annotate src/charset.h @ 18288:1e31ff4722f7

Added [ediff-misc] to menu-bar-tools-menu
author Michael Kifer <kifer@cs.stonybrook.edu>
date Wed, 18 Jun 1997 00:26:54 +0000
parents 59aa4a0772f6
children 33e78cc7f058
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
17052
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
1 /* Header for multilingual character handler.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
2 Ver.1.0
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
3 Copyright (C) 1995 Free Software Foundation, Inc.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
4 Copyright (C) 1995 Electrotechnical Laboratory, JAPAN.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
5
17071
70194012fb3a Fix FSF address in comment.
Karl Heuer <kwzh@gnu.org>
parents: 17052
diff changeset
6 This file is part of GNU Emacs.
70194012fb3a Fix FSF address in comment.
Karl Heuer <kwzh@gnu.org>
parents: 17052
diff changeset
7
70194012fb3a Fix FSF address in comment.
Karl Heuer <kwzh@gnu.org>
parents: 17052
diff changeset
8 GNU Emacs is free software; you can redistribute it and/or modify
70194012fb3a Fix FSF address in comment.
Karl Heuer <kwzh@gnu.org>
parents: 17052
diff changeset
9 it under the terms of the GNU General Public License as published by
70194012fb3a Fix FSF address in comment.
Karl Heuer <kwzh@gnu.org>
parents: 17052
diff changeset
10 the Free Software Foundation; either version 2, or (at your option)
70194012fb3a Fix FSF address in comment.
Karl Heuer <kwzh@gnu.org>
parents: 17052
diff changeset
11 any later version.
17052
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
12
17071
70194012fb3a Fix FSF address in comment.
Karl Heuer <kwzh@gnu.org>
parents: 17052
diff changeset
13 GNU Emacs is distributed in the hope that it will be useful,
70194012fb3a Fix FSF address in comment.
Karl Heuer <kwzh@gnu.org>
parents: 17052
diff changeset
14 but WITHOUT ANY WARRANTY; without even the implied warranty of
70194012fb3a Fix FSF address in comment.
Karl Heuer <kwzh@gnu.org>
parents: 17052
diff changeset
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
70194012fb3a Fix FSF address in comment.
Karl Heuer <kwzh@gnu.org>
parents: 17052
diff changeset
16 GNU General Public License for more details.
17052
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
17
17071
70194012fb3a Fix FSF address in comment.
Karl Heuer <kwzh@gnu.org>
parents: 17052
diff changeset
18 You should have received a copy of the GNU General Public License
70194012fb3a Fix FSF address in comment.
Karl Heuer <kwzh@gnu.org>
parents: 17052
diff changeset
19 along with GNU Emacs; see the file COPYING. If not, write to
70194012fb3a Fix FSF address in comment.
Karl Heuer <kwzh@gnu.org>
parents: 17052
diff changeset
20 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
70194012fb3a Fix FSF address in comment.
Karl Heuer <kwzh@gnu.org>
parents: 17052
diff changeset
21 Boston, MA 02111-1307, USA. */
17052
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
22
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
23 #ifndef _CHARSET_H
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
24 #define _CHARSET_H
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
25
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
26 /*** GENERAL NOTE on CHARACTER SET (CHARSET) ***
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
27
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
28 A character set ("charset" hereafter) is a meaningful collection
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
29 (i.e. language, culture, functionality, etc) of characters. Emacs
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
30 handles multiple charsets at once. Each charset corresponds to one
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
31 of ISO charsets (except for a special charset for composition
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
32 characters). Emacs identifies a charset by a unique identification
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
33 number, whereas ISO identifies a charset by a triplet of DIMENSION,
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
34 CHARS and FINAL-CHAR. So, hereafter, just saying "charset" means an
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
35 identification number (integer value).
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
36
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
37 The value range of charset is 0x00, 0x80..0xFE. There are four
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
38 kinds of charset depending on DIMENSION (1 or 2) and CHARS (94 or
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
39 96). For instance, a charset of DIMENSION2_CHARS94 contains 94x94
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
40
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
41
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
42 Within Emacs Lisp, a charset is treated as a symbol which has a
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
43 property `charset'. The property value is a vector containing
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
44 various information about the charset. For readability of C codes,
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
45 we use the following convention on C variable names:
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
46 charset_symbol: Emacs Lisp symbol of a charset
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
47 charset_id: Emacs Lisp integer of an identification number of a charset
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
48 charset: C integer of an identification number of a charset
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
49
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
50 Each charset (except for ASCII) is assigned a base leading-code
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
51 (range 0x80..0x9D). In addition, a charset of greater than 0xA0
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
52 (whose base leading-code is 0x9A..0x9D) is assigned an extended
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
53 leading-code (range 0xA0..0xFE). In this case, each base
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
54 leading-code specify the allowable range of extended leading-code as
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
55 shown in the table below. A leading-code is used to represent a
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
56 character in Emacs' buffer and string.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
57
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
58 We call a charset which has extended leading-code as "private
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
59 charset" because those are mainly for a charset which is not
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
60 registered by ISO. On the contrary, we call a charset which does
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
61 not have extended leading-code as "official charset".
