2 @c This is part of the XEmacs Lisp Reference Manual.
3 @c Copyright (C) 1996 Ben Wing.
4 @c See the file lispref.texi for copying conditions.
5 @setfilename ../../info/internationalization.info
6 @node MULE, Tips, Internationalization, top
9 @dfn{MULE} is the name originally given to the version of GNU Emacs
10 extended for multi-lingual (and in particular Asian-language) support.
11 ``MULE'' is short for ``MUlti-Lingual Emacs''. It is an extension and
12 complete rewrite of Nemacs (``Nihon Emacs'' where ``Nihon'' is the
13 Japanese word for ``Japan''), which only provided support for Japanese.
14 XEmacs refers to its multi-lingual support as @dfn{MULE support} since
15 it is based on @dfn{MULE}.
18 * Internationalization Terminology::
19 Definition of various internationalization terms.
20 * Charsets:: Sets of related characters.
21 * MULE Characters:: Working with characters in XEmacs/MULE.
22 * Composite Characters:: Making new characters by overstriking other ones.
23 * Coding Systems:: Ways of representing a string of chars using integers.
24 * CCL:: A special language for writing fast converters.
25 * Category Tables:: Subdividing charsets into groups.
28 @node Internationalization Terminology, Charsets, , MULE
29 @section Internationalization Terminology
31 In internationalization terminology, a string of text is divided up
32 into @dfn{characters}, which are the printable units that make up the
33 text. A single character is (for example) a capital @samp{A}, the
34 number @samp{2}, a Katakana character, a Hangul character, a Kanji
35 ideograph (an @dfn{ideograph} is a ``picture'' character, such as is
36 used in Japanese Kanji, Chinese Hanzi, and Korean Hanja; typically there
37 are thousands of such ideographs in each language), etc. The basic
38 property of a character is that it is the smallest unit of text with
39 semantic significance in text processing.
41 Human beings normally process text visually, so to a first approximation
42 a character may be identified with its shape. Note that the same
43 character may be drawn by two different people (or in two different
44 fonts) in slightly different ways, although the "basic shape" will be the
45 same. But consider the works of Scott Kim; human beings can recognize
46 hugely variant shapes as the "same" character. Sometimes, especially
47 where characters are extremely complicated to write, completely
48 different shapes may be defined as the "same" character in national
49 standards. The Taiwanese variant of Hanzi is generally the most
50 complicated; over the centuries, the Japanese, Koreans, and the People's
51 Republic of China have adopted simplifications of the shape, but the
52 line of descent from the original shape is recorded, and the meanings
53 and pronunciation of different forms of the same character are
54 considered to be identical within each language. (Of course, it may
55 take a specialist to recognize the related form; the point is that the
56 relations are standardized, despite the differing shapes.)
58 In some cases, the differences will be significant enough that it is
59 actually possible to identify two or more distinct shapes that both
60 represent the same character. For example, the lowercase letters
61 @samp{a} and @samp{g} each have two distinct possible shapes---the
62 @samp{a} can optionally have a curved tail projecting off the top, and
63 the @samp{g} can be formed either of two loops, or of one loop and a
64 tail hanging off the bottom. Such distinct possible shapes of a
65 character are called @dfn{glyphs}. The important characteristic of two
66 glyphs making up the same character is that the choice between one or
67 the other is purely stylistic and has no linguistic effect on a word
68 (this is the reason why a capital @samp{A} and lowercase @samp{a}
69 are different characters rather than different glyphs---e.g.
70 @samp{Aspen} is a city while @samp{aspen} is a kind of tree).
72 Note that @dfn{character} and @dfn{glyph} are used differently
73 here than elsewhere in XEmacs.
75 A @dfn{character set} is essentially a set of related characters. ASCII,
76 for example, is a set of 94 characters (or 128, if you count
77 non-printing characters). Other character sets are ISO8859-1 (ASCII
78 plus various accented characters and other international symbols),
79 JIS X 0201 (ASCII, more or less, plus half-width Katakana), JIS X 0208
80 (Japanese Kanji), JIS X 0212 (a second set of less-used Japanese Kanji),
81 GB2312 (Mainland Chinese Hanzi), etc.
83 The definition of a character set will implicitly or explicitly give
84 it an @dfn{ordering}, a way of assigning a number to each character in
85 the set. For many character sets, there is a natural ordering, for
86 example the ``ABC'' ordering of the Roman letters. But it is not clear
87 whether digits should come before or after the letters, and in fact
88 different European languages treat the ordering of accented characters
89 differently. It is useful to use the natural order where available, of
90 course. The number assigned to any particular character is called the
91 character's @dfn{code point}. (Within a given character set, each
92 character has a unique code point. Thus the word "set" is ill-chosen;
93 different orderings of the same characters are different character sets.
94 Identifying characters is simple enough for alphabetic character sets,
95 but the difference in ordering can cause great headaches when the same
96 thousands of characters are used by different cultures as in the Hanzi.)
98 A code point may be broken into a number of @dfn{position codes}. The
99 number of position codes required to index a particular character in a
100 character set is called the @dfn{dimension} of the character set. For
101 practical purposes, a position code may be thought of as a byte-sized
102 index. The printing characters of ASCII, being a relatively small
103 character set, is of dimension one, and each character in the set is
104 indexed using a single position code, in the range 1 through 94. Use of
105 this unusual range, rather than the familiar 33 through 126, is an
106 intentional abstraction; to understand the programming issues you must
107 break the equation between character sets and encodings.
109 JIS X 0208, i.e. Japanese Kanji, has thousands of characters, and is
110 of dimension two -- every character is indexed by two position codes,
111 each in the range 1 through 94. (This number ``94'' is not a
112 coincidence; we shall see that the JIS position codes were chosen so
113 that JIS kanji could be encoded without using codes that in ASCII are
114 associated with device control functions.) Note that the choice of the
115 range here is somewhat arbitrary. You could just as easily index the
116 printing characters in ASCII using numbers in the range 0 through 93, 2
117 through 95, 3 through 96, etc. In fact, the standardized
118 @emph{encoding} for the ASCII @emph{character set} uses the range 33
121 An @dfn{encoding} is a way of numerically representing characters from
122 one or more character sets into a stream of like-sized numerical values
123 called @dfn{words}; typically these are 8-bit, 16-bit, or 32-bit
124 quantities. If an encoding encompasses only one character set, then the
125 position codes for the characters in that character set could be used
126 directly. (This is the case with the trivial cipher used by children,
127 assigning 1 to `A', 2 to `B', and so on.) However, even with ASCII,
128 other considerations intrude. For example, why are the upper- and
129 lowercase alphabets separated by 8 characters? Why do the digits start
130 with `0' being assigned the code 48? In both cases because semantically
131 interesting operations (case conversion and numerical value extraction)
132 become convenient masking operations. Other artificial aspects (the
133 control characters being assigned to codes 0--31 and 127) are historical
134 accidents. (The use of 127 for @samp{DEL} is an artifact of the "punch
135 once" nature of paper tape, for example.)
137 Naive use of the position code is not possible, however, if more than
138 one character set is to be used in the encoding. For example, printed
139 Japanese text typically requires characters from multiple character sets
140 -- ASCII, JIS X 0208, and JIS X 0212, to be specific. Each of these is
141 indexed using one or more position codes in the range 1 through 94, so
142 the position codes could not be used directly or there would be no way
143 to tell which character was meant. Different Japanese encodings handle
144 this differently -- JIS uses special escape characters to denote
145 different character sets; EUC sets the high bit of the position codes
146 for JIS X 0208 and JIS X 0212, and puts a special extra byte before each
147 JIS X 0212 character; etc. (JIS, EUC, and most of the other encodings
148 you will encounter in files are 7-bit or 8-bit encodings. There is one
149 common 16-bit encoding, which is Unicode; this strives to represent all
150 the world's characters in a single large character set. 32-bit
151 encodings are often used internally in programs, such as XEmacs with
152 MULE support, to simplify the code that manipulates them; however, they
153 are not used externally because they are not very space-efficient.)
155 A general method of handling text using multiple character sets
156 (whether for multilingual text, or simply text in an extremely
157 complicated single language like Japanese) is defined in the
158 international standard ISO 2022. ISO 2022 will be discussed in more
159 detail later (@pxref{ISO 2022}), but for now suffice it to say that text
160 needs control functions (at least spacing), and if escape sequences are
161 to be used, an escape sequence introducer. It was decided to make all
162 text streams compatible with ASCII in the sense that the codes 0--31
163 (and 128-159) would always be control codes, never graphic characters,
164 and where defined by the character set the @samp{SPC} character would be
165 assigned code 32, and @samp{DEL} would be assigned 127. Thus there are
166 94 code points remaining if 7 bits are used. This is the reason that
167 most character sets are defined using position codes in the range 1
168 through 94. Then ISO 2022 compatible encodings are produced by shifting
169 the position codes 1 to 94 into character codes 33 to 126, or (if 8 bit
170 codes are available) into character codes 161 to 254.
172 Encodings are classified as either @dfn{modal} or @dfn{non-modal}. In
173 a @dfn{modal encoding}, there are multiple states that the encoding can
174 be in, and the interpretation of the values in the stream depends on the
175 current global state of the encoding. Special values in the encoding,
176 called @dfn{escape sequences}, are used to change the global state.
177 JIS, for example, is a modal encoding. The bytes @samp{ESC $ B}
178 indicate that, from then on, bytes are to be interpreted as position
179 codes for JIS X 0208, rather than as ASCII. This effect is cancelled
180 using the bytes @samp{ESC ( B}, which mean ``switch from whatever the
181 current state is to ASCII''. To switch to JIS X 0212, the escape
182 sequence @samp{ESC $ ( D}. (Note that here, as is common, the escape
183 sequences do in fact begin with @samp{ESC}. This is not necessarily the
184 case, however. Some encodings use control characters called "locking
185 shifts" (effect persists until cancelled) to switch character sets.)
187 A @dfn{non-modal encoding} has no global state that extends past the
188 character currently being interpreted. EUC, for example, is a
189 non-modal encoding. Characters in JIS X 0208 are encoded by setting
190 the high bit of the position codes, and characters in JIS X 0212 are
191 encoded by doing the same but also prefixing the character with the
194 The advantage of a modal encoding is that it is generally more
195 space-efficient, and is easily extendible because there are essentially
196 an arbitrary number of escape sequences that can be created. The
197 disadvantage, however, is that it is much more difficult to work with
198 if it is not being processed in a sequential manner. In the non-modal
199 EUC encoding, for example, the byte 0x41 always refers to the letter
200 @samp{A}; whereas in JIS, it could either be the letter @samp{A}, or
201 one of the two position codes in a JIS X 0208 character, or one of the
202 two position codes in a JIS X 0212 character. Determining exactly which
203 one is meant could be difficult and time-consuming if the previous
204 bytes in the string have not already been processed, or impossible if
205 they are drawn from an external stream that cannot be rewound.
