diff options
Diffstat (limited to 'Build/source/utils/texinfo/makeinfo/lang.c')
-rw-r--r-- | Build/source/utils/texinfo/makeinfo/lang.c | 1381 |
1 files changed, 1086 insertions, 295 deletions
diff --git a/Build/source/utils/texinfo/makeinfo/lang.c b/Build/source/utils/texinfo/makeinfo/lang.c index c72e8dbc631..a56cb617e57 100644 --- a/Build/source/utils/texinfo/makeinfo/lang.c +++ b/Build/source/utils/texinfo/makeinfo/lang.c @@ -1,13 +1,13 @@ /* lang.c -- language-dependent support. - $Id: lang.c,v 1.14 2004/11/22 23:57:33 karl Exp $ + $Id: lang.c,v 1.34 2007/12/03 01:38:43 karl Exp $ - Copyright (C) 1999, 2000, 2001, 2002, 2003, 2004 Free Software - Foundation, Inc. + Copyright (C) 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007 + Free Software Foundation, Inc. - This program is free software; you can redistribute it and/or modify + This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by - the Free Software Foundation; either version 2, or (at your option) - any later version. + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of @@ -15,8 +15,7 @@ GNU General Public License for more details. You should have received a copy of the GNU General Public License - along with this program; if not, write to the Free Software - Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. + along with this program. If not, see <http://www.gnu.org/licenses/>. Originally written by Karl Heinz Marbaise <kama@hippo.fido.de>. */ @@ -27,16 +26,21 @@ #include "makeinfo.h" #include "xml.h" +#include <assert.h> + /* Current document encoding. */ encoding_code_type document_encoding_code = no_encoding; /* Current language code; default is English. */ language_code_type language_code = en; +/* Language to use for translations that end up in the output. */ +char *document_language = "C"; + /* By default, unsupported encoding is an empty string. */ char *unknown_encoding = NULL; -static iso_map_type us_ascii_map [] = {{NULL, 0, 0}}; /* ASCII map is trivial */ +static iso_map_type asis_map [] = {{NULL, 0, 0}}; /* ASCII, etc. */ /* Translation table between HTML and ISO Codes. The last item is hopefully the Unicode. It might be possible that those Unicodes are @@ -74,79 +78,79 @@ static iso_map_type iso8859_1_map [] = { { "frac12", 0xBD, 0x00BD }, { "frac34", 0xBE, 0x00BE }, { "iquest", 0xBF, 0x00BF }, - { "Agrave", 0xC0, 0x00C0 }, - { "Aacute", 0xC1, 0x00C1 }, - { "Acirc", 0xC2, 0x00C2 }, - { "Atilde", 0xC3, 0x00C3 }, - { "Auml", 0xC4, 0x00C4 }, - { "Aring", 0xC5, 0x00C5 }, - { "AElig", 0xC6, 0x00C6 }, - { "Ccedil", 0xC7, 0x00C7 }, - { "Ccedil", 0xC7, 0x00C7 }, - { "Egrave", 0xC8, 0x00C8 }, - { "Eacute", 0xC9, 0x00C9 }, - { "Ecirc", 0xCA, 0x00CA }, - { "Euml", 0xCB, 0x00CB }, - { "Igrave", 0xCC, 0x00CC }, - { "Iacute", 0xCD, 0x00CD }, - { "Icirc", 0xCE, 0x00CE }, - { "Iuml", 0xCF, 0x00CF }, - { "ETH", 0xD0, 0x00D0 }, - { "Ntilde", 0xD1, 0x00D1 }, - { "Ograve", 0xD2, 0x00D2 }, - { "Oacute", 0xD3, 0x00D3 }, - { "Ocirc", 0xD4, 0x00D4 }, - { "Otilde", 0xD5, 0x00D5 }, - { "Ouml", 0xD6, 0x00D6 }, + { "Agrave", 0xC0, 0x00C0, "A" }, + { "Aacute", 0xC1, 0x00C1, "A" }, + { "Acirc", 0xC2, 0x00C2, "A" }, + { "Atilde", 0xC3, 0x00C3, "A" }, + { "Auml", 0xC4, 0x00C4, "A" }, + { "Aring", 0xC5, 0x00C5, "AA" }, + { "AElig", 0xC6, 0x00C6, "AE" }, + { "Ccedil", 0xC7, 0x00C7, "C" }, + { "Ccedil", 0xC7, 0x00C7, "C" }, + { "Egrave", 0xC8, 0x00C8, "E" }, + { "Eacute", 0xC9, 0x00C9, "E" }, + { "Ecirc", 0xCA, 0x00CA, "E" }, + { "Euml", 0xCB, 0x00CB, "E" }, + { "Igrave", 0xCC, 0x00CC, "I" }, + { "Iacute", 0xCD, 0x00CD, "I" }, + { "Icirc", 0xCE, 0x00CE, "I" }, + { "Iuml", 0xCF, 0x00CF, "I" }, + { "ETH", 0xD0, 0x00D0, "DH" }, + { "Ntilde", 0xD1, 0x00D1, "N" }, + { "Ograve", 0xD2, 0x00D2, "O" }, + { "Oacute", 0xD3, 0x00D3, "O" }, + { "Ocirc", 0xD4, 0x00D4, "O" }, + { "Otilde", 0xD5, 0x00D5, "O" }, + { "Ouml", 0xD6, 0x00D6, "O" }, { "times", 0xD7, 0x00D7 }, - { "Oslash", 0xD8, 0x00D8 }, - { "Ugrave", 0xD9, 0x00D9 }, - { "Uacute", 0xDA, 0x00DA }, - { "Ucirc", 0xDB, 0x00DB }, - { "Uuml", 0xDC, 0x00DC }, - { "Yacute", 0xDD, 0x00DD }, - { "THORN", 0xDE, 0x00DE }, - { "szlig", 0xDF, 0x00DF }, - { "agrave", 0xE0, 0x00E0 }, - { "aacute", 0xE1, 0x00E1 }, - { "acirc", 0xE2, 0x00E2 }, - { "atilde", 0xE3, 0x00E3 }, - { "auml", 0xE4, 0x00E4 }, - { "aring", 0xE5, 0x00E5 }, - { "aelig", 0xE6, 0x00E6 }, - { "ccedil", 0xE7, 0x00E7 }, - { "egrave", 0xE8, 0x00E8 }, - { "eacute", 0xE9, 0x00E9 }, - { "ecirc", 0xEA, 0x00EA }, - { "euml", 0xEB, 0x00EB }, - { "igrave", 0xEC, 0x00EC }, - { "iacute", 0xED, 0x00ED }, - { "icirc", 0xEE, 0x00EE }, - { "iuml", 0xEF, 0x00EF }, - { "eth", 0xF0, 0x00F0 }, - { "ntilde", 0xF1, 0x00F1 }, - { "ograve", 0xF2, 0x00F2 }, - { "oacute", 0xF3, 0x00F3 }, - { "ocirc", 0xF4, 0x00F4 }, - { "otilde", 0xF5, 0x00F5 }, - { "ouml", 0xF6, 0x00F6 }, + { "Oslash", 0xD8, 0x00D8, "OE" }, + { "Ugrave", 0xD9, 0x00D9, "U" }, + { "Uacute", 0xDA, 0x00DA, "U" }, + { "Ucirc", 0xDB, 0x00DB, "U" }, + { "Uuml", 0xDC, 0x00DC, "U" }, + { "Yacute", 0xDD, 0x00DD, "Y" }, + { "THORN", 0xDE, 0x00DE, "TH" }, + { "szlig", 0xDF, 0x00DF, "s" }, + { "agrave", 0xE0, 0x00E0, "a" }, + { "aacute", 0xE1, 0x00E1, "a" }, + { "acirc", 0xE2, 0x00E2, "a" }, + { "atilde", 0xE3, 0x00E3, "a" }, + { "auml", 0xE4, 0x00E4, "a" }, + { "aring", 0xE5, 0x00E5, "aa" }, + { "aelig", 0xE6, 0x00E6, "ae" }, + { "ccedil", 0xE7, 0x00E7, "c" }, + { "egrave", 0xE8, 0x00E8, "e" }, + { "eacute", 0xE9, 0x00E9, "e" }, + { "ecirc", 0xEA, 0x00EA, "e" }, + { "euml", 0xEB, 0x00EB, "e" }, + { "igrave", 0xEC, 0x00EC, "i" }, + { "iacute", 0xED, 0x00ED, "i" }, + { "icirc", 0xEE, 0x00EE, "i" }, + { "iuml", 0xEF, 0x00EF, "i" }, + { "eth", 0xF0, 0x00F0, "dh" }, + { "ntilde", 0xF1, 0x00F1, "n" }, + { "ograve", 0xF2, 0x00F2, "o"}, + { "oacute", 0xF3, 0x00F3, "o" }, + { "ocirc", 0xF4, 0x00F4, "o" }, + { "otilde", 0xF5, 0x00F5, "o" }, + { "ouml", 0xF6, 0x00F6, "o" }, { "divide", 0xF7, 0x00F7 }, - { "oslash", 0xF8, 0x00F8 }, - { "ugrave", 0xF9, 0x00F9 }, - { "uacute", 0xFA, 0x00FA }, - { "ucirc", 0xFB, 0x00FB }, - { "uuml", 0xFC, 0x00FC }, - { "yacute", 0xFD, 0x00FD }, - { "thorn", 0xFE, 0x00FE }, - { "yuml", 0xFF, 0x00FF }, + { "oslash", 0xF8, 0x00F8, "oe" }, + { "ugrave", 0xF9, 0x00F9, "u" }, + { "uacute", 0xFA, 0x00FA, "u" }, + { "ucirc", 0xFB, 0x00FB, "u" }, + { "uuml", 0xFC, 0x00FC, "u" }, + { "yacute", 0xFD, 0x00FD, "y" }, + { "thorn", 0xFE, 0x00FE, "th" }, + { "yuml", 0xFF, 0x00FF, "y" }, { NULL, 0, 0 } }; /* ISO 8859-15, also known as Latin 9, differs from Latin 1 in only a few positions. http://www.cs.tut.fi/~jkorpela/latin9.html has a good - explanation and listing, summarized here. The names are abbreviated - from the official Unicode names, to fit in a decent line length. + explanation and listing, summarized here. The names here are + abbreviated to fit in a decent line length. code position dec oct hex latin1 latin1 name latin9 latin9 name @@ -168,9 +172,9 @@ static iso_map_type iso8859_15_map [] = { { "pound", 0xA3, 0x00A3 }, { "euro", 0xA4, 0x20AC }, { "yen", 0xA5, 0x00A5 }, - { "Scaron", 0xA6, 0x0160 }, + { "Scaron", 0xA6, 0x0160, "S" }, { "sect", 0xA7, 0x00A7 }, - { "scaron", 0xA8, 0x0161 }, + { "scaron", 0xA8, 0x0161, "s" }, { "copy", 0xA9, 0x00A9 }, { "ordf", 0xAA, 0x00AA }, { "laquo", 0xAB, 0x00AB }, @@ -182,83 +186,82 @@ static iso_map_type iso8859_15_map [] = { { "plusmn", 0xB1, 0x00B1 }, { "sup2", 0xB2, 0x00B2 }, { "sup3", 