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
62
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
63 ---------------------------------------------------------------------------
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
64 charset dimension base leading-code extended leading-code
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
65 ---------------------------------------------------------------------------
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
66 0x00 official dim1 -- none -- -- none --
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
67 (ASCII)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
68 0x01..0x7F --never used--
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
69 0x80 COMPOSITION same as charset -- none --
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
70 0x81..0x8F official dim1 same as charset -- none --
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
71 0x90..0x99 official dim2 same as charset -- none --
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
72 0x9A..0x9F --never used--
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
73 0xA0..0xDF private dim1 0x9A same as charset
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
74 of 1-column width
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
75 0xE0..0xEF private dim1 0x9B same as charset
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
76 of 2-column width
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
77 0xF0..0xF4 private dim2 0x9C same as charset
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
78 of 1-column width
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
79 0xF5..0xFE private dim2 0x9D same as charset
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
80 of 2-column width
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
81 0xFF --never used--
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
82 ---------------------------------------------------------------------------
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
83
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
84 In the table, "COMPOSITION" means a charset for a composite
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
85 character which is a character composed from several (up to 16)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
86 non-composite characters (components). Although a composite
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
87 character can contain components of many charsets, a composite
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
88 character itself belongs to the charset CHARSET-COMPOSITION. See
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
89 the document "GENERAL NOTE on COMPOSITE CHARACTER" below for more
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
90 detail.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
91
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
92 */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
93
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
94 /* Definition of special leading-codes. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
95 /* Base leading-code. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
96 /* Special leading-code followed by components of a composite character. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
97 #define LEADING_CODE_COMPOSITION 0x80
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
98 /* Leading-code followed by extended leading-code. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
99 #define LEADING_CODE_PRIVATE_11 0x9A /* for private DIMENSION1 of 1-column */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
100 #define LEADING_CODE_PRIVATE_12 0x9B /* for private DIMENSION1 of 2-column */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
101 #define LEADING_CODE_PRIVATE_21 0x9C /* for private DIMENSION2 of 1-column */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
102 #define LEADING_CODE_PRIVATE_22 0x9D /* for private DIMENSION2o f 2-column */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
103
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
104 /* Extended leading-code. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
105 /* Start of each extended leading-codes. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
106 #define LEADING_CODE_EXT_11 0xA0 /* follows LEADING_CODE_PRIVATE_11 */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
107 #define LEADING_CODE_EXT_12 0xE0 /* follows LEADING_CODE_PRIVATE_12 */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
108 #define LEADING_CODE_EXT_21 0xF0 /* follows LEADING_CODE_PRIVATE_21 */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
109 #define LEADING_CODE_EXT_22 0xF5 /* follows LEADING_CODE_PRIVATE_22 */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
110 /* Maximum value of extended leading-codes. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
111 #define LEADING_CODE_EXT_MAX 0xFE
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
112
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
113 /* Definition of minimum/maximum charset of each DIMENSION. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
114 #define MIN_CHARSET_OFFICIAL_DIMENSION1 0x81
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
115 #define MAX_CHARSET_OFFICIAL_DIMENSION1 0x8F
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
116 #define MIN_CHARSET_OFFICIAL_DIMENSION2 0x90
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
117 #define MAX_CHARSET_OFFICIAL_DIMENSION2 0x99
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
118 #define MIN_CHARSET_PRIVATE_DIMENSION1 LEADING_CODE_EXT_11
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
119 #define MIN_CHARSET_PRIVATE_DIMENSION2 LEADING_CODE_EXT_21
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
120
17185
0d5a1bae9d73 (MAX_CHARSET): Definition changed to the actual
Kenichi Handa <handa@m17n.org>
parents: 17120
diff changeset
121 /* Maximum value of overall charset identification number. */
0d5a1bae9d73 (MAX_CHARSET): Definition changed to the actual
Kenichi Handa <handa@m17n.org>
parents: 17120
diff changeset
122 #define MAX_CHARSET 0xFE
0d5a1bae9d73 (MAX_CHARSET): Definition changed to the actual
Kenichi Handa <handa@m17n.org>
parents: 17120
diff changeset
123
17052
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
124 /* Definition of special charsets. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
125 #define CHARSET_ASCII 0
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
126 #define CHARSET_COMPOSITION 0x80
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
127
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
128 extern int charset_ascii; /* ASCII */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
129 extern int charset_composition; /* for a composite character */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
130 extern int charset_latin_iso8859_1; /* ISO8859-1 (Latin-1) */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
131 extern int charset_jisx0208_1978; /* JISX0208.1978 (Japanese Kanji old set) */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
132 extern int charset_jisx0208; /* JISX0208.1983 (Japanese Kanji) */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
133 extern int charset_katakana_jisx0201; /* JISX0201.Kana (Japanese Katakana) */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
134 extern int charset_latin_jisx0201; /* JISX0201.Roman (Japanese Roman) */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
135 extern int charset_big5_1; /* Big5 Level 1 (Chinese Traditional) */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
136 extern int charset_big5_2; /* Big5 Level 2 (Chinese Traditional) */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
137
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
138 /* Check if STR points the head of multi-byte form, i.e. *STR is an
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
139 ASCII character or a base leading-code. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
140 #define CHAR_HEAD_P(str) ((unsigned char) *(str) < 0xA0)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
141
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
142 /*** GENERAL NOTE on CHARACTER REPRESENTATION ***
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
143
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
144 At first, the term "character" or "char" is used for a multilingual
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
145 character (of course, including ASCII character), not for a byte in
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
146 computer memory. We use the term "code" or "byte" for the latter
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
147 case.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
148
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
149 A character is identified by charset and one or two POSITION-CODEs.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
150 POSITION-CODE is the position of the character in the charset. A
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
151 character of DIMENSION1 charset has one POSITION-CODE: POSITION-CODE-1.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
152 A character of DIMENSION2 charset has two POSITION-CODE:
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
153 POSITION-CODE-1 and POSITION-CODE-2. The code range of
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
154 POSITION-CODE is 0x20..0x7F.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
155
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
156 Emacs has two kinds of representation of a character: multi-byte
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
157 form (for buffer and string) and single-word form (for character
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
158 object in Emacs Lisp). The latter is called "character code" here
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
159 after. Both representation encode the information of charset and
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
160 POSITION-CODE but in a different way (for instance, MSB of
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
161 POSITION-CODE is set in multi-byte form).
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
162
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
163 For details of multi-byte form, see the section "2. Emacs internal
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
164 format handlers" of `coding.c'.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
165
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
166 Emacs uses 19 bits for a character code. The bits are divided into
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
167 3 fields: FIELD1(5bits):FIELD2(7bits):FIELD3(7bits).
