diff options
Diffstat (limited to 'gcc')
| -rw-r--r-- | gcc/ChangeLog | 35 | ||||
| -rw-r--r-- | gcc/c-opts.c | 4 | ||||
| -rw-r--r-- | gcc/c.opt | 5 | ||||
| -rw-r--r-- | gcc/cppcharset.c | 108 | ||||
| -rw-r--r-- | gcc/cppfiles.c | 55 | ||||
| -rw-r--r-- | gcc/cpphash.h | 9 | ||||
| -rw-r--r-- | gcc/cppinit.c | 10 | ||||
| -rw-r--r-- | gcc/cpplib.c | 4 | ||||
| -rw-r--r-- | gcc/doc/cppopts.texi | 10 | ||||
| -rw-r--r-- | gcc/testsuite/ChangeLog | 7 | ||||
| -rw-r--r-- | gcc/testsuite/gcc.c-torture/execute/wchar_t-1.c | 1 |
11 files changed, 130 insertions, 118 deletions
diff --git a/gcc/ChangeLog b/gcc/ChangeLog index 45b7679..740bc21 100644 --- a/gcc/ChangeLog +++ b/gcc/ChangeLog @@ -1,3 +1,23 @@ +2004-02-02 Eric Christopher <echristo@redhat.com> + Zack Weinberg <zack@codesourcery.com> + + * c-opts.c (c_common_handle_option): Add -finput-charset. + * c.opt: Ditto. + * cppcharset.c (one_iso88591_to_utf8): Remove. + (convert_iso88591_utf8): Ditto. + (conversion_tab): Remove 8859-1 converter. + (_cpp_input_to_utf8): Remove. + (_cpp_init_iconv_buffer): Ditto. + (_cpp_close_iconv_buffer): Ditto. + (_cpp_convert_input): New function. + (_cpp_default_encoding): Ditto. + * cpphash.h: Add/remove prototypes for above. + * cppfiles.c (read_file_guts): Use _cpp_convert_input. + * cppinit.c (cpp_create_reader): Use _cpp_default_encoding + for narrow execution and input character sets. + * cpplib.c (cpp_push_buffer): Delete uses of removed functions. + * doc/cppopts.texi: Document -finput-charset. + 2004-02-02 David Edelsohn <edelsohn@gnu.org> * rtlanal.c (refers_to_regno_p): Test regno, not inner_regno, @@ -86,7 +106,7 @@ * postreload.c (reload_combine, reload_combine_note_store, reload_combine_note_use, reload_cse_move2add, move2add_note_store): Likewise. * ra-colorize.c (combine, color_usable_p, get_free_reg, - calculate_dont_begin, calculate_dont_begin, colorize_one_web, + calculate_dont_begin, calculate_dont_begin, colorize_one_web, try_recolor_web, insert_coalesced_conflicts, check_colors, break_precolored_alias): Likewise. * ra-debug.c: Include regs.h @@ -210,7 +230,7 @@ constant. (write_node): Simplify comparisons against small constants before printing tests. - + 2004-01-31 Kazu Hirata <kazu@cs.umass.edu> * config/m32r/m32r.c (m32r_load_pic_register): Use GEN_INT @@ -435,8 +455,8 @@ (mark_constant_pool): Use new pool datastructures. 2004-01-30 Fariborz Jahanian <fjahanian@apple.com> - - * config/rs6000/rs6000.c (rs6000_emit_move): Remove #if 0. + + * config/rs6000/rs6000.c (rs6000_emit_move): Remove #if 0. Copy operands[1] to pseudo for simplify_gen_subreg. 2004-01-30 Kazu Hirata <kazu@cs.umass.edu> @@ -447,7 +467,7 @@ 2004-01-30 Andrew Pinski <pinskia@physics.uc.edu> * toplev.c: Include alloc-pool.h. - * Makefile.in (toplev.c): Update dependencies. + * Makefile.in (toplev.c): Update dependencies. 2004-01-30 Richard Kenner <kenner@vlsi1.ultra.nyu.edu> @@ -533,7 +553,7 @@ 2004-01-29 Devang Patel <dpatel@apple.com> * dwarf2out.c (gen_field_die): Do not equate decl number to die. - + 2004-01-28 Ian Lance Taylor <ian@wasabisystems.com> PR inline-asm/6162 @@ -682,7 +702,7 @@ target macros. 2004-01-28 Daniel Berlin <dberlin@dberlin.org> - + * timevar.c (timevar_print): Mention when checking is enabled. 