207 Non-modal encodings are further divided into @dfn{fixed-width} and
208 @dfn{variable-width} formats. A fixed-width encoding always uses
209 the same number of words per character, whereas a variable-width
210 encoding does not. EUC is a good example of a variable-width
211 encoding: one to three bytes are used per character, depending on
212 the character set. 16-bit and 32-bit encodings are nearly always
213 fixed-width, and this is in fact one of the main reasons for using
214 an encoding with a larger word size. The advantages of fixed-width
215 encodings should be obvious. The advantages of variable-width
216 encodings are that they are generally more space-efficient and allow
217 for compatibility with existing 8-bit encodings such as ASCII. (For
218 example, in Unicode ASCII characters are simply promoted to a 16-bit
219 representation. That means that every ASCII character contains a
220 @samp{NUL} byte; evidently all of the standard string manipulation
221 functions will lose badly in a fixed-width Unicode environment.)
223 The bytes in an 8-bit encoding are often referred to as @dfn{octets}
224 rather than simply as bytes. This terminology dates back to the days
225 before 8-bit bytes were universal, when some computers had 9-bit bytes,
226 others had 10-bit bytes, etc.
228 @node Charsets, MULE Characters, Internationalization Terminology, MULE
231 A @dfn{charset} in MULE is an object that encapsulates a
232 particular character set as well as an ordering of those characters.
233 Charsets are permanent objects and are named using symbols, like
236 @defun charsetp object
237 This function returns non-@code{nil} if @var{object} is a charset.
241 * Charset Properties:: Properties of a charset.
242 * Basic Charset Functions:: Functions for working with charsets.
243 * Charset Property Functions:: Functions for accessing charset properties.
244 * Predefined Charsets:: Predefined charset objects.
247 @node Charset Properties, Basic Charset Functions, , Charsets
248 @subsection Charset Properties
250 Charsets have the following properties:
254 A symbol naming the charset. Every charset must have a different name;
255 this allows a charset to be referred to using its name rather than
256 the actual charset object.
258 A documentation string describing the charset.
260 A regular expression matching the font registry field for this character
261 set. For example, both the @code{ascii} and @code{latin-iso8859-1}
262 charsets use the registry @code{"ISO8859-1"}. This field is used to
263 choose an appropriate font when the user gives a general font
264 specification such as @samp{-*-courier-medium-r-*-140-*}, i.e. a
265 14-point upright medium-weight Courier font.
267 Number of position codes used to index a character in the character set.
268 XEmacs/MULE can only handle character sets of dimension 1 or 2.
269 This property defaults to 1.
271 Number of characters in each dimension. In XEmacs/MULE, the only
272 allowed values are 94 or 96. (There are a couple of pre-defined
273 character sets, such as ASCII, that do not follow this, but you cannot
274 define new ones like this.) Defaults to 94. Note that if the dimension
275 is 2, the character set thus described is 94x94 or 96x96.
277 Number of columns used to display a character in this charset.
278 Only used in TTY mode. (Under X, the actual width of a character
279 can be derived from the font used to display the characters.)
280 If unspecified, defaults to the dimension. (This is almost
281 always the correct value, because character sets with dimension 2
282 are usually ideograph character sets, which need two columns to
283 display the intricate ideographs.)
285 A symbol, either @code{l2r} (left-to-right) or @code{r2l}
286 (right-to-left). Defaults to @code{l2r}. This specifies the
287 direction that the text should be displayed in, and will be
288 left-to-right for most charsets but right-to-left for Hebrew
289 and Arabic. (Right-to-left display is not currently implemented.)
291 Final byte of the standard ISO 2022 escape sequence designating this
292 charset. Must be supplied. Each combination of (@var{dimension},
293 @var{chars}) defines a separate namespace for final bytes, and each
294 charset within a particular namespace must have a different final byte.
295 Note that ISO 2022 restricts the final byte to the range 0x30 - 0x7E if
296 dimension == 1, and 0x30 - 0x5F if dimension == 2. Note also that final
297 bytes in the range 0x30 - 0x3F are reserved for user-defined (not
298 official) character sets. For more information on ISO 2022, see @ref{Coding
301 0 (use left half of font on output) or 1 (use right half of font on
302 output). Defaults to 0. This specifies how to convert the position
303 codes that index a character in a character set into an index into the
304 font used to display the character set. With @code{graphic} set to 0,
305 position codes 33 through 126 map to font indices 33 through 126; with
306 it set to 1, position codes 33 through 126 map to font indices 161
307 through 254 (i.e. the same number but with the high bit set). For
308 example, for a font whose registry is ISO8859-1, the left half of the
309 font (octets 0x20 - 0x7F) is the @code{ascii} charset, while the right
310 half (octets 0xA0 - 0xFF) is the @code{latin-iso8859-1} charset.
312 A compiled CCL program used to convert a character in this charset into
313 an index into the font. This is in addition to the @code{graphic}
314 property. If a CCL program is defined, the position codes of a
315 character will first be processed according to @code{graphic} and
316 then passed through the CCL program, with the resulting values used
319 This is used, for example, in the Big5 character set (used in Taiwan).
320 This character set is not ISO-2022-compliant, and its size (94x157) does
321 not fit within the maximum 96x96 size of ISO-2022-compliant character
322 sets. As a result, XEmacs/MULE splits it (in a rather complex fashion,
323 so as to group the most commonly used characters together) into two
324 charset objects (@code{big5-1} and @code{big5-2}), each of size 94x94,
325 and each charset object uses a CCL program to convert the modified
326 position codes back into standard Big5 indices to retrieve a character
330 Most of the above properties can only be set when the charset is
331 initialized, and cannot be changed later.
332 @xref{Charset Property Functions}.
334 @node Basic Charset Functions, Charset Property Functions, Charset Properties, Charsets
335 @subsection Basic Charset Functions
337 @defun find-charset charset-or-name
338 This function retrieves the charset of the given name. If
339 @var{charset-or-name} is a charset object, it is simply returned.
340 Otherwise, @var{charset-or-name} should be a symbol. If there is no
341 such charset, @code{nil} is returned. Otherwise the associated charset
345 @defun get-charset name
346 This function retrieves the charset of the given name. Same as
347 @code{find-charset} except an error is signalled if there is no such
348 charset instead of returning @code{nil}.
352 This function returns a list of the names of all defined charsets.
355 @defun make-charset name doc-string props
356 This function defines a new character set. This function is for use
357 with MULE support. @var{name} is a symbol, the name by which the
358 character set is normally referred. @var{doc-string} is a string
359 describing the character set. @var{props} is a property list,
360 describing the specific nature of the character set. The recognized
361 properties are @code{registry}, @code{dimension}, @code{columns},
362 @code{chars}, @code{final}, @code{graphic}, @code{direction}, and
363 @code{ccl-program}, as previously described.
366 @defun make-reverse-direction-charset charset new-name
367 This function makes a charset equivalent to @var{charset} but which goes
368 in the opposite direction. @var{new-name} is the name of the new
369 charset. The new charset is returned.
372 @defun charset-from-attributes dimension chars final &optional direction
373 This function returns a charset with the given @var{dimension},
374 @var{chars}, @var{final}, and @var{direction}. If @var{direction} is
375 omitted, both directions will be checked (left-to-right will be returned
376 if character sets exist for both directions).
379 @defun charset-reverse-direction-charset charset
380 This function returns the charset (if any) with the same dimension,
381 number of characters, and final byte as @var{charset}, but which is
382 displayed in the opposite direction.
385 @node Charset Property Functions, Predefined Charsets, Basic Charset Functions, Charsets
386 @subsection Charset Property Functions
388 All of these functions accept either a charset name or charset object.
390 @defun charset-property charset prop
391 This function returns property @var{prop} of @var{charset}.
392 @xref{Charset Properties}.
395 Convenience functions are also provided for retrieving individual
396 properties of a charset.
398 @defun charset-name charset
399 This function returns the name of @var{charset}. This will be a symbol.
402 @defun charset-description charset
403 This function returns the documentation string of @var{charset}.
406 @defun charset-registry charset
407 This function returns the registry of @var{charset}.
410 @defun charset-dimension charset
411 This function returns the dimension of @var{charset}.
414 @defun charset-chars charset
415 This function returns the number of characters per dimension of
419 @defun charset-width charset
420 This function returns the number of display columns per character (in
421 TTY mode) of @var{charset}.
424 @defun charset-direction charset
425 This function returns the display direction of @var{charset}---either
426 @code{l2r} or @code{r2l}.
429 @defun charset-iso-final-char charset
430 This function returns the final byte of the ISO 2022 escape sequence
431 designating @var{charset}.
434 @defun charset-iso-graphic-plane charset
435 This function returns either 0 or 1, depending on whether the position
436 codes of characters in @var{charset} map to the left or right half
437 of their font, respectively.
440 @defun charset-ccl-program charset
441 This function returns the CCL program, if any, for converting
442 position codes of characters in @var{charset} into font indices.
445 The two properties of a charset that can currently be set after the
446 charset has been created are the CCL program and the font registry.
