rename(src, dst)
    except OSError, err: # FIXME: check err (EEXIST ?)
        # 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 rename
        # destination to a temporary name, then delete that. then
        # rename is safe to do.
        temp = dst + "-force-rename"
        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 copyfile(src, dest):
    "copy a file, preserving mode and atime/mtime"
    if os.path.islink(src):
        try:
            os.unlink(dest)
        except:
            pass
        os.symlink(os.readlink(src), dest)
    else:
        try:
            shutil.copyfile(src, dest)
            shutil.copystat(src, dest)
        except shutil.Error, inst:
            raise Abort(str(inst))

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, kind in osutil.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)

class path_auditor(object):
    '''ensure that a filesystem path contains no banned components.
    the following properties of a path are checked:

    - under top-level .hg
    - starts at the root of a windows drive
    - contains ".."
    - traverses a symlink (e.g. a/symlink_here/b)
    - inside a nested repository'''

    def __init__(self, root):
        self.audited = set()
        self.auditeddir = set()
        self.root = root

    def __call__(self, path):
        if path in self.audited:
            return
        normpath = os.path.normcase(path)
        parts = splitpath(normpath)
        if (os.path.splitdrive(path)[0]
            or parts[0].lower() in ('.hg', '.hg.', '')
            or os.pardir in parts):
            raise Abort(_("path contains illegal component: %s") % path)
        if '.hg' in path.lower():
            lparts = [p.lower() for p in parts]
            for p in '.hg', '.hg.':
                if p in lparts[1:]:
                    pos = lparts.index(p)
                    base = os.path.join(*parts[:pos])
                    raise Abort(_('path %r is inside repo %r') % (path, base))
        def check(prefix):
            curpath = os.path.join(self.root, prefix)
            try:
                st = os.lstat(curpath)
            except OSError, err:
                # EINVAL can be raised as invalid path syntax under win32.
                # They must be ignored for patterns can be checked too.
                if err.errno not in (errno.ENOENT, errno.ENOTDIR, errno.EINVAL):
                    raise
            else:
                if stat.S_ISLNK(st.st_mode):
                    raise Abort(_('path %r traverses symbolic link %r') %
                                (path, prefix))
                elif (stat.S_ISDIR(st.st_mode) and
                      os.path.isdir(os.path.join(curpath, '.hg'))):
                    raise Abort(_('path %r is inside repo %r') %
                                (path, prefix))
        parts.pop()
        prefixes = []
        for n in range(len(parts)):
            prefix = os.sep.join(parts)
            if prefix in self.auditeddir:
                break
            check(prefix)
            prefixes.append(prefix)
            parts.pop()

        self.audited.add(path)
        # only add prefixes to the cache after checking everything: we don't
        # want to add "foo/bar/baz" before checking if there's a "foo/.hg"
        self.auditeddir.update(prefixes)

if os.name == 'nt':
    from windows import *
    def expand_glob(pats):
        '''On Windows, expand the implicit globs in a list of patterns'''
        ret = []
        for p in pats:
            kind, name = patkind(p, None)
            if kind is None:
                globbed = glob.glob(name)
                if globbed:
                    ret.extend(globbed)
                    continue
                # if we couldn't expand the glob, just keep it around
            ret.append(p)
        return ret
else:
    from posix import *

def makelock(info, pathname):
    try:
        return os.symlink(info, pathname)
    except OSError, why:
        if why.errno == errno.EEXIST:
            raise
    except AttributeError: # no symlink in os
        pass

    ld = os.open(pathname, os.O_CREAT | os.O_WRONLY | os.O_EXCL)
    os.write(ld, info)
    os.close(ld)

def readlock(pathname):
    try:
        return os.readlink(pathname)
    except OSError, why:
        if why.errno not in (errno.EINVAL, errno.ENOSYS):
            raise
    except AttributeError: # no symlink in os
        pass
    return posixfile(pathname).read()

def nlinks(pathname):
    """Return number of hardlinks for the given file."""
    return os.lstat(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)

# File system features

def checkcase(path):
    """
    Check whether the given path is on a case-sensitive filesystem

    Requires a path (like /foo/.hg) ending with a foldable final
    directory component.
    """
    s1 = os.stat(path)
    d, b = os.path.split(path)
    p2 = os.path.join(d, b.upper())
    if path == p2:
        p2 = os.path.join(d, b.lower())
    try:
        s2 = os.stat(p2)
        if s2 == s1:
            return False
        return True
    except:
        return True