0xB3, 0x00B3 }, - { "Zcaron", 0xB4, 0x017D }, + { "Zcaron", 0xB4, 0x017D, "Z" }, { "micro", 0xB5, 0x00B5 }, { "para", 0xB6, 0x00B6 }, { "middot", 0xB7, 0x00B7 }, - { "zcaron", 0xB8, 0x017E }, + { "zcaron", 0xB8, 0x017E, "z" }, { "sup1", 0xB9, 0x00B9 }, { "ordm", 0xBA, 0x00BA }, { "raquo", 0xBB, 0x00BB }, - { "OElig", 0xBC, 0x0152 }, - { "oelig", 0xBD, 0x0153 }, - { "Yuml", 0xBE, 0x0178 }, + { "OElig", 0xBC, 0x0152, "OE" }, + { "oelig", 0xBD, 0x0153, "oe" }, + { "Yuml", 0xBE, 0x0178, "y" }, { "iquest", 0xBF, 0x00BF }, - { "Agrave", 0xC0, 0x00C0 }, - { "Aacute", 0xC1, 0x00C1 }, - { "Acirc", 0xC2, 0x00C2 }, - { "Atilde", 0xC3, 0x00C3 }, - { "Auml", 0xC4, 0x00C4 }, - { "Aring", 0xC5, 0x00C5 }, - { "AElig", 0xC6, 0x00C6 }, - { "Ccedil", 0xC7, 0x00C7 }, - { "Ccedil", 0xC7, 0x00C7 }, - { "Egrave", 0xC8, 0x00C8 }, - { "Eacute", 0xC9, 0x00C9 }, - { "Ecirc", 0xCA, 0x00CA }, - { "Euml", 0xCB, 0x00CB }, - { "Igrave", 0xCC, 0x00CC }, - { "Iacute", 0xCD, 0x00CD }, - { "Icirc", 0xCE, 0x00CE }, - { "Iuml", 0xCF, 0x00CF }, - { "ETH", 0xD0, 0x00D0 }, - { "Ntilde", 0xD1, 0x00D1 }, - { "Ograve", 0xD2, 0x00D2 }, - { "Oacute", 0xD3, 0x00D3 }, - { "Ocirc", 0xD4, 0x00D4 }, - { "Otilde", 0xD5, 0x00D5 }, - { "Ouml", 0xD6, 0x00D6 }, + { "Agrave", 0xC0, 0x00C0, "A" }, + { "Aacute", 0xC1, 0x00C1, "A" }, + { "Acirc", 0xC2, 0x00C2, "A" }, + { "Atilde", 0xC3, 0x00C3, "A" }, + { "Auml", 0xC4, 0x00C4, "A" }, + { "Aring", 0xC5, 0x00C5, "AA" }, + { "AElig", 0xC6, 0x00C6, "AE" }, + { "Ccedil", 0xC7, 0x00C7, "C" }, + { "Egrave", 0xC8, 0x00C8, "E" }, + { "Eacute", 0xC9, 0x00C9, "E" }, + { "Ecirc", 0xCA, 0x00CA, "E" }, + { "Euml", 0xCB, 0x00CB, "E" }, + { "Igrave", 0xCC, 0x00CC, "I" }, + { "Iacute", 0xCD, 0x00CD, "I" }, + { "Icirc", 0xCE, 0x00CE, "I" }, + { "Iuml", 0xCF, 0x00CF, "I" }, + { "ETH", 0xD0, 0x00D0, "DH" }, + { "Ntilde", 0xD1, 0x00D1, "N" }, + { "Ograve", 0xD2, 0x00D2, "O" }, + { "Oacute", 0xD3, 0x00D3, "O" }, + { "Ocirc", 0xD4, 0x00D4, "O" }, + { "Otilde", 0xD5, 0x00D5, "O" }, + { "Ouml", 0xD6, 0x00D6, "O" }, { "times", 0xD7, 0x00D7 }, - { "Oslash", 0xD8, 0x00D8 }, - { "Ugrave", 0xD9, 0x00D9 }, - { "Uacute", 0xDA, 0x00DA }, - { "Ucirc", 0xDB, 0x00DB }, - { "Uuml", 0xDC, 0x00DC }, - { "Yacute", 0xDD, 0x00DD }, - { "THORN", 0xDE, 0x00DE }, - { "szlig", 0xDF, 0x00DF }, - { "agrave", 0xE0, 0x00E0 }, - { "aacute", 0xE1, 0x00E1 }, - { "acirc", 0xE2, 0x00E2 }, - { "atilde", 0xE3, 0x00E3 }, - { "auml", 0xE4, 0x00E4 }, - { "aring", 0xE5, 0x00E5 }, - { "aelig", 0xE6, 0x00E6 }, - { "ccedil", 0xE7, 0x00E7 }, - { "egrave", 0xE8, 0x00E8 }, - { "eacute", 0xE9, 0x00E9 }, - { "ecirc", 0xEA, 0x00EA }, - { "euml", 0xEB, 0x00EB }, - { "igrave", 0xEC, 0x00EC }, - { "iacute", 0xED, 0x00ED }, - { "icirc", 0xEE, 0x00EE }, - { "iuml", 0xEF, 0x00EF }, - { "eth", 0xF0, 0x00F0 }, - { "ntilde", 0xF1, 0x00F1 }, - { "ograve", 0xF2, 0x00F2 }, - { "oacute", 0xF3, 0x00F3 }, - { "ocirc", 0xF4, 0x00F4 }, - { "otilde", 0xF5, 0x00F5 }, - { "ouml", 0xF6, 0x00F6 }, + { "Oslash", 0xD8, 0x00D8, "OE" }, + { "Ugrave", 0xD9, 0x00D9, "U" }, + { "Uacute", 0xDA, 0x00DA, "U" }, + { "Ucirc", 0xDB, 0x00DB, "U" }, + { "Uuml", 0xDC, 0x00DC, "U" }, + { "Yacute", 0xDD, 0x00DD, "Y" }, + { "THORN", 0xDE, 0x00DE, "TH" }, + { "szlig", 0xDF, 0x00DF, "s" }, + { "agrave", 0xE0, 0x00E0, "a" }, + { "aacute", 0xE1, 0x00E1, "a" }, + { "acirc", 0xE2, 0x00E2, "a" }, + { "atilde", 0xE3, 0x00E3, "a" }, + { "auml", 0xE4, 0x00E4, "a" }, + { "aring", 0xE5, 0x00E5, "aa" }, + { "aelig", 0xE6, 0x00E6, "ae" }, + { "ccedil", 0xE7, 0x00E7, "c" }, + { "egrave", 0xE8, 0x00E8, "e" }, + { "eacute", 0xE9, 0x00E9, "e" }, + { "ecirc", 0xEA, 0x00EA, "e" }, + { "euml", 0xEB, 0x00EB, "e" }, + { "igrave", 0xEC, 0x00EC, "i" }, + { "iacute", 0xED, 0x00ED, "i" }, + { "icirc", 0xEE, 0x00EE, "i" }, + { "iuml", 0xEF, 0x00EF, "i" }, + { "eth", 0xF0, 0x00F0, "d" }, + { "ntilde", 0xF1, 0x00F1, "n" }, + { "ograve", 0xF2, 0x00F2, "o" }, + { "oacute", 0xF3, 0x00F3, "o" }, + { "ocirc", 0xF4, 0x00F4, "o" }, + { "otilde", 0xF5, 0x00F5, "o" }, + { "ouml", 0xF6, 0x00F6, "o" }, { "divide", 0xF7, 0x00F7 }, - { "oslash", 0xF8, 0x00F8 }, - { "ugrave", 0xF9, 0x00F9 }, - { "uacute", 0xFA, 0x00FA }, - { "ucirc", 0xFB, 0x00FB }, - { "uuml", 0xFC, 0x00FC }, - { "yacute", 0xFD, 0x00FD }, - { "thorn", 0xFE, 0x00FE }, - { "yuml", 0xFF, 0x00FF }, + { "oslash", 0xF8, 0x00F8, "oe" }, + { "ugrave", 0xF9, 0x00F9, "u" }, + { "uacute", 0xFA, 0x00FA, "u" }, + { "ucirc", 0xFB, 0x00FB, "u" }, + { "uuml", 0xFC, 0x00FC, "u" }, + { "yacute", 0xFD, 0x00FD, "y" }, + { "thorn", 0xFE, 0x00FE, "th" }, + { "yuml", 0xFF, 0x00FF, "y" }, { NULL, 0, 0 } }; @@ -277,7 +280,7 @@ static iso_map_type iso8859_15_map [] = { including the character table!!! (must see!) * ISO 8859-2 and even HTML entities !!! (must see!) - http://people.ssh.fi/mtr/genscript/88592.txt + 88592.txt in the GNU enscript distribution * (minor) http://www.agh.edu.pl/ogonki/plchars.html One more table, this time it includes even information about Polish @@ -286,107 +289,186 @@ static iso_map_type iso8859_15_map [] = { static iso_map_type iso8859_2_map [] = { { "nbsp", 0xA0, 0x00A0 }, /* NO-BREAK SPACE */ - { "", 0xA1, 0x0104 }, /* LATIN CAPITAL LETTER A WITH OGONEK */ + { "", 0xA1, 0x0104, "A" }, /* LATIN CAPITAL LETTER A WITH OGONEK */ { "", 0xA2, 0x02D8 }, /* BREVE */ - { "", 0xA3, 0x0141 }, /* LATIN CAPITAL LETTER L WITH STROKE */ + { "", 0xA3, 0x0141, "L" }, /* LATIN CAPITAL LETTER L WITH STROKE */ { "curren", 0xA4, 0x00A4 }, /* CURRENCY SIGN */ - { "", 0xA5, 0x013D }, /* LATIN CAPITAL LETTER L WITH CARON */ - { "", 0xA6, 0x015A }, /* LATIN CAPITAL LETTER S WITH ACUTE */ + { "", 0xA5, 0x013D, "L" }, /* LATIN CAPITAL LETTER L WITH CARON */ + { "", 0xA6, 0x015A, "S" }, /* LATIN CAPITAL LETTER S WITH ACUTE */ { "sect", 0xA7, 0x00A7 }, /* SECTION SIGN */ { "uml", 0xA8, 0x00A8 }, /* DIAERESIS */ - { "", 0xA9, 0x0160 }, /* LATIN CAPITAL LETTER S WITH CARON */ - { "", 0xAA, 0x015E }, /* LATIN CAPITAL LETTER S WITH CEDILLA */ - { "", 0xAB, 0x0164 }, /* LATIN CAPITAL LETTER T WITH CARON */ - { "", 0xAC, 0x0179 }, /* LATIN CAPITAL LETTER Z WITH ACUTE */ + { "#xa9", 0xA9, 0x0160, "S" }, /* LATIN CAPITAL LETTER S WITH CARON */ + { "", 0xAA, 0x015E, "S" }, /* LATIN CAPITAL LETTER S WITH CEDILLA */ + { "", 0xAB, 0x0164, "T" }, /* LATIN CAPITAL LETTER T WITH CARON */ + { "", 0xAC, 0x0179, "Z" }, /* LATIN CAPITAL LETTER Z WITH ACUTE */ { "shy", 0xAD, 0x00AD }, /* SOFT HYPHEN */ - { "", 0xAE, 0x017D }, /* LATIN CAPITAL LETTER Z WITH CARON */ - { "", 0xAF, 0x017B }, /* LATIN CAPITAL LETTER Z WITH DOT ABOVE */ + { "", 0xAE, 0x017D, "Z" }, /* LATIN CAPITAL LETTER Z WITH CARON */ + { "", 0xAF, 0x017B, "Z" }, /* LATIN CAPITAL LETTER Z WITH DOT ABOVE */ { "deg", 0xB0, 0x00B0 }, /* DEGREE SIGN */ - { "", 0xB1, 0x0105 }, /* LATIN SMALL LETTER A WITH OGONEK */ + { "", 0xB1, 0x0105, "a" }, /* LATIN SMALL LETTER A WITH OGONEK */ { "", 0xB2, 0x02DB }, /* OGONEK */ - { "", 0xB3, 0x0142 }, /* LATIN SMALL LETTER L WITH STROKE */ + { "", 0xB3, 0x0142, "l" }, /* LATIN SMALL LETTER L WITH STROKE */ { "acute", 0xB4, 0x00B4 }, /* ACUTE ACCENT */ - { "", 0xB5, 0x013E }, /* LATIN SMALL LETTER L WITH CARON */ - { "", 0xB6, 0x015B }, /* LATIN SMALL LETTER S WITH ACUTE */ + { "", 0xB5, 0x013E, "l" }, /* LATIN SMALL LETTER L WITH CARON */ + { "", 0xB6, 0x015B, "s" }, /* LATIN SMALL LETTER S WITH ACUTE */ { "", 0xB7, 0x02C7 }, /* CARON (Mandarin Chinese third tone) */ { "cedil", 0xB8, 0x00B8 }, /* CEDILLA */ - { "", 0xB9, 0x0161 }, /* LATIN SMALL LETTER S WITH CARON */ - { "", 0xBA, 0x015F }, /* LATIN SMALL LETTER S WITH CEDILLA */ - { "", 0xBB, 0x0165 }, /* LATIN SMALL LETTER T WITH CARON */ - { "", 0xBC, 0x017A }, /* LATIN SMALL LETTER Z WITH ACUTE */ + { "", 0xB9, 0x0161, "s" }, /* LATIN SMALL LETTER S WITH CARON */ + { "", 0xBA, 0x015F, "s" }, /* LATIN SMALL LETTER S WITH CEDILLA */ + { "", 0xBB, 0x0165, "t" }, /* LATIN SMALL LETTER T WITH CARON */ + { "", 0xBC, 0x017A, "z" }, /* LATIN SMALL LETTER Z WITH ACUTE */ { "", 0xBD, 0x02DD }, /* DOUBLE ACUTE ACCENT */ - { "", 0xBE, 0x017E }, /* LATIN SMALL LETTER Z WITH CARON */ - { "", 0xBF, 0x017C }, /* LATIN SMALL LETTER Z WITH DOT ABOVE */ - { "", 0xC0, 0x0154 }, /* LATIN CAPITAL LETTER R WITH ACUTE */ - { "", 0xC1, 0x00C1 }, /* LATIN CAPITAL LETTER A WITH ACUTE */ - { "", 0xC2, 0x00C2 }, /* LATIN CAPITAL LETTER A WITH CIRCUMFLEX */ - { "", 0xC3, 0x0102 }, /* LATIN CAPITAL LETTER A WITH BREVE */ - { "", 0xC4, 0x00C4 }, /* LATIN CAPITAL LETTER A WITH DIAERESIS */ - { "", 0xC5, 0x0139 }, /* LATIN CAPITAL LETTER L WITH ACUTE */ - { "", 0xC6, 0x0106 }, /* LATIN CAPITAL LETTER C WITH ACUTE */ - { "", 0xC7, 0x00C7 }, /* LATIN CAPITAL LETTER C WITH CEDILLA */ - { "", 0xC8, 0x010C }, /* LATIN CAPITAL LETTER C WITH CARON */ - { "", 0xC9, 0x00C9 }, /* LATIN CAPITAL LETTER E WITH ACUTE */ - { "", 0xCA, 0x0118 }, /* LATIN CAPITAL LETTER E WITH OGONEK */ - { "", 0xCB, 0x00CB }, /* LATIN CAPITAL LETTER E WITH DIAERESIS */ - { "", 0xCC, 0x011A }, /* LATIN CAPITAL LETTER E WITH CARON */ - { "", 0xCD, 0x00CD }, /* LATIN CAPITAL LETTER I WITH ACUTE */ - { "", 0xCE, 0x00CE }, /* LATIN CAPITAL LETTER I WITH CIRCUMFLEX */ - { "", 0xCF, 0x010E }, /* LATIN CAPITAL LETTER D WITH CARON */ - { "", 0xD0, 0x0110 }, /* LATIN CAPITAL LETTER D WITH STROKE */ - { "", 0xD1, 0x0143 }, /* LATIN CAPITAL LETTER N WITH ACUTE */ - { "", 0xD2, 0x0147 }, /* LATIN CAPITAL LETTER N WITH CARON */ - { "", 0xD3, 0x00D3 }, /* LATIN CAPITAL LETTER O WITH ACUTE */ - { "", 0xD4, 0x00D4 }, /* LATIN CAPITAL LETTER O WITH CIRCUMFLEX */ - { "", 0xD5, 0x0150 }, /* LATIN CAPITAL LETTER O WITH DOUBLE ACUTE */ - { "", 0xD6, 0x00D6 }, /* LATIN CAPITAL LETTER O WITH DIAERESIS */ + { "", 0xBE, 0x017E, "z" }, /* LATIN SMALL LETTER Z WITH CARON */ + { "", 0xBF, 0x017C, "z" }, /* LATIN SMALL LETTER Z WITH DOT ABOVE */ + { "", 0xC0, 0x0154, "R" }, /* LATIN CAPITAL LETTER R WITH ACUTE */ + { "Aacute", 0xC1, 0x00C1, "A" }, /* LATIN CAPITAL LETTER A WITH ACUTE */ + { "Acirc", 0xC2, 0x00C2, "A" }, /* LATIN CAPITAL LETTER A WITH CIRCUMFLEX */ + { "", 0xC3, 0x0102, "A" }, /* LATIN CAPITAL LETTER A WITH BREVE */ + { "Auml", 0xC4, 0x00C4, "A" }, /* LATIN CAPITAL LETTER A WITH DIAERESIS */ + { "", 0xC5, 0x0139, "L" }, /* LATIN CAPITAL LETTER L WITH ACUTE */ + { "", 0xC6, 0x0106, "C" }, /* LATIN CAPITAL LETTER C WITH ACUTE */ + { "Ccedil", 0xC7, 0x00C7, "C" }, /* LATIN CAPITAL LETTER C WITH CEDILLA */ + { "", 0xC8, 0x010C, "C" }, /* LATIN CAPITAL LETTER C WITH CARON */ + { "Eacute", 0xC9, 0x00C9, "E" }, /* LATIN CAPITAL LETTER E WITH ACUTE */ + { "", 0xCA, 0x0118, "E" }, /* LATIN CAPITAL LETTER E WITH OGONEK */ + { "Euml", 0xCB, 0x00CB, "E" }, /* LATIN CAPITAL LETTER E WITH DIAERESIS */ + { "", 0xCC, 0x011A, "E" }, /* LATIN CAPITAL LETTER E WITH CARON */ + { "Iacute", 0xCD, 0x00CD, "I" }, /* LATIN CAPITAL LETTER I WITH ACUTE */ + { "Icirc", 0xCE, 0x00CE, "I" }, /* LATIN CAPITAL LETTER I WITH CIRCUMFLEX */ + { "", 0xCF, 0x010E, "D" }, /* LATIN CAPITAL LETTER D WITH CARON */ + { "ETH", 0xD0, 0x0110, "D" }, /* LATIN CAPITAL LETTER D WITH STROKE */ + { "", 0xD1, 0x0143, "N" }, /* LATIN CAPITAL LETTER N WITH ACUTE */ + { "", 0xD2, 0x0147, "N" }, /* LATIN CAPITAL LETTER N WITH CARON */ + { "Oacute", 0xD3, 0x00D3, "O" }, /* LATIN CAPITAL LETTER O WITH ACUTE */ + { "Ocirc", 0xD4, 0x00D4, "O" }, /* LATIN CAPITAL LETTER O WITH CIRCUMFLEX */ + { "", 0xD5, 0x0150, "O" }, /* LATIN CAPITAL LETTER O WITH DOUBLE ACUTE */ + { "Ouml", 0xD6, 0x00D6, "O" }, /* LATIN CAPITAL LETTER O WITH DIAERESIS */ { "times", 0xD7, 0x00D7 }, /* MULTIPLICATION SIGN */ - { "", 0xD8, 0x0158 }, /* LATIN CAPITAL LETTER R WITH CARON */ - { "", 0xD9, 0x016E }, /* LATIN CAPITAL LETTER U WITH RING ABOVE */ - { "", 0xDA, 0x00DA }, /* LATIN CAPITAL LETTER U WITH ACUTE */ - { "", 0xDB, 0x0170 }, /* LATIN CAPITAL LETTER U WITH DOUBLE ACUTE */ - { "", 0xDC, 0x00DC }, /* LATIN CAPITAL LETTER U WITH DIAERESIS */ - { "", 0xDD, 0x00DD }, /* LATIN CAPITAL LETTER Y WITH ACUTE */ - { "", 0xDE, 0x0162 }, /* LATIN CAPITAL LETTER T WITH CEDILLA */ - { "", 0xDF, 0x00DF }, /* LATIN SMALL LETTER SHARP S (German) */ - { "", 0xE0, 0x0155 }, /* LATIN SMALL LETTER R WITH ACUTE */ - { "", 0xE1, 0x00E1 }, /* LATIN SMALL LETTER A WITH ACUTE */ - { "", 0xE2, 0x00E2 }, /* LATIN SMALL LETTER A WITH CIRCUMFLEX */ - { "", 0xE3, 0x0103 }, /* LATIN SMALL LETTER A WITH BREVE */ - { "", 0xE4, 0x00E4 }, /* LATIN SMALL LETTER A WITH DIAERESIS */ - { "", 0xE5, 0x013A }, /* LATIN SMALL LETTER L WITH ACUTE */ - { "", 0xE6, 0x0107 }, /* LATIN SMALL LETTER C WITH ACUTE */ - { "", 0xE7, 0x00E7 }, /* LATIN SMALL LETTER C WITH CEDILLA */ - { "", 0xE8, 0x010D }, /* LATIN SMALL LETTER C WITH CARON */ - { "", 0xE9, 0x00E9 }, /* LATIN SMALL LETTER E WITH ACUTE */ - { "", 0xEA, 0x0119 }, /* LATIN SMALL LETTER E WITH OGONEK */ - { "", 0xEB, 0x00EB }, /* LATIN SMALL LETTER E WITH DIAERESIS */ - { "", 0xEC, 0x011B }, /* LATIN SMALL LETTER E WITH CARON */ - { "", 0xED, 0x00ED }, /* LATIN SMALL LETTER I WITH ACUTE */ - { "", 0xEE, 0x00EE }, /* LATIN SMALL LETTER I WITH CIRCUMFLEX */ - { "", 0xEF, 0x010F }, /* LATIN SMALL LETTER D WITH CARON */ - { "", 0xF0, 0x0111 }, /* LATIN SMALL LETTER D WITH STROKE */ - { "", 0xF1, 0x0144 }, /* LATIN SMALL LETTER N WITH ACUTE */ - { "", 0xF2, 0x0148 }, /* LATIN SMALL LETTER N WITH CARON */ - { "", 0xF3, 0x00F3 }, /* LATIN SMALL LETTER O WITH ACUTE */ - { "", 0xF4, 0x00F4 }, /* LATIN SMALL LETTER O WITH CIRCUMFLEX */ - { "", 0xF5, 0x0151 }, /* LATIN SMALL LETTER O WITH DOUBLE ACUTE */ - { "", 0xF6, 0x00F6 }, /* LATIN SMALL LETTER O WITH DIAERESIS */ + { "", 0xD8, 0x0158, "R" }, /* LATIN CAPITAL LETTER R WITH CARON */ + { "", 0xD9, 0x016E, "U" }, /* LATIN CAPITAL LETTER U WITH RING ABOVE */ + { "Uacute", 0xDA, 0x00DA, "U" }, /* LATIN CAPITAL LETTER U WITH ACUTE */ + { "", 0xDB, 0x0170, "U" }, /* LATIN CAPITAL LETTER U WITH DOUBLE ACUTE */ + { "Uuml", 0xDC, 0x00DC, "U" }, /* LATIN CAPITAL LETTER U WITH DIAERESIS */ + { "Yacute", 0xDD, 0x00DD, "Y" }, /* LATIN CAPITAL LETTER Y WITH ACUTE */ + { "", 0xDE, 0x0162, "T" }, /* LATIN CAPITAL LETTER T WITH CEDILLA */ + { "szlig", 0xDF, 0x00DF, "ss" }, /* LATIN SMALL LETTER SHARP S (German) */ + { "", 0xE0, 0x0155, "s" }, /* LATIN SMALL LETTER R WITH ACUTE */ + { "aacute", 0xE1, 0x00E1, "a" }, /* LATIN SMALL LETTER A WITH ACUTE */ + { "acirc", 0xE2, 0x00E2, "a" }, /* LATIN SMALL LETTER A WITH CIRCUMFLEX */ + { "", 0xE3, 0x0103, "a" }, /* LATIN SMALL LETTER A WITH BREVE */ + { "auml", 0xE4, 0x00E4, "a" }, /* LATIN SMALL LETTER A WITH DIAERESIS */ + { "", 0xE5, 0x013A, "l" }, /* LATIN SMALL LETTER L WITH ACUTE */ + { "", 0xE6, 0x0107, "c" }, /* LATIN SMALL LETTER C WITH ACUTE */ + { "ccedil", 0xE7, 0x00E7, "c" }, /* LATIN SMALL LETTER C WITH CEDILLA */ + { "", 0xE8, 0x010D, "c" }, /* LATIN SMALL LETTER C WITH CARON */ + { "eacute", 0xE9, 