doctest: do not embed non-ascii characters in docstring

Since the outer docstring is parsed as a unicode on Python 3, we have to
either double-escape or construct non-ascii string from ascii string.

mercurial/encoding.py

@@ -1,585 +1,586
 # encoding.py - character transcoding support for Mercurial
 #
 #  Copyright 2005-2009 Matt Mackall <[e-mail]> and others
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import
 
 import io
 import locale
 import os
 import unicodedata
 
 from . import (
     error,
     policy,
     pycompat,
 )
 
 from .pure import (
     charencode as charencodepure,
 )
 
 charencode = policy.importmod(r'charencode')
 
 isasciistr = charencode.isasciistr
 asciilower = charencode.asciilower
 asciiupper = charencode.asciiupper
 _jsonescapeu8fast = charencode.jsonescapeu8fast
 
 _sysstr = pycompat.sysstr
 
 if pycompat.ispy3:
     unichr = chr
 
 # These unicode characters are ignored by HFS+ (Apple Technote 1150,
 # "Unicode Subtleties"), so we need to ignore them in some places for
 # sanity.
 _ignore = [unichr(int(x, 16)).encode("utf-8") for x in
            "200c 200d 200e 200f 202a 202b 202c 202d 202e "
            "206a 206b 206c 206d 206e 206f feff".split()]
 # verify the next function will work
 assert all(i.startswith(("\xe2", "\xef")) for i in _ignore)
 
 def hfsignoreclean(s):
     """Remove codepoints ignored by HFS+ from s.
 
     >>> hfsignoreclean(u'.h\u200cg'.encode('utf-8'))
     '.hg'
     >>> hfsignoreclean(u'.h\ufeffg'.encode('utf-8'))
     '.hg'
     """
     if "\xe2" in s or "\xef" in s:
         for c in _ignore:
             s = s.replace(c, '')
     return s
 
 # encoding.environ is provided read-only, which may not be used to modify
 # the process environment
 _nativeenviron = (not pycompat.ispy3 or os.supports_bytes_environ)
 if not pycompat.ispy3:
     environ = os.environ  # re-exports
 elif _nativeenviron:
     environ = os.environb  # re-exports
 else:
     # preferred encoding isn't known yet; use utf-8 to avoid unicode error
     # and recreate it once encoding is settled
     environ = dict((k.encode(u'utf-8'), v.encode(u'utf-8'))
                    for k, v in os.environ.items())  # re-exports
 
 _encodingfixers = {
     '646': lambda: 'ascii',
     'ANSI_X3.4-1968': lambda: 'ascii',
 }
 
 try:
     encoding = environ.get("HGENCODING")
     if not encoding:
         encoding = locale.getpreferredencoding().encode('ascii') or 'ascii'
         encoding = _encodingfixers.get(encoding, lambda: encoding)()
 except locale.Error:
     encoding = 'ascii'
 encodingmode = environ.get("HGENCODINGMODE", "strict")
 fallbackencoding = 'ISO-8859-1'
 
 class localstr(bytes):
     '''This class allows strings that are unmodified to be
     round-tripped to the local encoding and back'''
     def __new__(cls, u, l):
         s = bytes.__new__(cls, l)
         s._utf8 = u
         return s
     def __hash__(self):
         return hash(self._utf8) # avoid collisions in local string space
 
 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.
 
     The localstr class is used to cache the known UTF-8 encoding of
     strings next to their local representation to allow lossless
     round-trip conversion back to UTF-8.
 
     >>> u = b'foo: \\xc3\\xa4' # utf-8
     >>> l = tolocal(u)
     >>> l
     'foo: ?'
     >>> fromlocal(l)
     'foo: \\xc3\\xa4'
     >>> u2 = b'foo: \\xc3\\xa1'
     >>> d = { l: 1, tolocal(u2): 2 }
     >>> len(d) # no collision
     2
     >>> b'foo: ?' in d
     False
     >>> l1 = b'foo: \\xe4' # historical latin1 fallback
     >>> l = tolocal(l1)
     >>> l
     'foo: ?'
     >>> fromlocal(l) # magically in utf-8
     'foo: \\xc3\\xa4'
     """
 
     if isasciistr(s):
         return s
 
     try:
         try:
             # make sure string is actually stored in UTF-8
             u = s.decode('UTF-8')
             if encoding == 'UTF-8':
                 # fast path
                 return s
             r = u.encode(_sysstr(encoding), u"replace")
             if u == r.decode(_sysstr(encoding)):
                 # r is a safe, non-lossy encoding of s
                 return r
             return localstr(s, r)
         except UnicodeDecodeError:
             # we should only get here if we're looking at an ancient changeset
             try:
                 u = s.decode(_sysstr(fallbackencoding))
                 r = u.encode(_sysstr(encoding), u"replace")
                 if u == r.decode(_sysstr(encoding)):
                     # r is a safe, non-lossy encoding of s
                     return r
                 return localstr(u.encode('UTF-8'), r)
             except UnicodeDecodeError:
                 u = s.decode("utf-8", "replace") # last ditch
                 # can't round-trip
                 return u.encode(_sysstr(encoding), u"replace")
     except LookupError as k:
         raise error.Abort(k, hint="please check your locale settings")
 