448 @defun set-charset-ccl-program charset ccl-program
449 This function sets the @code{ccl-program} property of @var{charset} to
453 @defun set-charset-registry charset registry
454 This function sets the @code{registry} property of @var{charset} to
458 @node Predefined Charsets, , Charset Property Functions, Charsets
459 @subsection Predefined Charsets
461 The following charsets are predefined in the C code.
464 Name Type Fi Gr Dir Registry
465 --------------------------------------------------------------
466 ascii 94 B 0 l2r ISO8859-1
467 control-1 94 0 l2r ---
468 latin-iso8859-1 94 A 1 l2r ISO8859-1
469 latin-iso8859-2 96 B 1 l2r ISO8859-2
470 latin-iso8859-3 96 C 1 l2r ISO8859-3
471 latin-iso8859-4 96 D 1 l2r ISO8859-4
472 cyrillic-iso8859-5 96 L 1 l2r ISO8859-5
473 arabic-iso8859-6 96 G 1 r2l ISO8859-6
474 greek-iso8859-7 96 F 1 l2r ISO8859-7
475 hebrew-iso8859-8 96 H 1 r2l ISO8859-8
476 latin-iso8859-9 96 M 1 l2r ISO8859-9
477 thai-tis620 96 T 1 l2r TIS620
478 katakana-jisx0201 94 I 1 l2r JISX0201.1976
479 latin-jisx0201 94 J 0 l2r JISX0201.1976
480 japanese-jisx0208-1978 94x94 @@ 0 l2r JISX0208.1978
481 japanese-jisx0208 94x94 B 0 l2r JISX0208.19(83|90)
482 japanese-jisx0212 94x94 D 0 l2r JISX0212
483 chinese-gb2312 94x94 A 0 l2r GB2312
484 chinese-cns11643-1 94x94 G 0 l2r CNS11643.1
485 chinese-cns11643-2 94x94 H 0 l2r CNS11643.2
486 chinese-big5-1 94x94 0 0 l2r Big5
487 chinese-big5-2 94x94 1 0 l2r Big5
488 korean-ksc5601 94x94 C 0 l2r KSC5601
489 composite 96x96 0 l2r ---
492 The following charsets are predefined in the Lisp code.
495 Name Type Fi Gr Dir Registry
496 --------------------------------------------------------------
497 arabic-digit 94 2 0 l2r MuleArabic-0
498 arabic-1-column 94 3 0 r2l MuleArabic-1
499 arabic-2-column 94 4 0 r2l MuleArabic-2
500 sisheng 94 0 0 l2r sisheng_cwnn\|OMRON_UDC_ZH
501 chinese-cns11643-3 94x94 I 0 l2r CNS11643.1
502 chinese-cns11643-4 94x94 J 0 l2r CNS11643.1
503 chinese-cns11643-5 94x94 K 0 l2r CNS11643.1
504 chinese-cns11643-6 94x94 L 0 l2r CNS11643.1
505 chinese-cns11643-7 94x94 M 0 l2r CNS11643.1
506 ethiopic 94x94 2 0 l2r Ethio
507 ascii-r2l 94 B 0 r2l ISO8859-1
508 ipa 96 0 1 l2r MuleIPA
509 vietnamese-viscii-lower 96 1 1 l2r VISCII1.1
510 vietnamese-viscii-upper 96 2 1 l2r VISCII1.1
513 For all of the above charsets, the dimension and number of columns are
516 Note that ASCII, Control-1, and Composite are handled specially.
517 This is why some of the fields are blank; and some of the filled-in
518 fields (e.g. the type) are not really accurate.
520 @node MULE Characters, Composite Characters, Charsets, MULE
521 @section MULE Characters
523 @defun make-char charset arg1 &optional arg2
524 This function makes a multi-byte character from @var{charset} and octets
525 @var{arg1} and @var{arg2}.
528 @defun char-charset character
529 This function returns the character set of char @var{character}.
532 @defun char-octet character &optional n
533 This function returns the octet (i.e. position code) numbered @var{n}
534 (should be 0 or 1) of char @var{character}. @var{n} defaults to 0 if omitted.
537 @defun find-charset-region start end &optional buffer
538 This function returns a list of the charsets in the region between
539 @var{start} and @var{end}. @var{buffer} defaults to the current buffer
543 @defun find-charset-string string
544 This function returns a list of the charsets in @var{string}.
547 @node Composite Characters, Coding Systems, MULE Characters, MULE
548 @section Composite Characters
550 Composite characters are not yet completely implemented.
552 @defun make-composite-char string
553 This function converts a string into a single composite character. The
554 character is the result of overstriking all the characters in the
558 @defun composite-char-string character
559 This function returns a string of the characters comprising a composite
563 @defun compose-region start end &optional buffer
564 This function composes the characters in the region from @var{start} to
565 @var{end} in @var{buffer} into one composite character. The composite
566 character replaces the composed characters. @var{buffer} defaults to
567 the current buffer if omitted.
570 @defun decompose-region start end &optional buffer
571 This function decomposes any composite characters in the region from
572 @var{start} to @var{end} in @var{buffer}. This converts each composite
573 character into one or more characters, the individual characters out of
574 which the composite character was formed. Non-composite characters are
575 left as-is. @var{buffer} defaults to the current buffer if omitted.
578 @node Coding Systems, CCL, Composite Characters, MULE
579 @section Coding Systems
581 A coding system is an object that defines how text containing multiple
582 character sets is encoded into a stream of (typically 8-bit) bytes. The
583 coding system is used to decode the stream into a series of characters
584 (which may be from multiple charsets) when the text is read from a file
585 or process, and is used to encode the text back into the same format
586 when it is written out to a file or process.
588 For example, many ISO-2022-compliant coding systems (such as Compound
589 Text, which is used for inter-client data under the X Window System) use
590 escape sequences to switch between different charsets -- Japanese Kanji,
591 for example, is invoked with @samp{ESC $ ( B}; ASCII is invoked with
592 @samp{ESC ( B}; and Cyrillic is invoked with @samp{ESC - L}. See
593 @code{make-coding-system} for more information.
595 Coding systems are normally identified using a symbol, and the symbol is
596 accepted in place of the actual coding system object whenever a coding
597 system is called for. (This is similar to how faces and charsets work.)
599 @defun coding-system-p object
600 This function returns non-@code{nil} if @var{object} is a coding system.
604 * Coding System Types:: Classifying coding systems.
605 * ISO 2022:: An international standard for
606 charsets and encodings.
607 * EOL Conversion:: Dealing with different ways of denoting
609 * Coding System Properties:: Properties of a coding system.
610 * Basic Coding System Functions:: Working with coding systems.
611 * Coding System Property Functions:: Retrieving a coding system's properties.
612 * Encoding and Decoding Text:: Encoding and decoding text.
613 * Detection of Textual Encoding:: Determining how text is encoded.
614 * Big5 and Shift-JIS Functions:: Special functions for these non-standard
616 * Predefined Coding Systems:: Coding systems implemented by MULE.
619 @node Coding System Types, ISO 2022, , Coding Systems
620 @subsection Coding System Types
622 The coding system type determines the basic algorithm XEmacs will use to
623 decode or encode a data stream. Character encodings will be converted
624 to the MULE encoding, escape sequences processed, and newline sequences
625 converted to XEmacs's internal representation. There are three basic
626 classes of coding system type: no-conversion, ISO-2022, and special.
628 No conversion allows you to look at the file's internal representation.
629 Since XEmacs is basically a text editor, "no conversion" does convert
630 newline conventions by default. (Use the 'binary coding-system if this
633 ISO 2022 (@pxref{ISO 2022}) is the basic international standard regulating
634 use of "coded character sets for the exchange of data", ie, text
635 streams. ISO 2022 contains functions that make it possible to encode
636 text streams to comply with restrictions of the Internet mail system and
637 de facto restrictions of most file systems (eg, use of the separator
638 character in file names). Coding systems which are not ISO 2022
639 conformant can be difficult to handle. Perhaps more important, they are
640 not adaptable to multilingual information interchange, with the obvious
641 exception of ISO 10646 (Unicode). (Unicode is partially supported by
642 XEmacs with the addition of the Lisp package ucs-conv.)
644 The special class of coding systems includes automatic detection, CCL (a
645 "little language" embedded as an interpreter, useful for translating
646 between variants of a single character set), non-ISO-2022-conformant
647 encodings like Unicode, Shift JIS, and Big5, and MULE internal coding.
648 (NB: this list is based on XEmacs 21.2. Terminology may vary slightly
649 for other versions of XEmacs and for GNU Emacs 20.)
653 No conversion, for binary files, and a few special cases of non-ISO-2022
654 coding systems where conversion is done by hook functions (usually
655 implemented in CCL). On output, graphic characters that are not in
656 ASCII or Latin-1 will be replaced by a @samp{?}. (For a
657 no-conversion-encoded buffer, these characters will only be present if
658 you explicitly insert them.)
660 Any ISO-2022-compliant encoding. Among others, this includes JIS (the
661 Japanese encoding commonly used for e-mail), national variants of EUC
662 (the standard Unix encoding for Japanese and other languages), and
663 Compound Text (an encoding used in X11). You can specify more specific
664 information about the conversion with the @var{flags} argument.
666 ISO 10646 UCS-4 encoding. A 31-bit fixed-width superset of Unicode.
668 ISO 10646 UTF-8 encoding. A ``file system safe'' transformation format
669 that can be used with both UCS-4 and Unicode.
671 Automatic conversion. XEmacs attempts to detect the coding system used
674 Shift-JIS (a Japanese encoding commonly used in PC operating systems).
676 Big5 (the encoding commonly used for Taiwanese).
678 The conversion is performed using a user-written pseudo-code program.
679 CCL (Code Conversion Language) is the name of this pseudo-code. For
680 example, CCL is used to map KOI8-R characters (an encoding for Russian
681 Cyrillic) to ISO8859-5 (the form used internally by MULE).
683 Write out or read in the raw contents of the memory representing the
684 buffer's text. This is primarily useful for debugging purposes, and is
685 only enabled when XEmacs has been compiled with @code{DEBUG_XEMACS} set
686 (the @samp{--debug} configure option). @strong{Warning}: Reading in a
687 file using @code{internal} conversion can result in an internal
688 inconsistency in the memory representing a buffer's text, which will
689 produce unpredictable results and may cause XEmacs to crash. Under
690 normal circumstances you should never use @code{internal} conversion.
693 @node ISO 2022, EOL Conversion, Coding System Types, Coding Systems
696 This section briefly describes the ISO 2022 encoding standard. A more
697 thorough treatment is available in the original document of ISO
698 2022 as well as various national standards (such as JIS X 0202).
700 Character sets (@dfn{charsets}) are classified into the following four
701 categories, according to the number of characters in the charset:
702 94-charset, 96-charset, 94x94-charset, and 96x96-charset. This means
703 that although an ISO 2022 coding system may have variable width
704 characters, each charset used is fixed-width (in contrast to the MULE
705 character set and UTF-8, for example).
707 ISO 2022 provides for switching between character sets via escape
708 sequences. This switching is somewhat complicated, because ISO 2022
709 provides for both legacy applications like Internet mail that accept
710 only 7 significant bits in some contexts (RFC 822 headers, for example),
711 and more modern "8-bit clean" applications. It also provides for
712 compact and transparent representation of languages like Japanese which
713 mix ASCII and a national script (even outside of computer programs).
715 First, ISO 2022 codified prevailing practice by dividing the code space
716 into "control" and "graphic" regions. The code points 0x00-0x1F and
717 0x80-0x9F are reserved for "control characters", while "graphic
718 characters" must be assigned to code points in the regions 0x20-0x7F and
719 0xA0-0xFF. The positions 0x20 and 0x7F are special, and under some
720 circumstances must be assigned the graphic character "ASCII SPACE" and
721 the control character "ASCII DEL" respectively.
723 The various regions are given the name C0 (0x00-0x1F), GL (0x20-0x7F),
724 C1 (0x80-0x9F), and GR (0xA0-0xFF). GL and GR stand for "graphic left"
725 and "graphic right", respectively, because of the standard method of
726 displaying graphic character sets in tables with the high byte indexing
727 columns and the low byte indexing rows. I don't find it very intuitive,
728 but these are called "registers".