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
168
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
169 A character code of DIMENSION1 character uses FIELD2 to hold charset
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
170 and FIELD3 to hold POSITION-CODE-1. A character code of DIMENSION2
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
171 character uses FIELD1 to hold charset, FIELD2 and FIELD3 to hold
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
172 POSITION-CODE-1 and POSITION-CODE-2 respectively.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
173
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
174 More precisely...
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
175
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
176 FIELD2 of DIMENSION1 character (except for ASCII) is "charset - 0x70".
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
177 This is to make all character codes except for ASCII greater than
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
178 256 (ASCII's FIELD2 is 0). So, the range of FIELD2 of DIMENSION1
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
179 character is 0 or 0x11..0x7F.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
180
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
181 FIELD1 of DIMENSION2 character is "charset - 0x8F" for official
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
182 charset and "charset - 0xE0" for private charset. So, the range of
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
183 FIELD1 of DIMENSION2 character is 0x01..0x1E.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
184
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
185 -----------------------------------------------------------------------
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
186 charset FIELD1 (5-bit) FIELD2 (7-bit) FIELD3 (7-bit)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
187 -----------------------------------------------------------------------
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
188 ASCII 0 0 POSITION-CODE-1
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
189 DIMENSION1 0 charset - 0x70 POSITION-CODE-1
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
190 DIMENSION2(o) charset - 0x8F POSITION-CODE-1 POSITION-CODE-2
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
191 DIMENSION2(p) charset - 0xE0 POSITION-CODE-1 POSITION-CODE-2
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
192 -----------------------------------------------------------------------
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
193 "(o)": official, "(p)": private
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
194 -----------------------------------------------------------------------
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
195
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
196 */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
197
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
198 /*** GENERAL NOTE on COMPOSITE CHARACTER ***
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
199
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
200 A composite character is a character composed from several (up to
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
201 16) non-composite characters (components). Although each components
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
202 can belong to any charset, a composite character itself belongs to
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
203 the charset `charset-composition' and is assigned a special
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
204 leading-code `LEADING_CODE_COMPOSITION' for multi-byte form. See
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
205 the document "2. Emacs internal format handlers" in `coding.c' for
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
206 more detail about multi-byte form.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
207
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
208 A character code of composite character has special format. In the
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
209 above document, FIELD1 of a composite character is 0x1F. Each
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
210 composite character is assigned a sequential number CMPCHAR-ID.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
211 FIELD2 and FIELD3 are combined to make 14bits field for holding
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
212 CMPCHAR-ID, which means that Emacs can handle at most 2^14 (= 16384)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
213 composite characters at once.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
214
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
215 -----------------------------------------------------------------------
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
216 charset FIELD1 (5-bit) FIELD2&3 (14-bit)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
217 -----------------------------------------------------------------------
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
218 CHARSET-COMPOSITION 0x1F CMPCHAR-ID
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
219 -----------------------------------------------------------------------
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
220
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
221 Emacs assigns CMPCHAR-ID to a composite character only when it
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
222 requires the character code of the composite character (e.g. while
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
223 displaying the composite character).
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
224
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
225 */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
226
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
227 /* Masks of each field of character code. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
228 #define CHAR_FIELD1_MASK (0x1F << 14)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
229 #define CHAR_FIELD2_MASK (0x7F << 7)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
230 #define CHAR_FIELD3_MASK 0x7F
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
231
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
232 /* Macros to access each field of character C. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
233 #define CHAR_FIELD1(c) (((c) & CHAR_FIELD1_MASK) >> 14)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
234 #define CHAR_FIELD2(c) (((c) & CHAR_FIELD2_MASK) >> 7)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
235 #define CHAR_FIELD3(c) ((c) & CHAR_FIELD3_MASK)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
236
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
237 /* Minimum character code of character of each DIMENSION. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
238 #define MIN_CHAR_OFFICIAL_DIMENSION1 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
239 ((MIN_CHARSET_OFFICIAL_DIMENSION1 - 0x70) << 7)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
240 #define MIN_CHAR_PRIVATE_DIMENSION1 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
241 ((MIN_CHARSET_PRIVATE_DIMENSION1 - 0x70) << 7)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
242 #define MIN_CHAR_OFFICIAL_DIMENSION2 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
243 ((MIN_CHARSET_OFFICIAL_DIMENSION2 - 0x8F) << 14)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
244 #define MIN_CHAR_PRIVATE_DIMENSION2 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
245 ((MIN_CHARSET_PRIVATE_DIMENSION2 - 0xE0) << 14)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
246 #define MIN_CHAR_COMPOSITION \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
247 (0x1F << 14)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
248
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
249 /* 1 if C is an ASCII character, else 0. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
250 #define SINGLE_BYTE_CHAR_P(c) ((c) < 0x100)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
251 /* 1 if C is an composite character, else 0. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
252 #define COMPOSITE_CHAR_P(c) ((c) >= MIN_CHAR_COMPOSITION)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
253
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
254 /* A char-table containing information of each character set.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
255
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
256 Unlike ordinary char-tables, this doesn't contain any nested table.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
257 Only the top level elements are used. Each element is a vector of
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
258 the following information:
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
259 CHARSET-ID, BYTES, DIMENSION, CHARS, WIDTH, DIRECTION,
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
260 LEADING-CODE-BASE, LEADING-CODE-EXT,
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
261 ISO-FINAL-CHAR, ISO-GRAPHIC-PLANE,
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
262 REVERSE-CHARSET, SHORT-NAME, LONG-NAME, DESCRIPTION,
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
263 PLIST.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
264
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
265 CHARSET-ID (integer) is the identification number of the charset.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
266
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
267 BYTE (integer) is the length of multi-byte form of a character in
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
268 the charset: one of 1, 2, 3, and 4.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
269
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
270 DIMENSION (integer) is the number of bytes to represent a character: 1 or 2.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
271
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
272 CHARS (integer) is the number of characters in a dimension: 94 or 96.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
273
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
274 WIDTH (integer) is the number of columns a character in the charset
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
275 occupies on the screen: one of 0, 1, and 2.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
276
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
277 DIRECTION (integer) is the rendering direction of characters in the
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
278 charset when rendering. If 0, render from right to left, else
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
279 render from left to right.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
280
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
281 LEADING-CODE-BASE (integer) is the base leading-code for the
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
282 charset.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
283
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
284 LEADING-CODE-EXT (integer) is the extended leading-code for the
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
285 charset. All charsets of less than 0xA0 has the value 0.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
286
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
287 ISO-FINAL-CHAR (character) is the final character of the
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
288 corresponding ISO 2022 charset.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
289
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
290 ISO-GRAPHIC-PLANE (integer) is the graphic plane to be invoked
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
291 while encoding to variants of ISO 2022 coding system, one of the
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
292 following: 0/graphic-plane-left(GL), 1/graphic-plane-right(GR).