_fspathcache = {}
def fspath(name, root):
    '''Get name in the case stored in the filesystem

    The name is either relative to root, or it is an absolute path starting
    with root. Note that this function is unnecessary, and should not be
    called, for case-sensitive filesystems (simply because it's expensive).
    '''
    # If name is absolute, make it relative
    if name.lower().startswith(root.lower()):
        l = len(root)
        if name[l] == os.sep or name[l] == os.altsep:
            l = l + 1
        name = name[l:]

    if not os.path.exists(os.path.join(root, name)):
        return None

    seps = os.sep
    if os.altsep:
        seps = seps + os.altsep
    # Protect backslashes. This gets silly very quickly.
    seps.replace('\\','\\\\')
    pattern = re.compile(r'([^%s]+)|([%s]+)' % (seps, seps))
    dir = os.path.normcase(os.path.normpath(root))
    result = []
    for part, sep in pattern.findall(name):
        if sep:
            result.append(sep)
            continue

        if dir not in _fspathcache:
            _fspathcache[dir] = os.listdir(dir)
        contents = _fspathcache[dir]

        lpart = part.lower()
        for n in contents:
            if n.lower() == lpart:
                result.append(n)
                break
        else:
            # Cannot happen, as the file exists!
            result.append(part)
        dir = os.path.join(dir, lpart)

    return ''.join(result)

def checkexec(path):
    """
    Check whether the given path is on a filesystem with UNIX-like exec flags

    Requires a directory (like /foo/.hg)
    """

    # VFAT on some Linux versions can flip mode but it doesn't persist
    # a FS remount. Frequently we can detect it if files are created
    # with exec bit on.

    try:
        EXECFLAGS = stat.S_IXUSR | stat.S_IXGRP | stat.S_IXOTH
        fh, fn = tempfile.mkstemp("", "", path)
        try:
            os.close(fh)
            m = os.stat(fn).st_mode & 0777
            new_file_has_exec = m & EXECFLAGS
            os.chmod(fn, m ^ EXECFLAGS)
            exec_flags_cannot_flip = ((os.stat(fn).st_mode & 0777) == m)
        finally:
            os.unlink(fn)
    except (IOError, OSError):
        # we don't care, the user probably won't be able to commit anyway
        return False
    return not (new_file_has_exec or exec_flags_cannot_flip)

def checklink(path):
    """check whether the given path is on a symlink-capable filesystem"""
    # mktemp is not racy because symlink creation will fail if the
    # file already exists
    name = tempfile.mktemp(dir=path)
    try:
        os.symlink(".", name)
        os.unlink(name)
        return True
    except (OSError, AttributeError):
        return False

def needbinarypatch():
    """return True if patches should be applied in binary mode by default."""
    return os.name == 'nt'

def endswithsep(path):
    '''Check path ends with os.sep or os.altsep.'''
    return path.endswith(os.sep) or os.altsep and path.endswith(os.altsep)

def splitpath(path):
    '''Split path by os.sep.
    Note that this function does not use os.altsep because this is
    an alternative of simple "xxx.split(os.sep)".
    It is recommended to use os.path.normpath() before using this
    function if need.'''
    return path.split(os.sep)

def gui():
    '''Are we running in a GUI?'''
    return os.name == "nt" or os.name == "mac" or os.environ.get("DISPLAY")

def lookup_reg(key, name=None, scope=None):
    return None

def mktempcopy(name, emptyok=False, createmode=None):
    """Create a temporary file with the same contents from name

    The permission bits are copied from the original file.

    If the temporary file is going to be truncated immediately, you
    can use emptyok=True as an optimization.

    Returns the name of the temporary file.
    """
    d, fn = os.path.split(name)
    fd, temp = tempfile.mkstemp(prefix='.%s-' % fn, dir=d)
    os.close(fd)
    # Temporary files are created with mode 0600, which is usually not
    # what we want.  If the original file already exists, just copy
    # its mode.  Otherwise, manually obey umask.
    try:
        st_mode = os.lstat(name).st_mode & 0777
    except OSError, inst:
        if inst.errno != errno.ENOENT:
            raise
        st_mode = createmode
        if st_mode is None:
            st_mode = ~umask
        st_mode &= 0666
    os.chmod(temp, st_mode)
    if emptyok:
        return temp
    try:
        try:
            ifp = posixfile(name, "rb")
        except IOError, inst:
            if inst.errno == errno.ENOENT:
                return temp
            if not getattr(inst, 'filename', None):
                inst.filename = name
            raise
        ofp = posixfile(temp, "wb")
        for chunk in filechunkiter(ifp):
            ofp.write(chunk)
        ifp.close()
        ofp.close()
    except:
        try: os.unlink(temp)
        except: pass
        raise
    return temp

class atomictempfile(posixfile):
    """file-like object that atomically updates a file