730 An ISO 2022-conformant encoding for a graphic character set must use a
731 fixed number of bytes per character, and the values must fit into a
732 single register; that is, each byte must range over either 0x20-0x7F, or
733 0xA0-0xFF. It is not allowed to extend the range of the repertoire of a
734 character set by using both ranges at the same. This is why a standard
735 character set such as ISO 8859-1 is actually considered by ISO 2022 to
736 be an aggregation of two character sets, ASCII and LATIN-1, and why it
737 is technically incorrect to refer to ISO 8859-1 as "Latin 1". Also, a
738 single character's bytes must all be drawn from the same register; this
739 is why Shift JIS (for Japanese) and Big 5 (for Chinese) are not ISO
740 2022-compatible encodings.
742 The reason for this restriction becomes clear when you attempt to define
743 an efficient, robust encoding for a language like Japanese. Like ISO
744 8859, Japanese encodings are aggregations of several character sets. In
745 practice, the vast majority of characters are drawn from the "JIS Roman"
746 character set (a derivative of ASCII; it won't hurt to think of it as
747 ASCII) and the JIS X 0208 standard "basic Japanese" character set
748 including not only ideographic characters ("kanji") but syllabic
749 Japanese characters ("kana"), a wide variety of symbols, and many
750 alphabetic characters (Roman, Greek, and Cyrillic) as well. Although
751 JIS X 0208 includes the whole Roman alphabet, as a 2-byte code it is not
752 suited to programming; thus the inclusion of ASCII in the standard
755 For normal Japanese text such as in newspapers, a broad repertoire of
756 approximately 3000 characters is used. Evidently this won't fit into
757 one byte; two must be used. But much of the text processed by Japanese
758 computers is computer source code, nearly all of which is ASCII. A not
759 insignificant portion of ordinary text is English (as such or as
760 borrowed Japanese vocabulary) or other languages which can represented
761 at least approximately in ASCII, as well. It seems reasonable then to
762 represent ASCII in one byte, and JIS X 0208 in two. And this is exactly
763 what the Extended Unix Code for Japanese (EUC-JP) does. ASCII is
764 invoked to the GL register, and JIS X 0208 is invoked to the GR
765 register. Thus, each byte can be tested for its character set by
766 looking at the high bit; if set, it is Japanese, if clear, it is ASCII.
767 Furthermore, since control characters like newline can never be part of
768 a graphic character, even in the case of corruption in transmission the
769 stream will be resynchronized at every line break, on the order of 60-80
770 bytes. This coding system requires no escape sequences or special
771 control codes to represent 99.9% of all Japanese text.
773 Note carefully the distinction between the character sets (ASCII and JIS
774 X 0208), the encoding (EUC-JP), and the coding system (ISO 2022). The
775 JIS X 0208 character set is used in three different encodings for
776 Japanese, but in ISO-2022-JP it is invoked into GL (so the high bit is
777 always clear), in EUC-JP it is invoked into GR (setting the high bit in
778 the process), and in Shift JIS the high bit may be set or reset, and the
779 significant bits are shifted within the 16-bit character so that the two
780 main character sets can coexist with a third (the "halfwidth katakana"
781 of JIS X 0201). As the name implies, the ISO-2022-JP encoding is also a
782 version of the ISO-2022 coding system.
784 In order to systematically treat subsidiary character sets (like the
785 "halfwidth katakana" already mentioned, and the "supplementary kanji" of
786 JIS X 0212), four further registers are defined: G0, G1, G2, and G3.
787 Unlike GL and GR, they are not logically distinguished by internal
788 format. Instead, the process of "invocation" mentioned earlier is
789 broken into two steps: first, a character set is @dfn{designated} to one
790 of the registers G0-G3 by use of an @dfn{escape sequence} of the form:
793 ESC [@var{I}] @var{I} @var{F}
796 where @var{I} is an intermediate character or characters in the range
797 0x20 - 0x3F, and @var{F}, from the range 0x30-0x7Fm is the final
798 character identifying this charset. (Final characters in the range
799 0x30-0x3F are reserved for private use and will never have a publicly
802 Then that register is @dfn{invoked} to either GL or GR, either
803 automatically (designations to G0 normally involve invocation to GL as
804 well), or by use of shifting (affecting only the following character in
805 the data stream) or locking (effective until the next designation or
806 locking) control sequences. An encoding conformant to ISO 2022 is
807 typically defined by designating the initial contents of the G0-G3
808 registers, specifying a 7 or 8 bit environment, and specifying whether
809 further designations will be recognized.
811 Some examples of character sets and the registered final characters
812 @var{F} used to designate them:
817 ASCII (B), left (J) and right (I) half of JIS X 0201, ...
819 Latin-1 (A), Latin-2 (B), Latin-3 (C), ...
821 GB2312 (A), JIS X 0208 (B), KSC5601 (C), ...
826 The meanings of the various characters in these sequences, where not
827 specified by the ISO 2022 standard (such as the ESC character), are
828 assigned by @dfn{ECMA}, the European Computer Manufacturers Association.
830 The meaning of intermediate characters are:
834 $ [0x24]: indicate charset of dimension 2 (94x94 or 96x96).
835 ( [0x28]: designate to G0 a 94-charset whose final byte is @var{F}.
836 ) [0x29]: designate to G1 a 94-charset whose final byte is @var{F}.
837 * [0x2A]: designate to G2 a 94-charset whose final byte is @var{F}.
838 + [0x2B]: designate to G3 a 94-charset whose final byte is @var{F}.
839 , [0x2C]: designate to G0 a 96-charset whose final byte is @var{F}.
840 - [0x2D]: designate to G1 a 96-charset whose final byte is @var{F}.
841 . [0x2E]: designate to G2 a 96-charset whose final byte is @var{F}.
842 / [0x2F]: designate to G3 a 96-charset whose final byte is @var{F}.
846 The comma may be used in files read and written only by MULE, as a MULE
847 extension, but this is illegal in ISO 2022. (The reason is that in ISO
848 2022 G0 must be a 94-member character set, with 0x20 assigned the value
849 SPACE, and 0x7F assigned the value DEL.)
851 Here are examples of designations:
855 ESC ( B : designate to G0 ASCII
856 ESC - A : designate to G1 Latin-1
857 ESC $ ( A or ESC $ A : designate to G0 GB2312
858 ESC $ ( B or ESC $ B : designate to G0 JISX0208
859 ESC $ ) C : designate to G1 KSC5601
863 (The short forms used to designate GB2312 and JIS X 0208 are for
864 backwards compatibility; the long forms are preferred.)
866 To use a charset designated to G2 or G3, and to use a charset designated
867 to G1 in a 7-bit environment, you must explicitly invoke G1, G2, or G3
868 into GL. There are two types of invocation, Locking Shift (forever) and
869 Single Shift (one character only).
871 Locking Shift is done as follows:
874 LS0 or SI (0x0F): invoke G0 into GL
875 LS1 or SO (0x0E): invoke G1 into GL
876 LS2: invoke G2 into GL
877 LS3: invoke G3 into GL
878 LS1R: invoke G1 into GR
879 LS2R: invoke G2 into GR
880 LS3R: invoke G3 into GR
883 Single Shift is done as follows:
887 SS2 or ESC N: invoke G2 into GL
888 SS3 or ESC O: invoke G3 into GL
892 The shift functions (such as LS1R and SS3) are represented by control
893 characters (from C1) in 8 bit environments and by escape sequences in 7
896 (#### Ben says: I think the above is slightly incorrect. It appears that
897 SS2 invokes G2 into GR and SS3 invokes G3 into GR, whereas ESC N and
898 ESC O behave as indicated. The above definitions will not parse
899 EUC-encoded text correctly, and it looks like the code in mule-coding.c
900 has similar problems.)
902 Evidently there are a lot of ISO-2022-compliant ways of encoding
903 multilingual text. Now, in the world, there exist many coding systems
904 such as X11's Compound Text, Japanese JUNET code, and so-called EUC
905 (Extended UNIX Code); all of these are variants of ISO 2022.
907 In MULE, we characterize a version of ISO 2022 by the following
912 The character sets initially designated to G0 thru G3.
914 Whether short form designations are allowed for Japanese and Chinese.
916 Whether ASCII should be designated to G0 before control characters.
918 Whether ASCII should be designated to G0 at the end of line.
920 7-bit environment or 8-bit environment.
922 Whether Locking Shifts are used or not.
924 Whether to use ASCII or the variant JIS X 0201-1976-Roman.
926 Whether to use JIS X 0208-1983 or the older version JIS X 0208-1976.
929 (The last two are only for Japanese.)
931 By specifying these attributes, you can create any variant
934 Here are several examples:
938 ISO-2022-JP -- Coding system used in Japanese email (RFC 1463 #### check).
939 1. G0 <- ASCII, G1..3 <- never used
946 8. Use JIS X 0208-1983
950 ctext -- X11 Compound Text
951 1. G0 <- ASCII, G1 <- Latin-1, G2,3 <- never used.
955 5. 8-bit environment.
958 8. Use JIS X 0208-1983.
962 euc-china -- Chinese EUC. Often called the "GB encoding", but that is
963 technically incorrect.
964 1. G0 <- ASCII, G1 <- GB 2312, G2,3 <- never used.
968 5. 8-bit environment.
971 8. Use JIS X 0208-1983.
975 ISO-2022-KR -- Coding system used in Korean email.
976 1. G0 <- ASCII, G1 <- KSC 5601, G2,3 <- never used.
980 5. 7-bit environment.
983 8. Use JIS X 0208-1983.
987 MULE creates all of these coding systems by default.
989 @node EOL Conversion, Coding System Properties, ISO 2022, Coding Systems
990 @subsection EOL Conversion
994 Automatically detect the end-of-line type (LF, CRLF, or CR). Also
995 generate subsidiary coding systems named @code{@var{name}-unix},
996 @code{@var{name}-dos}, and @code{@var{name}-mac}, that are identical to
997 this coding system but have an EOL-TYPE value of @code{lf}, @code{crlf},
998 and @code{cr}, respectively.
1000 The end of a line is marked externally using ASCII LF. Since this is
1001 also the way that XEmacs represents an end-of-line internally,
1002 specifying this option results in no end-of-line conversion. This is
1003 the standard format for Unix text files.
1005 The end of a line is marked externally using ASCII CRLF. This is the
1006 standard format for MS-DOS text files.
1008 The end of a line is marked externally using ASCII CR. This is the
1009 standard format for Macintosh text files.
1011 Automatically detect the end-of-line type but do not generate subsidiary
1012 coding systems. (This value is converted to @code{nil} when stored
1013 internally, and @code{coding-system-property} will return @code{nil}.)
