X-Git-Url: https://git.hcoop.net/bpt/emacs.git/blobdiff_plain/ae0d461554a9351a0d897ce0e60b47fc9670431e..2bfa3d3e1fb347ba76bddf77f3e288049635821d:/src/charset.c diff --git a/src/charset.c b/src/charset.c index fdb8eebde8..97d6346614 100644 --- a/src/charset.c +++ b/src/charset.c @@ -1,13 +1,15 @@ /* Basic character set support. - Copyright (C) 2001-2013 Free Software Foundation, Inc. - Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, - 2005, 2006, 2007, 2008, 2009, 2010, 2011 - National Institute of Advanced Industrial Science and Technology (AIST) - Registration Number H14PRO021 - Copyright (C) 2003, 2004 - National Institute of Advanced Industrial Science and Technology (AIST) - Registration Number H13PRO009 +Copyright (C) 2001-2014 Free Software Foundation, Inc. + +Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, + 2005, 2006, 2007, 2008, 2009, 2010, 2011 + National Institute of Advanced Industrial Science and Technology (AIST) + Registration Number H14PRO021 + +Copyright (C) 2003, 2004 + National Institute of Advanced Industrial Science and Technology (AIST) + Registration Number H13PRO009 This file is part of GNU Emacs. @@ -26,8 +28,7 @@ along with GNU Emacs. If not, see . */ #include -#define CHARSET_INLINE EXTERN_INLINE - +#include #include #include #include @@ -388,12 +389,12 @@ load_charset_map (struct charset *charset, struct charset_map_entries *entries, { if (ascii_compatible_p) { - if (! ASCII_BYTE_P (from_c)) + if (! ASCII_CHAR_P (from_c)) { if (from_c < nonascii_min_char) nonascii_min_char = from_c; } - else if (! ASCII_BYTE_P (to_c)) + else if (! ASCII_CHAR_P (to_c)) { nonascii_min_char = 0x80; } @@ -477,32 +478,39 @@ read_hex (FILE *fp, bool *eof, bool *overflow) `file-name-handler-alist' to avoid running any Lisp code. */ static void -load_charset_map_from_file (struct charset *charset, Lisp_Object mapfile, int control_flag) +load_charset_map_from_file (struct charset *charset, Lisp_Object mapfile, + int control_flag) { unsigned min_code = CHARSET_MIN_CODE (charset); unsigned max_code = CHARSET_MAX_CODE (charset); int fd; - FILE *fp; + FILE *fp = NULL; Lisp_Object suffixes; struct charset_map_entries *head, *entries; int n_entries; ptrdiff_t count; - USE_SAFE_ALLOCA; - suffixes = Fcons (build_string (".map"), - Fcons (build_string (".TXT"), Qnil)); + suffixes = list2 (build_string (".map"), build_string (".TXT")); - count = SPECPDL_INDEX (); - specbind (Qfile_name_handler_alist, Qnil); - fd = openp (Vcharset_map_path, mapfile, suffixes, NULL, Qnil); - unbind_to (count, Qnil); - if (fd < 0 - || ! (fp = fdopen (fd, "r"))) - error ("Failure in loading charset map: %s", SDATA (mapfile)); + dynwind_begin (); + record_unwind_protect_ptr (fclose_ptr_unwind, &fp); + { + dynwind_begin (); + specbind (Qfile_name_handler_alist, Qnil); + fd = openp (Vcharset_map_path, mapfile, suffixes, NULL, Qnil, false); + fp = fd < 0 ? 0 : fdopen (fd, "r"); + if (!fp) + { + int open_errno = errno; + emacs_close (fd); + report_file_errno ("Loading charset map", mapfile, open_errno); + } + dynwind_end (); + } - /* Use SAFE_ALLOCA instead of alloca, as `charset_map_entries' is - large (larger than MAX_ALLOCA). */ - head = SAFE_ALLOCA (sizeof *head); + /* Use record, as `charset_map_entries' is large (larger than + MAX_ALLOCA). */ + head = xmalloc (sizeof *head); entries = head; memset (entries, 0, sizeof (struct charset_map_entries)); @@ -531,9 +539,9 @@ load_charset_map_from_file (struct charset *charset, Lisp_Object