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
293
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
294 REVERSE-CHARSET (integer) is the charset which differs only in
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
295 LEFT-TO-RIGHT value from the charset. If there's no such a
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
296 charset, the value is -1.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
297
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
298 SHORT-NAME (string) is the short name to refer to the charset.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
299
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
300 LONG-NAME (string) is the long name to refer to the charset.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
301
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
302 DESCRIPTION (string) is the description string of the charset.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
303
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
304 PLIST (property list) may contain any type of information a user
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
305 want to put and get by functions `put-charset-property' and
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
306 `get-charset-property' respectively. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
307 extern Lisp_Object Vcharset_table;
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
308
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
309 /* Macros to access various information of CHARSET in Vcharset_table.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
310 We provide these macros for efficiency. No range check of CHARSET. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
311
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
312 /* Return entry of CHARSET (lisp integer) in Vcharset_table. */
17321
9f837bea89e3 (CHARSET_TABLE_ENTRY): Handle ASCII charset correctly.
Kenichi Handa <handa@m17n.org>
parents: 17185
diff changeset
313 #define CHARSET_TABLE_ENTRY(charset) \
9f837bea89e3 (CHARSET_TABLE_ENTRY): Handle ASCII charset correctly.
Kenichi Handa <handa@m17n.org>
parents: 17185
diff changeset
314 XCHAR_TABLE (Vcharset_table)->contents[((charset) == CHARSET_ASCII \
9f837bea89e3 (CHARSET_TABLE_ENTRY): Handle ASCII charset correctly.
Kenichi Handa <handa@m17n.org>
parents: 17185
diff changeset
315 ? 0 : (charset) + 128)]
17052
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
316
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
317 /* Return information INFO-IDX of CHARSET. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
318 #define CHARSET_TABLE_INFO(charset, info_idx) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
319 XVECTOR (CHARSET_TABLE_ENTRY (charset))->contents[info_idx]
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
320
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
321 #define CHARSET_ID_IDX (0)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
322 #define CHARSET_BYTES_IDX (1)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
323 #define CHARSET_DIMENSION_IDX (2)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
324 #define CHARSET_CHARS_IDX (3)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
325 #define CHARSET_WIDTH_IDX (4)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
326 #define CHARSET_DIRECTION_IDX (5)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
327 #define CHARSET_LEADING_CODE_BASE_IDX (6)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
328 #define CHARSET_LEADING_CODE_EXT_IDX (7)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
329 #define CHARSET_ISO_FINAL_CHAR_IDX (8)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
330 #define CHARSET_ISO_GRAPHIC_PLANE_IDX (9)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
331 #define CHARSET_REVERSE_CHARSET_IDX (10)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
332 #define CHARSET_SHORT_NAME_IDX (11)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
333 #define CHARSET_LONG_NAME_IDX (12)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
334 #define CHARSET_DESCRIPTION_IDX (13)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
335 #define CHARSET_PLIST_IDX (14)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
336 /* Size of a vector of each entry of Vcharset_table. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
337 #define CHARSET_MAX_IDX (15)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
338
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
339 /* And several more macros to be used frequently. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
340 #define CHARSET_BYTES(charset) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
341 XFASTINT (CHARSET_TABLE_INFO (charset, CHARSET_BYTES_IDX))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
342 #define CHARSET_DIMENSION(charset) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
343 XFASTINT (CHARSET_TABLE_INFO (charset, CHARSET_DIMENSION_IDX))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
344 #define CHARSET_CHARS(charset) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
345 XFASTINT (CHARSET_TABLE_INFO (charset, CHARSET_CHARS_IDX))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
346 #define CHARSET_WIDTH(charset) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
347 XFASTINT (CHARSET_TABLE_INFO (charset, CHARSET_WIDTH_IDX))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
348 #define CHARSET_DIRECTION(charset) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
349 XFASTINT (CHARSET_TABLE_INFO (charset, CHARSET_DIRECTION_IDX))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
350 #define CHARSET_LEADING_CODE_BASE(charset) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
351 XFASTINT (CHARSET_TABLE_INFO (charset, CHARSET_LEADING_CODE_BASE_IDX))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
352 #define CHARSET_LEADING_CODE_EXT(charset) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
353 XFASTINT (CHARSET_TABLE_INFO (charset, CHARSET_LEADING_CODE_EXT_IDX))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
354 #define CHARSET_ISO_FINAL_CHAR(charset) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
355 XFASTINT (CHARSET_TABLE_INFO (charset, CHARSET_ISO_FINAL_CHAR_IDX))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
356 #define CHARSET_ISO_GRAPHIC_PLANE(charset) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
357 XFASTINT (CHARSET_TABLE_INFO (charset, CHARSET_ISO_GRAPHIC_PLANE_IDX))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
358 #define CHARSET_REVERSE_CHARSET(charset) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
359 XINT (CHARSET_TABLE_INFO (charset, CHARSET_REVERSE_CHARSET_IDX))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
360
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
361 /* Macros to specify direction of a charset. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
362 #define CHARSET_DIRECTION_LEFT_TO_RIGHT 0
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
363 #define CHARSET_DIRECTION_RIGHT_TO_LEFT 1
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
364
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
365 /* A vector of charset symbol indexed by charset-id. This is used
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
366 only for returning charset symbol from C functions. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
367 extern Lisp_Object Vcharset_symbol_table;
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
368
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
369 /* Return symbol of CHARSET. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
370 #define CHARSET_SYMBOL(charset) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
371 XVECTOR (Vcharset_symbol_table)->contents[charset]
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
372
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
373 /* 1 if CHARSET is valid, else 0. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
374 #define CHARSET_VALID_P(charset) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
375 ((charset) == 0 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
376 || ((charset) >= 0x80 && (charset) <= MAX_CHARSET_OFFICIAL_DIMENSION2) \
