diff options
| author | Antoine Pitrou <solipsis@pitrou.net> | 2011-11-11 12:29:12 (GMT) |
|---|---|---|
| committer | Antoine Pitrou <solipsis@pitrou.net> | 2011-11-11 12:29:12 (GMT) |
| commit | 0290c7a811afee488f809bc8327485a55a3792cc (patch) | |
| tree | 62a02f27149cfaa9d9a5669424df0748547e3b77 | |
| parent | 28a08205c5112bde6585ec60bfdf48c8363c15b6 (diff) | |
| download | cpython-0290c7a811afee488f809bc8327485a55a3792cc.zip cpython-0290c7a811afee488f809bc8327485a55a3792cc.tar.gz cpython-0290c7a811afee488f809bc8327485a55a3792cc.tar.bz2 | |
Fix regression on 2-byte wchar_t systems (Windows)
| -rw-r--r-- | Objects/unicodeobject.c | 19 |
1 files changed, 12 insertions, 7 deletions
diff --git a/Objects/unicodeobject.c b/Objects/unicodeobject.c index a4d210b..6267dd3 100644 --- a/Objects/unicodeobject.c +++ b/Objects/unicodeobject.c @@ -6252,15 +6252,18 @@ _PyUnicode_DecodeUnicodeInternal(const char *s, end = s + size; while (s < end) { + Py_UNICODE uch; Py_UCS4 ch; /* We copy the raw representation one byte at a time because the pointer may be unaligned (see test_codeccallbacks). */ - ((char *) &ch)[0] = s[0]; - ((char *) &ch)[1] = s[1]; + ((char *) &uch)[0] = s[0]; + ((char *) &uch)[1] = s[1]; #ifdef Py_UNICODE_WIDE - ((char *) &ch)[2] = s[2]; - ((char *) &ch)[3] = s[3]; + ((char *) &uch)[2] = s[2]; + ((char *) &uch)[3] = s[3]; #endif + ch = uch; + /* We have to sanity check the raw data, otherwise doom looms for some malformed UCS-4 data. */ if ( @@ -6292,10 +6295,12 @@ _PyUnicode_DecodeUnicodeInternal(const char *s, #ifndef Py_UNICODE_WIDE if (ch >= 0xD800 && ch <= 0xDBFF && s < end) { - Py_UCS4 ch2 = *(Py_UNICODE*)s; - if (ch2 >= 0xDC00 && ch2 <= 0xDFFF) + Py_UNICODE uch2; + ((char *) &uch2)[0] = s[0]; + ((char *) &uch2)[1] = s[1]; + if (uch2 >= 0xDC00 && uch2 <= 0xDFFF) { - ch = (((ch & 0x3FF)<<10) | (ch2 & 0x3FF)) + 0x10000; + ch = (((uch & 0x3FF)<<10) | (uch2 & 0x3FF)) + 0x10000; s += Py_UNICODE_SIZE; } } |