mapfile, int co if (from < min_code || to > max_code || from > to || c > MAX_CHAR) continue; - if (n_entries > 0 && (n_entries % 0x10000) == 0) + if (n_entries == 0x10000) { - entries->next = SAFE_ALLOCA (sizeof *entries->next); + entries->next = xmalloc (sizeof *entries->next); entries = entries->next; memset (entries, 0, sizeof (struct charset_map_entries)); n_entries = 0; @@ -545,9 +553,10 @@ load_charset_map_from_file (struct charset *charset, Lisp_Object mapfile, int co n_entries++; } fclose (fp); + fp = NULL; load_charset_map (charset, head, n_entries, control_flag); - SAFE_FREE (); + dynwind_end (); } static void @@ -907,7 +916,7 @@ usage: (define-charset-internal ...) */) if (! charset.code_linear_p) { - charset.code_space_mask = xzalloc (256); + charset.code_space_mask = xzalloc_atomic (256); for (i = 0; i < 4; i++) for (j = charset.code_space[i * 4]; j <= charset.code_space[i * 4 + 1]; j++) @@ -1178,7 +1187,7 @@ usage: (define-charset-internal ...) */) charset.iso_final) = id; if (new_definition_p) Viso_2022_charset_list = nconc2 (Viso_2022_charset_list, - Fcons (make_number (id), Qnil)); + list1 (make_number (id))); if (ISO_CHARSET_TABLE (1, 0, 'J') == id) charset_jisx0201_roman = id; else if (ISO_CHARSET_TABLE (2, 0, '@') == id) @@ -1198,7 +1207,7 @@ usage: (define-charset-internal ...) */) emacs_mule_bytes[charset.emacs_mule_id] = charset.dimension + 2; if (new_definition_p) Vemacs_mule_charset_list = nconc2 (Vemacs_mule_charset_list, - Fcons (make_number (id), Qnil)); + list1 (make_number (id))); } if (new_definition_p) @@ -1206,7 +1215,7 @@ usage: (define-charset-internal ...) */) Vcharset_list = Fcons (args[charset_arg_name], Vcharset_list); if (charset.supplementary_p) Vcharset_ordered_list = nconc2 (Vcharset_ordered_list, - Fcons (make_number (id), Qnil)); + list1 (make_number (id))); else { Lisp_Object tail; @@ -1223,7 +1232,7 @@ usage: (define-charset-internal ...) */) Vcharset_ordered_list); else if (NILP (tail)) Vcharset_ordered_list = nconc2 (Vcharset_ordered_list, - Fcons (make_number (id), Qnil)); + list1 (make_number (id))); else { val = Fcons (XCAR (tail), XCDR (tail)); @@ -1515,7 +1524,7 @@ find_charsets_in_text (const unsigned char *ptr, ptrdiff_t nchars, if (!NILP (table)) c = translate_char (table, c); - if (ASCII_BYTE_P (c)) + if (ASCII_CHAR_P (c)) ASET (charsets, charset_ascii, Qt); else ASET (charsets, charset_eight_bit, Qt); @@ -1855,10 +1864,7 @@ DEFUN ("decode-char", Fdecode_char, Sdecode_char, 2, 3, 0, doc: /* Decode the pair of CHARSET and CODE-POINT into a character. Return nil if CODE-POINT is not valid in CHARSET. -CODE-POINT may be a cons (HIGHER-16-BIT-VALUE . LOWER-16-BIT-VALUE). - -Optional argument RESTRICTION specifies a way to map the pair of CCS -and CODE-POINT to a character. Currently not supported and just ignored. */) +CODE-POINT may be a cons (HIGHER-16-BIT-VALUE . LOWER-16-BIT-VALUE). */) (Lisp_Object charset, Lisp_Object code_point, Lisp_Object restriction) { int c, id; @@ -1875,10 +1881,7 @@ and CODE-POINT to a character. Currently not supported and just ignored. */) DEFUN ("encode-char", Fencode_char, Sencode_char, 2, 3, 0, doc: /* Encode the character CH into a code-point of CHARSET. -Return nil if CHARSET doesn't include CH. - -Optional argument RESTRICTION specifies a way to map CH to a -code-point in CCS. Currently not supported and just ignored. */) +Return nil if CHARSET doesn't include CH. */) (Lisp_Object ch, Lisp_Object charset, Lisp_Object