 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.
     """
 
     # can we do a lossless round-trip?
     if isinstance(s, localstr):
         return s._utf8
     if isasciistr(s):
         return s
 
     try:
         u = s.decode(_sysstr(encoding), _sysstr(encodingmode))
         return u.encode("utf-8")
     except UnicodeDecodeError as inst:
         sub = s[max(0, inst.start - 10):inst.start + 10]
         raise error.Abort("decoding near '%s': %s!" % (sub, inst))
     except LookupError as k:
         raise error.Abort(k, hint="please check your locale settings")
 
 def unitolocal(u):
     """Convert a unicode string to a byte string of local encoding"""
     return tolocal(u.encode('utf-8'))
 
 def unifromlocal(s):
     """Convert a byte string of local encoding to a unicode string"""
     return fromlocal(s).decode('utf-8')
 
 def unimethod(bytesfunc):
     """Create a proxy method that forwards __unicode__() and __str__() of
     Python 3 to __bytes__()"""
     def unifunc(obj):
         return unifromlocal(bytesfunc(obj))
     return unifunc
 
 # converter functions between native str and byte string. use these if the
 # character encoding is not aware (e.g. exception message) or is known to
 # be locale dependent (e.g. date formatting.)
 if pycompat.ispy3:
     strtolocal = unitolocal
     strfromlocal = unifromlocal
     strmethod = unimethod
 else:
     strtolocal = pycompat.identity
     strfromlocal = pycompat.identity
     strmethod = pycompat.identity
 
 if not _nativeenviron:
     # now encoding and helper functions are available, recreate the environ
     # dict to be exported to other modules
     environ = dict((tolocal(k.encode(u'utf-8')), tolocal(v.encode(u'utf-8')))
                    for k, v in os.environ.items())  # re-exports
 
 # How to treat ambiguous-width characters. Set to 'wide' to treat as wide.
 _wide = _sysstr(environ.get("HGENCODINGAMBIGUOUS", "narrow") == "wide"
                 and "WFA" or "WF")
 
 def colwidth(s):
     "Find the column width of a string for display in the local encoding"
     return ucolwidth(s.decode(_sysstr(encoding), u'replace'))
 
 def ucolwidth(d):
     "Find the column width of a Unicode string for display"
     eaw = getattr(unicodedata, 'east_asian_width', None)
     if eaw is not None:
         return sum([eaw(c) in _wide and 2 or 1 for c in d])
     return len(d)
 
 def getcols(s, start, c):
     '''Use colwidth to find a c-column substring of s starting at byte
     index start'''
     for x in xrange(start + c, len(s)):
         t = s[start:x]
         if colwidth(t) == c:
             return t
 
 def trim(s, width, ellipsis='', leftside=False):
     """Trim string 's' to at most 'width' columns (including 'ellipsis').
 
     If 'leftside' is True, left side of string 's' is trimmed.
     'ellipsis' is always placed at trimmed side.
 
+    >>> from .node import bin
     >>> ellipsis = b'+++'
     >>> from . import encoding
     >>> encoding.encoding = b'utf-8'
     >>> t = b'1234567890'
     >>> print trim(t, 12, ellipsis=ellipsis)
     1234567890
     >>> print trim(t, 10, ellipsis=ellipsis)
     1234567890
     >>> print trim(t, 8, ellipsis=ellipsis)
     12345+++
     >>> print trim(t, 8, ellipsis=ellipsis, leftside=True)
     +++67890
     >>> print trim(t, 8)
     12345678
     >>> print trim(t, 8, leftside=True)
     34567890
     >>> print trim(t, 3, ellipsis=ellipsis)
     +++
     >>> print trim(t, 1, ellipsis=ellipsis)
     +
     >>> u = u'\u3042\u3044\u3046\u3048\u304a' # 2 x 5 = 10 columns
     >>> t = u.encode(pycompat.sysstr(encoding.encoding))
     >>> print trim(t, 12, ellipsis=ellipsis)
     \xe3\x81\x82\xe3\x81\x84\xe3\x81\x86\xe3\x81\x88\xe3\x81\x8a
     >>> print trim(t, 10, ellipsis=ellipsis)
     \xe3\x81\x82\xe3\x81\x84\xe3\x81\x86\xe3\x81\x88\xe3\x81\x8a
     >>> print trim(t, 8, ellipsis=ellipsis)
     \xe3\x81\x82\xe3\x81\x84+++
     >>> print trim(t, 8, ellipsis=ellipsis, leftside=True)
     +++\xe3\x81\x88\xe3\x81\x8a
     >>> print trim(t, 5)
     \xe3\x81\x82\xe3\x81\x84
     >>> print trim(t, 5, leftside=True)
     \xe3\x81\x88\xe3\x81\x8a
     >>> print trim(t, 4, ellipsis=ellipsis)
     +++
     >>> print trim(t, 4, ellipsis=ellipsis, leftside=True)
     +++
-    >>> t = b'\x11\x22\x33\x44\x55\x66\x77\x88\x99\xaa' # invalid byte sequence
+    >>> t = bin(b'112233445566778899aa') # invalid byte sequence
     >>> print trim(t, 12, ellipsis=ellipsis)
     \x11\x22\x33\x44\x55\x66\x77\x88\x99\xaa
     >>> print trim(t, 10, ellipsis=ellipsis)
     \x11\x22\x33\x44\x55\x66\x77\x88\x99\xaa
     >>> print trim(t, 8, ellipsis=ellipsis)
     \x11\x22\x33\x44\x55+++
     >>> print trim(t, 8, ellipsis=ellipsis, leftside=True)
     +++\x66\x77\x88\x99\xaa
     >>> print trim(t, 8)
     \x11\x22\x33\x44\x55\x66\x77\x88
     >>> print trim(t, 8, leftside=True)
     \x33\x44\x55\x66\x77\x88\x99\xaa
     >>> print trim(t, 3, ellipsis=ellipsis)
     +++
     >>> print trim(t, 1, ellipsis=ellipsis)
     +
     """
     try:
         u = s.decode(_sysstr(encoding))
     except UnicodeDecodeError:
         if len(s) <= width: # trimming is not needed
             return s
         width -= len(ellipsis)
         if width <= 0: # no enough room even for ellipsis
             return ellipsis[:width + len(ellipsis)]
         if leftside:
             return ellipsis + s[-width:]
         return s[:width] + ellipsis
 