17185
0d5a1bae9d73 (MAX_CHARSET): Definition changed to the actual
Kenichi Handa <handa@m17n.org>
parents: 17120
diff changeset
377 || ((charset) >= MIN_CHARSET_PRIVATE_DIMENSION1 && (charset) <= MAX_CHARSET))
17052
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
378
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
379 /* 1 if CHARSET is already defined, else 0. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
380 #define CHARSET_DEFINED_P(charset) \
17185
0d5a1bae9d73 (MAX_CHARSET): Definition changed to the actual
Kenichi Handa <handa@m17n.org>
parents: 17120
diff changeset
381 (((charset) >= 0) && ((charset) <= MAX_CHARSET) \
17052
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
382 && !NILP (CHARSET_TABLE_ENTRY (charset)))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
383
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
384 /* Since the information CHARSET-BYTES and CHARSET-WIDTH of
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
385 Vcharset_table can be retrieved only from the first byte of
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
386 multi-byte form (an ASCII code or a base leading-code), we provide
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
387 here tables to be used by macros BYTES_BY_CHAR_HEAD and
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
388 WIDTH_BY_CHAR_HEAD for faster information retrieval. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
389 extern int bytes_by_char_head[256];
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
390 extern int width_by_char_head[256];
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
391
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
392 #define BYTES_BY_CHAR_HEAD(char_head) bytes_by_char_head[char_head]
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
393 #define WIDTH_BY_CHAR_HEAD(char_head) width_by_char_head[char_head]
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
394
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
395 /* Charset of the character C. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
396 #define CHAR_CHARSET(c) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
397 (SINGLE_BYTE_CHAR_P (c) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
398 ? CHARSET_ASCII \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
399 : ((c) < MIN_CHAR_OFFICIAL_DIMENSION2 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
400 ? CHAR_FIELD2 (c) + 0x70 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
401 : ((c) < MIN_CHAR_PRIVATE_DIMENSION2 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
402 ? CHAR_FIELD1 (c) + 0x8F \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
403 : ((c) < MIN_CHAR_COMPOSITION \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
404 ? CHAR_FIELD1 (c) + 0xE0 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
405 : CHARSET_COMPOSITION))))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
406
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
407 /* Return charset at the place pointed by P. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
408 #define CHARSET_AT(p) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
409 (*(p) < 0x80 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
410 ? CHARSET_ASCII \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
411 : (*(p) == LEADING_CODE_COMPOSITION \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
412 ? CHARSET_COMPOSITION \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
413 : (*(p) < LEADING_CODE_PRIVATE_11 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
414 ? (int)*(p) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
415 : (*(p) <= LEADING_CODE_PRIVATE_22 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
416 ? (int)*((p) + 1) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
417 : -1))))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
418
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
419 /* Same as `CHARSET_AT ()' but perhaps runs faster because of an
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
420 additional argument C which is the code (byte) at P. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
421 #define FIRST_CHARSET_AT(p, c) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
422 ((c) < 0x80 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
423 ? CHARSET_ASCII \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
424 : ((c) == LEADING_CODE_COMPOSITION \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
425 ? CHARSET_COMPOSITION \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
426 : ((c) < LEADING_CODE_PRIVATE_11 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
427 ? (int)(c) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
428 : ((c) <= LEADING_CODE_PRIVATE_22 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
429 ? (int)*((p) + 1) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
430 : -1))))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
431
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
432 /* Check if two characters C1 and C2 belong to the same charset.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
433 Always return 0 for composite characters. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
434 #define SAME_CHARSET_P(c1, c2) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
435 (c1 < MIN_CHAR_COMPOSITION \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
436 && (SINGLE_BYTE_CHAR_P (c1) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
437 ? SINGLE_BYTE_CHAR_P (c2) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
438 : (c1 < MIN_CHAR_OFFICIAL_DIMENSION2 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
439 ? (c1 & CHAR_FIELD2_MASK) == (c2 & CHAR_FIELD2_MASK) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
440 : (c1 & CHAR_FIELD1_MASK) == (c2 & CHAR_FIELD1_MASK))))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
441
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
442 /* Return a non-ASCII character of which charset is CHARSET and
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
443 position-codes are C1 and C2. DIMENSION1 character ignores C2. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
444 #define MAKE_NON_ASCII_CHAR(charset, c1, c2) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
445 ((charset) == CHARSET_COMPOSITION \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
446 ? MAKE_COMPOSITE_CHAR (((c1) << 7) + (c2)) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
447 : (CHARSET_DIMENSION (charset) == 1 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
448 ? (((charset) - 0x70) << 7) | (c1) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
449 : ((charset) < MIN_CHARSET_PRIVATE_DIMENSION2 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
450 ? (((charset) - 0x8F) << 14) | ((c1) << 7) | (c2) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
451 : (((charset) - 0xE0) << 14) | ((c1) << 7) | (c2))))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
452
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
453 /* Return a composite character of which CMPCHAR-ID is ID. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
454 #define MAKE_COMPOSITE_CHAR(id) (MIN_CHAR_COMPOSITION + (id))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
455
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
456 /* Return CMPCHAR-ID of a composite character C. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
457 #define COMPOSITE_CHAR_ID(c) ((c) - MIN_CHAR_COMPOSITION)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
458
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
459 /* Return a character of which charset is CHARSET and position-codes
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
460 are C1 and C2. DIMENSION1 character ignores C2. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
461 #define MAKE_CHAR(charset, c1, c2) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
462 ((charset) == CHARSET_ASCII \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
463 ? (c1) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
464 : MAKE_NON_ASCII_CHAR ((charset), (c1) & 0x7F, (c2) & 0x7F))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
465
17833
59aa4a0772f6 (VALID_CHAR_P): New macro.