1016 @node Coding System Properties, Basic Coding System Functions, EOL Conversion, Coding Systems
1017 @subsection Coding System Properties
1021 String to be displayed in the modeline when this coding system is
1025 End-of-line conversion to be used. It should be one of the types
1026 listed in @ref{EOL Conversion}.
1029 The coding system which is the same as this one, except that it uses the
1030 Unix line-breaking convention.
1033 The coding system which is the same as this one, except that it uses the
1034 DOS line-breaking convention.
1037 The coding system which is the same as this one, except that it uses the
1038 Macintosh line-breaking convention.
1040 @item post-read-conversion
1041 Function called after a file has been read in, to perform the decoding.
1042 Called with two arguments, @var{start} and @var{end}, denoting a region of
1043 the current buffer to be decoded.
1045 @item pre-write-conversion
1046 Function called before a file is written out, to perform the encoding.
1047 Called with two arguments, @var{start} and @var{end}, denoting a region of
1048 the current buffer to be encoded.
1051 The following additional properties are recognized if @var{type} is
1059 The character set initially designated to the G0 - G3 registers.
1060 The value should be one of
1064 A charset object (designate that character set)
1066 @code{nil} (do not ever use this register)
1068 @code{t} (no character set is initially designated to the register, but
1069 may be later on; this automatically sets the corresponding
1070 @code{force-g*-on-output} property)
1073 @item force-g0-on-output
1074 @itemx force-g1-on-output
1075 @itemx force-g2-on-output
1076 @itemx force-g3-on-output
1077 If non-@code{nil}, send an explicit designation sequence on output
1078 before using the specified register.
1081 If non-@code{nil}, use the short forms @samp{ESC $ @@}, @samp{ESC $ A},
1082 and @samp{ESC $ B} on output in place of the full designation sequences
1083 @samp{ESC $ ( @@}, @samp{ESC $ ( A}, and @samp{ESC $ ( B}.
1086 If non-@code{nil}, don't designate ASCII to G0 at each end of line on
1087 output. Setting this to non-@code{nil} also suppresses other
1088 state-resetting that normally happens at the end of a line.
1091 If non-@code{nil}, don't designate ASCII to G0 before control chars on
1095 If non-@code{nil}, use 7-bit environment on output. Otherwise, use 8-bit
1099 If non-@code{nil}, use locking-shift (SO/SI) instead of single-shift or
1100 designation by escape sequence.
1103 If non-@code{nil}, don't use ISO6429's direction specification.
1106 If non-@code{nil}, literal control characters that are the same as the
1107 beginning of a recognized ISO 2022 or ISO 6429 escape sequence (in
1108 particular, ESC (0x1B), SO (0x0E), SI (0x0F), SS2 (0x8E), SS3 (0x8F),
1109 and CSI (0x9B)) are ``quoted'' with an escape character so that they can
1110 be properly distinguished from an escape sequence. (Note that doing
1111 this results in a non-portable encoding.) This encoding flag is used for
1112 byte-compiled files. Note that ESC is a good choice for a quoting
1113 character because there are no escape sequences whose second byte is a
1114 character from the Control-0 or Control-1 character sets; this is
1115 explicitly disallowed by the ISO 2022 standard.
1117 @item input-charset-conversion
1118 A list of conversion specifications, specifying conversion of characters
1119 in one charset to another when decoding is performed. Each
1120 specification is a list of two elements: the source charset, and the
1121 destination charset.
1123 @item output-charset-conversion
1124 A list of conversion specifications, specifying conversion of characters
1125 in one charset to another when encoding is performed. The form of each
1126 specification is the same as for @code{input-charset-conversion}.
1129 The following additional properties are recognized (and required) if
1130 @var{type} is @code{ccl}:
1134 CCL program used for decoding (converting to internal format).
1137 CCL program used for encoding (converting to external format).
1140 The following properties are used internally: @var{eol-cr},
1141 @var{eol-crlf}, @var{eol-lf}, and @var{base}.
1143 @node Basic Coding System Functions, Coding System Property Functions, Coding System Properties, Coding Systems
1144 @subsection Basic Coding System Functions
1146 @defun find-coding-system coding-system-or-name
1147 This function retrieves the coding system of the given name.
1149 If @var{coding-system-or-name} is a coding-system object, it is simply
1150 returned. Otherwise, @var{coding-system-or-name} should be a symbol.
1151 If there is no such coding system, @code{nil} is returned. Otherwise
1152 the associated coding system object is returned.
1155 @defun get-coding-system name
1156 This function retrieves the coding system of the given name. Same as
1157 @code{find-coding-system} except an error is signalled if there is no
1158 such coding system instead of returning @code{nil}.
1161 @defun coding-system-list
1162 This function returns a list of the names of all defined coding systems.
1165 @defun coding-system-name coding-system
1166 This function returns the name of the given coding system.
1169 @defun coding-system-base coding-system
1170 Returns the base coding system (undecided EOL convention)
1174 @defun make-coding-system name type &optional doc-string props
1175 This function registers symbol @var{name} as a coding system.
1177 @var{type} describes the conversion method used and should be one of
1178 the types listed in @ref{Coding System Types}.
1180 @var{doc-string} is a string describing the coding system.
1182 @var{props} is a property list, describing the specific nature of the
1183 character set. Recognized properties are as in @ref{Coding System
1187 @defun copy-coding-system old-coding-system new-name
1188 This function copies @var{old-coding-system} to @var{new-name}. If
1189 @var{new-name} does not name an existing coding system, a new one will
1193 @defun subsidiary-coding-system coding-system eol-type
1194 This function returns the subsidiary coding system of
1195 @var{coding-system} with eol type @var{eol-type}.
1198 @node Coding System Property Functions, Encoding and Decoding Text, Basic Coding System Functions, Coding Systems
1199 @subsection Coding System Property Functions
1201 @defun coding-system-doc-string coding-system
1202 This function returns the doc string for @var{coding-system}.
1205 @defun coding-system-type coding-system
1206 This function returns the type of @var{coding-system}.
1209 @defun coding-system-property coding-system prop
1210 This function returns the @var{prop} property of @var{coding-system}.
1213 @node Encoding and Decoding Text, Detection of Textual Encoding, Coding System Property Functions, Coding Systems
1214 @subsection Encoding and Decoding Text
1216 @defun decode-coding-region start end coding-system &optional buffer
1217 This function decodes the text between @var{start} and @var{end} which
1218 is encoded in @var{coding-system}. This is useful if you've read in
1219 encoded text from a file without decoding it (e.g. you read in a
1220 JIS-formatted file but used the @code{binary} or @code{no-conversion} coding
1221 system, so that it shows up as @samp{^[$B!<!+^[(B}). The length of the
1222 encoded text is returned. @var{buffer} defaults to the current buffer
1226 @defun encode-coding-region start end coding-system &optional buffer
1227 This function encodes the text between @var{start} and @var{end} using
1228 @var{coding-system}. This will, for example, convert Japanese
1229 characters into stuff such as @samp{^[$B!<!+^[(B} if you use the JIS
1230 encoding. The length of the encoded text is returned. @var{buffer}
1231 defaults to the current buffer if unspecified.
1234 @node Detection of Textual Encoding, Big5 and Shift-JIS Functions, Encoding and Decoding Text, Coding Systems
1235 @subsection Detection of Textual Encoding
1237 @defun coding-category-list
1238 This function returns a list of all recognized coding categories.
1241 @defun set-coding-priority-list list
1242 This function changes the priority order of the coding categories.
1243 @var{list} should be a list of coding categories, in descending order of
1244 priority. Unspecified coding categories will be lower in priority than
1245 all specified ones, in the same relative order they were in previously.
1248 @defun coding-priority-list
1249 This function returns a list of coding categories in descending order of
1253 @defun set-coding-category-system coding-category coding-system
1254 This function changes the coding system associated with a coding category.
1257 @defun coding-category-system coding-category
1258 This function returns the coding system associated with a coding category.
1261 @defun detect-coding-region start end &optional buffer
1262 This function detects coding system of the text in the region between
1263 @var{start} and @var{end}. Returned value is a list of possible coding
1264 systems ordered by priority. If only ASCII characters are found, it
1265 returns @code{autodetect} or one of its subsidiary coding systems
1266 according to a detected end-of-line type. Optional arg @var{buffer}
1267 defaults to the current buffer.
1270 @node Big5 and Shift-JIS Functions, Predefined Coding Systems, Detection of Textual Encoding, Coding Systems
1271 @subsection Big5 and Shift-JIS Functions
1273 These are special functions for working with the non-standard
1274 Shift-JIS and Big5 encodings.
1276 @defun decode-shift-jis-char code
1277 This function decodes a JIS X 0208 character of Shift-JIS coding-system.
1278 @var{code} is the character code in Shift-JIS as a cons of type bytes.
1279 The corresponding character is returned.
1282 @defun encode-shift-jis-char character
1283 This function encodes a JIS X 0208 character @var{character} to
1284 SHIFT-JIS coding-system. The corresponding character code in SHIFT-JIS
1285 is returned as a cons of two bytes.
1288 @defun decode-big5-char code
1289 This function decodes a Big5 character @var{code} of BIG5 coding-system.
1290 @var{code} is the character code in BIG5. The corresponding character
1294 @defun encode-big5-char character
1295 This function encodes the Big5 character @var{character} to BIG5
1296 coding-system. The corresponding character code in Big5 is returned.
1299 @node Predefined Coding Systems, , Big5 and Shift-JIS Functions, Coding Systems
1300 @subsection Coding Systems Implemented
1302 MULE initializes most of the commonly used coding systems at XEmacs's
1303 startup. A few others are initialized only when the relevant language
1304 environment is selected and support libraries are loaded. (NB: The
1305 following list is based on XEmacs 21.2.19, the development branch at the
1306 time of writing. The list may be somewhat different for other
1307 versions. Recent versions of GNU Emacs 20 implement a few more rare
1308 coding systems; work is being done to port these to XEmacs.)
1310 Unfortunately, there is not a consistent naming convention for character
1311 sets, and for practical purposes coding systems often take their name
1312 from their principal character sets (ASCII, KOI8-R, Shift JIS). Others
1313 take their names from the coding system (ISO-2022-JP, EUC-KR), and a few
1314 from their non-text usages (internal, binary). To provide for this, and
1315 for the fact that many coding systems have several common names, an
1316 aliasing system is provided. Finally, some effort has been made to use
1317 names that are registered as MIME charsets (this is why the name
1318 'shift_jis contains that un-Lisp-y underscore).
1320 There is a systematic naming convention regarding end-of-line (EOL)
1321 conventions for different systems. A coding system whose name ends in
1322 "-unix" forces the assumptions that lines are broken by newlines (0x0A).