2004-01-28 Giovanni Bajo <giovannibajo@gcc.gnu.org> @@ -3984,4 +4004,3 @@ * invoke.texi (-O1): Document change. See ChangeLog.10 for earlier changes. - diff --git a/gcc/c-opts.c b/gcc/c-opts.c index 2df0a7e..c89ba00 100644 --- a/gcc/c-opts.c +++ b/gcc/c-opts.c @@ -904,6 +904,10 @@ c_common_handle_option (size_t scode, const char *arg, int value) cpp_opts->wide_charset = arg; break; + case OPT_finput_charset_: + cpp_opts->input_charset = arg; + break; + case OPT_ftemplate_depth_: max_tinst_depth = value; break; @@ -482,6 +482,11 @@ fexec-charset= C ObjC C++ ObjC++ Joined RejectNegative -fexec-charset=<cset> Convert all strings and character constants to character set <cset> +finput-charset= +C ObjC C++ ObjC++ Joined RejectNegative +-finput-charset=<cset> Specify the default character set for source files. + + fexternal-templates C++ ObjC++ diff --git a/gcc/cppcharset.c b/gcc/cppcharset.c index 9fc180a..b73128d 100644 --- a/gcc/cppcharset.c +++ b/gcc/cppcharset.c @@ -446,31 +446,6 @@ one_utf16_to_utf8 (iconv_t bigend, const uchar **inbufp, size_t *inbytesleftp, return 0; } -/* The first 256 code points of ISO 8859.1 have the same numeric - values as the first 256 code points of Unicode, therefore the - incoming ISO 8859.1 character can be passed directly to - one_cppchar_to_utf8 (which expects a Unicode value). */ - -static int -one_iso88591_to_utf8 (iconv_t bigend ATTRIBUTE_UNUSED, const uchar **inbufp, - size_t *inbytesleftp, uchar **outbufp, size_t *outbytesleftp) -{ - const uchar *inbuf = *inbufp; - int rval; - - if (*inbytesleftp > 1) - return EINVAL; - - rval = one_cppchar_to_utf8 ((cppchar_t)*inbuf, outbufp, outbytesleftp); - if (rval) - return rval; - - *inbufp += 1; - *inbytesleftp -= 1; - - return 0; -} - /* Helper routine for the next few functions. The 'const' on one_conversion means that we promise not to modify what function is pointed to, which lets the inliner see through it. */ @@ -554,14 +529,6 @@ convert_utf32_utf8 (iconv_t cd, const uchar *from, size_t flen, return conversion_loop (one_utf32_to_utf8, cd, from, flen, to); } -static bool -convert_iso88591_utf8 (iconv_t cd, const uchar *from, size_t flen, - struct _cpp_strbuf *to) -{ - return conversion_loop (one_iso88591_to_utf8, cd, from, flen, to); -} - - /* Identity conversion, used when we have no alternative. */ static bool convert_no_conversion (iconv_t cd ATTRIBUTE_UNUSED, @@ -639,7 +606,6 @@ static const struct conversion conversion_tab[] = { { "UTF-32BE/UTF-8", convert_utf32_utf8, (iconv_t)1 }, { "UTF-16LE/UTF-8", convert_utf16_utf8, (iconv_t)0 }, { "UTF-16BE/UTF-8", convert_utf16_utf8, (iconv_t)1 }, - { "ISO-8859-1/UTF-8", convert_iso88591_utf8, (iconv_t)0 }, }; /* Subroutine of cpp_init_iconv: initialize and return a @@ -1388,44 +1354,58 @@ cpp_interpret_charconst (cpp_reader *pfile, const cpp_token *token, } uchar * -_cpp_input_to_utf8 (cpp_reader *pfile, const uchar *input, cppchar_t length) +_cpp_convert_input (cpp_reader *pfile, const char *input_charset, + uchar *input, size_t size, size_t len, off_t *st_size) { - struct _cpp_strbuf tbuf; - struct cset_converter cvt = pfile->buffer->input_cset_desc; + struct