    All writes will be redirected to a temporary copy of the original
    file.  When rename is called, the copy is renamed to the original
    name, making the changes visible.
    """
    def __init__(self, name, mode, createmode):
        self.__name = name
        self.temp = mktempcopy(name, emptyok=('w' in mode),
                               createmode=createmode)
        posixfile.__init__(self, self.temp, mode)

    def rename(self):
        if not self.closed:
            posixfile.close(self)
            rename(self.temp, localpath(self.__name))

    def __del__(self):
        if not self.closed:
            try:
                os.unlink(self.temp)
            except: pass
            posixfile.close(self)

def makedirs(name, mode=None):
    """recursive directory creation with parent mode inheritance"""
    try:
        os.mkdir(name)
        if mode is not None:
            os.chmod(name, mode)
        return
    except OSError, err:
        if err.errno == errno.EEXIST:
            return
        if err.errno != errno.ENOENT:
            raise
    parent = os.path.abspath(os.path.dirname(name))
    makedirs(parent, mode)
    makedirs(name, mode)

class opener(object):
    """Open files relative to a base directory

    This class is used to hide the details of COW semantics and
    remote file access from higher level code.
    """
    def __init__(self, base, audit=True):
        self.base = base
        if audit:
            self.audit_path = path_auditor(base)
        else:
            self.audit_path = always
        self.createmode = None

    def __getattr__(self, name):
        if name == '_can_symlink':
            self._can_symlink = checklink(self.base)
            return self._can_symlink
        raise AttributeError(name)

    def _fixfilemode(self, name):
        if self.createmode is None:
            return
        os.chmod(name, self.createmode & 0666)

    def __call__(self, path, mode="r", text=False, atomictemp=False):
        self.audit_path(path)
        f = os.path.join(self.base, path)

        if not text and "b" not in mode:
            mode += "b" # for that other OS

        nlink = -1
        if mode not in ("r", "rb"):
            try:
                nlink = nlinks(f)
            except OSError:
                nlink = 0
                d = os.path.dirname(f)
                if not os.path.isdir(d):
                    makedirs(d, self.createmode)
            if atomictemp:
                return atomictempfile(f, mode, self.createmode)
            if nlink > 1:
                rename(mktempcopy(f), f)
        fp = posixfile(f, mode)
        if nlink == 0:
            self._fixfilemode(f)
        return fp

    def symlink(self, src, dst):
        self.audit_path(dst)
        linkname = os.path.join(self.base, dst)
        try:
            os.unlink(linkname)
        except OSError:
            pass

        dirname = os.path.dirname(linkname)
        if not os.path.exists(dirname):
            makedirs(dirname, self.createmode)

        if self._can_symlink:
            try:
                os.symlink(src, linkname)
            except OSError, err:
                raise OSError(err.errno, _('could not symlink to %r: %s') %
                              (src, err.strerror), linkname)
        else:
            f = self(dst, "w")
            f.write(src)
            f.close()
            self._fixfilemode(dst)

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):
        """in_iter is the iterator that's iterating over the input chunks.
        targetsize is how big a buffer to try to maintain."""
        self.iter = iter(in_iter)
        self.buf = ''
        self.targetsize = 2**16

    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 self.iter:
            # Clamp to a multiple of self.targetsize
            targetsize = max(l, self.targetsize)
            collector = cStringIO.StringIO()
            collector.write(self.buf)
            collected = len(self.buf)
            for chunk in self.iter:
                collector.write(chunk)
                collected += len(chunk)
                if collected >= targetsize:
                    break
            if collected < targetsize:
                self.iter = False
            self.buf = collector.getvalue()
        if len(self.buf) == l:
            s, self.buf = str(self.buf), ''
        else:
            s, self.buf = self.buf[:l], buffer(self.buf, l)
        return s

def filechunkiter(f, size=65536, limit=None):
    """Create a generator that produces the data in the file size
    (default 65536) bytes at a time, up to optional limit (default is
    to read all data).  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."""
    assert size >= 0
    assert limit is None or limit >= 0
    while True:
        if limit is None: nbytes = size
        else: nbytes = min(limit, size)
        s = nbytes and f.read(nbytes)
        if not s: break
        if limit: limit -= len(s)
        yield s