1323 A coding system whose name ends in "-mac" forces the assumptions that
1324 lines are broken by ASCII CRs (0x0D). A coding system whose name ends
1325 in "-dos" forces the assumptions that lines are broken by CRLF sequences
1326 (0x0D 0x0A). These subsidiary coding systems are automatically derived
1327 from a base coding system. Use of the base coding system implies
1328 autodetection of the text file convention. (The fact that the -unix,
1329 -mac, and -dos are derived from a base system results in them showing up
1330 as "aliases" in `list-coding-systems'.) These subsidiaries have a
1331 consistent modeline indicator as well. "-dos" coding systems have ":T"
1332 appended to their modeline indicator, while "-mac" coding systems have
1333 ":t" appended (eg, "ISO8:t" for iso-2022-8-mac).
1335 In the following table, each coding system is given with its mode line
1336 indicator in parentheses. Non-textual coding systems are listed first,
1337 followed by textual coding systems and their aliases. (The coding system
1338 subsidiary modeline indicators ":T" and ":t" will be omitted from the
1339 table of coding systems.)
1341 ### SJT 1999-08-23 Maybe should order these by language? Definitely
1342 need language usage for the ISO-8859 family.
1344 Note that although true coding system aliases have been implemented for
1345 XEmacs 21.2, the coding system initialization has not yet been converted
1346 as of 21.2.19. So coding systems described as aliases have the same
1347 properties as the aliased coding system, but will not be equal as Lisp
1352 @item automatic-conversion
1354 @itemx undecided-dos
1355 @itemx undecided-mac
1356 @itemx undecided-unix
1358 Modeline indicator: @code{Auto}. A type @code{undecided} coding system.
1359 Attempts to determine an appropriate coding system from file contents or
1363 @itemx no-conversion
1366 @itemx raw-text-unix
1367 @itemx no-conversion-dos
1368 @itemx no-conversion-mac
1369 @itemx no-conversion-unix
1371 Modeline indicator: @code{Raw}. A type @code{no-conversion} coding system,
1372 which converts only line-break-codes. An implementation quirk means
1373 that this coding system is also used for ISO8859-1.
1376 Modeline indicator: @code{Binary}. A type @code{no-conversion} coding
1377 system which does no character coding or EOL conversions. An alias for
1378 @code{raw-text-unix}.
1381 @itemx alternativnyj-dos
1382 @itemx alternativnyj-mac
1383 @itemx alternativnyj-unix
1385 Modeline indicator: @code{Cy.Alt}. A type @code{ccl} coding system used for
1386 Alternativnyj, an encoding of the Cyrillic alphabet.
1393 Modeline indicator: @code{Zh/Big5}. A type @code{big5} coding system used for
1394 BIG5, the most common encoding of traditional Chinese as used in Taiwan.
1397 @itemx cn-gb-2312-dos
1398 @itemx cn-gb-2312-mac
1399 @itemx cn-gb-2312-unix
1401 Modeline indicator: @code{Zh-GB/EUC}. A type @code{iso2022} coding system used
1402 for simplified Chinese (as used in the People's Republic of China), with
1403 the @code{ascii} (G0), @code{chinese-gb2312} (G1), and @code{sisheng}
1404 (G2) character sets initially designated. Chinese EUC (Extended Unix
1408 @itemx ctext-hebrew-dos
1409 @itemx ctext-hebrew-mac
1410 @itemx ctext-hebrew-unix
1412 Modeline indicator: @code{CText/Hbrw}. A type @code{iso2022} coding system
1413 with the @code{ascii} (G0) and @code{hebrew-iso8859-8} (G1) character
1414 sets initially designated for Hebrew.
1421 Modeline indicator: @code{CText}. A type @code{iso2022} 8-bit coding system
1422 with the @code{ascii} (G0) and @code{latin-iso8859-1} (G1) character
1423 sets initially designated. X11 Compound Text Encoding. Often
1424 mistakenly recognized instead of EUC encodings; usual cause is
1425 inappropriate setting of @code{coding-priority-list}.
1429 Modeline indicator: @code{ESC/Quot}. A type @code{iso2022} 8-bit coding
1430 system with the @code{ascii} (G0) and @code{latin-iso8859-1} (G1)
1431 character sets initially designated and escape quoting. Unix EOL
1432 conversion (ie, no conversion). It is used for .ELC files.
1439 Modeline indicator: @code{Ja/EUC}. A type @code{iso2022} 8-bit coding system
1440 with @code{ascii} (G0), @code{japanese-jisx0208} (G1),
1441 @code{katakana-jisx0201} (G2), and @code{japanese-jisx0212} (G3)
1442 initially designated. Japanese EUC (Extended Unix Code).
1449 Modeline indicator: @code{ko/EUC}. A type @code{iso2022} 8-bit coding system
1450 with @code{ascii} (G0) and @code{korean-ksc5601} (G1) initially
1451 designated. Korean EUC (Extended Unix Code).
1454 Modeline indicator: @code{Zh-GB/Hz}. A type @code{no-conversion} coding
1455 system with Unix EOL convention (ie, no conversion) using
1456 post-read-decode and pre-write-encode functions to translate the Hz/ZW
1457 coding system used for Chinese.
1460 @itemx iso-2022-7bit-unix
1461 @itemx iso-2022-7bit-dos
1462 @itemx iso-2022-7bit-mac
1465 Modeline indicator: @code{ISO7}. A type @code{iso2022} 7-bit coding system
1466 with @code{ascii} (G0) initially designated. Other character sets must
1467 be explicitly designated to be used.
1469 @item iso-2022-7bit-ss2
1470 @itemx iso-2022-7bit-ss2-dos
1471 @itemx iso-2022-7bit-ss2-mac
1472 @itemx iso-2022-7bit-ss2-unix
1474 Modeline indicator: @code{ISO7/SS}. A type @code{iso2022} 7-bit coding system
1475 with @code{ascii} (G0) initially designated. Other character sets must
1476 be explicitly designated to be used. SS2 is used to invoke a
1477 96-charset, one character at a time.
1480 @itemx iso-2022-8-dos
1481 @itemx iso-2022-8-mac
1482 @itemx iso-2022-8-unix
1484 Modeline indicator: @code{ISO8}. A type @code{iso2022} 8-bit coding system
1485 with @code{ascii} (G0) and @code{latin-iso8859-1} (G1) initially
1486 designated. Other character sets must be explicitly designated to be
1487 used. No single-shift or locking-shift.
1489 @item iso-2022-8bit-ss2
1490 @itemx iso-2022-8bit-ss2-dos
1491 @itemx iso-2022-8bit-ss2-mac
1492 @itemx iso-2022-8bit-ss2-unix
1494 Modeline indicator: @code{ISO8/SS}. A type @code{iso2022} 8-bit coding system
1495 with @code{ascii} (G0) and @code{latin-iso8859-1} (G1) initially
1496 designated. Other character sets must be explicitly designated to be
1497 used. SS2 is used to invoke a 96-charset, one character at a time.
1499 @item iso-2022-int-1
1500 @itemx iso-2022-int-1-dos
1501 @itemx iso-2022-int-1-mac
1502 @itemx iso-2022-int-1-unix
1504 Modeline indicator: @code{INT-1}. A type @code{iso2022} 7-bit coding system
1505 with @code{ascii} (G0) and @code{korean-ksc5601} (G1) initially
1506 designated. ISO-2022-INT-1.
1508 @item iso-2022-jp-1978-irv
1509 @itemx iso-2022-jp-1978-irv-dos
1510 @itemx iso-2022-jp-1978-irv-mac
1511 @itemx iso-2022-jp-1978-irv-unix
1513 Modeline indicator: @code{Ja-78/7bit}. A type @code{iso2022} 7-bit coding
1514 system. For compatibility with old Japanese terminals; if you need to
1515 know, look at the source.
1518 @itemx iso-2022-jp-2 (ISO7/SS)
1519 @itemx iso-2022-jp-dos
1520 @itemx iso-2022-jp-mac
1521 @itemx iso-2022-jp-unix
1522 @itemx iso-2022-jp-2-dos
1523 @itemx iso-2022-jp-2-mac
1524 @itemx iso-2022-jp-2-unix
1526 Modeline indicator: @code{MULE/7bit}. A type @code{iso2022} 7-bit coding
1527 system with @code{ascii} (G0) initially designated, and complex
1528 specifications to insure backward compatibility with old Japanese
1529 systems. Used for communication with mail and news in Japan. The "-2"
1530 versions also use SS2 to invoke a 96-charset one character at a time.
1533 Modeline indicator: @code{Ko/7bit} A type @code{iso2022} 7-bit coding
1534 system with @code{ascii} (G0) and @code{korean-ksc5601} (G1) initially
1535 designated. Used for e-mail in Korea.
1538 @itemx iso-2022-lock-dos
1539 @itemx iso-2022-lock-mac
1540 @itemx iso-2022-lock-unix
1542 Modeline indicator: @code{ISO7/Lock}. A type @code{iso2022} 7-bit coding
1543 system with @code{ascii} (G0) initially designated, using Locking-Shift
1544 to invoke a 96-charset.
1547 @itemx iso-8859-1-dos
1548 @itemx iso-8859-1-mac
1549 @itemx iso-8859-1-unix
1551 Due to implementation, this is not a type @code{iso2022} coding system,
1552 but rather an alias for the @code{raw-text} coding system.
1555 @itemx iso-8859-2-dos
1556 @itemx iso-8859-2-mac
1557 @itemx iso-8859-2-unix
1559 Modeline indicator: @code{MIME/Ltn-2}. A type @code{iso2022} coding
1560 system with @code{ascii} (G0) and @code{latin-iso8859-2} (G1) initially
1564 @itemx iso-8859-3-dos
1565 @itemx iso-8859-3-mac
1566 @itemx iso-8859-3-unix
1568 Modeline indicator: @code{MIME/Ltn-3}. A type @code{iso2022} coding system
1569 with @code{ascii} (G0) and @code{latin-iso8859-3} (G1) initially
1573 @itemx iso-8859-4-dos
1574 @itemx iso-8859-4-mac
1575 @itemx iso-8859-4-unix
1577 Modeline indicator: @code{MIME/Ltn-4}. A type @code{iso2022} coding system
1578 with @code{ascii} (G0) and @code{latin-iso8859-4} (G1) initially
1582 @itemx iso-8859-5-dos
1583 @itemx iso-8859-5-mac
1584 @itemx iso-8859-5-unix
1586 Modeline indicator: @code{ISO8/Cyr}. A type @code{iso2022} coding system with
1587 @code{ascii} (G0) and @code{cyrillic-iso8859-5} (G1) initially invoked.
1590 @itemx iso-8859-7-dos
1591 @itemx iso-8859-7-mac
1592 @itemx iso-8859-7-unix
1594 Modeline indicator: @code{Grk}. A type @code{iso2022} coding system with
1595 @code{ascii} (G0) and @code{greek-iso8859-7} (G1) initially invoked.