cset_converter input_cset; + struct _cpp_strbuf to; - tbuf.asize = MAX (OUTBUF_BLOCK_SIZE, length); - tbuf.text = xmalloc (tbuf.asize); - tbuf.len = 0; + input_cset = init_iconv_desc (pfile, SOURCE_CHARSET, input_charset); + if (input_cset.func == convert_no_conversion) + { + to.text = input; + to.asize = size; + to.len = len; + } + else + { + to.asize = MAX (65536, len); + to.text = xmalloc (to.asize); + to.len = 0; - if (!APPLY_CONVERSION (cvt, input, length, &tbuf)) - { - cpp_error (pfile, CPP_DL_ERROR, "converting input to source character set."); - return NULL; - } + if (!APPLY_CONVERSION (input_cset, input, len, &to)) + cpp_error (pfile, CPP_DL_ERROR, + "failure to convert %s to %s", + CPP_OPTION (pfile, input_charset), SOURCE_CHARSET); - if (length) - tbuf.text[tbuf.len] = '\n'; - else - tbuf.text[0] = '\n'; + free (input); + } - return tbuf.text; -} + /* Clean up the mess. */ + if (input_cset.func == convert_using_iconv) + iconv_close (input_cset.cd); - /* Check the input file format. At present assuming the input file - is in iso-8859-1 format. Convert this input character set to - source character set format (UTF-8). */ + /* Resize buffer if we allocated substantially too much, or if we + haven't enough space for the \n-terminator. */ + if (to.len + 4096 < to.asize || to.len >= to.asize) + to.text = xrealloc (to.text, to.len + 1); -void -_cpp_init_iconv_buffer (cpp_reader *pfile, const char *from) -{ - pfile->buffer->input_cset_desc = init_iconv_desc (pfile, SOURCE_CHARSET, - from); + to.text[to.len] = '\n'; + *st_size = to.len; + return to.text; } -void -_cpp_close_iconv_buffer (cpp_reader *pfile) +const char * +_cpp_default_encoding (void) { - if (HAVE_ICONV - && pfile->buffer->input_cset_desc.func == convert_using_iconv) - iconv_close (pfile->buffer->input_cset_desc.cd); + const char *current_encoding = NULL; + +#if defined (HAVE_LOCALE_H) && defined (HAVE_LANGINFO_CODESET) + setlocale (LC_CTYPE, ""); + current_encoding = nl_langinfo (CODESET); +#endif + if (current_encoding == NULL || *current_encoding == '\0') + current_encoding = SOURCE_CHARSET; + + return current_encoding; } diff --git a/gcc/cppfiles.c b/gcc/cppfiles.c index cd735f5..c886c89 100644 --- a/gcc/cppfiles.c +++ b/gcc/cppfiles.c @@ -312,7 +312,7 @@ pch_open_file (cpp_reader *pfile, _cpp_file *file, bool *invalid_pch) /* Try to open the path FILE->name appended to FILE->dir. This is where remap and PCH intercept the file lookup process. Return true - if the file was found, whether or not the open was successful. + if the file was found, whether or not the open was successful. Set *INVALID_PCH to true if a PCH file is found but wasn't valid. */ static bool @@ -398,10 +398,10 @@ _cpp_find_file (cpp_reader *pfile, const char *fname, cpp_dir *start_dir, bool f open_file_failed (pfile, file); if (invalid_pch) { - cpp_error (pfile, CPP_DL_ERROR, + cpp_error (pfile, CPP_DL_ERROR, "one or more PCH files were found, but they were invalid"); if (!cpp_get_options (pfile)->warn_invalid_pch) - cpp_error (pfile, CPP_DL_ERROR, + cpp_error (pfile, CPP_DL_ERROR, "use -Winvalid-pch for more information"); } break; @@ -457,7 +457,7 @@ read_file_guts (cpp_reader *pfile, _cpp_file *file) ssize_t size, total, count; uchar *buf; bool regular; - + if (S_ISBLK (file->st.st_mode)) { cpp_error (pfile, CPP_DL_ERROR, "%s is a block device", file->path); @@ -514,15 +514,8 @@ read_file_guts (cpp_reader *pfile, _cpp_file *file) cpp_error (pfile, CPP_DL_WARNING, "%s is shorter than expected", file->path); - /* Shrink buffer if we allocated substantially too much. */ - if (total + 4096 < size) - buf = xrealloc (buf, total + 1); - - /* The lexer requires that the buffer be \n-terminated. */ - buf[total] = '\n'; - - file->buffer = buf; - file->st.st_size = total; + file->buffer = _cpp_convert_input (pfile, CPP_OPTION (pfile, input_charset), + buf, size, total, &file->st.st_size); file->buffer_valid = true; return true; @@ -566,7 +559,7 @@ should_stack_file (cpp_reader *pfile, _cpp_file *file, bool import) if (file->once_only) return false; - /* We must mark the file once-only if #import now, before header + /* We must mark the file once-only if #import now, before header guard checks. Otherwise, undefining the header guard might cause the file to be re-stacked. */ if (import) @@ -1283,7 +1276,7 @@ struct pchf_data { This is used as an optimisation, it means we don't have to search the structure if we're processing a regular #include. */ bool have_once_only; - + struct pchf_entry { /* The size of this file. This is used to save running a MD5 checksum if the sizes don't match. */ @@ -1298,7 +1291,7 @@ struct pchf_data { static struct pchf_data *pchf; /* Data for pchf_addr. */ -struct pchf_adder_info +struct pchf_adder_info { cpp_reader *pfile; struct pchf_data *d; @@ -1322,11 +1315,11 @@ pchf_adder (void **slot, void *data) the PCH file shouldn't be written... */ if (f->dont_read || f->err_no) return 1; - + d->entries[count].once_only = f->once_only; d->have_once_only |= f->once_only; if (f->buffer_valid) - md5_buffer ((const char *)f->buffer, + md5_buffer ((const char *)f->buffer, f->st.st_size, d->entries[count].sum); else { @@ -1365,22 +1358,22 @@ _cpp_save_file_entries (cpp_reader *pfile, FILE *f) struct pchf_data *result; size_t result_size; struct pchf_adder_info pai; - + count = htab_elements (pfile->file_hash); - result_size = (sizeof (struct pchf_data) + result_size = (sizeof (struct pchf_data) + sizeof (struct pchf_entry) * (count - 1)); result = xcalloc (result_size, 1); - + result->count = 0; result->have_once_only = false; - + pai.pfile = pfile; pai.d = result; htab_traverse (pfile->file_hash, pchf_adder, &pai); result_size = (sizeof (struct pchf_data) + sizeof (struct pchf_entry) * (result->count - 1)); - + qsort (result->entries, result->count, sizeof (struct pchf_entry), pchf_save_compare); @@ -1393,11 +1386,11 @@ bool _cpp_read_file_entries (cpp_reader *pfile ATTRIBUTE_UNUSED, FILE *f) { struct pchf_data d; - + if (fread (&d, sizeof (struct pchf_data) - sizeof (struct pchf_entry), 1, f) != 1) return false; - + pchf = xmalloc (sizeof (struct pchf_data) + sizeof (struct pchf_entry) * (d.count - 1)); memcpy (pchf, &d, sizeof (struct pchf_data) - sizeof (struct pchf_entry)); @@ -1422,7 +1415,7 @@ struct pchf_compare_data /* Do we need to worry about entries that don't have ONCE_ONLY set? */ bool check_included; - + /* The