Kenichi Handa <handa@m17n.org>
parents: 17726
diff changeset
466 /* 1 if C is in the range of possible character code Emacs can have. */
59aa4a0772f6 (VALID_CHAR_P): New macro.
Kenichi Handa <handa@m17n.org>
parents: 17726
diff changeset
467 #define VALID_CHAR_P(c) \
59aa4a0772f6 (VALID_CHAR_P): New macro.
Kenichi Handa <handa@m17n.org>
parents: 17726
diff changeset
468 ((c) >= 0 \
59aa4a0772f6 (VALID_CHAR_P): New macro.
Kenichi Handa <handa@m17n.org>
parents: 17726
diff changeset
469 && (SINGLE_BYTE_CHAR_P (c) \
59aa4a0772f6 (VALID_CHAR_P): New macro.
Kenichi Handa <handa@m17n.org>
parents: 17726
diff changeset
470 || ((c) < MIN_CHAR_COMPOSITION \
59aa4a0772f6 (VALID_CHAR_P): New macro.
Kenichi Handa <handa@m17n.org>
parents: 17726
diff changeset
471 ? ((c) & CHAR_FIELD1_MASK \
59aa4a0772f6 (VALID_CHAR_P): New macro.
Kenichi Handa <handa@m17n.org>
parents: 17726
diff changeset
472 ? (CHAR_FIELD2 (c) >= 32 && CHAR_FIELD3 (c) >= 32) \
59aa4a0772f6 (VALID_CHAR_P): New macro.
Kenichi Handa <handa@m17n.org>
parents: 17726
diff changeset
473 : (CHAR_FIELD2 (c) >= 16 && CHAR_FIELD3 (c) >= 32)) \
59aa4a0772f6 (VALID_CHAR_P): New macro.
Kenichi Handa <handa@m17n.org>
parents: 17726
diff changeset
474 : (c) < MIN_CHAR_COMPOSITION + n_cmpchars)))
59aa4a0772f6 (VALID_CHAR_P): New macro.
Kenichi Handa <handa@m17n.org>
parents: 17726
diff changeset
475
17052
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
476 /* The charset of non-ASCII character C is set to CHARSET, and the
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
477 position-codes of C are set to C1 and C2. C2 of DIMENSION1 character
17321
9f837bea89e3 (CHARSET_TABLE_ENTRY): Handle ASCII charset correctly.
Kenichi Handa <handa@m17n.org>
parents: 17185
diff changeset
478 is -1. */
17052
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
479 #define SPLIT_NON_ASCII_CHAR(c, charset, c1, c2) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
480 ((c) < MIN_CHAR_OFFICIAL_DIMENSION2 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
481 ? (charset = CHAR_FIELD2 (c) + 0x70, \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
482 c1 = CHAR_FIELD3 (c), \
17321
9f837bea89e3 (CHARSET_TABLE_ENTRY): Handle ASCII charset correctly.
Kenichi Handa <handa@m17n.org>
parents: 17185
diff changeset
483 c2 = -1) \
17052
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
484 : (charset = ((c) < MIN_CHAR_COMPOSITION \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
485 ? (CHAR_FIELD1 (c) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
486 + ((c) < MIN_CHAR_PRIVATE_DIMENSION2 ? 0x8F : 0xE0)) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
487 : CHARSET_COMPOSITION), \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
488 c1 = CHAR_FIELD2 (c), \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
489 c2 = CHAR_FIELD3 (c)))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
490
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
491 /* The charset of character C is set to CHARSET, and the
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
492 position-codes of C are set to C1 and C2. C2 of DIMENSION1 character
17321
9f837bea89e3 (CHARSET_TABLE_ENTRY): Handle ASCII charset correctly.
Kenichi Handa <handa@m17n.org>
parents: 17185
diff changeset
493 is -1. */
17052
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
494 #define SPLIT_CHAR(c, charset, c1, c2) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
495 (SINGLE_BYTE_CHAR_P (c) \
17321
9f837bea89e3 (CHARSET_TABLE_ENTRY): Handle ASCII charset correctly.
Kenichi Handa <handa@m17n.org>
parents: 17185
diff changeset
496 ? charset = CHARSET_ASCII, c1 = (c), c2 = -1 \
17052
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
497 : SPLIT_NON_ASCII_CHAR (c, charset, c1, c2))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
498
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
499 /* The charset of the character at STR is set to CHARSET, and the
17321
9f837bea89e3 (CHARSET_TABLE_ENTRY): Handle ASCII charset correctly.