     if ucolwidth(u) <= width: # trimming is not needed
         return s
 
     width -= len(ellipsis)
     if width <= 0: # no enough room even for ellipsis
         return ellipsis[:width + len(ellipsis)]
 
     if leftside:
         uslice = lambda i: u[i:]
         concat = lambda s: ellipsis + s
     else:
         uslice = lambda i: u[:-i]
         concat = lambda s: s + ellipsis
     for i in xrange(1, len(u)):
         usub = uslice(i)
         if ucolwidth(usub) <= width:
             return concat(usub.encode(_sysstr(encoding)))
     return ellipsis # no enough room for multi-column characters
 
 def lower(s):
     "best-effort encoding-aware case-folding of local string s"
     try:
         return asciilower(s)
     except UnicodeDecodeError:
         pass
     try:
         if isinstance(s, localstr):
             u = s._utf8.decode("utf-8")
         else:
             u = s.decode(_sysstr(encoding), _sysstr(encodingmode))
 
         lu = u.lower()
         if u == lu:
             return s # preserve localstring
         return lu.encode(_sysstr(encoding))
     except UnicodeError:
         return s.lower() # we don't know how to fold this except in ASCII
     except LookupError as k:
         raise error.Abort(k, hint="please check your locale settings")
 
 def upper(s):
     "best-effort encoding-aware case-folding of local string s"
     try:
         return asciiupper(s)
     except UnicodeDecodeError:
         return upperfallback(s)
 
 def upperfallback(s):
     try:
         if isinstance(s, localstr):
             u = s._utf8.decode("utf-8")
         else:
             u = s.decode(_sysstr(encoding), _sysstr(encodingmode))
 
         uu = u.upper()
         if u == uu:
             return s # preserve localstring
         return uu.encode(_sysstr(encoding))
     except UnicodeError:
         return s.upper() # we don't know how to fold this except in ASCII
     except LookupError as k:
         raise error.Abort(k, hint="please check your locale settings")
 
 class normcasespecs(object):
     '''what a platform's normcase does to ASCII strings
 
     This is specified per platform, and should be consistent with what normcase
     on that platform actually does.
 
     lower: normcase lowercases ASCII strings
     upper: normcase uppercases ASCII strings
     other: the fallback function should always be called
 
     This should be kept in sync with normcase_spec in util.h.'''
     lower = -1
     upper = 1
     other = 0
 
 def jsonescape(s, paranoid=False):
     '''returns a string suitable for JSON
 
     JSON is problematic for us because it doesn't support non-Unicode
     bytes. To deal with this, we take the following approach:
 
     - localstr objects are converted back to UTF-8
     - valid UTF-8/ASCII strings are passed as-is
     - other strings are converted to UTF-8b surrogate encoding
     - apply JSON-specified string escaping
 
     (escapes are doubled in these tests)
 
     >>> jsonescape(b'this is a test')
     'this is a test'
     >>> jsonescape(b'escape characters: \\0 \\x0b \\x7f')
     'escape characters: \\\\u0000 \\\\u000b \\\\u007f'
     >>> jsonescape(b'escape characters: \\b \\t \\n \\f \\r \\" \\\\')
     'escape characters: \\\\b \\\\t \\\\n \\\\f \\\\r \\\\" \\\\\\\\'
     >>> jsonescape(b'a weird byte: \\xdd')
     'a weird byte: \\xed\\xb3\\x9d'
     >>> jsonescape(b'utf-8: caf\\xc3\\xa9')
     'utf-8: caf\\xc3\\xa9'
     >>> jsonescape(b'')
     ''
 
     If paranoid, non-ascii and common troublesome characters are also escaped.
     This is suitable for web output.
 
