diff options
author | Matthew Hiller <hiller@gcc.gnu.org> | 2002-04-24 22:18:25 +0000 |
---|---|---|
committer | Matthew Hiller <hiller@gcc.gnu.org> | 2002-04-24 22:18:25 +0000 |
commit | 64cdc38350d29ed0e2d5fa0b4bdd074ec69135dd (patch) | |
tree | c295949959dc8ecb1998758d055af141fb272257 /gcc | |
parent | 67f0a6bf0132909f392c07fa5eab27b3e8c6e483 (diff) | |
download | gcc-64cdc38350d29ed0e2d5fa0b4bdd074ec69135dd.zip gcc-64cdc38350d29ed0e2d5fa0b4bdd074ec69135dd.tar.gz gcc-64cdc38350d29ed0e2d5fa0b4bdd074ec69135dd.tar.bz2 |
cpplex.c: Remove conditional #undef of MULTIBYTE_CHARS.
2002-04-24 Matt Hiller <hiller@redhat.com>
* cpplex.c: Remove conditional #undef of MULTIBYTE_CHARS.
* c-lex.c: Ditto.
* cpplex.c (skip_line_comment): Process comment one multibyte
character at a time rather than one char at a time, if
appropriate.
(parse_string): Process string one multibyte character at a time
rather than one char at a time, if appropriate.
* c-lex.c (lex_string): Lex and copy multibyte strings
appropriately.
* cpplib.h (cppchar_t): Change to unsigned.
From-SVN: r52737
Diffstat (limited to 'gcc')
-rw-r--r-- | gcc/c-lex.c | 21 | ||||
-rw-r--r-- | gcc/cpplex.c | 71 | ||||
-rw-r--r-- | gcc/cpplib.h | 6 |
3 files changed, 76 insertions, 22 deletions
diff --git a/gcc/c-lex.c b/gcc/c-lex.c index 8a76dd3..f2f8c8c 100644 --- a/gcc/c-lex.c +++ b/gcc/c-lex.c @@ -40,13 +40,6 @@ Software Foundation, 59 Temple Place - Suite 330, Boston, MA #include "splay-tree.h" #include "debug.h" -/* MULTIBYTE_CHARS support only works for native compilers. - ??? Ideally what we want is to model widechar support after - the current floating point support. */ -#ifdef CROSS_COMPILE -#undef MULTIBYTE_CHARS -#endif - #ifdef MULTIBYTE_CHARS #include "mbchar.h" #include <locale.h> @@ -1282,8 +1275,8 @@ lex_string (str, len, wide) c = cpp_parse_escape (parse_in, &p, limit, mask); } - /* Add this single character into the buffer either as a wchar_t - or as a single byte. */ + /* Add this single character into the buffer either as a wchar_t, + a multibyte sequence, or as a single byte. */ if (wide) { unsigned charwidth = TYPE_PRECISION (char_type_node); @@ -1304,6 +1297,16 @@ lex_string (str, len, wide) } q += WCHAR_BYTES; } +#ifdef MULTIBYTE_CHARS + else if (char_len > 1) + { + /* We're dealing with a multibyte character. */ + for ( ; char_len >0; --char_len) + { + *q++ = *(p - char_len); + } + } +#endif else { *q++ = c; diff --git a/gcc/cpplex.c b/gcc/cpplex.c index eed9ffd..6a44b86 100644 --- a/gcc/cpplex.c +++ b/gcc/cpplex.c @@ -25,13 +25,6 @@ Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ #include "cpplib.h" #include "cpphash.h" -/* MULTIBYTE_CHARS support only works for native compilers. - ??? Ideally what we want is to model widechar support after - the current floating point support. */ -#ifdef CROSS_COMPILE -#undef MULTIBYTE_CHARS -#endif - #ifdef MULTIBYTE_CHARS #include "mbchar.h" #include <locale.h> @@ -312,14 +305,39 @@ skip_line_comment (pfile) cpp_buffer *buffer = pfile->buffer; unsigned int orig_line = pfile->line; cppchar_t c; +#ifdef MULTIBYTE_CHARS + wchar_t wc; + int char_len; +#endif pfile->state.lexing_comment = 1; +#ifdef MULTIBYTE_CHARS + /* Reset multibyte conversion state. */ + (void) local_mbtowc (NULL, NULL, 0); +#endif do { if (buffer->cur == buffer->rlimit) goto at_eof; +#ifdef MULTIBYTE_CHARS + char_len = local_mbtowc (&wc, (const char *) buffer->cur, + buffer->rlimit - buffer->cur); + if (char_len == -1) + { + cpp_error (pfile, DL_WARNING, + "ignoring invalid multibyte character"); + char_len = 1; + c = *buffer->cur++; + } + else + { + buffer->cur += char_len; + c = wc; + } +#else c = *buffer->cur++; +#endif if (c == '?' || c == '\\') c = skip_escaped_newlines (pfile); } @@ -617,10 +635,18 @@ parse_string (pfile, token, terminator) unsigned char *dest, *limit; cppchar_t c; bool warned_nulls = false; +#ifdef MULTIBYTE_CHARS + wchar_t wc; + int char_len; +#endif dest = BUFF_FRONT (pfile->u_buff); limit = BUFF_LIMIT (pfile->u_buff); +#ifdef MULTIBYTE_CHARS + /* Reset multibyte conversion state. */ + (void) local_mbtowc (NULL, NULL, 0); +#endif for (;;) { /* We need room for another char, possibly the terminating NUL. */ @@ -632,8 +658,26 @@ parse_string (pfile, token, terminator) limit = BUFF_LIMIT (pfile->u_buff); } - /* Handle trigraphs, escaped newlines etc. */ +#ifdef MULTIBYTE_CHARS + char_len = local_mbtowc (&wc, (const char *) buffer->cur, + buffer->rlimit - buffer->cur); + if (char_len == -1) + { + cpp_error (pfile, DL_WARNING, + "ignoring invalid multibyte character"); + char_len = 1; + c = *buffer->cur++; + } + else + { + buffer->cur += char_len; + c = wc; + } +#else c = *buffer->cur++; +#endif + + /* Handle trigraphs, escaped newlines etc. */ if (c == '?' || c == '\\') c = skip_escaped_newlines (pfile); @@ -666,8 +710,15 @@ parse_string (pfile, token, terminator) "null character(s) preserved in literal"); } } - - *dest++ = c; +#ifdef MULTIBYTE_CHARS + if (char_len > 1) + { + for ( ; char_len > 0; --char_len) + *dest++ = (*buffer->cur - char_len); + } + else +#endif + *dest++ = c; } *dest = '\0'; diff --git a/gcc/cpplib.h b/gcc/cpplib.h index 359326e..765d65a 100644 --- a/gcc/cpplib.h +++ b/gcc/cpplib.h @@ -187,9 +187,9 @@ struct cpp_token } val; }; -/* A standalone character. We may want to make it unsigned for the - same reason we use unsigned char - to avoid signedness issues. */ -typedef int cppchar_t; +/* A standalone character. It is unsigned for the same reason we use + unsigned char - to avoid signedness issues. */ +typedef unsigned int cppchar_t; /* Values for opts.dump_macros. dump_only means inhibit output of the preprocessed text |