Kenichi Handa <handa@m17n.org>
parents: 17185
diff changeset
500 position-codes are set to C1 and C2. C2 of DIMENSION1 character is -1.
17052
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
501 If the character is a composite character, the upper 7-bit and
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
502 lower 7-bit of CMPCHAR-ID are set in C1 and C2 respectively. No
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
503 range checking. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
504 #define SPLIT_STRING(str, len, charset, c1, c2) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
505 ((BYTES_BY_CHAR_HEAD ((unsigned char) *(str)) < 2 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
506 || BYTES_BY_CHAR_HEAD ((unsigned char) *(str)) > len \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
507 || split_non_ascii_string (str, len, &charset, &c1, &c2, 0) < 0) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
508 ? c1 = *(str), charset = CHARSET_ASCII \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
509 : charset)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
510
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
511 /* Mapping table from ISO2022's charset (specified by DIMENSION,
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
512 CHARS, and FINAL_CHAR) to Emacs' charset. Should be accessed by
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
513 macro ISO_CHARSET_TABLE (DIMENSION, CHARS, FINAL_CHAR). */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
514 extern int iso_charset_table[2][2][128];
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
515
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
516 #define ISO_CHARSET_TABLE(dimension, chars, final_char) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
517 iso_charset_table[XINT (dimension) - 1][XINT (chars) > 94][XINT (final_char)]
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
518
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
519 #define BASE_LEADING_CODE_P(c) (BYTES_BY_CHAR_HEAD ((unsigned char) (c)) > 1)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
520
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
521 /* The following two macros CHAR_STRING and STRING_CHAR are the main
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
522 entry points to convert between Emacs two types of character
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
523 representations: multi-byte form and single-word form (character
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
524 code). */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
525
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
526 /* Set STR a pointer to the multi-byte form of the character C. If C
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
527 is not a composite character, the multi-byte form is set in WORKBUF
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
528 and STR points WORKBUF. The caller should allocate at least 4-byte
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
529 area at WORKBUF in advance. Returns the length of the multi-byte
17833
59aa4a0772f6 (VALID_CHAR_P): New macro.
Kenichi Handa <handa@m17n.org>
parents: 17726
diff changeset
530 form. If C is an invalid character code, signal an error. */
17052
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
531
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
532 #define CHAR_STRING(c, workbuf, str) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
533 (SINGLE_BYTE_CHAR_P (c) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
534 ? *(str = workbuf) = (unsigned char)(c), 1 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
535 : non_ascii_char_to_string (c, workbuf, &str))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
536
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
537 /* Return a character code of the character of which multi-byte form
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
538 is at STR and the length is LEN. If STR doesn't contain valid
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
539 multi-byte form, only the first byte in STR is returned. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
540
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
541 #define STRING_CHAR(str, len) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
542 ((BYTES_BY_CHAR_HEAD ((unsigned char) *(str)) == 1 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
543 || BYTES_BY_CHAR_HEAD ((unsigned char) *(str)) > (len)) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
544 ? (unsigned char) *(str) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
545 : string_to_non_ascii_char (str, len, 0))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
546
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
547 /* This is like STRING_CHAR but the third arg ACTUAL_LEN is set to
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
548 the length of the multi-byte form. Just to know the length, use
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
549 MULTIBYTE_FORM_LENGTH. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
550
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
551 #define STRING_CHAR_AND_LENGTH(str, len, actual_len) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
552 ((BYTES_BY_CHAR_HEAD ((unsigned char) *(str)) == 1 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
553 || BYTES_BY_CHAR_HEAD ((unsigned char) *(str)) > (len)) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
554 ? (actual_len = 1), (unsigned char) *(str) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
555 : string_to_non_ascii_char (str, len, &actual_len))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
556
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
557 /* Return the length of the multi-byte form at string STR of length LEN. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
558
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
559 #define MULTIBYTE_FORM_LENGTH(str, len) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
560 ((BYTES_BY_CHAR_HEAD (*(unsigned char *)(str)) == 1 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
561 || BYTES_BY_CHAR_HEAD (*(unsigned char *)(str)) > (len)) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
562 ? 1 \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
563 : multibyte_form_length (str, len))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
564
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
565 /* Set C a (possibly multibyte) character at P. P points into a
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
566 string which is the virtual concatenation of STR1 (which ends at
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
567 END1) or STR2 (which ends at END2). */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
568
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
569 #define GET_CHAR_AFTER_2(c, p, str1, end1, str2, end2) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
570 do { \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
571 const char *dtemp = (p) == (end1) ? (str2) : (p); \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
572 const char *dlimit = ((p) >= (str1) && (p) < (end1)) ? (end1) : (end2); \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
573 c = STRING_CHAR (dtemp, dlimit - dtemp); \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
574 } while (0)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
575
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
576 /* Set C a (possibly multibyte) character before P. P points into a
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
577 string which is the virtual concatenation of STR1 (which ends at
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
578 END1) or STR2 (which ends at END2). */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
579
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
580 #define GET_CHAR_BEFORE_2(c, p, str1, end1, str2, end2) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
581 do { \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
582 const char *dtemp = (p); \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
583 const char *dlimit = ((p) > (str2) && (p) <= (end2)) ? (str2) : (str1); \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
584 while (dtemp-- > dlimit && (unsigned char) *dtemp >= 0xA0); \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
