Show More
@@ -1,79 +1,79 b'' | |||
|
1 | 1 | # charencode.py - miscellaneous character encoding |
|
2 | 2 | # |
|
3 | 3 | # Copyright 2005-2009 Matt Mackall <mpm@selenic.com> and others |
|
4 | 4 | # |
|
5 | 5 | # This software may be used and distributed according to the terms of the |
|
6 | 6 | # GNU General Public License version 2 or any later version. |
|
7 | 7 | |
|
8 | 8 | from __future__ import absolute_import |
|
9 | 9 | |
|
10 | 10 | import array |
|
11 | 11 | |
|
12 | 12 | from .. import ( |
|
13 | 13 | pycompat, |
|
14 | 14 | ) |
|
15 | 15 | |
|
16 | 16 | def isasciistr(s): |
|
17 | 17 | try: |
|
18 | 18 | s.decode('ascii') |
|
19 | 19 | return True |
|
20 | 20 | except UnicodeDecodeError: |
|
21 | 21 | return False |
|
22 | 22 | |
|
23 | 23 | def asciilower(s): |
|
24 | 24 | '''convert a string to lowercase if ASCII |
|
25 | 25 | |
|
26 | 26 | Raises UnicodeDecodeError if non-ASCII characters are found.''' |
|
27 | 27 | s.decode('ascii') |
|
28 | 28 | return s.lower() |
|
29 | 29 | |
|
30 | 30 | def asciiupper(s): |
|
31 | 31 | '''convert a string to uppercase if ASCII |
|
32 | 32 | |
|
33 | 33 | Raises UnicodeDecodeError if non-ASCII characters are found.''' |
|
34 | 34 | s.decode('ascii') |
|
35 | 35 | return s.upper() |
|
36 | 36 | |
|
37 | 37 | _jsonmap = [] |
|
38 | 38 | _jsonmap.extend("\\u%04x" % x for x in range(32)) |
|
39 | 39 | _jsonmap.extend(pycompat.bytechr(x) for x in range(32, 127)) |
|
40 | 40 | _jsonmap.append('\\u007f') |
|
41 | 41 | _jsonmap[0x09] = '\\t' |
|
42 | 42 | _jsonmap[0x0a] = '\\n' |
|
43 | 43 | _jsonmap[0x22] = '\\"' |
|
44 | 44 | _jsonmap[0x5c] = '\\\\' |
|
45 | 45 | _jsonmap[0x08] = '\\b' |
|
46 | 46 | _jsonmap[0x0c] = '\\f' |
|
47 | 47 | _jsonmap[0x0d] = '\\r' |
|
48 | 48 | _paranoidjsonmap = _jsonmap[:] |
|
49 | 49 | _paranoidjsonmap[0x3c] = '\\u003c' # '<' (e.g. escape "</script>") |
|
50 | 50 | _paranoidjsonmap[0x3e] = '\\u003e' # '>' |
|
51 | 51 | _jsonmap.extend(pycompat.bytechr(x) for x in range(128, 256)) |
|
52 | 52 | |
|
53 | 53 | def jsonescapeu8fast(u8chars, paranoid): |
|
54 | 54 | """Convert a UTF-8 byte string to JSON-escaped form (fast path) |
|
55 | 55 | |
|
56 | 56 | Raises ValueError if non-ASCII characters have to be escaped. |
|
57 | 57 | """ |
|
58 | 58 | if paranoid: |
|
59 | 59 | jm = _paranoidjsonmap |
|
60 | 60 | else: |
|
61 | 61 | jm = _jsonmap |
|
62 | 62 | try: |
|
63 | 63 | return ''.join(jm[x] for x in bytearray(u8chars)) |
|
64 | 64 | except IndexError: |
|
65 | 65 | raise ValueError |
|
66 | 66 | |
|
67 | 67 | def jsonescapeu8fallback(u8chars, paranoid): |
|
68 | 68 | """Convert a UTF-8 byte string to JSON-escaped form (slow path) |
|
69 | 69 | |
|
70 | 70 | Escapes all non-ASCII characters no matter if paranoid is False. |
|
71 | 71 | """ |
|
72 | 72 | if paranoid: |
|
73 | 73 | jm = _paranoidjsonmap |
|
74 | 74 | else: |
|
75 | 75 | jm = _jsonmap |
|
76 | 76 | # non-BMP char is represented as UTF-16 surrogate pair |
|
77 | u16codes = array.array('H', u8chars.decode('utf-8').encode('utf-16')) | |
|
77 | u16codes = array.array(r'H', u8chars.decode('utf-8').encode('utf-16')) | |
|
78 | 78 | u16codes.pop(0) # drop BOM |
|
79 | 79 | return ''.join(jm[x] if x < 128 else '\\u%04x' % x for x in u16codes) |
General Comments 0
You need to be logged in to leave comments.
Login now