restriction) { int c, id; @@ -2046,6 +2049,8 @@ CH in the charset. */) DEFUN ("char-charset", Fchar_charset, Schar_charset, 1, 2, 0, doc: /* Return the charset of highest priority that contains CH. +ASCII characters are an exception: for them, this function always +returns `ascii'. If optional 2nd arg RESTRICTION is non-nil, it is a list of charsets from which to find the charset. It may also be a coding system. In that case, find the charset from what supported by that coding system. */) @@ -2302,13 +2307,15 @@ init_charset (void) obscure problem (eg bug#6401), so better abort. */ fprintf (stderr, "Error: charsets directory not found:\n\ %s\n\ -Emacs will not function correctly without the character map files.\n\ +Emacs will not function correctly without the character map files.\n%s\ Please check your installation!\n", - SDATA (tempdir)); + SDATA (tempdir), + egetenv("EMACSDATA") ? "The EMACSDATA environment \ +variable is set, maybe it has the wrong value?\n" : ""); exit (1); } - Vcharset_map_path = Fcons (tempdir, Qnil); + Vcharset_map_path = list1 (tempdir); } @@ -2348,6 +2355,8 @@ static struct charset charset_table_init[180]; void syms_of_charset (void) { +#include "charset.x" + DEFSYM (Qcharsetp, "charsetp"); DEFSYM (Qascii, "ascii"); @@ -2381,33 +2390,9 @@ syms_of_charset (void) } charset_table = charset_table_init; - charset_table_size = sizeof charset_table_init / sizeof *charset_table_init; + charset_table_size = ARRAYELTS (charset_table_init); charset_table_used = 0; - defsubr (&Scharsetp); - defsubr (&Smap_charset_chars); - defsubr (&Sdefine_charset_internal); - defsubr (&Sdefine_charset_alias); - defsubr (&Scharset_plist); - defsubr (&Sset_charset_plist); - defsubr (&Sunify_charset); - defsubr (&Sget_unused_iso_final_char); - defsubr (&Sdeclare_equiv_charset); - defsubr (&Sfind_charset_region); - defsubr (&Sfind_charset_string); - defsubr (&Sdecode_char); - defsubr (&Sencode_char); - defsubr (&Ssplit_char); - defsubr (&Smake_char); - defsubr (&Schar_charset); - defsubr (&Scharset_after); - defsubr (&Siso_charset); - defsubr (&Sclear_charset_maps); - defsubr (&Scharset_priority_list); - defsubr (&Sset_charset_priority); - defsubr (&Scharset_id_internal); - defsubr (&Ssort_charsets); - DEFVAR_LISP ("charset-map-path", Vcharset_map_path, doc: /* List of directories to search for charset map files. */); Vcharset_map_path = Qnil; @@ -2427,19 +2412,19 @@ the value may be a list of mnemonics. */); Vcurrent_iso639_language = Qnil; charset_ascii - = define_charset_internal (Qascii, 1, "\x00\x7F\x00\x00\x00\x00", + = define_charset_internal (Qascii, 1, "\x00\x7F\0\0\0\0\0", 0, 127, 'B', -1, 0, 1, 0, 0); charset_iso_8859_1 - = define_charset_internal (Qiso_8859_1, 1, "\x00\xFF\x00\x00\x00\x00", + = define_charset_internal (Qiso_8859_1, 1, "\x00\xFF\0\0\0\0\0", 0, 255, -1, -1, -1, 1, 0, 0); charset_unicode - = define_charset_internal (Qunicode, 3, "\x00\xFF\x00\xFF\x00\x10", + = define_charset_internal (Qunicode, 3, "\x00\xFF\x00\xFF\x00\x10\0", 0, MAX_UNICODE_CHAR, -1, 0, -1, 1, 0, 0); charset_emacs - = define_charset_internal (Qemacs, 3, "\x00\xFF\x00\xFF\x00\x3F", + = define_charset_internal (Qemacs, 3, "\x00\xFF\x00\xFF\x00\x3F\0", 0, MAX_5_BYTE_CHAR, -1, 0, -1, 1, 1, 0); charset_eight_bit - = define_charset_internal (Qeight_bit, 1, "\x80\xFF\x00\x00\x00\x00", + = define_charset_internal (Qeight_bit, 1, "\x80\xFF\0\0\0\0\0", 128, 255, -1, 0, -1, 0, 1, MAX_5_BYTE_CHAR + 1); charset_unibyte = charset_iso_8859_1;