585 c = STRING_CHAR (dtemp, p - dtemp); \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
586 } while (0)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
587
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
588 #ifdef emacs
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
589
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
590 /* Increase the buffer point POS of the current buffer to the next
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
591 character boundary. This macro relies on the fact that *GPT_ADDR
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
592 and *Z_ADDR are always accessible and the values are '\0'. No
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
593 range checking of POS. */
17120
c9c1bc92b8f8 (INC_POS, DEC_POS): Don't increase or decrease too
Kenichi Handa <handa@m17n.org>
parents: 17071
diff changeset
594 #define INC_POS(pos) \
c9c1bc92b8f8 (INC_POS, DEC_POS): Don't increase or decrease too
Kenichi Handa <handa@m17n.org>
parents: 17071
diff changeset
595 do { \
c9c1bc92b8f8 (INC_POS, DEC_POS): Don't increase or decrease too
Kenichi Handa <handa@m17n.org>
parents: 17071
diff changeset
596 unsigned char *p = POS_ADDR (pos); \
c9c1bc92b8f8 (INC_POS, DEC_POS): Don't increase or decrease too
Kenichi Handa <handa@m17n.org>
parents: 17071
diff changeset
597 pos++; \
c9c1bc92b8f8 (INC_POS, DEC_POS): Don't increase or decrease too
Kenichi Handa <handa@m17n.org>
parents: 17071
diff changeset
598 if (*p++ >= 0x80) \
c9c1bc92b8f8 (INC_POS, DEC_POS): Don't increase or decrease too
Kenichi Handa <handa@m17n.org>
parents: 17071
diff changeset
599 while (!CHAR_HEAD_P (p)) p++, pos++; \
17052
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
600 } while (0)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
601
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
602 /* Decrease the buffer point POS of the current buffer to the previous
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
603 character boundary. No range checking of POS. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
604 #define DEC_POS(pos) \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
605 do { \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
606 unsigned char *p, *p_min; \
17120
c9c1bc92b8f8 (INC_POS, DEC_POS): Don't increase or decrease too
Kenichi Handa <handa@m17n.org>
parents: 17071
diff changeset
607 int pos_saved = --pos; \
c9c1bc92b8f8 (INC_POS, DEC_POS): Don't increase or decrease too
Kenichi Handa <handa@m17n.org>
parents: 17071
diff changeset
608 if (pos < GPT) \
17052
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
609 p = BEG_ADDR + pos - 1, p_min = BEG_ADDR; \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
610 else \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
611 p = BEG_ADDR + GAP_SIZE + pos - 1, p_min = GAP_END_ADDR; \
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
612 while (p > p_min && !CHAR_HEAD_P (p)) p--, pos--; \
17120
c9c1bc92b8f8 (INC_POS, DEC_POS): Don't increase or decrease too
Kenichi Handa <handa@m17n.org>
parents: 17071
diff changeset
613 if (*p < 0x80 && pos != pos_saved) pos = pos_saved; \
17052
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
614 } while (0)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
615
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
616 #endif /* emacs */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
617
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
618 /* Maximum counts of components in one composite character. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
619 #define MAX_COMPONENT_COUNT 16
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
620
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
621 /* Structure to hold information of a composite character. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
622 struct cmpchar_info {
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
623 /* Byte length of the composite character. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
624 int len;
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
625
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
626 /* Multi-byte form of the composite character. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
627 unsigned char *data;
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
628
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
629 /* Length of glyph codes. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
630 int glyph_len;
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
631
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
632 /* Width of the overall glyph of the composite character. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
633 int width;
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
634
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
635 /* Pointer to an array of glyph codes of the composite character.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
636 This actually contains only character code, no face. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
637 GLYPH *glyph;
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
638
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
639 /* Pointer to an array of composition rules. The value has the form:
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
640 (0xA0 + ((GLOBAL-REF-POINT << 2) | NEW-REF-POINT))
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
641 where each XXX-REF-POINT is 0..8. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
642 unsigned char *cmp_rule;
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
643
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
644 /* Pointer to an array of x-axis offset of left edge of glyphs
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
645 relative to the left of of glyph[0] except for the first element
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
646 which is the absolute offset from the left edge of overall glyph.
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
647 The actual pixel offset should be calculated by multiplying each
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
648 frame's one column width by this value:
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
649 (i.e. FONT_WIDTH (f->output_data.x->font) * col_offset[N]). */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
650 float *col_offset;
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
651
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
652 /* Work slot used by `dumpglyphs' (xterm.c). */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
653 int face_work;
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
654 };
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
655
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
656 /* Table of pointers to the structure `cmpchar_info' indexed by
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
657 CMPCHAR-ID. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
658 extern struct cmpchar_info **cmpchar_table;
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
659 /* Number of the current composite characters. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
660 extern int n_cmpchars;
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
661
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
662 /* This is the maximum length of multi-byte form. */
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
663 #define MAX_LENGTH_OF_MULTI_BYTE_FORM (MAX_COMPONENT_COUNT * 6)
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
664
17185
0d5a1bae9d73 (MAX_CHARSET): Definition changed to the actual
Kenichi Handa <handa@m17n.org>
parents: 17120
diff changeset
665 /* Maximum character code currently used. */
0d5a1bae9d73 (MAX_CHARSET): Definition changed to the actual
Kenichi Handa <handa@m17n.org>
parents: 17120
diff changeset
666 #define MAX_CHAR (MIN_CHAR_COMPOSITION + n_cmpchars)
0d5a1bae9d73 (MAX_CHARSET): Definition changed to the actual
Kenichi Handa <handa@m17n.org>
parents: 17120
diff changeset
667
17726
b5f2a7d6a2d0 (unify_char): Extern it.
Kenichi Handa <handa@m17n.org>
parents: 17321
diff changeset
668 extern int unify_char ();
b5f2a7d6a2d0 (unify_char): Extern it.
Kenichi Handa <handa@m17n.org>
parents: 17321
diff changeset
669
17052
d0d7b244b1d0 Initial revision
Karl Heuer <kwzh@gnu.org>
parents:
diff changeset
670 #endif /* _CHARSET_H */