file that we're searching for. */ _cpp_file *f; }; @@ -1435,15 +1428,15 @@ pchf_compare (const void *d_p, const void *e_p) const struct pchf_entry *e = (const struct pchf_entry *)e_p; struct pchf_compare_data *d = (struct pchf_compare_data *)d_p; int result; - + result = memcmp (&d->size, &e->size, sizeof (off_t)); if (result != 0) return result; - + if (! d->sum_computed) { _cpp_file *const f = d->f; - + md5_buffer ((const char *)f->buffer, f->st.st_size, d->sum); d->sum_computed = true; } @@ -1458,7 +1451,7 @@ pchf_compare (const void *d_p, const void *e_p) return 1; } -/* Check that F is not in a list read from a PCH file (if any). +/* Check that F is not in a list read from a PCH file (if any). Assumes that f->buffer_valid is true. Return TRUE if the file should not be read. */ @@ -1468,7 +1461,7 @@ check_file_against_entries (cpp_reader *pfile ATTRIBUTE_UNUSED, bool check_included) { struct pchf_compare_data d; - + if (pchf == NULL || (! check_included && ! pchf->have_once_only)) return false; diff --git a/gcc/cpphash.h b/gcc/cpphash.h index 4c044c1..6606746 100644 --- a/gcc/cpphash.h +++ b/gcc/cpphash.h @@ -563,9 +563,6 @@ extern void _cpp_init_internal_pragmas (cpp_reader *); extern void _cpp_do_file_change (cpp_reader *, enum lc_reason, const char *, unsigned int, unsigned int); extern void _cpp_pop_buffer (cpp_reader *); -extern uchar *_cpp_input_to_utf8 (cpp_reader *, const unsigned char *, cppchar_t); -extern void _cpp_init_iconv_buffer (cpp_reader *, const char *); -extern void _cpp_close_iconv_buffer (cpp_reader *); /* In cpptrad.c. */ extern bool _cpp_scan_out_logical_line (cpp_reader *, cpp_macro *); @@ -582,8 +579,12 @@ extern size_t _cpp_replacement_text_len (const cpp_macro *); extern cppchar_t _cpp_valid_ucn (cpp_reader *, const uchar **, const uchar *, int); extern void _cpp_destroy_iconv (cpp_reader *); -extern bool _cpp_interpret_string_notranslate (cpp_reader *, const cpp_string *, +extern bool _cpp_interpret_string_notranslate (cpp_reader *, + const cpp_string *, cpp_string *); +extern uchar *_cpp_convert_input (cpp_reader *, const char *, uchar *, + size_t, size_t, off_t *); +extern const char *_cpp_default_encoding (void); /* Utility routines and macros. */ #define DSC(str) (const uchar *)str, sizeof str - 1 diff --git a/gcc/cppinit.c b/gcc/cppinit.c index ab2331d..d6509fd 100644 --- a/gcc/cppinit.c +++ b/gcc/cppinit.c @@ -159,11 +159,11 @@ cpp_create_reader (enum c_lang lang, hash_table *table, CPP_OPTION (pfile, bytes_big_endian) = 1; /* does not matter */ /* Default to no charset conversion. */ - CPP_OPTION (pfile, narrow_charset) = 0; + CPP_OPTION (pfile, narrow_charset) = _cpp_default_encoding (); CPP_OPTION (pfile, wide_charset) = 0; - /* Default the input character set to iso-8859-1 for now. */ - CPP_OPTION (pfile, input_charset) = "ISO-8859-1"; + /* Default the input character set to UTF-8. */ + CPP_OPTION (pfile, input_charset) = _cpp_default_encoding (); /* A fake empty "directory" used as the starting point for files looked up without a search path. Name cannot be '/' because we @@ -579,7 +579,7 @@ read_original_directory (cpp_reader *pfile) debugdir[token->val.str.len - 4] = '\0'; pfile->cb.dir_change (pfile, debugdir); - } + } /* We want to process the fake line changes as regular changes, to get them output. */ @@ -591,7 +591,7 @@ read_original_directory (cpp_reader *pfile) /* This is called at the end of preprocessing. It pops the last buffer and writes dependency output, and returns the number of errors. - + Maybe it should also reset state, such that you could call cpp_start_read with a new filename to restart processing. */ int diff --git a/gcc/cpplib.c b/gcc/cpplib.c index 33b2de6..fa16859 100644 --- a/gcc/cpplib.c +++ b/gcc/cpplib.c @@ -1925,7 +1925,6 @@ cpp_push_buffer (cpp_reader *pfile, const uchar *buffer, size_t len, int from_stage3) { cpp_buffer *new = xobnew (&pfile->buffer_ob, cpp_buffer); - const char *input = CPP_OPTION (pfile, input_charset); /* Clears, amongst other things, if_stack and mi_cmacro. */ memset (new, 0, sizeof (cpp_buffer)); @@ -1937,7 +1936,6 @@ cpp_push_buffer (cpp_reader *pfile, const uchar *buffer, size_t len, new->need_line = true; pfile->buffer = new; - _cpp_init_iconv_buffer (pfile, input); return new; } @@ -1960,8 +1958,6 @@ _cpp_pop_buffer (cpp_reader *pfile) /* In case of a missing #endif. */ pfile->state.skipping = 0; - _cpp_close_iconv_buffer (pfile); - /* _cpp_do_file_change expects pfile->buffer to be the new one. */ pfile->buffer = buffer->prev; diff --git a/gcc/doc/cppopts.texi b/gcc/doc/cppopts.texi index 653e2e7..4649e6ec 100644 --- a/gcc/doc/cppopts.texi +++ b/gcc/doc/cppopts.texi @@ -511,6 +511,16 @@ corresponds to the width of @code{wchar_t}. As with by the system's @code{iconv} library routine; however, you will have problems with encodings that do not fit exactly in @code{wchar_t}. +@item -finput-charset=@var{charset} +@opindex finput-charset +Set the input character set, used for translation from the character +set of the input file to the source character set used by GCC. If the +locale does not specify, or GCC cannot get this information from the +locale, the default is UTF-8. This can be overriden by either the locale +or this command line option. Currently the command line option takes +precedence if there's a conflict. @var{charset} can be any encoding +supported by the system's @code{iconv} library routine. + @item -fworking-directory @opindex fworking-directory @opindex fno-working-directory diff --git a/gcc/testsuite/ChangeLog b/gcc/testsuite/ChangeLog index 56a7c0c..84a9692 100644 --- a/gcc/testsuite/ChangeLog +++ b/gcc/testsuite/ChangeLog @@ -1,3 +1,8 @@ +2004-02-02 Eric Christopher <echristo@redhat.com> + Zack Weinberg <zack@codesourcery.com> + + * gcc.c-torture/execute/wchar_t-1.c: Add -finput-charset. + 2004-02-02 Zack Weinberg <zack@codesourcery.com> * g++.dg/eh/forced1.C, g++.dg/eh/forced2.C, g++.dg/eh/forced3.C @@ -20556,5 +20561,3 @@ rlsruhe.de> correspond to c-torture 1.11. * New file. - - diff --git a/gcc/testsuite/gcc.c-torture/execute/wchar_t-1.c b/gcc/testsuite/gcc.c-torture/execute/wchar_t-1.c index 3efdcf5..7e90cc0 100644 --- a/gcc/testsuite/gcc.c-torture/execute/wchar_t-1.c +++ b/gcc/testsuite/gcc.c-torture/execute/wchar_t-1.c @@ -1,3 +1,4 @@ +/* { dg-options "-finput-charset=utf-8" } */ typedef __WCHAR_TYPE__ wchar_t; wchar_t x[] = L"Ä"; wchar_t y = L'Ä'; |