21 _encodingmode = os.environ.get("HGENCODINGMODE", "strict")

    22 

    23 def tolocal(s):

    24     """

    25     Convert a string from internal UTF-8 to local encoding

    26 

    27     All internal strings should be UTF-8 but some repos before the

    28     implementation of locale support may contain latin1 or possibly

    29     other character sets. We attempt to decode everything strictly

    30     using UTF-8, then Latin-1, and failing that, we use UTF-8 and

    31     replace unknown characters.

    32     """

    33     for e in "utf-8 latin1".split():

    34         try:

    35             u = s.decode(e) # attempt strict decoding

    36             return u.encode(_encoding, "replace")

    37         except UnicodeDecodeError:

    38             pass

    39     u = s.decode("utf-8", "replace") # last ditch

    40     return u.encode(_encoding, "replace")

    41 

    42 def fromlocal(s):

    43     """

    44     Convert a string from the local character encoding to UTF-8

    45 

    46     We attempt to decode strings using the encoding mode set by

    47     HG_ENCODINGMODE, which defaults to 'strict'. In this mode, unknown

    48     characters will cause an error message. Other modes include

    49     'replace', which replaces unknown characters with a special

    50     Unicode character, and 'ignore', which drops the character.

    51     """

    52     try:

    53         return s.decode(_encoding, _encodingmode).encode("utf-8")

    54     except UnicodeDecodeError, inst:

    55         sub = s[max(0, inst.start-10):inst.start+10]

    56         raise Abort("decoding near '%s': %s!\n" % (sub, inst))

    57 

    58 def locallen(s):

    59     """Find the length in characters of a local string"""

    60     return len(s.decode(_encoding, "replace"))

    61 

    62 def localsub(s, a, b=None):

    63     try:

    64         u = s.decode(_encoding, _encodingmode)

    65         if b is not None:

    66             u = u[a:b]

    67         else:

    68             u = u[:a]

    69         return u.encode(_encoding, _encodingmode)

    70     except UnicodeDecodeError, inst:

    71         sub = s[max(0, inst.start-10), inst.start+10]

    72         raise Abort("decoding near '%s': %s!\n" % (sub, inst))