1598 @itemx iso-8859-8-dos
1599 @itemx iso-8859-8-mac
1600 @itemx iso-8859-8-unix
1602 Modeline indicator: @code{MIME/Hbrw}. A type @code{iso2022} coding system with
1603 @code{ascii} (G0) and @code{hebrew-iso8859-8} (G1) initially invoked.
1606 @itemx iso-8859-9-dos
1607 @itemx iso-8859-9-mac
1608 @itemx iso-8859-9-unix
1610 Modeline indicator: @code{MIME/Ltn-5}. A type @code{iso2022} coding system
1611 with @code{ascii} (G0) and @code{latin-iso8859-9} (G1) initially
1619 Modeline indicator: @code{KOI8}. A type @code{ccl} coding-system used for
1620 KOI8-R, an encoding of the Cyrillic alphabet.
1623 @itemx shift_jis-dos
1624 @itemx shift_jis-mac
1625 @itemx shift_jis-unix
1627 Modeline indicator: @code{Ja/SJIS}. A type @code{shift-jis} coding-system
1628 implementing the Shift-JIS encoding for Japanese. The underscore is to
1629 conform to the MIME charset implementing this encoding.
1636 Modeline indicator: @code{TIS620}. A type @code{ccl} encoding for Thai. The
1637 external encoding is defined by TIS620, the internal encoding is
1638 peculiar to MULE, and called @code{thai-xtis}.
1642 Modeline indicator: @code{VIQR}. A type @code{no-conversion} coding
1643 system with Unix EOL convention (ie, no conversion) using
1644 post-read-decode and pre-write-encode functions to translate the VIQR
1645 coding system for Vietnamese.
1652 Modeline indicator: @code{VISCII}. A type @code{ccl} coding-system used
1653 for VISCII 1.1 for Vietnamese. Differs slightly from VSCII; VISCII is
1654 given priority by XEmacs.
1661 Modeline indicator: @code{VSCII}. A type @code{ccl} coding-system used
1662 for VSCII 1.1 for Vietnamese. Differs slightly from VISCII, which is
1663 given priority by XEmacs. Use
1664 @code{(prefer-coding-system 'vietnamese-vscii)} to give priority to VSCII.
1668 @node CCL, Category Tables, Coding Systems, MULE
1671 CCL (Code Conversion Language) is a simple structured programming
1672 language designed for character coding conversions. A CCL program is
1673 compiled to CCL code (represented by a vector of integers) and executed
1674 by the CCL interpreter embedded in Emacs. The CCL interpreter
1675 implements a virtual machine with 8 registers called @code{r0}, ...,
1676 @code{r7}, a number of control structures, and some I/O operators. Take
1677 care when using registers @code{r0} (used in implicit @dfn{set}
1678 statements) and especially @code{r7} (used internally by several
1679 statements and operations, especially for multiple return values and I/O
1682 CCL is used for code conversion during process I/O and file I/O for
1683 non-ISO2022 coding systems. (It is the only way for a user to specify a
1684 code conversion function.) It is also used for calculating the code
1685 point of an X11 font from a character code. However, since CCL is
1686 designed as a powerful programming language, it can be used for more
1687 generic calculation where efficiency is demanded. A combination of
1688 three or more arithmetic operations can be calculated faster by CCL than
1691 @strong{Warning:} The code in @file{src/mule-ccl.c} and
1692 @file{$packages/lisp/mule-base/mule-ccl.el} is the definitive
1693 description of CCL's semantics. The previous version of this section
1694 contained several typos and obsolete names left from earlier versions of
1695 MULE, and many may remain. (I am not an experienced CCL programmer; the
1696 few who know CCL well find writing English painful.)
1698 A CCL program transforms an input data stream into an output data
1699 stream. The input stream, held in a buffer of constant bytes, is left
1700 unchanged. The buffer may be filled by an external input operation,
1701 taken from an Emacs buffer, or taken from a Lisp string. The output
1702 buffer is a dynamic array of bytes, which can be written by an external
1703 output operation, inserted into an Emacs buffer, or returned as a Lisp
1706 A CCL program is a (Lisp) list containing two or three members. The
1707 first member is the @dfn{buffer magnification}, which indicates the
1708 required minimum size of the output buffer as a multiple of the input
1709 buffer. It is followed by the @dfn{main block} which executes while
1710 there is input remaining, and an optional @dfn{EOF block} which is
1711 executed when the input is exhausted. Both the main block and the EOF
1712 block are CCL blocks.
1714 A @dfn{CCL block} is either a CCL statement or list of CCL statements.
1715 A @dfn{CCL statement} is either a @dfn{set statement} (either an integer
1716 or an @dfn{assignment}, which is a list of a register to receive the
1717 assignment, an assignment operator, and an expression) or a @dfn{control
1718 statement} (a list starting with a keyword, whose allowable syntax
1719 depends on the keyword).
1722 * CCL Syntax:: CCL program syntax in BNF notation.
1723 * CCL Statements:: Semantics of CCL statements.
1724 * CCL Expressions:: Operators and expressions in CCL.
1725 * Calling CCL:: Running CCL programs.
1726 * CCL Examples:: The encoding functions for Big5 and KOI-8.
1729 @node CCL Syntax, CCL Statements, , CCL
1730 @comment Node, Next, Previous, Up
1731 @subsection CCL Syntax
1733 The full syntax of a CCL program in BNF notation:
1737 (BUFFER_MAGNIFICATION
1741 BUFFER_MAGNIFICATION := integer
1742 CCL_MAIN_BLOCK := CCL_BLOCK
1743 CCL_EOF_BLOCK := CCL_BLOCK
1746 STATEMENT | (STATEMENT [STATEMENT ...])
1748 SET | IF | BRANCH | LOOP | REPEAT | BREAK | READ | WRITE
1753 | (REG ASSIGNMENT_OPERATOR EXPRESSION)
1756 EXPRESSION := ARG | (EXPRESSION OPERATOR ARG)
1758 IF := (if EXPRESSION CCL_BLOCK [CCL_BLOCK])
1759 BRANCH := (branch EXPRESSION CCL_BLOCK [CCL_BLOCK ...])
1760 LOOP := (loop STATEMENT [STATEMENT ...])
1764 | (write-repeat [REG | integer | string])
1765 | (write-read-repeat REG [integer | ARRAY])
1768 | (read-if (REG OPERATOR ARG) CCL_BLOCK CCL_BLOCK)
1769 | (read-branch REG CCL_BLOCK [CCL_BLOCK ...])
1772 | (write EXPRESSION)
1773 | (write integer) | (write string) | (write REG ARRAY)
1775 CALL := (call ccl-program-name)
1778 REG := r0 | r1 | r2 | r3 | r4 | r5 | r6 | r7
1779 ARG := REG | integer
1781 + | - | * | / | % | & | '|' | ^ | << | >> | <8 | >8 | //
1782 | < | > | == | <= | >= | != | de-sjis | en-sjis
1783 ASSIGNMENT_OPERATOR :=
1784 += | -= | *= | /= | %= | &= | '|=' | ^= | <<= | >>=
1785 ARRAY := '[' integer ... ']'
1788 @node CCL Statements, CCL Expressions, CCL Syntax, CCL
1789 @comment Node, Next, Previous, Up
1790 @subsection CCL Statements
1792 The Emacs Code Conversion Language provides the following statement
1793 types: @dfn{set}, @dfn{if}, @dfn{branch}, @dfn{loop}, @dfn{repeat},
1794 @dfn{break}, @dfn{read}, @dfn{write}, @dfn{call}, and @dfn{end}.
1796 @heading Set statement:
1798 The @dfn{set} statement has three variants with the syntaxes
1799 @samp{(@var{reg} = @var{expression})},
1800 @samp{(@var{reg} @var{assignment_operator} @var{expression})}, and
1801 @samp{@var{integer}}. The assignment operator variation of the
1802 @dfn{set} statement works the same way as the corresponding C expression
1803 statement does. The assignment operators are @code{+=}, @code{-=},
1804 @code{*=}, @code{/=}, @code{%=}, @code{&=}, @code{|=}, @code{^=},
1805 @code{<<=}, and @code{>>=}, and they have the same meanings as in C. A
1806 "naked integer" @var{integer} is equivalent to a @var{set} statement of
1807 the form @code{(r0 = @var{integer})}.
1809 @heading I/O statements:
1811 The @dfn{read} statement takes one or more registers as arguments. It
1812 reads one byte (a C char) from the input into each register in turn.
1814 The @dfn{write} takes several forms. In the form @samp{(write @var{reg}
1815 ...)} it takes one or more registers as arguments and writes each in
1816 turn to the output. The integer in a register (interpreted as an
1817 Emchar) is encoded to multibyte form (ie, Bufbytes) and written to the
1818 current output buffer. If it is less than 256, it is written as is.
1819 The forms @samp{(write @var{expression})} and @samp{(write
1820 @var{integer})} are treated analogously. The form @samp{(write
1821 @var{string})} writes the constant string to the output. A
1822 "naked string" @samp{@var{string}} is equivalent to the statement @samp{(write
1823 @var{string})}. The form @samp{(write @var{reg} @var{array})} writes
1824 the @var{reg}th element of the @var{array} to the output.
1826 @heading Conditional statements:
1828 The @dfn{if} statement takes an @var{expression}, a @var{CCL block}, and
1829 an optional @var{second CCL block} as arguments. If the
1830 @var{expression} evaluates to non-zero, the first @var{CCL block} is
1831 executed. Otherwise, if there is a @var{second CCL block}, it is
1834 The @dfn{read-if} variant of the @dfn{if} statement takes an
1835 @var{expression}, a @var{CCL block}, and an optional @var{second CCL
1836 block} as arguments. The @var{expression} must have the form
1837 @code{(@var{reg} @var{operator} @var{operand})} (where @var{operand} is
1838 a register or an integer). The @code{read-if} statement first reads
1839 from the input into the first register operand in the @var{expression},
1840 then conditionally executes a CCL block just as the @code{if} statement
1843 The @dfn{branch} statement takes an @var{expression} and one or more CCL
1844 blocks as arguments. The CCL blocks are treated as a zero-indexed
1845 array, and the @code{branch} statement uses the @var{expression} as the
1846 index of the CCL block to execute. Null CCL blocks may be used as
1847 no-ops, continuing execution with the statement following the
1848 @code{branch} statement in the containing CCL block. Out-of-range
1849 values for the @var{expression} are also treated as no-ops.
1851 The @dfn{read-branch} variant of the @dfn{branch} statement takes an
1852 @var{register}, a @var{CCL block}, and an optional @var{second CCL
1853 block} as arguments. The @code{read-branch} statement first reads from
1854 the input into the @var{register}, then conditionally executes a CCL
1855 block just as the @code{branch} statement does.
