Mercurial > public > mercurial-scm > hg-stable
annotate mercurial/encoding.py @ 8226:8b2cd04a6e97
put license and copyright info into comment blocks
author | Martin Geisler <mg@lazybytes.net> |
---|---|
date | Sun, 26 Apr 2009 01:13:08 +0200 |
parents | 46293a0c7e9f |
children | b87a50b7125c |
rev | line source |
---|---|
8226
8b2cd04a6e97
put license and copyright info into comment blocks
Martin Geisler <mg@lazybytes.net>
parents:
8225
diff
changeset
|
1 # encoding.py - character transcoding support for Mercurial |
8b2cd04a6e97
put license and copyright info into comment blocks
Martin Geisler <mg@lazybytes.net>
parents:
8225
diff
changeset
|
2 # |
8b2cd04a6e97
put license and copyright info into comment blocks
Martin Geisler <mg@lazybytes.net>
parents:
8225
diff
changeset
|
3 # Copyright 2005-2009 Matt Mackall <mpm@selenic.com> and others |
8b2cd04a6e97
put license and copyright info into comment blocks
Martin Geisler <mg@lazybytes.net>
parents:
8225
diff
changeset
|
4 # |
8b2cd04a6e97
put license and copyright info into comment blocks
Martin Geisler <mg@lazybytes.net>
parents:
8225
diff
changeset
|
5 # This software may be used and distributed according to the terms of the |
8b2cd04a6e97
put license and copyright info into comment blocks
Martin Geisler <mg@lazybytes.net>
parents:
8225
diff
changeset
|
6 # GNU General Public License version 2, incorporated herein by reference. |
7948
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
7 |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
8 import sys, unicodedata, locale, os, error |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
9 |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
10 _encodingfixup = {'646': 'ascii', 'ANSI_X3.4-1968': 'ascii'} |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
11 |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
12 try: |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
13 encoding = os.environ.get("HGENCODING") |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
14 if sys.platform == 'darwin' and not encoding: |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
15 # On darwin, getpreferredencoding ignores the locale environment and |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
16 # always returns mac-roman. We override this if the environment is |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
17 # not C (has been customized by the user). |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
18 locale.setlocale(locale.LC_CTYPE, '') |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
19 encoding = locale.getlocale()[1] |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
20 if not encoding: |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
21 encoding = locale.getpreferredencoding() or 'ascii' |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
22 encoding = _encodingfixup.get(encoding, encoding) |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
23 except locale.Error: |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
24 encoding = 'ascii' |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
25 encodingmode = os.environ.get("HGENCODINGMODE", "strict") |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
26 fallbackencoding = 'ISO-8859-1' |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
27 |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
28 def tolocal(s): |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
29 """ |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
30 Convert a string from internal UTF-8 to local encoding |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
31 |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
32 All internal strings should be UTF-8 but some repos before the |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
33 implementation of locale support may contain latin1 or possibly |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
34 other character sets. We attempt to decode everything strictly |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
35 using UTF-8, then Latin-1, and failing that, we use UTF-8 and |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
36 replace unknown characters. |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
37 """ |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
38 for e in ('UTF-8', fallbackencoding): |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
39 try: |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
40 u = s.decode(e) # attempt strict decoding |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
41 return u.encode(encoding, "replace") |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
42 except LookupError, k: |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
43 raise error.Abort("%s, please check your locale settings" % k) |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
44 except UnicodeDecodeError: |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
45 pass |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
46 u = s.decode("utf-8", "replace") # last ditch |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
47 return u.encode(encoding, "replace") |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
48 |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
49 def fromlocal(s): |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
50 """ |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
51 Convert a string from the local character encoding to UTF-8 |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
52 |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
53 We attempt to decode strings using the encoding mode set by |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
54 HGENCODINGMODE, which defaults to 'strict'. In this mode, unknown |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
55 characters will cause an error message. Other modes include |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
56 'replace', which replaces unknown characters with a special |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
57 Unicode character, and 'ignore', which drops the character. |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
58 """ |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
59 try: |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
60 return s.decode(encoding, encodingmode).encode("utf-8") |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
61 except UnicodeDecodeError, inst: |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
62 sub = s[max(0, inst.start-10):inst.start+10] |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
63 raise error.Abort("decoding near '%s': %s!" % (sub, inst)) |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
64 except LookupError, k: |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
65 raise error.Abort("%s, please check your locale settings" % k) |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
66 |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
67 def colwidth(s): |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
68 "Find the column width of a UTF-8 string for display" |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
69 d = s.decode(encoding, 'replace') |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
70 if hasattr(unicodedata, 'east_asian_width'): |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
71 w = unicodedata.east_asian_width |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
72 return sum([w(c) in 'WF' and 2 or 1 for c in d]) |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
73 return len(d) |
de377b1a9a84
move encoding bits from util to encoding
Matt Mackall <mpm@selenic.com>
parents:
diff
changeset
|
74 |