0x00E9, "e" }, /* LATIN SMALL LETTER E WITH ACUTE */ + { "", 0xEA, 0x0119, "e" }, /* LATIN SMALL LETTER E WITH OGONEK */ + { "euml", 0xEB, 0x00EB, "e" }, /* LATIN SMALL LETTER E WITH DIAERESIS */ + { "", 0xEC, 0x011B, "e" }, /* LATIN SMALL LETTER E WITH CARON */ + { "iacute", 0xED, 0x00ED, "i" }, /* LATIN SMALL LETTER I WITH ACUTE */ + { "icirc", 0xEE, 0x00EE, "i" }, /* LATIN SMALL LETTER I WITH CIRCUMFLEX */ + { "", 0xEF, 0x010F, "d" }, /* LATIN SMALL LETTER D WITH CARON */ + { "", 0xF0, 0x0111, "d" }, /* LATIN SMALL LETTER D WITH STROKE */ + { "", 0xF1, 0x0144, "n" }, /* LATIN SMALL LETTER N WITH ACUTE */ + { "", 0xF2, 0x0148, "n" }, /* LATIN SMALL LETTER N WITH CARON */ + { "oacute", 0xF3, 0x00F3, "o" }, /* LATIN SMALL LETTER O WITH ACUTE */ + { "ocirc", 0xF4, 0x00F4, "o" }, /* LATIN SMALL LETTER O WITH CIRCUMFLEX */ + { "", 0xF5, 0x0151, "o" }, /* LATIN SMALL LETTER O WITH DOUBLE ACUTE */ + { "ouml", 0xF6, 0x00F6, "o" }, /* LATIN SMALL LETTER O WITH DIAERESIS */ { "divide", 0xF7, 0x00F7 }, /* DIVISION SIGN */ - { "", 0xF8, 0x0159 }, /* LATIN SMALL LETTER R WITH CARON */ - { "", 0xF9, 0x016F }, /* LATIN SMALL LETTER U WITH RING ABOVE */ - { "", 0xFA, 0x00FA }, /* LATIN SMALL LETTER U WITH ACUTE */ - { "", 0xFB, 0x0171 }, /* LATIN SMALL LETTER U WITH DOUBLE ACUTE */ - { "", 0xFC, 0x00FC }, /* LATIN SMALL LETTER U WITH DIAERESIS */ - { "", 0xFD, 0x00FD }, /* LATIN SMALL LETTER Y WITH ACUTE */ - { "", 0xFE, 0x0163 }, /* LATIN SMALL LETTER T WITH CEDILLA */ + { "", 0xF8, 0x0159, "r" }, /* LATIN SMALL LETTER R WITH CARON */ + { "", 0xF9, 0x016F, "u" }, /* LATIN SMALL LETTER U WITH RING ABOVE */ + { "uacute", 0xFA, 0x00FA, "u" }, /* LATIN SMALL LETTER U WITH ACUTE */ + { "", 0xFB, 0x0171, "u" }, /* LATIN SMALL LETTER U WITH DOUBLE ACUTE */ + { "uuml", 0xFC, 0x00FC, "u" }, /* LATIN SMALL LETTER U WITH DIAERESIS */ + { "yacute", 0xFD, 0x00FD, "y" }, /* LATIN SMALL LETTER Y WITH ACUTE */ + { "", 0xFE, 0x0163, "t" }, /* LATIN SMALL LETTER T WITH CEDILLA */ { "", 0xFF, 0x02D9 }, /* DOT ABOVE (Mandarin Chinese light tone) */ { NULL, 0, 0 } }; +/* Common map for koi8-u, koi8-r */ +static iso_map_type koi8_map [] = { + { "", 0xa3, 0x0415, "io"}, /* CYRILLIC SMALL LETTER IO */ + { "", 0xa4, 0x0454, "ie"}, /* CYRILLIC SMALL LETTER UKRAINIAN IE */ + { "", 0xa6, 0x0456, "i"}, /* CYRILLIC SMALL LETTER BYELORUSSIAN-UKRAINIAN I */ + { "", 0xa7, 0x0457, "yi"}, /* CYRILLIC SMALL LETTER YI */ + + { "", 0xb3, 0x04d7, "IO"}, /* CYRILLIC CAPITAL LETTER IO */ + { "", 0xb4, 0x0404, "IE"}, /* CYRILLIC CAPITAL LETTER UKRAINIAN IE */ + { "", 0xb6, 0x0406, "I"}, /* CYRILLIC CAPITAL LETTER BYELORUSSIAN-UKRAINIAN I */ + { "", 0xb7, 0x0407, "YI"}, /* CYRILLIC CAPITAL LETTER YI */ +/* { "", 0xbf, 0x}, / * CYRILLIC COPYRIGHT SIGN */ + { "", 0xc0, 0x042e, "yu"}, /* CYRILLIC SMALL LETTER YU */ + { "", 0xc1, 0x0430, "a"}, /* CYRILLIC SMALL LETTER A */ + { "", 0xc2, 0x0431, "b"}, /* CYRILLIC SMALL LETTER BE */ + { "", 0xc3, 0x0446, "c"}, /* CYRILLIC SMALL LETTER TSE */ + { "", 0xc4, 0x0434, "d"}, /* CYRILLIC SMALL LETTER DE */ + { "", 0xc5, 0x0435, "e"}, /* CYRILLIC SMALL LETTER IE */ + { "", 0xc6, 0x0444, "f"}, /* CYRILLIC SMALL LETTER EF */ + { "", 0xc7, 0x0433, "g"}, /* CYRILLIC SMALL LETTER GHE */ + { "", 0xc8, 0x0445, "h"}, /* CYRILLIC SMALL LETTER HA */ + { "", 0xc9, 0x0438, "i"}, /* CYRILLIC SMALL LETTER I */ + { "", 0xca, 0x0439, "i"}, /* CYRILLIC SMALL LETTER SHORT I */ + { "", 0xcb, 0x043a, "k"}, /* CYRILLIC SMALL LETTER KA */ + { "", 0xcc, 0x043b, "l"}, /* CYRILLIC SMALL LETTER EL */ + { "", 0xcd, 0x043c, "m"}, /* CYRILLIC SMALL LETTER EM */ + { "", 0xce, 0x043d, "n"}, /* CYRILLIC SMALL LETTER EN */ + { "", 0xcf, 0x043e, "o"}, /* CYRILLIC SMALL LETTER O */ + { "", 0xd0, 0x043f, "p"}, /* CYRILLIC SMALL LETTER PE */ + { "", 0xd1, 0x044f, "ya"}, /* CYRILLIC SMALL LETTER YA */ + { "", 0xd2, 0x0440, "r"}, /* CYRILLIC SMALL LETTER ER */ + { "", 0xd3, 0x0441, "s"}, /* CYRILLIC SMALL LETTER ES */ + { "", 0xd4, 0x0442, "t"}, /* CYRILLIC SMALL LETTER TE */ + { "", 0xd5, 0x0443, "u"}, /* CYRILLIC SMALL LETTER U */ + { "", 0xd6, 0x0436, "zh"}, /* CYRILLIC SMALL LETTER ZHE */ + { "", 0xd7, 0x0432, "v"}, /* CYRILLIC SMALL LETTER VE */ + { "", 0xd8, 0x044c, "x"}, /* CYRILLIC SMALL LETTER SOFT SIGN */ + { "", 0xd9, 0x044b, "y"}, /* CYRILLIC SMALL LETTER YERU */ + { "", 0xda, 0x0437, "z"}, /* CYRILLIC SMALL LETTER ZE */ + { "", 0xdb, 0x0448, "sh"}, /* CYRILLIC SMALL LETTER SHA */ + { "", 0xdc, 0x044d, "e"}, /* CYRILLIC SMALL LETTER E */ + { "", 0xdd, 0x0449, "shch"}, /* CYRILLIC SMALL LETTER SHCHA */ + { "", 0xde, 0x0447, "ch"}, /* CYRILLIC SMALL LETTER CHA */ + { "", 0xdf, 0x044a, "w"}, /* CYRILLIC SMALL LETTER HARD SIGN */ + { "", 0xe0, 0x042d, "YU"}, /* CYRILLIC CAPITAL LETTER YU */ + { "", 0xe1, 0x0410, "A"}, /* CYRILLIC CAPITAL LETTER A */ + { "", 0xe2, 0x0411, "B"}, /* CYRILLIC CAPITAL LETTER BE */ + { "", 0xe3, 0x0426, "C"}, /* CYRILLIC CAPITAL LETTER TSE */ + { "", 0xe4, 0x0414, "D"}, /* CYRILLIC CAPITAL LETTER DE */ + { "", 0xe5, 0x0415, "E"}, /* CYRILLIC CAPITAL LETTER IE */ + { "", 0xe6, 0x0424, "F"}, /* CYRILLIC CAPITAL LETTER EF */ + { "", 0xe7, 0x0413, "G"}, /* CYRILLIC CAPITAL LETTER GHE */ + { "", 0xe8, 0x0425, "H"}, /* CYRILLIC CAPITAL LETTER HA */ + { "", 0xe9, 0x0418, "I"}, /* CYRILLIC CAPITAL LETTER I */ + { "", 0xea, 0x0419, "I"}, /* CYRILLIC CAPITAL LETTER SHORT I */ + { "", 0xeb, 0x041a, "K"}, /* CYRILLIC CAPITAL LETTER KA */ + { "", 0xec, 0x041b, "L"}, /* CYRILLIC CAPITAL LETTER EL */ + { "", 0xed, 0x041c, "M"}, /* CYRILLIC CAPITAL LETTER EM */ + { "", 0xee, 0x041d, "N"}, /* CYRILLIC CAPITAL LETTER EN */ + { "", 0xef, 0x041e, "O"}, /* CYRILLIC CAPITAL LETTER O */ + { "", 0xf0, 0x041f, "P"}, /* CYRILLIC CAPITAL LETTER PE */ + { "", 0xf1, 0x042f, "YA"}, /* CYRILLIC CAPITAL LETTER YA */ + { "", 0xf2, 0x0420, "R"}, /* CYRILLIC CAPITAL LETTER ER */ + { "", 0xf3, 0x0421, "S"}, /* CYRILLIC CAPITAL LETTER ES */ + { "", 0xf4, 0x0422, "T"}, /* CYRILLIC CAPITAL LETTER TE */ + { "", 0xf5, 0x0423, "U"}, /* CYRILLIC CAPITAL LETTER U */ + { "", 0xf6, 0x0416, "ZH"}, /* CYRILLIC CAPITAL LETTER ZHE */ + { "", 0xf7, 0x0412, "V"}, /* CYRILLIC CAPITAL LETTER VE */ + { "", 0xf8, 0x042c, "X"}, /* CYRILLIC CAPITAL LETTER SOFT SIGN */ + { "", 0xf9, 0x042b, "Y"}, /* CYRILLIC CAPITAL LETTER YERU */ + { "", 0xfa, 0x0417, "Z"}, /* CYRILLIC CAPITAL LETTER ZE */ + { "", 0xfb, 0x0428, "SH"}, /* CYRILLIC CAPITAL LETTER SHA */ + { "", 0xfc, 0x042d, "E"}, /* CYRILLIC CAPITAL LETTER E */ + { "", 0xfd, 0x0429, "SHCH"}, /* CYRILLIC CAPITAL LETTER SHCHA */ + { "", 0xfe, 0x0427, "CH"}, /* CYRILLIC CAPITAL LETTER CHE */ + { "", 0xff, 0x042a, "W"}, /* CYRILLIC CAPITAL LETTER HARD SIGN */ + { NULL, 0, 0 } +}; + encoding_type encoding_table[] = { { no_encoding, "(no encoding)", NULL }, - { US_ASCII, "US-ASCII", us_ascii_map }, + { US_ASCII, "US-ASCII", asis_map }, { ISO_8859_1, "iso-8859-1", (iso_map_type *) iso8859_1_map }, { ISO_8859_2, "iso-8859-2", (iso_map_type *) iso8859_2_map }, { ISO_8859_3, "iso-8859-3", NULL }, @@ -402,81 +484,389 @@ encoding_type encoding_table[] = { { ISO_8859_13, "iso-8859-13", NULL }, { ISO_8859_14, "iso-8859-14", NULL }, { ISO_8859_15, "iso-8859-15", (iso_map_type *) iso8859_15_map }, + { KOI8_R, "koi8-r", (iso_map_type *) koi8_map }, + { KOI8_U, "koi8-u", (iso_map_type *) koi8_map }, + { UTF_8, "utf-8", asis_map }, /* specific code for this below */ { last_encoding_code, NULL, NULL } }; + +/* List of HTML entities. */ +static struct { const char *html; unsigned int unicode; } unicode_map[] = { +/* Extracted from http://www.w3.org/TR/html401/sgml/entities.html through + sed -n -e 's|<!ENTITY \([^ ][^ ]*\) *CDATA "[&]#\([0-9][0-9]*\);".