JehanneOS/newlib
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

* strfuncs.cc (sys_cp_mbstowcs): Treat src as unsigned char *. Convert failure

Browse Source 
of f_mbtowc into a single malformed utf-16 value.
This commit is contained in:
Christopher Faylor 2009-05-31 03:59:38 +00:00
parent 025e4859af
commit e44b9069cd
2 changed files with 30 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
winsup/cygwin/ChangeLog
View File

@ -1,3 +1,8 @@
2009-05-30 Christopher Faylor <me+cygwin@cgf.cx>
* strfuncs.cc (sys_cp_mbstowcs): Treat src as unsigned char *. Convert
failure of f_mbtowc into a single malformed utf-16 value.
2009-05-30 Christopher Faylor <me+cygwin@cgf.cx> 2009-05-30 Christopher Faylor <me+cygwin@cgf.cx>
* cygwin/include/sys/termios.h: Make default erase character "^?". * cygwin/include/sys/termios.h: Make default erase character "^?".

32
winsup/cygwin/strfuncs.cc
View File

@ -523,7 +523,7 @@ sys_cp_mbstowcs (mbtowc_p f_mbtowc, char *charset, wchar_t *dst, size_t dlen,
const char *src, size_t nms) const char *src, size_t nms)
{ {
wchar_t *ptr = dst; wchar_t *ptr = dst;
char *pmbs = (char *) src; unsigned const char *pmbs = (unsigned const char *) src;
size_t count = 0; size_t count = 0;
size_t len = dlen; size_t len = dlen;
int bytes; int bytes;
@ -539,12 +539,13 @@ sys_cp_mbstowcs (mbtowc_p f_mbtowc, char *charset, wchar_t *dst, size_t dlen,
UTF-8 sequence, there must be at least two more bytes left, and the UTF-8 sequence, there must be at least two more bytes left, and the
next byte must be a valid UTF-8 start byte. If the charset isn't next byte must be a valid UTF-8 start byte. If the charset isn't
UTF-8 anyway, try to convert the following bytes as UTF-8 sequence. */ UTF-8 anyway, try to convert the following bytes as UTF-8 sequence. */
if (*pmbs == 0x0e && nms > 2 && *(unsigned char *) (pmbs + 1) >= 0xc2 if (*pmbs == 0x0e && nms > 2 && pmbs[1] >= 0xc2
&& *(unsigned char *) (pmbs + 1) <= 0xf4 && *charset != 'U'/*TF-8*/) && pmbs[1] <= 0xf4 && *charset != 'U'/*TF-8*/)
{ {
pmbs++; pmbs++;
--nms; --nms;
bytes = __utf8_mbtowc (_REENT, ptr, pmbs, nms, charset, &ps); bytes = __utf8_mbtowc (_REENT, ptr, (const char *) pmbs, nms,
charset, &ps);
if (bytes < 0) if (bytes < 0)
{ {
/* Invalid UTF-8 sequence? Treat the ASCII SO character as /* Invalid UTF-8 sequence? Treat the ASCII SO character as
@ -560,7 +561,7 @@ sys_cp_mbstowcs (mbtowc_p f_mbtowc, char *charset, wchar_t *dst, size_t dlen,
if (ps.__count == 4) /* First half of a surrogate. */ if (ps.__count == 4) /* First half of a surrogate. */
{ {
wchar_t *ptr2 = dst ? ptr + 1 : NULL; wchar_t *ptr2 = dst ? ptr + 1 : NULL;
int bytes2 = __utf8_mbtowc (_REENT, ptr2, pmbs + bytes, int bytes2 = __utf8_mbtowc (_REENT, ptr2, (const char *) pmbs + bytes,
nms - bytes, charset, &ps); nms - bytes, charset, &ps);
if (bytes2 < 0) if (bytes2 < 0)
break; break;
@ -571,8 +572,25 @@ sys_cp_mbstowcs (mbtowc_p f_mbtowc, char *charset, wchar_t *dst, size_t dlen,
--len; --len;
} }
} }
else else if ((bytes = f_mbtowc (_REENT, ptr, (const char *) pmbs, nms, charset, &ps)) < 0 && *pmbs > '\x80')
bytes = f_mbtowc (_REENT, ptr, pmbs, nms, charset, &ps); {
/* This should probably be handled in f_mbtowc which can operate
on sequences rather than individual characters.
The technique is based on a discussion here:
http://www.mail-archive.com/linux-utf8@nl.linux.org/msg00080.html
This is hardly perfect. Windows doesn't do anything sensical with
characters converted to this format and (currently) we don't convert
them back into their original single byte form. It does allow
processing of src to continue, however, which, since there is no
way to signal decoding errors, seems like the best we can do.
*/
*ptr = L'\xdc80' | *pmbs;
bytes = 1;
}
if (bytes > 0) if (bytes > 0)
{ {
pmbs += bytes; pmbs += bytes;