     >>> s = b'escape characters: \\0 \\x0b \\x7f'
     >>> assert jsonescape(s) == jsonescape(s, paranoid=True)
     >>> s = b'escape characters: \\b \\t \\n \\f \\r \\" \\\\'
     >>> assert jsonescape(s) == jsonescape(s, paranoid=True)
     >>> jsonescape(b'escape boundary: \\x7e \\x7f \\xc2\\x80', paranoid=True)
     'escape boundary: ~ \\\\u007f \\\\u0080'
     >>> jsonescape(b'a weird byte: \\xdd', paranoid=True)
     'a weird byte: \\\\udcdd'
     >>> jsonescape(b'utf-8: caf\\xc3\\xa9', paranoid=True)
     'utf-8: caf\\\\u00e9'
     >>> jsonescape(b'non-BMP: \\xf0\\x9d\\x84\\x9e', paranoid=True)
     'non-BMP: \\\\ud834\\\\udd1e'
     >>> jsonescape(b'<[e-mail]>', paranoid=True)
     '\\\\[e-mail]\\\\u003e'
     '''
 
     u8chars = toutf8b(s)
     try:
         return _jsonescapeu8fast(u8chars, paranoid)
     except ValueError:
         pass
     return charencodepure.jsonescapeu8fallback(u8chars, paranoid)
 
 _utf8len = [0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 3, 4]
 
 def getutf8char(s, pos):
     '''get the next full utf-8 character in the given string, starting at pos
 
     Raises a UnicodeError if the given location does not start a valid
     utf-8 character.
     '''
 
     # find how many bytes to attempt decoding from first nibble
     l = _utf8len[ord(s[pos]) >> 4]
     if not l: # ascii
         return s[pos]
 
     c = s[pos:pos + l]
     # validate with attempted decode
     c.decode("utf-8")
     return c
 
 def toutf8b(s):
     '''convert a local, possibly-binary string into UTF-8b
 
     This is intended as a generic method to preserve data when working
     with schemes like JSON and XML that have no provision for
     arbitrary byte strings. As Mercurial often doesn't know
     what encoding data is in, we use so-called UTF-8b.
 
     If a string is already valid UTF-8 (or ASCII), it passes unmodified.
     Otherwise, unsupported bytes are mapped to UTF-16 surrogate range,
     uDC00-uDCFF.
 
     Principles of operation:
 
     - ASCII and UTF-8 data successfully round-trips and is understood
       by Unicode-oriented clients
     - filenames and file contents in arbitrary other encodings can have
       be round-tripped or recovered by clueful clients
     - local strings that have a cached known UTF-8 encoding (aka
       localstr) get sent as UTF-8 so Unicode-oriented clients get the
       Unicode data they want
     - because we must preserve UTF-8 bytestring in places such as
       filenames, metadata can't be roundtripped without help
 
     (Note: "UTF-8b" often refers to decoding a mix of valid UTF-8 and
     arbitrary bytes into an internal Unicode format that can be
     re-encoded back into the original. Here we are exposing the
     internal surrogate encoding as a UTF-8 string.)
     '''
 
     if not isinstance(s, localstr) and isasciistr(s):
         return s
     if "\xed" not in s:
         if isinstance(s, localstr):
             return s._utf8
         try:
             s.decode('utf-8')
             return s
         except UnicodeDecodeError:
             pass
 
     r = ""
     pos = 0
     l = len(s)
     while pos < l:
         try:
             c = getutf8char(s, pos)
             if "\xed\xb0\x80" <= c <= "\xed\xb3\xbf":
                 # have to re-escape existing U+DCxx characters
                 c = unichr(0xdc00 + ord(s[pos])).encode('utf-8')
                 pos += 1
             else:
                 pos += len(c)
         except UnicodeDecodeError:
             c = unichr(0xdc00 + ord(s[pos])).encode('utf-8')
             pos += 1
         r += c
     return r
 
 def fromutf8b(s):
     '''Given a UTF-8b string, return a local, possibly-binary string.
 
     return the original binary string. This
     is a round-trip process for strings like filenames, but metadata
     that's was passed through tolocal will remain in UTF-8.
 