*| { "\1", \2 },|p' + | LC_ALL=C sort -k2 */ + { "AElig", 198 }, + { "Aacute", 193 }, + { "Acirc", 194 }, + { "Agrave", 192 }, + { "Alpha", 913 }, + { "Aring", 197 }, + { "Atilde", 195 }, + { "Auml", 196 }, + { "Beta", 914 }, + { "Ccedil", 199 }, + { "Chi", 935 }, + { "Dagger", 8225 }, + { "Delta", 916 }, + { "ETH", 208 }, + { "Eacute", 201 }, + { "Ecirc", 202 }, + { "Egrave", 200 }, + { "Epsilon", 917 }, + { "Eta", 919 }, + { "Euml", 203 }, + { "Gamma", 915 }, + { "Iacute", 205 }, + { "Icirc", 206 }, + { "Igrave", 204 }, + { "Iota", 921 }, + { "Iuml", 207 }, + { "Kappa", 922 }, + { "Lambda", 923 }, + { "Mu", 924 }, + { "Ntilde", 209 }, + { "Nu", 925 }, + { "OElig", 338 }, + { "Oacute", 211 }, + { "Ocirc", 212 }, + { "Ograve", 210 }, + { "Omega", 937 }, + { "Omicron", 927 }, + { "Oslash", 216 }, + { "Otilde", 213 }, + { "Ouml", 214 }, + { "Phi", 934 }, + { "Pi", 928 }, + { "Prime", 8243 }, + { "Psi", 936 }, + { "Rho", 929 }, + { "Scaron", 352 }, + { "Sigma", 931 }, + { "THORN", 222 }, + { "Tau", 932 }, + { "Theta", 920 }, + { "Uacute", 218 }, + { "Ucirc", 219 }, + { "Ugrave", 217 }, + { "Upsilon", 933 }, + { "Uuml", 220 }, + { "Xi", 926 }, + { "Yacute", 221 }, + { "Yuml", 376 }, + { "Zeta", 918 }, + { "aacute", 225 }, + { "acirc", 226 }, + { "acute", 180 }, + { "aelig", 230 }, + { "agrave", 224 }, + { "alefsym", 8501 }, + { "alpha", 945 }, + { "amp", 38 }, + { "and", 8743 }, + { "ang", 8736 }, + { "aring", 229 }, + { "asymp", 8776 }, + { "atilde", 227 }, + { "auml", 228 }, + { "bdquo", 8222 }, + { "beta", 946 }, + { "brvbar", 166 }, + { "bull", 8226 }, + { "cap", 8745 }, + { "ccedil", 231 }, + { "cedil", 184 }, + { "cent", 162 }, + { "chi", 967 }, + { "circ", 710 }, + { "clubs", 9827 }, + { "cong", 8773 }, + { "copy", 169 }, + { "crarr", 8629 }, + { "cup", 8746 }, + { "curren", 164 }, + { "dArr", 8659 }, + { "dagger", 8224 }, + { "darr", 8595 }, + { "deg", 176 }, + { "delta", 948 }, + { "diams", 9830 }, + { "divide", 247 }, + { "eacute", 233 }, + { "ecirc", 234 }, + { "egrave", 232 }, + { "empty", 8709 }, + { "emsp", 8195 }, + { "ensp", 8194 }, + { "epsilon", 949 }, + { "equiv", 8801 }, + { "eta", 951 }, + { "eth", 240 }, + { "euml", 235 }, + { "euro", 8364 }, + { "exist", 8707 }, + { "fnof", 402 }, + { "forall", 8704 }, + { "frac12", 189 }, + { "frac14", 188 }, + { "frac34", 190 }, + { "frasl", 8260 }, + { "gamma", 947 }, + { "ge", 8805 }, + { "gt", 62 }, + { "hArr", 8660 }, + { "harr", 8596 }, + { "hearts", 9829 }, + { "hellip", 8230 }, + { "iacute", 237 }, + { "icirc", 238 }, + { "iexcl", 161 }, + { "igrave", 236 }, + { "image", 8465 }, + { "infin", 8734 }, + { "int", 8747 }, + { "iota", 953 }, + { "iquest", 191 }, + { "isin", 8712 }, + { "iuml", 239 }, + { "kappa", 954 }, + { "lArr", 8656 }, + { "lambda", 955 }, + { "lang", 9001 }, + { "laquo", 171 }, + { "larr", 8592 }, + { "lceil", 8968 }, + { "ldquo", 8220 }, + { "le", 8804 }, + { "lfloor", 8970 }, + { "lowast", 8727 }, + { "loz", 9674 }, + { "lrm", 8206 }, + { "lsaquo", 8249 }, + { "lsquo", 8216 }, + { "lt", 60 }, + { "macr", 175 }, + { "mdash", 8212 }, + { "micro", 181 }, + { "middot", 183 }, + { "minus", 8722 }, + { "mu", 956 }, + { "nabla", 8711 }, + { "nbsp", 160 }, + { "ndash", 8211 }, + { "ne", 8800 }, + { "ni", 8715 }, + { "not", 172 }, + { "notin", 8713 }, + { "nsub", 8836 }, + { "ntilde", 241 }, + { "nu", 957 }, + { "oacute", 243 }, + { "ocirc", 244 }, + { "oelig", 339 }, + { "ograve", 242 }, + { "oline", 8254 }, + { "omega", 969 }, + { "omicron", 959 }, + { "oplus", 8853 }, + { "or", 8744 }, + { "ordf", 170 }, + { "ordm", 186 }, + { "oslash", 248 }, + { "otilde", 245 }, + { "otimes", 8855 }, + { "ouml", 246 }, + { "para", 182 }, + { "part", 8706 }, + { "permil", 8240 }, + { "perp", 8869 }, + { "phi", 966 }, + { "pi", 960 }, + { "piv", 982 }, + { "plusmn", 177 }, + { "pound", 163 }, + { "prime", 8242 }, + { "prod", 8719 }, + { "prop", 8733 }, + { "psi", 968 }, + { "quot", 34 }, + { "rArr", 8658 }, + { "radic", 8730 }, + { "rang", 9002 }, + { "raquo", 187 }, + { "rarr", 8594 }, + { "rceil", 8969 }, + { "rdquo", 8221 }, + { "real", 8476 }, + { "reg", 174 }, + { "rfloor", 8971 }, + { "rho", 961 }, + { "rlm", 8207 }, + { "rsaquo", 8250 }, + { "rsquo", 8217 }, + { "sbquo", 8218 }, + { "scaron", 353 }, + { "sdot", 8901 }, + { "sect", 167 }, + { "shy", 173 }, + { "sigma", 963 }, + { "sigmaf", 962 }, + { "sim", 8764 }, + { "spades", 9824 }, + { "sub", 8834 }, + { "sube", 8838 }, + { "sum", 8721 }, + { "sup", 8835 }, + { "sup1", 185 }, + { "sup2", 178 }, + { "sup3", 179 }, + { "supe", 8839 }, + { "szlig", 223 }, + { "tau", 964 }, + { "there4", 8756 }, + { "theta", 952 }, + { "thetasym", 977 }, + { "thinsp", 8201 }, + { "thorn", 254 }, + { "tilde", 732 }, + { "times", 215 }, + { "trade", 8482 }, + { "uArr", 8657 }, + { "uacute", 250 }, + { "uarr", 8593 }, + { "ucirc", 251 }, + { "ugrave", 249 }, + { "uml", 168 }, + { "upsih", 978 }, + { "upsilon", 965 }, + { "uuml", 252 }, + { "weierp", 8472 }, + { "xi", 958 }, + { "yacute", 253 }, + { "yen", 165 }, + { "yuml", 255 }, + { "zeta", 950 }, + { "zwj", 8205 }, + { "zwnj", 8204 } +}; + + + +/* To update this list of language codes, download the current data from + http://www.loc.gov/standards/iso639-2; specifically, + http://www.loc.gov/standards/iso639-2/ISO-639-2_values_8bits-8559-1.txt. + Run cut -d\| -f 3,4 <ISO-639-2_values_8bits-8559-1.txt | sort -u >/tmp/639.2 + + Extract the C table below to a file, say /tmp/lc, and run: + cut -c 10- /tmp/lc| sed -e 's/", "/|/' -e 's,".*,,'>/tmp/lang.2 + + Then: comm -3 /tmp/639.2 /tmp/lang.c. + + Also update the enum in lang.h. */ language_type language_table[] = { { aa, "aa", "Afar" }, { ab, "ab", "Abkhazian" }, + { ae, "ae", "Avestan" }, { af, "af", "Afrikaans" }, + { ak, "ak", "Akan" }, { am, "am", "Amharic" }, + { an, "an", "Aragonese" }, { ar, "ar", "Arabic" }, { as, "as", "Assamese" }, + { av, "av", "Avaric" }, { ay, "ay", "Aymara" }, { az, "az", "Azerbaijani" }, { ba, "ba", "Bashkir" }, - { be, "be", "Byelorussian" }, + { be, "be", "Belarusian" }, { bg, "bg", "Bulgarian" }, { bh, "bh", "Bihari" }, { bi, "bi", "Bislama" }, - { bn, "bn", "Bengali; Bangla" }, + { bm, "bm", "Bambara" }, + { bn, "bn", "Bengali" }, { bo, "bo", "Tibetan" }, { br, "br", "Breton" }, - { ca, "ca", "Catalan" }, + { bs, "bs", "Bosnian" }, + { ca, "ca", "Catalan; Valencian" }, + { ce, "ce", "Chechen" }, + { ch, "ch", "Chamorro" }, { co, "co", "Corsican" }, + { cr, "cr", "Cree" }, { cs, "cs", "Czech" }, + { cu, "cu", "Church Slavic; Old Slavonic; Church Slavonic; Old Bulgarian; Old Church Slavonic" }, + { cv, "cv", "Chuvash" }, { cy, "cy", "Welsh" }, { da, "da", "Danish" }, { de, "de", "German" }, - { dz, "dz", "Bhutani" }, - { el, "el", "Greek" }, + { dv, "dv", "Divehi; Dhivehi; Maldivian" }, + { dz, "dz", "Dzongkha" }, + { ee, "ee", "Ewe" }, + { el, "el", "Greek, Modern (1453-)" }, { en, "en", "English" }, { eo, "eo", "Esperanto" }, - { es, "es", "Spanish" }, + { es, "es", "Spanish; Castilian" }, { et, "et", "Estonian" }, { eu, "eu", "Basque" }, { fa, "fa", "Persian" }, + { ff, "ff", "Fulah" }, { fi, "fi", "Finnish" }, - { fj, "fj", "Fiji" }, + { fj, "fj", "Fijian" }, { fo, "fo", "Faroese" }, { fr, "fr", "French" }, - { fy, "fy", "Frisian" }, + { fy, "fy", "Western