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 %1%2'):
    """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()
    if "%1" in format or "%2" in format:
        sign = (tz > 0) and "-" or "+"
        minutes = abs(tz) / 60
        format = format.replace("%1", "%c%02d" % (sign, minutes / 60))
        format = format.replace("%2", "%02d" % (minutes % 60))
    s = time.strftime(format, time.gmtime(float(t) - tz))
    return s

def shortdate(date=None):
    """turn (timestamp, tzoff) tuple into iso 8631 date."""
    return datestr(date, format='%Y-%m-%d')

def strdate(string, format, defaults=[]):
    """parse a localized time string and return a (unixtime, offset) tuple.
    if the string cannot be parsed, ValueError is raised."""
    def timezone(string):
        tz = string.split()[-1]
        if tz[0] in "+-" and len(tz) == 5 and tz[1:].isdigit():
            sign = (tz[0] == "+") and 1 or -1
            hours = int(tz[1:3])
            minutes = int(tz[3:5])
            return -sign * (hours * 60 + minutes) * 60
        if tz == "GMT" or tz == "UTC":
            return 0
        return None

    # NOTE: unixtime = localunixtime + offset
    offset, date = timezone(string), string
    if offset != None:
        date = " ".join(string.split()[:-1])

    # add missing elements from defaults
    for part in defaults:
        found = [True for p in part if ("%"+p) in format]
        if not found:
            date += "@" + defaults[part]
            format += "@%" + part[0]

    timetuple = time.strptime(date, format)
    localunixtime = int(calendar.timegm(timetuple))
    if offset is None:
        # local timezone
        unixtime = int(time.mktime(timetuple))
        offset = unixtime - localunixtime
    else:
        unixtime = localunixtime + offset
    return unixtime, offset

def parsedate(date, formats=None, defaults=None):
    """parse a localized date/time string and return a (unixtime, offset) tuple.

    The date may be a "unixtime offset" string or in one of the specified
    formats. If the date already is a (unixtime, offset) tuple, it is returned.
    """
    if not date:
        return 0, 0
    if isinstance(date, tuple) and len(date) == 2:
        return date
    if not formats:
        formats = defaultdateformats
    date = date.strip()
    try:
        when, offset = map(int, date.split(' '))
    except ValueError:
        # fill out defaults
        if not defaults:
            defaults = {}
        now = makedate()
        for part in "d mb yY HI M S".split():
            if part not in defaults:
                if part[0] in "HMS":
                    defaults[part] = "00"
                else:
                    defaults[part] = datestr(now, "%" + part[0])

        for format in formats:
            try:
                when, offset = strdate(date, format, defaults)
            except (ValueError, OverflowError):
                pass
            else:
                break
        else:
            raise Abort(_('invalid date: %r ') % date)
    # validate explicit (probably user-specified) date and
    # time zone offset. values must fit in signed 32 bits for
    # current 32-bit linux runtimes. timezones go from UTC-12
    # to UTC+14
    if abs(when) > 0x7fffffff:
        raise Abort(_('date exceeds 32 bits: %d') % when)
    if offset < -50400 or offset > 43200:
        raise Abort(_('impossible time zone offset: %d') % offset)
    return when, offset

def matchdate(date):
    """Return a function that matches a given date match specifier

    Formats include:

    '{date}' match a given date to the accuracy provided

    '<{date}' on or before a given date

    '>{date}' on or after a given date

    """

    def lower(date):
        d = dict(mb="1", d="1")
        return parsedate(date, extendeddateformats, d)[0]

    def upper(date):
        d = dict(mb="12", HI="23", M="59", S="59")
        for days in "31 30 29".split():
            try:
                d["d"] = days
                return parsedate(date, extendeddateformats, d)[0]
            except:
                pass
        d["d"] = "28"
        return parsedate(date, extendeddateformats, d)[0]

    if date[0] == "<":
        when = upper(date[1:])
        return lambda x: x <= when
    elif date[0] == ">":
        when = lower(date[1:])
        return lambda x: x >= when
    elif date[0] == "-":
        try:
            days = int(date[1:])
        except ValueError:
            raise Abort(_("invalid day spec: %s") % date[1:])
        when = makedate()[0] - days * 3600 * 24
        return lambda x: x >= when
    elif " to " in date:
        a, b = date.split(" to ")
        start, stop = lower(a), upper(b)
        return lambda x: x >= start and x <= stop
    else:
        start, stop = lower(date), upper(date)
        return lambda x: x >= start and x <= stop

