encoding: drop circular import by proxying through '<policy>.charencode'

I decided not to split charencode.c to new C extension module because it
would duplicate binary codes unnecessarily.

mercurial/pure/charencode.py

@@ -0,0 +1,22 b''
+# charencode.py - miscellaneous character encoding
+#
+#  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
+
+def asciilower(s):
+    '''convert a string to lowercase if ASCII
+
+    Raises UnicodeDecodeError if non-ASCII characters are found.'''
+    s.decode('ascii')
+    return s.lower()
+
+def asciiupper(s):
+    '''convert a string to uppercase if ASCII
+
+    Raises UnicodeDecodeError if non-ASCII characters are found.'''
+    s.decode('ascii')
+    return s.upper()

mercurial/encoding.py

@@ -1,602 +1,575 b''
 # 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 array
 import locale
 import os
 import unicodedata
 
 from . import (
     error,
     policy,
     pycompat,
 )
 
+charencode = policy.importmod(r'charencode')
+
+asciilower = charencode.asciilower
+asciiupper = charencode.asciiupper
+
 _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(str):
     '''This class allows strings that are unmodified to be
     round-tripped to the local encoding and back'''
     def __new__(cls, u, l):
         s = str.__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 = 'foo: \\xc3\\xa4' # utf-8
     >>> l = tolocal(u)
     >>> l
     'foo: ?'
     >>> fromlocal(l)
     'foo: \\xc3\\xa4'
     >>> u2 = 'foo: \\xc3\\xa1'
     >>> d = { l: 1, tolocal(u2): 2 }
     >>> len(d) # no collision
     2
     >>> 'foo: ?' in d
     False
     >>> l1 = 'foo: \\xe4' # historical latin1 fallback
     >>> l = tolocal(l1)
     >>> l
     'foo: ?'
     >>> fromlocal(l) # magically in utf-8
     'foo: \\xc3\\xa4'
     """
 
     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
 
     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.
 
     >>> ellipsis = '+++'
     >>> from . import encoding
     >>> encoding.encoding = 'utf-8'
     >>> t= '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(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 = '\x11\x22\x33\x44\x55\x66\x77\x88\x99\xaa' # 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 _asciilower(s):
-    '''convert a string to lowercase if ASCII
-
-    Raises UnicodeDecodeError if non-ASCII characters are found.'''
-    s.decode('ascii')
-    return s.lower()
-
-def asciilower(s):
-    # delay importing avoids cyclic dependency around "parsers" in
-    # pure Python build (util => i18n => encoding => parsers => util)
-    parsers = policy.importmod(r'parsers')
-    impl = getattr(parsers, 'asciilower', _asciilower)
-    global asciilower
-    asciilower = impl
-    return impl(s)
-
-def _asciiupper(s):
-    '''convert a string to uppercase if ASCII
-
-    Raises UnicodeDecodeError if non-ASCII characters are found.'''
-    s.decode('ascii')
-    return s.upper()
-
-def asciiupper(s):
-    # delay importing avoids cyclic dependency around "parsers" in
-    # pure Python build (util => i18n => encoding => parsers => util)
-    parsers = policy.importmod(r'parsers')
-    impl = getattr(parsers, 'asciiupper', _asciiupper)
-    global asciiupper
-    asciiupper = impl
-    return impl(s)
-
 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
 
 _jsonmap = []
 _jsonmap.extend("\\u%04x" % x for x in range(32))
 _jsonmap.extend(pycompat.bytechr(x) for x in range(32, 127))
 _jsonmap.append('\\u007f')
 _jsonmap[0x09] = '\\t'
 _jsonmap[0x0a] = '\\n'
 _jsonmap[0x22] = '\\"'
 _jsonmap[0x5c] = '\\\\'
 _jsonmap[0x08] = '\\b'
 _jsonmap[0x0c] = '\\f'
 _jsonmap[0x0d] = '\\r'
 _paranoidjsonmap = _jsonmap[:]
 _paranoidjsonmap[0x3c] = '\\u003c'  # '<' (e.g. escape "</script>")
 _paranoidjsonmap[0x3e] = '\\u003e'  # '>'
 _jsonmap.extend(pycompat.bytechr(x) for x in range(128, 256))
 
 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('this is a test')
     'this is a test'
     >>> jsonescape('escape characters: \\0 \\x0b \\x7f')
     'escape characters: \\\\u0000 \\\\u000b \\\\u007f'
     >>> jsonescape('escape characters: \\t \\n \\r \\" \\\\')
     'escape characters: \\\\t \\\\n \\\\r \\\\" \\\\\\\\'
     >>> jsonescape('a weird byte: \\xdd')
     'a weird byte: \\xed\\xb3\\x9d'
     >>> jsonescape('utf-8: caf\\xc3\\xa9')
     'utf-8: caf\\xc3\\xa9'
     >>> jsonescape('')
     ''
 
     If paranoid, non-ascii and common troublesome characters are also escaped.
     This is suitable for web output.
 