Frisian" }, { ga, "ga", "Irish" }, - { gd, "gd", "Scots Gaelic" }, + { gd, "gd", "Gaelic; Scottish Gaelic" }, { gl, "gl", "Galician" }, { gn, "gn", "Guarani" }, { gu, "gu", "Gujarati" }, + { gv, "gv", "Manx" }, { ha, "ha", "Hausa" }, { he, "he", "Hebrew" } /* (formerly iw) */, { hi, "hi", "Hindi" }, + { ho, "ho", "Hiri Motu" }, { hr, "hr", "Croatian" }, + { ht, "ht", "Haitian; Haitian Creole" }, { hu, "hu", "Hungarian" }, { hy, "hy", "Armenian" }, - { ia, "ia", "Interlingua" }, + { hz, "hz", "Herero" }, + { ia, "ia", "Interlingua (International Auxiliary Language Association)" }, { id, "id", "Indonesian" } /* (formerly in) */, { ie, "ie", "Interlingue" }, - { ik, "ik", "Inupiak" }, + { ig, "ig", "Igbo" }, + { ii, "ii", "Sichuan Yi" }, + { ik, "ik", "Inupiaq" }, + { io, "io", "Ido" }, { is, "is", "Icelandic" }, { it, "it", "Italian" }, { iu, "iu", "Inuktitut" }, { ja, "ja", "Japanese" }, - { jw, "jw", "Javanese" }, + { jv, "jv", "Javanese" }, { ka, "ka", "Georgian" }, + { kg, "kg", "Kongo" }, + { ki, "ki", "Kikuyu; Gikuyu" }, + { kj, "kj", "Kuanyama; Kwanyama" }, { kk, "kk", "Kazakh" }, - { kl, "kl", "Greenlandic" }, - { km, "km", "Cambodian" }, + { kl, "kl", "Kalaallisut; Greenlandic" }, + { km, "km", "Central Khmer" }, { kn, "kn", "Kannada" }, { ko, "ko", "Korean" }, + { kr, "kr", "Kanuri" }, { ks, "ks", "Kashmiri" }, { ku, "ku", "Kurdish" }, - { ky, "ky", "Kirghiz" }, + { kv, "kv", "Komi" }, + { kw, "kw", "Cornish" }, + { ky, "ky", "Kirghiz; Kyrgyz" }, { la, "la", "Latin" }, + { lb, "lb", "Luxembourgish; Letzeburgesch" }, + { lg, "lg", "Ganda" }, + { li, "li", "Limburgan; Limburger; Limburgish" }, { ln, "ln", "Lingala" }, - { lo, "lo", "Laothian" }, + { lo, "lo", "Lao" }, { lt, "lt", "Lithuanian" }, - { lv, "lv", "Latvian, Lettish" }, + { lu, "lu", "Luba-Katanga" }, + { lv, "lv", "Latvian" }, { mg, "mg", "Malagasy" }, + { mh, "mh", "Marshallese" }, { mi, "mi", "Maori" }, { mk, "mk", "Macedonian" }, { ml, "ml", "Malayalam" }, @@ -487,27 +877,38 @@ language_type language_table[] = { { mt, "mt", "Maltese" }, { my, "my", "Burmese" }, { na, "na", "Nauru" }, + { nb, "nb", "Bokmål, Norwegian; Norwegian Bokmål" }, + { nd, "nd", "Ndebele, North; North Ndebele" }, { ne, "ne", "Nepali" }, - { nl, "nl", "Dutch" }, + { ng, "ng", "Ndonga" }, + { nl, "nl", "Dutch; Flemish" }, + { nn, "nn", "Norwegian Nynorsk; Nynorsk, Norwegian" }, { no, "no", "Norwegian" }, - { oc, "oc", "Occitan" }, - { om, "om", "(Afan) Oromo" }, + { nr, "nr", "Ndebele, South; South Ndebele" }, + { nv, "nv", "Navajo; Navaho" }, + { ny, "ny", "Chichewa; Chewa; Nyanja" }, + { oc, "oc", "Occitan (post 1500); Provençal" }, + { oj, "oj", "Ojibwa" }, + { om, "om", "Oromo" }, { or, "or", "Oriya" }, - { pa, "pa", "Punjabi" }, + { os, "os", "Ossetian; Ossetic" }, + { pa, "pa", "Panjabi; Punjabi" }, + { pi, "pi", "Pali" }, { pl, "pl", "Polish" }, - { ps, "ps", "Pashto, Pushto" }, + { ps, "ps", "Pushto" }, { pt, "pt", "Portuguese" }, { qu, "qu", "Quechua" }, - { rm, "rm", "Rhaeto-Romance" }, - { rn, "rn", "Kirundi" }, + { rm, "rm", "Romansh" }, + { rn, "rn", "Rundi" }, { ro, "ro", "Romanian" }, { ru, "ru", "Russian" }, { rw, "rw", "Kinyarwanda" }, { sa, "sa", "Sanskrit" }, + { sc, "sc", "Sardinian" }, { sd, "sd", "Sindhi" }, - { sg, "sg", "Sangro" }, - { sh, "sh", "Serbo-Croatian" }, - { si, "si", "Sinhalese" }, + { se, "se", "Northern Sami" }, + { sg, "sg", "Sango" }, + { si, "si", "Sinhala; Sinhalese" }, { sk, "sk", "Slovak" }, { sl, "sl", "Slovenian" }, { sm, "sm", "Samoan" }, @@ -515,8 +916,8 @@ language_type language_table[] = { { so, "so", "Somali" }, { sq, "sq", "Albanian" }, { sr, "sr", "Serbian" }, - { ss, "ss", "Siswati" }, - { st, "st", "Sesotho" }, + { ss, "ss", "Swati" }, + { st, "st", "Sotho, Souther" }, { su, "su", "Sundanese" }, { sv, "sv", "Swedish" }, { sw, "sw", "Swahili" }, @@ -527,48 +928,348 @@ language_type language_table[] = { { ti, "ti", "Tigrinya" }, { tk, "tk", "Turkmen" }, { tl, "tl", "Tagalog" }, - { tn, "tn", "Setswana" }, - { to, "to", "Tonga" }, + { tn, "tn", "Tswana" }, + { to, "to", "Tonga (Tonga Islands)" }, { tr, "tr", "Turkish" }, { ts, "ts", "Tsonga" }, { tt, "tt", "Tatar" }, { tw, "tw", "Twi" }, - { ug, "ug", "Uighur" }, + { ty, "ty", "Tahitian" }, + { ug, "ug", "Uighur; Uyghur" }, { uk, "uk", "Ukrainian" }, { ur, "ur", "Urdu" }, { uz, "uz", "Uzbek" }, + { ve, "ve", "Venda" }, { vi, "vi", "Vietnamese" }, - { vo, "vo", "Volapuk" }, + { vo, "vo", "Volapük" }, + { wa, "wa", "Walloon" }, { wo, "wo", "Wolof" }, { xh, "xh", "Xhosa" }, { yi, "yi", "Yiddish" } /* (formerly ji) */, { yo, "yo", "Yoruba" }, - { za, "za", "Zhuang" }, + { za, "za", "Zhuang; Chuang" }, { zh, "zh", "Chinese" }, { zu, "zu", "Zulu" }, { last_language_code, NULL, NULL } }; + +static const char * +cm_search_iso_map_char (byte_t ch) +{ + int i; + iso_map_type *iso = encoding_table[document_encoding_code].isotab; + + /* If no conversion table for this encoding, quit. */ + if (!iso) + return NULL; + + for (i = 0; iso[i].html; i++) + if (iso[i].bytecode == ch) + return iso[i].translit; + + return NULL; +} + +const char * +lang_transliterate_char (byte_t ch) +{ + if (transliterate_file_names + && document_encoding_code != no_encoding) + return cm_search_iso_map_char (ch); + return NULL; +} + + -/* @documentlanguage. Maybe we'll do something useful with this in the - future. For now, we just recognize it. */ +/* Given a language code LL_CODE, return a "default" country code (in + new memory). We use the same table as gettext, and return LL_CODE + uppercased in the absence of any better possibility, with a warning. + (gettext silently defaults to the C locale, but we want to give users + a shot at fixing ambiguities.) We also return en_US for en, while + gettext does not. */ + +#define SIZEOF(a) (sizeof(a) / sizeof(a[0])) + +static char * +default_country_for_lang (const char *ll_code) +{ + /* This table comes from msginit.c in gettext (0.16.1). It is + intentionally copied verbatim even though the format is not the + most convenient for our purposes, to make updates simple. */ + static const char *locales_with_principal_territory[] = { + /* Language Main territory */ + "ace_ID", /* Achinese Indonesia */ + "af_ZA", /* Afrikaans South Africa */ + "ak_GH", /* Akan Ghana */ + "am_ET", /* Amharic Ethiopia */ + "an_ES", /* Aragonese Spain */ + "ang_GB", /* Old English Britain */ + "as_IN", /* Assamese India */ + "av_RU", /* Avaric Russia */ + "awa_IN", /* Awadhi India */ + "az_AZ", /* Azerbaijani Azerbaijan */ + "bad_CF", /* Banda Central African Republic */ + "ban_ID", /* Balinese Indonesia */ + "be_BY", /* Belarusian Belarus */ + "bem_ZM", /* Bemba Zambia */ + "bg_BG", /* Bulgarian Bulgaria */ + "bho_IN", /* Bhojpuri India */ + "bik_PH", /* Bikol Philippines */ + "bin_NG", /* Bini Nigeria */ + "bm_ML", /* Bambara Mali */ + "bn_IN", /* Bengali India */ + "bo_CN", /* Tibetan China */ + "br_FR", /* Breton France */ + "bs_BA", /* Bosnian Bosnia */ + "btk_ID", /* Batak Indonesia */ + "bug_ID", /* Buginese Indonesia */ + "ca_ES", /* Catalan Spain */ + "ce_RU", /* Chechen Russia */ + "ceb_PH", /* Cebuano Philippines */ + "co_FR", /* Corsican France */ + "cr_CA", /* Cree Canada */ + "cs_CZ", /* Czech Czech Republic */ + "csb_PL", /* Kashubian Poland */ + "cy_GB", /* Welsh Britain */ + "da_DK", /* Danish Denmark */ + "de_DE", /* German Germany */ + "din_SD", /* Dinka Sudan */ + "doi_IN", /* Dogri India */ + "dv_MV", /* Divehi Maldives */ + "dz_BT", /* Dzongkha Bhutan */ + "ee_GH", /* Éwé Ghana */ + "el_GR", /* Greek Greece */ + /* Don't put "en_GB" or "en_US" here. That would be asking for fruitless + political discussion. */ + "es_ES", /* Spanish Spain */ + "et_EE", /* Estonian Estonia */ + "fa_IR", /* Persian Iran */ + "fi_FI", /* Finnish Finland */ + "fil_PH", /* Filipino Philippines */ + "fj_FJ", /* Fijian Fiji */ + "fo_FO", /* Faroese Faeroe Islands */ + "fon_BJ", /* Fon Benin */ + "fr_FR", /* French France */ + "fy_NL", /* Western Frisian Netherlands */ + "ga_IE", /* Irish Ireland */ + "gd_GB", /* Scots Britain */ + "gon_IN", /* Gondi India */ + "gsw_CH", /* Swiss German Switzerland */ + "gu_IN", /* Gujarati India */ + "he_IL", /* Hebrew Israel */ + "hi_IN", /* Hindi India */ + "hil_PH", /* Hiligaynon Philippines */ + "hr_HR", /* Croatian Croatia */ + "ht_HT", /* Haitian Haiti */ + "hu_HU", /* Hungarian Hungary */ + "hy_AM", /* Armenian Armenia */ + "id_ID", /* Indonesian Indonesia */ + "ig_NG", /* Igbo Nigeria */ + "ii_CN", /* Sichuan Yi China */ + "ilo_PH", /* Iloko Philippines */ + "is_IS", /* Icelandic Iceland */ + "it_IT", /* Italian Italy */ + "ja_JP", /* Japanese Japan */ + "jab_NG", /* Hyam Nigeria */ + "jv_ID", /* Javanese Indonesia */ + "ka_GE", /* Georgian Georgia */ + "kab_DZ", /* Kabyle Algeria */ + "kaj_NG", /* Jju Nigeria */ + "kam_KE", /* Kamba Kenya */ + "kmb_AO", /* Kimbundu Angola */ + "kcg_NG", /* Tyap Nigeria */ + "kdm_NG", /* Kagoma Nigeria */ + "kg_CD", /* Kongo Democratic Republic of Congo */ + "kk_KZ", /* Kazakh Kazakhstan */ + "kl_GL", /* Kalaallisut Greenland */ + "km_KH", /* Khmer Cambodia */ + "kn_IN", /* Kannada India */ + "ko_KR", /* Korean Korea (South) */ + "kok_IN", /* Konkani India */ + "kr_NG", /* Kanuri Nigeria */ + "kru_IN", /* Kurukh India */ + "lg_UG", /* Ganda Uganda */ + "li_BE", /* Limburgish Belgium */ + "lo_LA", /* Laotian Laos */ + "lt_LT", /* Lithuanian Lithuania */ + "lu_CD", /* Luba-Katanga Democratic Republic of Congo */ + "lua_CD", /* Luba-Lulua Democratic Republic of Congo */ + "luo_KE", /* Luo Kenya */ + "lv_LV", /* Latvian Latvia */ + "mad_ID", /* Madurese Indonesia */ + "mag_IN", /* Magahi India */ + "mai_IN", /* Maithili India */ + "mak_ID", /* Makasar Indonesia */ + "man_ML", /* Mandingo Mali */ + "men_SL", /* Mende Sierra Leone */ + "mg_MG", /* Malagasy Madagascar */ + "min_ID", /* Minangkabau Indonesia */ + "mk_MK", /* Macedonian Macedonia */ + "ml_IN", /* Malayalam India */ + "mn_MN", /* Mongolian Mongolia */ + "mni_IN", /* Manipuri India */ + "mos_BF", /* Mossi Burkina Faso */ + "mr_IN", /* Marathi India */ + "ms_MY", /* Malay Malaysia */ + "mt_MT", /* Maltese Malta */ + "mwr_IN", /* Marwari India */ + "my_MM", /* Burmese Myanmar */ + "na_NR", /* Nauru Nauru */ + "nah_MX", /* Nahuatl Mexico */ + "nap_IT", /* Neapolitan Italy */ + "nb_NO", /* Norwegian BokmÃ¥l Norway */ + "nds_DE", /* Low Saxon Germany */ + "ne_NP", /* Nepali Nepal */ + "nl_NL", /* Dutch Netherlands */ + "nn_NO", /* Norwegian Nynorsk Norway */ + "no_NO", /* Norwegian Norway */ + "nr_ZA", /* South Ndebele South Africa */ + "nso_ZA", /* Northern Sotho South Africa */ + "nym_TZ", /* Nyamwezi Tanzania */ + "nyn_UG", /* Nyankole Uganda */ + "oc_FR", /* Occitan France */ + "oj_CA", /* Ojibwa Canada */ + "or_IN", /* Oriya India */ + "pa_IN", /* Punjabi India */ + "pag_PH", /* Pangasinan Philippines */ + "pam_PH", /* Pampanga Philippines */ + "pbb_CO", /* Páez Colombia */ + "pl_PL", /* Polish Poland */ + "ps_AF", /* Pashto Afghanistan */ + "pt_PT", /* Portuguese Portugal */ + "raj_IN", /* Rajasthani India */ + "rm_CH", /* Rhaeto-Roman Switzerland */ + "rn_BI", /* Kirundi Burundi */ + "ro_RO", /* Romanian Romania */ + "ru_RU", /* Russian Russia */ + "sa_IN", /* Sanskrit India */ + "sas_ID", /* Sasak Indonesia */ + "sat_IN", /* Santali India */ + "sc_IT", /* Sardinian Italy */ + "scn_IT", /* Sicilian Italy */ + "sg_CF", /* Sango Central African Republic */ + "shn_MM", /* Shan Myanmar */ + "si_LK", /* Sinhala Sri Lanka */ + "sid_ET", /* Sidamo Ethiopia */ + "sk_SK", /* Slovak Slovakia */ + "sl_SI", /* Slovenian Slovenia */ + "so_SO", /* Somali Somalia */ + "sq_AL", /* Albanian Albania */ + "sr_RS", /* Serbian Serbia */ + "sr_YU", /* Serbian Yugoslavia */ + "srr_SN", /* Serer Senegal */ + "suk_TZ", /* Sukuma Tanzania */ + "sus_GN", /* Susu Guinea */ + "sv_SE", /* Swedish Sweden */ + "te_IN", /* Telugu India */ + "tem_SL", /* Timne Sierra Leone */ + "tet_ID", /* Tetum Indonesia */ + "tg_TJ", /* Tajik Tajikistan */ + "th_TH", /* Thai Thailand */ + "tiv_NG", /* Tiv Nigeria */ + "tk_TM", /* Turkmen Turkmenistan */ + "tl_PH", /* Tagalog Philippines */ + "to_TO", /* Tonga Tonga */ + "tr_TR", /* Turkish Turkey */ + "tum_MW", /* Tumbuka Malawi */ + "uk_UA", /* Ukrainian Ukraine */ + "umb_AO", /* Umbundu Angola */ + "ur_PK", /* Urdu Pakistan */ + "uz_UZ", /* Uzbek Uzbekistan */ + "ve_ZA", /* Venda South Africa */ + "vi_VN", /* Vietnamese Vietnam */ + "wa_BE", /* Walloon Belgium */ + "wal_ET", /* Walamo Ethiopia */ + "war_PH", /* Waray Philippines */ + "wen_DE", /* Sorbian Germany */ + "yao_MW", /* Yao Malawi */ + "zap_MX" /* Zapotec Mexico */ + }; + int c; + int cc_len; + int ll_len = strlen (ll_code); + char *cc_code = xmalloc (ll_len + 1 + 1); + int principal_len = SIZEOF (locales_with_principal_territory); + + strcpy (cc_code, ll_code); + strcat (cc_code, "_"); + cc_len = ll_len + 1; + + for (c = 0; c < principal_len; c++) + { + const char *principal_locale = locales_with_principal_territory[c]; + if (strncmp (principal_locale, cc_code, cc_len) == 0) + { + const char *underscore = strchr (principal_locale, '_'); + /* should always be there, but in case ... */ + free (cc_code); + cc_code = xstrdup (underscore ? underscore + 1 : principal_locale); + break; + } + } + + /* If we didn't find one to copy, warn and duplicate. */ + if (c == principal_len) + { + if (mbscasecmp (ll_code, "en") == 0) + cc_code = xstrdup ("en_US"); + else + { + warning (_("no default territory known for language `%s'"), ll_code); + for (c = 0; c < strlen (ll_code); c++) + cc_code[c] = toupper (ll_code[c]); + cc_code[c] = 0; + /* We're probably wasting a byte, oops. */ + } + } + + return cc_code; +} + + + +/* @documentlanguage ARG. We set the global `document_language' (which + `getdocumenttext' uses) to the gettextish string, and + `language_code' to the corresponding enum value. */ -/* XML documents can make use of this data. Unfortunately, it clashes with - the structure currently used. So instead of enclosing content into - a language block, we just output an empty element. Anyways, a stream based - parser can make good use of it. */ void cm_documentlanguage (void) { language_code_type c; - char *lang_arg; + char *lang_arg, *ll_part, *cc_part, *locale_string; + char *underscore; /* Read the line with the language code on it. */ get_rest_of_line (0, &lang_arg); + /* What we're passed might be either just a language code LL (we'll + also need it with the usual _CC appended, if there is one), or a + locale name as used by gettext, LL_CC (we'll also need its + constituents separately). */ + underscore = strchr (lang_arg, '_'); + if (underscore) + { + ll_part = substring (lang_arg, underscore); + cc_part = xstrdup (underscore + 1); + locale_string = xstrdup (lang_arg); + } + else + { + ll_part = xstrdup (lang_arg); + cc_part = default_country_for_lang (ll_part); + locale_string = xmalloc (strlen (ll_part) + 1 + strlen (cc_part) + 1); + strcpy (locale_string, ll_part); + strcat (locale_string, "_"); + strcat (locale_string, cc_part); + } + + /* Done with analysis of user-supplied string. */ + free (lang_arg); + /* Linear search is fine these days. */ for (c = aa; c != last_language_code; c++) { - if (strcmp (lang_arg, language_table[c].abbrev) == 0) + if (strcmp (ll_part, language_table[c].abbrev) == 0) { /* Set current language code. */ language_code = c; break; @@ -577,15 +1278,30 @@ cm_documentlanguage (void) /* If we didn't find this code, complain. */ if (c == last_language_code) - warning (_("%s is not a valid ISO 639 language code"), lang_arg); + warning (_("%s is not a valid ISO 639 language code"), ll_part); + /* Set the language our `getdocumenttext' function uses for + translating document strings. */ + document_language = xstrdup (locale_string); + free (locale_string); + if (xml && !docbook) - { - xml_insert_element_with_attribute (DOCUMENTLANGUAGE, START, "xml:lang=\"%s\"", lang_arg); + { /* According to http://www.opentag.com/xfaq_lang.htm, xml:lang + takes an ISO 639 language code, optionally followed by a dash + (not underscore) and an ISO 3166 country code. So we have + to make another version with - instead of _. */ + char *xml_locale = xmalloc (strlen (ll_part) + 1 + strlen (cc_part) + 1); + strcpy (xml_locale, ll_part); + strcat (xml_locale, "-"); + strcat (xml_locale, cc_part); + xml_insert_element_with_attribute (DOCUMENTLANGUAGE, START, + "xml:lang=\"%s\"", xml_locale); xml_insert_element (DOCUMENTLANGUAGE, END); + free (xml_locale); } - free (lang_arg); + free (ll_part); + free (cc_part); } @@ -596,20 +1312,48 @@ cm_documentlanguage (void) static int cm_search_iso_map (char *html) { - int i; - iso_map_type *iso = encoding_table[document_encoding_code].isotab; - - /* If no conversion table for this encoding, quit. */ - if (!iso) - return -1; - - for (i = 0; iso[i].html; i++) + if (document_encoding_code == UTF_8) { - if (strcmp (html, iso[i].html) == 0) - return i; + /* Binary search in unicode_map. */ + size_t low = 0; + size_t high = sizeof (unicode_map) / sizeof (unicode_map[0]); + + /* At each loop iteration, low < high; for indices < low the values are + smaller than HTML; for indices >= high the values are greater than HTML. + So, if HTML occurs in the list, it is at low <= position < high. */ + do + { + size_t mid = low + (high - low) / 2; /* low <= mid < high */ + int cmp = strcmp (unicode_map[mid].html, html); + + if (cmp < 0) + low = mid + 1; + else if (cmp > 0) + high = mid; + else /* cmp == 0 */ + return unicode_map[mid].unicode; + } + while (low < high); + + return -1; } + else + { + int i; + iso_map_type *iso = encoding_table[document_encoding_code].isotab; + + /* If no conversion table for this encoding, quit. */ + if (!iso) + return -1; + + for (i = 0; iso[i].html; i++) + { + if (strcmp (html, iso[i].html) == 0) + return i; + } - return -1; + return -1; + } } @@ -642,7 +1386,7 @@ cm_documentencoding (void) /* See if we have this encoding. */ for (enc = no_encoding+1; enc != last_encoding_code; enc++) { - if (strcasecmp (enc_arg, encoding_table[enc].encname) == 0) + if (mbscasecmp (enc_arg, encoding_table[enc].encname) == 0) { document_encoding_code = enc; break; @@ -691,6 +1435,39 @@ current_document_encoding (void) } +/* Add RC per the current encoding. */ + +static void +add_encoded_char_from_code (int rc) +{ + if (document_encoding_code == UTF_8) + { + if (rc < 0x80) + add_char (rc); + else if (rc < 0x800) + { + add_char (0xc0 | (rc >> 6)); + add_char (0x80 | (rc & 0x3f)); + } + else if (rc < 0x10000) + { + add_char (0xe0 | (rc >> 12)); + add_char (0x80 | ((rc >> 6) & 0x3f)); + add_char (0x80 | (rc & 0x3f)); + } + else + { + add_char (0xf0 | (rc >> 18)); + add_char (0x80 | ((rc >> 12) & 0x3f)); + add_char (0x80 | ((rc >> 6) & 0x3f)); + add_char (0x80 | (rc & 0x3f)); + } + } + else + add_char (encoding_table[document_encoding_code].isotab[rc].bytecode); +} + + /* If html or xml output, add &HTML_STR; to the output. If not html and the user requested encoded output, add the real 8-bit character corresponding to HTML_STR from the translation tables. Otherwise, @@ -703,13 +1480,13 @@ add_encoded_char (char *html_str, char *info_str) add_word_args ("&%s;", html_str); else if (xml) xml_insert_entity (html_str); - else if (enable_encoding) + else if (enable_encoding && document_encoding_code != no_encoding) { /* Look for HTML_STR in the current translation table. */ int rc = cm_search_iso_map (html_str); if (rc >= 0) /* We found it, add the real character. */ - add_char (encoding_table[document_encoding_code].isotab[rc].bytecode); + add_encoded_char_from_code (rc); else { /* We didn't find it, that seems bad. */ warning (_("invalid encoded character `%s'"), html_str); @@ -789,11 +1566,16 @@ cm_accent_generic_html (int arg, int start, int end, char *html_supported, } +/* If END is zero, there is nothing in the paragraph to accent. This + can happen when we're in a menu with an accent command and + --no-headers is given, so the base character is not added. In this + case we're not producing any output anyway, so just forget it. + Otherwise, produce the ASCII version of the accented char. */ static void cm_accent_generic_no_headers (int arg, int start, int end, int single, char *html_solo) { - if (arg == END) + if (arg == END && end > 0) { if (no_encoding) add_char (single); @@ -801,19 +1583,20 @@ cm_accent_generic_no_headers (int arg, int start, int end, int single, { int rc; char *buffer = xmalloc (1 + strlen (html_solo) + 1); + assert (end > 0); buffer[0] = output_paragraph[end - 1]; buffer[1] = 0; strcat (buffer, html_solo); rc = cm_search_iso_map (buffer); if (rc >= 0) - /* A little bit tricky ;-) - Here we replace the character which has - been inserted in read_command with - the value we have found in converting table - Does there exist a better way to do this? kama. */ - output_paragraph[end - 1] - = encoding_table[document_encoding_code].isotab[rc].bytecode; + { + /* Here we replace the character which has + been inserted in read_command with + the value we have found in the conversion table. */ + output_paragraph_offset--; + add_encoded_char_from_code (rc); + } else { /* If we didn't find a translation for this character, put the single instead. E.g., &Xuml; does not exist so X¨ @@ -976,6 +1759,8 @@ cm_special_char (int arg) add_encoded_char ("ordf" , "a"); else if (strcmp (command, "ordm") == 0) add_encoded_char ("ordm" , "o"); + else if (strcmp (command, "textdegree") == 0) + add_encoded_char ("deg" , "o"); else if (strcmp (command, "AE") == 0) add_encoded_char ("AElig", command); else if (strcmp (command, "ae") == 0) @@ -990,6 +1775,12 @@ cm_special_char (int arg) add_encoded_char ("aring", command); else if (strcmp (command, "ss") == 0) add_encoded_char ("szlig", command); + else if (strcmp (command, "guillemetleft") == 0 + || strcmp (command, "guillemotleft") == 0) + add_encoded_char ("laquo", "<<"); + else if (strcmp (command, "guillemetright") == 0 + || strcmp (command, "guillemotright") == 0) + add_encoded_char ("raquo", ">>"); else line_error ("cm_special_char internal error: command=@%s", command); } |