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:]
    f = user.find(' ')
    if f >= 0:
        user = user[:f]
    f = user.find('.')
    if f >= 0:
        user = user[:f]
    return user

def email(author):
    '''get email of author.'''
    r = author.find('>')
    if r == -1: r = None
    return author[author.find('<')+1:r]

def ellipsis(text, maxlength=400):
    """Trim string to at most maxlength (default: 400) characters."""
    if len(text) <= maxlength:
        return text
    else:
        return "%s..." % (text[:maxlength-3])

def walkrepos(path, followsym=False, seen_dirs=None, recurse=False):
    '''yield every hg repository under path, recursively.'''
    def errhandler(err):
        if err.filename == path:
            raise err
    if followsym and hasattr(os.path, 'samestat'):
        def _add_dir_if_not_there(dirlst, dirname):
            match = False
            samestat = os.path.samestat
            dirstat = os.stat(dirname)
            for lstdirstat in dirlst:
                if samestat(dirstat, lstdirstat):
                    match = True
                    break
            if not match:
                dirlst.append(dirstat)
            return not match
    else:
        followsym = False

    if (seen_dirs is None) and followsym:
        seen_dirs = []
        _add_dir_if_not_there(seen_dirs, path)
    for root, dirs, files in os.walk(path, topdown=True, onerror=errhandler):
        if '.hg' in dirs:
            yield root # found a repository
            qroot = os.path.join(root, '.hg', 'patches')
            if os.path.isdir(os.path.join(qroot, '.hg')):
                yield qroot # we have a patch queue repo here
            if recurse:
                # avoid recursing inside the .hg directory
                dirs.remove('.hg')
            else:
                dirs[:] = [] # don't descend further
        elif followsym:
            newdirs = []
            for d in dirs:
                fname = os.path.join(root, d)
                if _add_dir_if_not_there(seen_dirs, fname):
                    if os.path.islink(fname):
                        for hgname in walkrepos(fname, True, seen_dirs):
                            yield hgname
                    else:
                        newdirs.append(d)
            dirs[:] = newdirs

_rcpath = None

def os_rcpath():
    '''return default os-specific hgrc search path'''
    path = system_rcpath()
    path.extend(user_rcpath())
    path = [os.path.normpath(f) for f in path]
    return path

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, kind in osutil.listdir(p):
                        if f.endswith('.rc'):
                            _rcpath.append(os.path.join(p, f))
                else:
                    _rcpath.append(p)
        else:
            _rcpath = os_rcpath()
    return _rcpath

def bytecount(nbytes):
    '''return byte count formatted as readable string, with units'''

    units = (
        (100, 1<<30, _('%.0f GB')),
        (10, 1<<30, _('%.1f GB')),
        (1, 1<<30, _('%.2f GB')),
        (100, 1<<20, _('%.0f MB')),
        (10, 1<<20, _('%.1f MB')),
        (1, 1<<20, _('%.2f MB')),
        (100, 1<<10, _('%.0f KB')),
        (10, 1<<10, _('%.1f KB')),
        (1, 1<<10, _('%.2f KB')),
        (1, 1, _('%.0f bytes')),
        )

    for multiplier, divisor, format in units:
        if nbytes >= divisor * multiplier:
            return format % (nbytes / float(divisor))
    return units[-1][2] % nbytes

def drop_scheme(scheme, path):
    sc = scheme + ':'
    if path.startswith(sc):
        path = path[len(sc):]
        if path.startswith('//'):
            path = path[2:]
    return path

def uirepr(s):
    # Avoid double backslash in Windows path repr()
    return repr(s).replace('\\\\', '\\')

def termwidth():
    if 'COLUMNS' in os.environ:
        try:
            return int(os.environ['COLUMNS'])
        except ValueError:
            pass
    try:
        import termios, array, fcntl
        for dev in (sys.stdout, sys.stdin):
            try:
                fd = dev.fileno()
                if not os.isatty(fd):
                    continue
                arri = fcntl.ioctl(fd, termios.TIOCGWINSZ, '\0' * 8)
                return array.array('h', arri)[1]
            except ValueError:
                pass
    except ImportError:
        pass
    return 80

def iterlines(iterator):
    for chunk in iterator:
        for line in chunk.splitlines():
            yield line