1857 @heading Loop control statements:
1859 The @dfn{loop} statement creates a block with an implied jump from the
1860 end of the block back to its head. The loop is exited on a @code{break}
1861 statement, and continued without executing the tail by a @code{repeat}
1864 The @dfn{break} statement, written @samp{(break)}, terminates the
1865 current loop and continues with the next statement in the current
1868 The @dfn{repeat} statement has three variants, @code{repeat},
1869 @code{write-repeat}, and @code{write-read-repeat}. Each continues the
1870 current loop from its head, possibly after performing I/O.
1871 @code{repeat} takes no arguments and does no I/O before jumping.
1872 @code{write-repeat} takes a single argument (a register, an
1873 integer, or a string), writes it to the output, then jumps.
1874 @code{write-read-repeat} takes one or two arguments. The first must
1875 be a register. The second may be an integer or an array; if absent, it
1876 is implicitly set to the first (register) argument.
1877 @code{write-read-repeat} writes its second argument to the output, then
1878 reads from the input into the register, and finally jumps. See the
1879 @code{write} and @code{read} statements for the semantics of the I/O
1880 operations for each type of argument.
1882 @heading Other control statements:
1884 The @dfn{call} statement, written @samp{(call @var{ccl-program-name})},
1885 executes a CCL program as a subroutine. It does not return a value to
1886 the caller, but can modify the register status.
1888 The @dfn{end} statement, written @samp{(end)}, terminates the CCL
1889 program successfully, and returns to caller (which may be a CCL
1890 program). It does not alter the status of the registers.
1892 @node CCL Expressions, Calling CCL, CCL Statements, CCL
1893 @comment Node, Next, Previous, Up
1894 @subsection CCL Expressions
1896 CCL, unlike Lisp, uses infix expressions. The simplest CCL expressions
1897 consist of a single @var{operand}, either a register (one of @code{r0},
1898 ..., @code{r0}) or an integer. Complex expressions are lists of the
1899 form @code{( @var{expression} @var{operator} @var{operand} )}. Unlike
1900 C, assignments are not expressions.
1902 In the following table, @var{X} is the target resister for a @dfn{set}.
1903 In subexpressions, this is implicitly @code{r7}. This means that
1904 @code{>8}, @code{//}, @code{de-sjis}, and @code{en-sjis} cannot be used
1905 freely in subexpressions, since they return parts of their values in
1906 @code{r7}. @var{Y} may be an expression, register, or integer, while
1907 @var{Z} must be a register or an integer.
1909 @multitable @columnfractions .22 .14 .09 .55
1910 @item Name @tab Operator @tab Code @tab C-like Description
1911 @item CCL_PLUS @tab @code{+} @tab 0x00 @tab X = Y + Z
1912 @item CCL_MINUS @tab @code{-} @tab 0x01 @tab X = Y - Z
1913 @item CCL_MUL @tab @code{*} @tab 0x02 @tab X = Y * Z
1914 @item CCL_DIV @tab @code{/} @tab 0x03 @tab X = Y / Z
1915 @item CCL_MOD @tab @code{%} @tab 0x04 @tab X = Y % Z
1916 @item CCL_AND @tab @code{&} @tab 0x05 @tab X = Y & Z
1917 @item CCL_OR @tab @code{|} @tab 0x06 @tab X = Y | Z
1918 @item CCL_XOR @tab @code{^} @tab 0x07 @tab X = Y ^ Z
1919 @item CCL_LSH @tab @code{<<} @tab 0x08 @tab X = Y << Z
1920 @item CCL_RSH @tab @code{>>} @tab 0x09 @tab X = Y >> Z
1921 @item CCL_LSH8 @tab @code{<8} @tab 0x0A @tab X = (Y << 8) | Z
1922 @item CCL_RSH8 @tab @code{>8} @tab 0x0B @tab X = Y >> 8, r[7] = Y & 0xFF
1923 @item CCL_DIVMOD @tab @code{//} @tab 0x0C @tab X = Y / Z, r[7] = Y % Z
1924 @item CCL_LS @tab @code{<} @tab 0x10 @tab X = (X < Y)
1925 @item CCL_GT @tab @code{>} @tab 0x11 @tab X = (X > Y)
1926 @item CCL_EQ @tab @code{==} @tab 0x12 @tab X = (X == Y)
1927 @item CCL_LE @tab @code{<=} @tab 0x13 @tab X = (X <= Y)
1928 @item CCL_GE @tab @code{>=} @tab 0x14 @tab X = (X >= Y)
1929 @item CCL_NE @tab @code{!=} @tab 0x15 @tab X = (X != Y)
1930 @item CCL_ENCODE_SJIS @tab @code{en-sjis} @tab 0x16 @tab X = HIGHER_BYTE (SJIS (Y, Z))
1931 @item @tab @tab @tab r[7] = LOWER_BYTE (SJIS (Y, Z)
1932 @item CCL_DECODE_SJIS @tab @code{de-sjis} @tab 0x17 @tab X = HIGHER_BYTE (DE-SJIS (Y, Z))
1933 @item @tab @tab @tab r[7] = LOWER_BYTE (DE-SJIS (Y, Z))
1936 The CCL operators are as in C, with the addition of CCL_LSH8, CCL_RSH8,
1937 CCL_DIVMOD, CCL_ENCODE_SJIS, and CCL_DECODE_SJIS. The CCL_ENCODE_SJIS
1938 and CCL_DECODE_SJIS treat their first and second bytes as the high and
1939 low bytes of a two-byte character code. (SJIS stands for Shift JIS, an
1940 encoding of Japanese characters used by Microsoft. CCL_ENCODE_SJIS is a
1941 complicated transformation of the Japanese standard JIS encoding to
1942 Shift JIS. CCL_DECODE_SJIS is its inverse.) It is somewhat odd to
1943 represent the SJIS operations in infix form.
1945 @node Calling CCL, CCL Examples, CCL Expressions, CCL
1946 @comment Node, Next, Previous, Up
1947 @subsection Calling CCL
1949 CCL programs are called automatically during Emacs buffer I/O when the
1950 external representation has a coding system type of @code{shift-jis},
1951 @code{big5}, or @code{ccl}. The program is specified by the coding
1952 system (@pxref{Coding Systems}). You can also call CCL programs from
1953 other CCL programs, and from Lisp using these functions:
1955 @defun ccl-execute ccl-program status
1956 Execute @var{ccl-program} with registers initialized by
1957 @var{status}. @var{ccl-program} is a vector of compiled CCL code
1958 created by @code{ccl-compile}. It is an error for the program to try to
1959 execute a CCL I/O command. @var{status} must be a vector of nine
1960 values, specifying the initial value for the R0, R1 .. R7 registers and
1961 for the instruction counter IC. A @code{nil} value for a register
1962 initializer causes the register to be set to 0. A @code{nil} value for
1963 the IC initializer causes execution to start at the beginning of the
1964 program. When the program is done, @var{status} is modified (by
1965 side-effect) to contain the ending values for the corresponding
1969 @defun ccl-execute-on-string ccl-program status string &optional continue
1970 Execute @var{ccl-program} with initial @var{status} on
1971 @var{string}. @var{ccl-program} is a vector of compiled CCL code
1972 created by @code{ccl-compile}. @var{status} must be a vector of nine
1973 values, specifying the initial value for the R0, R1 .. R7 registers and
1974 for the instruction counter IC. A @code{nil} value for a register
1975 initializer causes the register to be set to 0. A @code{nil} value for
1976 the IC initializer causes execution to start at the beginning of the
1977 program. An optional fourth argument @var{continue}, if non-@code{nil}, causes
1979 remain on the unsatisfied read operation if the program terminates due
1980 to exhaustion of the input buffer. Otherwise the IC is set to the end
1981 of the program. When the program is done, @var{status} is modified (by
1982 side-effect) to contain the ending values for the corresponding
1983 registers and IC. Returns the resulting string.
1986 To call a CCL program from another CCL program, it must first be
1989 @defun register-ccl-program name ccl-program
1990 Register @var{name} for CCL program @var{ccl-program} in
1991 @code{ccl-program-table}. @var{ccl-program} should be the compiled form of
1992 a CCL program, or @code{nil}. Return index number of the registered CCL
1996 Information about the processor time used by the CCL interpreter can be
1997 obtained using these functions:
1999 @defun ccl-elapsed-time
2000 Returns the elapsed processor time of the CCL interpreter as cons of
2001 user and system time, as
2002 floating point numbers measured in seconds. If only one
2003 overall value can be determined, the return value will be a cons of that
2007 @defun ccl-reset-elapsed-time
2008 Resets the CCL interpreter's internal elapsed time registers.
2011 @node CCL Examples, , Calling CCL, CCL
2012 @comment Node, Next, Previous, Up
2013 @subsection CCL Examples
2015 This section is not yet written.
2017 @node Category Tables, , CCL, MULE
2018 @section Category Tables
2020 A category table is a type of char table used for keeping track of
2021 categories. Categories are used for classifying characters for use in
2022 regexps---you can refer to a category rather than having to use a
2023 complicated [] expression (and category lookups are significantly
2026 There are 95 different categories available, one for each printable
2027 character (including space) in the ASCII charset. Each category is
2028 designated by one such character, called a @dfn{category designator}.
2029 They are specified in a regexp using the syntax @samp{\cX}, where X is a
2030 category designator. (This is not yet implemented.)
2032 A category table specifies, for each character, the categories that
2033 the character is in. Note that a character can be in more than one
2034 category. More specifically, a category table maps from a character to
2035 either the value @code{nil} (meaning the character is in no categories)
2036 or a 95-element bit vector, specifying for each of the 95 categories
2037 whether the character is in that category.
2039 Special Lisp functions are provided that abstract this, so you do not
2040 have to directly manipulate bit vectors.
2042 @defun category-table-p object
2043 This function returns @code{t} if @var{object} is a category table.
2046 @defun category-table &optional buffer
2047 This function returns the current category table. This is the one
2048 specified by the current buffer, or by @var{buffer} if it is
2052 @defun standard-category-table
2053 This function returns the standard category table. This is the one used
2057 @defun copy-category-table &optional category-table
2058 This function returns a new category table which is a copy of
2059 @var{category-table}, which defaults to the standard category table.
2062 @defun set-category-table category-table &optional buffer
2063 This function selects @var{category-table} as the new category table for
2064 @var{buffer}. @var{buffer} defaults to the current buffer if omitted.
2067 @defun category-designator-p object
2068 This function returns @code{t} if @var{object} is a category designator (a
2069 char in the range @samp{' '} to @samp{'~'}).
2072 @defun category-table-value-p object
2073 This function returns @code{t} if @var{object} is a category table value.
2074 Valid values are @code{nil} or a bit vector of size 95.