     >>> roundtrip = lambda x: fromutf8b(toutf8b(x)) == x
     >>> m = b"\\xc3\\xa9\\x99abcd"
     >>> toutf8b(m)
     '\\xc3\\xa9\\xed\\xb2\\x99abcd'
     >>> roundtrip(m)
     True
     >>> roundtrip(b"\\xc2\\xc2\\x80")
     True
     >>> roundtrip(b"\\xef\\xbf\\xbd")
     True
     >>> roundtrip(b"\\xef\\xef\\xbf\\xbd")
     True
     >>> roundtrip(b"\\xf1\\x80\\x80\\x80\\x80")
     True
     '''
 
     if isasciistr(s):
         return s
     # fast path - look for uDxxx prefixes in s
     if "\xed" not in s:
         return s
 
     # We could do this with the unicode type but some Python builds
     # use UTF-16 internally (issue5031) which causes non-BMP code
     # points to be escaped. Instead, we use our handy getutf8char
     # helper again to walk the string without "decoding" it.
 
     r = ""
     pos = 0
     l = len(s)
     while pos < l:
         c = getutf8char(s, pos)
         pos += len(c)
         # unescape U+DCxx characters
         if "\xed\xb0\x80" <= c <= "\xed\xb3\xbf":
             c = chr(ord(c.decode("utf-8")) & 0xff)
         r += c
     return r
 
 if pycompat.ispy3:
     class strio(io.TextIOWrapper):
         """Wrapper around TextIOWrapper that respects hg's encoding assumptions.
 
         Also works around Python closing streams.
         """
 
         def __init__(self, buffer):
             super(strio, self).__init__(buffer, encoding=_sysstr(encoding))
 
         def __del__(self):
             """Override __del__ so it doesn't close the underlying stream."""
 else:
     strio = pycompat.identity

mercurial/store.py

@@ -1,575 +1,575
 # store.py - repository store handling for Mercurial
 #
 # Copyright 2008 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import
 
 import errno
 import hashlib
 import os
 import stat
 
 from .i18n import _
 from . import (
     error,
     policy,
     pycompat,
     util,
     vfs as vfsmod,
 )
 
 parsers = policy.importmod(r'parsers')
 
 # This avoids a collision between a file named foo and a dir named
 # foo.i or foo.d
 def _encodedir(path):
     '''
     >>> _encodedir(b'data/foo.i')
     'data/foo.i'
     >>> _encodedir(b'data/foo.i/bla.i')
     'data/foo.i.hg/bla.i'
     >>> _encodedir(b'data/foo.i.hg/bla.i')
     'data/foo.i.hg.hg/bla.i'
     >>> _encodedir(b'data/foo.i\\ndata/foo.i/bla.i\\ndata/foo.i.hg/bla.i\\n')
     'data/foo.i\\ndata/foo.i.hg/bla.i\\ndata/foo.i.hg.hg/bla.i\\n'
     '''
     return (path
             .replace(".hg/", ".hg.hg/")
             .replace(".i/", ".i.hg/")
             .replace(".d/", ".d.hg/"))
 
 encodedir = getattr(parsers, 'encodedir', _encodedir)
 
 def decodedir(path):
     '''
     >>> decodedir(b'data/foo.i')
     'data/foo.i'
     >>> decodedir(b'data/foo.i.hg/bla.i')
     'data/foo.i/bla.i'
     >>> decodedir(b'data/foo.i.hg.hg/bla.i')
     'data/foo.i.hg/bla.i'
     '''
     if ".hg/" not in path:
         return path
     return (path
             .replace(".d.hg/", ".d/")
             .replace(".i.hg/", ".i/")
             .replace(".hg.hg/", ".hg/"))
 
 def _reserved():
     ''' characters that are problematic for filesystems
 
     * ascii escapes (0..31)
     * ascii hi (126..255)
     * windows specials
 
     these characters will be escaped by encodefunctions
     '''
     winreserved = [ord(x) for x in u'\\:*?"<>|']
     for x in range(32):
         yield x
     for x in range(126, 256):
         yield x
     for x in winreserved:
         yield x
 
 def _buildencodefun():
     '''
     >>> enc, dec = _buildencodefun()
 
     >>> enc(b'nothing/special.txt')
     'nothing/special.txt'
     >>> dec(b'nothing/special.txt')
     'nothing/special.txt'
 
     >>> enc(b'HELLO')
     '_h_e_l_l_o'
     >>> dec(b'_h_e_l_l_o')
     'HELLO'
 
     >>> enc(b'hello:world?')
     'hello~3aworld~3f'
     >>> dec(b'hello~3aworld~3f')
     'hello:world?'
 
-    >>> enc(b'the\x07quick\xADshot')
+    >>> enc(b'the\\x07quick\\xADshot')
     'the~07quick~adshot'
     >>> dec(b'the~07quick~adshot')
     'the\\x07quick\\xadshot'
     '''
     e = '_'
     xchr = pycompat.bytechr
     asciistr = list(map(xchr, range(127)))
     capitals = list(range(ord("A"), ord("Z") + 1))
 
     cmap = dict((x, x) for x in asciistr)
     for x in _reserved():
         cmap[xchr(x)] = "~%02x" % x
     for x in capitals + [ord(e)]:
         cmap[xchr(x)] = e + xchr(x).lower()
 
     dmap = {}
     for k, v in cmap.iteritems():
         dmap[v] = k
     def decode(s):
         i = 0
         while i < len(s):
             for l in xrange(1, 4):
                 try:
                     yield dmap[s[i:i + l]]
                     i += l
                     break
                 except KeyError:
                     pass
             else:
                 raise KeyError
     return (lambda s: ''.join([cmap[s[c:c + 1]] for c in xrange(len(s))]),
             lambda s: ''.join(list(decode(s))))
 