     >>> jsonescape('escape boundary: \\x7e \\x7f \\xc2\\x80', paranoid=True)
     'escape boundary: ~ \\\\u007f \\\\u0080'
     >>> jsonescape('a weird byte: \\xdd', paranoid=True)
     'a weird byte: \\\\udcdd'
     >>> jsonescape('utf-8: caf\\xc3\\xa9', paranoid=True)
     'utf-8: caf\\\\u00e9'
     >>> jsonescape('non-BMP: \\xf0\\x9d\\x84\\x9e', paranoid=True)
     'non-BMP: \\\\ud834\\\\udd1e'
     >>> jsonescape('<[e-mail]>', paranoid=True)
     '\\\\[e-mail]\\\\u003e'
     '''
 
     if paranoid:
         jm = _paranoidjsonmap
     else:
         jm = _jsonmap
 
     u8chars = toutf8b(s)
     try:
         return ''.join(jm[x] for x in bytearray(u8chars))  # fast path
     except IndexError:
         pass
     # non-BMP char is represented as UTF-16 surrogate pair
     u16codes = array.array('H', u8chars.decode('utf-8').encode('utf-16'))
     u16codes.pop(0)  # drop BOM
     return ''.join(jm[x] if x < 128 else '\\u%04x' % x for x in u16codes)
 
 _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 "\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 = "\\xc3\\xa9\\x99abcd"
     >>> toutf8b(m)
     '\\xc3\\xa9\\xed\\xb2\\x99abcd'
     >>> roundtrip(m)
     True
     >>> roundtrip("\\xc2\\xc2\\x80")
     True
     >>> roundtrip("\\xef\\xbf\\xbd")
     True
     >>> roundtrip("\\xef\\xef\\xbf\\xbd")
     True
     >>> roundtrip("\\xf1\\x80\\x80\\x80\\x80")
     True
     '''
 
     # 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

mercurial/policy.py

@@ -1,114 +1,116 b''
 # policy.py - module policy logic for Mercurial.
 #
 # Copyright 2015 Gregory Szorc <[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 os
 import sys
 
 # Rules for how modules can be loaded. Values are:
 #
 #    c - require C extensions
 #    allow - allow pure Python implementation when C loading fails
 #    cffi - required cffi versions (implemented within pure module)
 #    cffi-allow - allow pure Python implementation if cffi version is missing
 #    py - only load pure Python modules
 #
 # By default, fall back to the pure modules so the in-place build can
 # run without recompiling the C extensions. This will be overridden by
 # __modulepolicy__ generated by setup.py.
 policy = b'allow'
 _packageprefs = {
     # policy: (versioned package, pure package)
     b'c': (r'cext', None),
     b'allow': (r'cext', r'pure'),
     b'cffi': (r'cffi', None),
     b'cffi-allow': (r'cffi', r'pure'),
     b'py': (None, r'pure'),
 }
 
 try:
     from . import __modulepolicy__
     policy = __modulepolicy__.modulepolicy
 except ImportError:
     pass
 
 # PyPy doesn't load C extensions.
 #
 # The canonical way to do this is to test platform.python_implementation().
 # But we don't import platform and don't bloat for it here.
 if r'__pypy__' in sys.builtin_module_names:
     policy = b'cffi'
 
 # Our C extensions aren't yet compatible with Python 3. So use pure Python
 # on Python 3 for now.
 if sys.version_info[0] >= 3:
     policy = b'py'
 
 # Environment variable can always force settings.
 if sys.version_info[0] >= 3:
     if r'HGMODULEPOLICY' in os.environ:
         policy = os.environ[r'HGMODULEPOLICY'].encode(r'utf-8')
 else:
     policy = os.environ.get(r'HGMODULEPOLICY', policy)
 
 def _importfrom(pkgname, modname):
     # from .<pkgname> import <modname> (where . is looked through this module)
     fakelocals = {}
     pkg = __import__(pkgname, globals(), fakelocals, [modname], level=1)
     try:
         fakelocals[modname] = mod = getattr(pkg, modname)
     except AttributeError:
         raise ImportError(r'cannot import name %s' % modname)
     # force import; fakelocals[modname] may be replaced with the real module
     getattr(mod, r'__doc__', None)
     return fakelocals[modname]
 
 # keep in sync with "version" in C modules
 _cextversions = {
     (r'cext', r'base85'): 1,
     (r'cext', r'bdiff'): 1,
     (r'cext', r'diffhelpers'): 1,
     (r'cext', r'mpatch'): 1,
     (r'cext', r'osutil'): 1,
     (r'cext', r'parsers'): 1,
 }
 
 # map import request to other package or module
 _modredirects = {
+    (r'cext', r'charencode'): (r'cext', r'parsers'),
     (r'cffi', r'base85'): (r'pure', r'base85'),
+    (r'cffi', r'charencode'): (r'pure', r'charencode'),
     (r'cffi', r'diffhelpers'): (r'pure', r'diffhelpers'),
     (r'cffi', r'parsers'): (r'pure', r'parsers'),
 }
 
 def _checkmod(pkgname, modname, mod):
     expected = _cextversions.get((pkgname, modname))
     actual = getattr(mod, r'version', None)
     if actual != expected:
         raise ImportError(r'cannot import module %s.%s '
                           r'(expected version: %d, actual: %r)'
                           % (pkgname, modname, expected, actual))
 
 def importmod(modname):
     """Import module according to policy and check API version"""
     try:
         verpkg, purepkg = _packageprefs[policy]
     except KeyError:
         raise ImportError(r'invalid HGMODULEPOLICY %r' % policy)
     assert verpkg or purepkg
     if verpkg:
         pn, mn = _modredirects.get((verpkg, modname), (verpkg, modname))
         try:
             mod = _importfrom(pn, mn)
             if pn == verpkg:
                 _checkmod(pn, mn, mod)
             return mod
         except ImportError:
             if not purepkg:
                 raise
     pn, mn = _modredirects.get((purepkg, modname), (purepkg, modname))
     return _importfrom(pn, mn)