 _encodefname, _decodefname = _buildencodefun()
 
 def encodefilename(s):
     '''
     >>> encodefilename(b'foo.i/bar.d/bla.hg/hi:world?/HELLO')
     'foo.i.hg/bar.d.hg/bla.hg.hg/hi~3aworld~3f/_h_e_l_l_o'
     '''
     return _encodefname(encodedir(s))
 
 def decodefilename(s):
     '''
     >>> decodefilename(b'foo.i.hg/bar.d.hg/bla.hg.hg/hi~3aworld~3f/_h_e_l_l_o')
     'foo.i/bar.d/bla.hg/hi:world?/HELLO'
     '''
     return decodedir(_decodefname(s))
 
 def _buildlowerencodefun():
     '''
     >>> f = _buildlowerencodefun()
     >>> f(b'nothing/special.txt')
     'nothing/special.txt'
     >>> f(b'HELLO')
     'hello'
     >>> f(b'hello:world?')
     'hello~3aworld~3f'
-    >>> f(b'the\x07quick\xADshot')
+    >>> f(b'the\\x07quick\\xADshot')
     'the~07quick~adshot'
     '''
     cmap = dict([(chr(x), chr(x)) for x in xrange(127)])
     for x in _reserved():
         cmap[chr(x)] = "~%02x" % x
     for x in range(ord("A"), ord("Z") + 1):
         cmap[chr(x)] = chr(x).lower()
     return lambda s: "".join([cmap[c] for c in s])
 
 lowerencode = getattr(parsers, 'lowerencode', None) or _buildlowerencodefun()
 
 # Windows reserved names: con, prn, aux, nul, com1..com9, lpt1..lpt9
 _winres3 = ('aux', 'con', 'prn', 'nul') # length 3
 _winres4 = ('com', 'lpt')               # length 4 (with trailing 1..9)
 def _auxencode(path, dotencode):
     '''
     Encodes filenames containing names reserved by Windows or which end in
     period or space. Does not touch other single reserved characters c.
     Specifically, c in '\\:*?"<>|' or ord(c) <= 31 are *not* encoded here.
     Additionally encodes space or period at the beginning, if dotencode is
     True. Parameter path is assumed to be all lowercase.
     A segment only needs encoding if a reserved name appears as a
     basename (e.g. "aux", "aux.foo"). A directory or file named "foo.aux"
     doesn't need encoding.
 
     >>> s = b'.foo/aux.txt/txt.aux/con/prn/nul/foo.'
     >>> _auxencode(s.split(b'/'), True)
     ['~2efoo', 'au~78.txt', 'txt.aux', 'co~6e', 'pr~6e', 'nu~6c', 'foo~2e']
     >>> s = b'.com1com2/lpt9.lpt4.lpt1/conprn/com0/lpt0/foo.'
     >>> _auxencode(s.split(b'/'), False)
     ['.com1com2', 'lp~749.lpt4.lpt1', 'conprn', 'com0', 'lpt0', 'foo~2e']
     >>> _auxencode([b'foo. '], True)
     ['foo.~20']
     >>> _auxencode([b' .foo'], True)
     ['~20.foo']
     '''
     for i, n in enumerate(path):
         if not n:
             continue
         if dotencode and n[0] in '. ':
             n = "~%02x" % ord(n[0:1]) + n[1:]
             path[i] = n
         else:
             l = n.find('.')
             if l == -1:
                 l = len(n)
             if ((l == 3 and n[:3] in _winres3) or
                 (l == 4 and n[3:4] <= '9' and n[3:4] >= '1'
                         and n[:3] in _winres4)):
                 # encode third letter ('aux' -> 'au~78')
                 ec = "~%02x" % ord(n[2:3])
                 n = n[0:2] + ec + n[3:]
                 path[i] = n
         if n[-1] in '. ':
             # encode last period or space ('foo...' -> 'foo..~2e')
             path[i] = n[:-1] + "~%02x" % ord(n[-1:])
     return path
 
 _maxstorepathlen = 120
 _dirprefixlen = 8
 _maxshortdirslen = 8 * (_dirprefixlen + 1) - 4
 
 def _hashencode(path, dotencode):
     digest = hashlib.sha1(path).hexdigest()
     le = lowerencode(path[5:]).split('/') # skips prefix 'data/' or 'meta/'
     parts = _auxencode(le, dotencode)
     basename = parts[-1]
     _root, ext = os.path.splitext(basename)
     sdirs = []
     sdirslen = 0
     for p in parts[:-1]:
         d = p[:_dirprefixlen]
         if d[-1] in '. ':
             # Windows can't access dirs ending in period or space
             d = d[:-1] + '_'
         if sdirslen == 0:
             t = len(d)
         else:
             t = sdirslen + 1 + len(d)
             if t > _maxshortdirslen:
                 break
         sdirs.append(d)
         sdirslen = t
     dirs = '/'.join(sdirs)
     if len(dirs) > 0:
         dirs += '/'
     res = 'dh/' + dirs + digest + ext
     spaceleft = _maxstorepathlen - len(res)
     if spaceleft > 0:
         filler = basename[:spaceleft]
         res = 'dh/' + dirs + filler + digest + ext
     return res
 
 def _hybridencode(path, dotencode):
     '''encodes path with a length limit
 
     Encodes all paths that begin with 'data/', according to the following.
 
     Default encoding (reversible):
 
     Encodes all uppercase letters 'X' as '_x'. All reserved or illegal
     characters are encoded as '~xx', where xx is the two digit hex code
     of the character (see encodefilename).
     Relevant path components consisting of Windows reserved filenames are
     masked by encoding the third character ('aux' -> 'au~78', see _auxencode).
 
     Hashed encoding (not reversible):
 
     If the default-encoded path is longer than _maxstorepathlen, a
     non-reversible hybrid hashing of the path is done instead.
     This encoding uses up to _dirprefixlen characters of all directory
     levels of the lowerencoded path, but not more levels than can fit into
     _maxshortdirslen.
     Then follows the filler followed by the sha digest of the full path.
     The filler is the beginning of the basename of the lowerencoded path
     (the basename is everything after the last path separator). The filler
     is as long as possible, filling in characters from the basename until
     the encoded path has _maxstorepathlen characters (or all chars of the
     basename have been taken).
     The extension (e.g. '.i' or '.d') is preserved.
 
     The string 'data/' at the beginning is replaced with 'dh/', if the hashed
     encoding was used.
     '''
     path = encodedir(path)
     ef = _encodefname(path).split('/')
     res = '/'.join(_auxencode(ef, dotencode))
     if len(res) > _maxstorepathlen:
         res = _hashencode(path, dotencode)
     return res
 
 def _pathencode(path):
     de = encodedir(path)
     if len(path) > _maxstorepathlen:
         return _hashencode(de, True)
     ef = _encodefname(de).split('/')
     res = '/'.join(_auxencode(ef, True))
     if len(res) > _maxstorepathlen:
         return _hashencode(de, True)
     return res
 
 _pathencode = getattr(parsers, 'pathencode', _pathencode)
 
 def _plainhybridencode(f):
     return _hybridencode(f, False)
 
 def _calcmode(vfs):
     try:
         # files in .hg/ will be created using this mode
         mode = vfs.stat().st_mode
             # avoid some useless chmods
         if (0o777 & ~util.umask) == (0o777 & mode):
             mode = None
     except OSError:
         mode = None
     return mode
 
 _data = ('data meta 00manifest.d 00manifest.i 00changelog.d 00changelog.i'
          ' phaseroots obsstore')
 
 class basicstore(object):
     '''base class for local repository stores'''
     def __init__(self, path, vfstype):
         vfs = vfstype(path)
         self.path = vfs.base
         self.createmode = _calcmode(vfs)
         vfs.createmode = self.createmode
         self.rawvfs = vfs
         self.vfs = vfsmod.filtervfs(vfs, encodedir)
         self.opener = self.vfs
 
     def join(self, f):
         return self.path + '/' + encodedir(f)
 
     def _walk(self, relpath, recurse):
         '''yields (unencoded, encoded, size)'''
         path = self.path
         if relpath:
             path += '/' + relpath
         striplen = len(self.path) + 1
         l = []
         if self.rawvfs.isdir(path):
             visit = [path]
             readdir = self.rawvfs.readdir
             while visit:
                 p = visit.pop()
                 for f, kind, st in readdir(p, stat=True):
                     fp = p + '/' + f
                     if kind == stat.S_IFREG and f[-2:] in ('.d', '.i'):
                         n = util.pconvert(fp[striplen:])
                         l.append((decodedir(n), n, st.st_size))
                     elif kind == stat.S_IFDIR and recurse:
                         visit.append(fp)
         l.sort()
         return l
 
     def datafiles(self):
         return self._walk('data', True) + self._walk('meta', True)
 
     def topfiles(self):
         # yield manifest before changelog
         return reversed(self._walk('', False))
 
     def walk(self):
         '''yields (unencoded, encoded, size)'''
         # yield data files first
         for x in self.datafiles():
             yield x
         for x in self.topfiles():
             yield x
 
     def copylist(self):
         return ['requires'] + _data.split()
 
     def write(self, tr):
         pass
 
     def invalidatecaches(self):
         pass
 
     def markremoved(self, fn):
         pass
 
     def __contains__(self, path):
         '''Checks if the store contains path'''
         path = "/".join(("data", path))
         # file?
         if self.vfs.exists(path + ".i"):
             return True
         # dir?
         if not path.endswith("/"):
             path = path + "/"
         return self.vfs.exists(path)
 
 class encodedstore(basicstore):
     def __init__(self, path, vfstype):
         vfs = vfstype(path + '/store')
         self.path = vfs.base
         self.createmode = _calcmode(vfs)
         vfs.createmode = self.createmode
         self.rawvfs = vfs
         self.vfs = vfsmod.filtervfs(vfs, encodefilename)
         self.opener = self.vfs
 
     def datafiles(self):
         for a, b, size in super(encodedstore, self).datafiles():
             try:
                 a = decodefilename(a)
             except KeyError:
                 a = None
             yield a, b, size
 
     def join(self, f):
         return self.path + '/' + encodefilename(f)
 
     def copylist(self):
         return (['requires', '00changelog.i'] +
                 ['store/' + f for f in _data.split()])
 
 class fncache(object):
     # the filename used to be partially encoded
     # hence the encodedir/decodedir dance
     def __init__(self, vfs):
         self.vfs = vfs
         self.entries = None
         self._dirty = False
 
     def _load(self):
         '''fill the entries from the fncache file'''
         self._dirty = False
         try:
             fp = self.vfs('fncache', mode='rb')
         except IOError:
             # skip nonexistent file
             self.entries = set()
             return
         self.entries = set(decodedir(fp.read()).splitlines())
         if '' in self.entries:
             fp.seek(0)
             for n, line in enumerate(util.iterfile(fp)):
                 if not line.rstrip('\n'):
                     t = _('invalid entry in fncache, line %d') % (n + 1)
                     raise error.Abort(t)
         fp.close()
 
     def write(self, tr):
         if self._dirty:
             tr.addbackup('fncache')
             fp = self.vfs('fncache', mode='wb', atomictemp=True)
             if self.entries:
                 fp.write(encodedir('\n'.join(self.entries) + '\n'))
             fp.close()
             self._dirty = False
 
     def add(self, fn):
         if self.entries is None:
             self._load()
         if fn not in self.entries:
             self._dirty = True
             self.entries.add(fn)
 
     def remove(self, fn):
         if self.entries is None:
             self._load()
         try:
             self.entries.remove(fn)
             self._dirty = True
         except KeyError:
             pass
 
     def __contains__(self, fn):
         if self.entries is None:
             self._load()
         return fn in self.entries
 
     def __iter__(self):
         if self.entries is None:
             self._load()
         return iter(self.entries)
 
 class _fncachevfs(vfsmod.abstractvfs, vfsmod.proxyvfs):
     def __init__(self, vfs, fnc, encode):
         vfsmod.proxyvfs.__init__(self, vfs)
         self.fncache = fnc
         self.encode = encode
 
     def __call__(self, path, mode='r', *args, **kw):
         if mode not in ('r', 'rb') and (path.startswith('data/') or
                                         path.startswith('meta/')):
             self.fncache.add(path)
         return self.vfs(self.encode(path), mode, *args, **kw)
 
     def join(self, path):
         if path:
             return self.vfs.join(self.encode(path))
         else:
             return self.vfs.join(path)
 
 class fncachestore(basicstore):
     def __init__(self, path, vfstype, dotencode):
         if dotencode:
             encode = _pathencode
         else:
             encode = _plainhybridencode
         self.encode = encode
         vfs = vfstype(path + '/store')
         self.path = vfs.base
         self.pathsep = self.path + '/'
         self.createmode = _calcmode(vfs)
         vfs.createmode = self.createmode
         self.rawvfs = vfs
         fnc = fncache(vfs)
         self.fncache = fnc
         self.vfs = _fncachevfs(vfs, fnc, encode)
         self.opener = self.vfs
 
     def join(self, f):
         return self.pathsep + self.encode(f)
 
     def getsize(self, path):
         return self.rawvfs.stat(path).st_size
 
     def datafiles(self):
         for f in sorted(self.fncache):
             ef = self.encode(f)
             try:
                 yield f, ef, self.getsize(ef)
             except OSError as err:
                 if err.errno != errno.ENOENT:
                     raise
 
     def copylist(self):
         d = ('data meta dh fncache phaseroots obsstore'
              ' 00manifest.d 00manifest.i 00changelog.d 00changelog.i')
         return (['requires', '00changelog.i'] +
                 ['store/' + f for f in d.split()])
 
     def write(self, tr):
         self.fncache.write(tr)
 
     def invalidatecaches(self):
         self.fncache.entries = None
 
     def markremoved(self, fn):
         self.fncache.remove(fn)
 
     def _exists(self, f):
         ef = self.encode(f)
         try:
             self.getsize(ef)
             return True
         except OSError as err:
             if err.errno != errno.ENOENT:
                 raise
             # nonexistent entry
             return False
 
     def __contains__(self, path):
         '''Checks if the store contains path'''
         path = "/".join(("data", path))
         # check for files (exact match)
         e = path + '.i'
         if e in self.fncache and self._exists(e):
             return True
         # now check for directories (prefix match)
         if not path.endswith('/'):
             path += '/'
         for e in self.fncache:
             if e.startswith(path) and self._exists(e):
                 return True
         return False
 
 def store(requirements, path, vfstype):
     if 'store' in requirements:
         if 'fncache' in requirements:
             return fncachestore(path, vfstype, 'dotencode' in requirements)
         return encodedstore(path, vfstype)
     return basicstore(path, vfstype)