(CCL_WRITE_CHAR): Check if CH is valid or not. If
[bpt/emacs.git] / src / ccl.c
1 /* CCL (Code Conversion Language) interpreter.
2 Copyright (C) 1995, 1997 Electrotechnical Laboratory, JAPAN.
3 Licensed to the Free Software Foundation.
4
5 This file is part of GNU Emacs.
6
7 GNU Emacs is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 2, or (at your option)
10 any later version.
11
12 GNU Emacs is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
16
17 You should have received a copy of the GNU General Public License
18 along with GNU Emacs; see the file COPYING. If not, write to
19 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
20 Boston, MA 02111-1307, USA. */
21
22 #ifdef emacs
23 #include <config.h>
24 #endif
25
26 #include <stdio.h>
27
28 #ifdef emacs
29
30 #include "lisp.h"
31 #include "charset.h"
32 #include "ccl.h"
33 #include "coding.h"
34
35 #else /* not emacs */
36
37 #include "mulelib.h"
38
39 #endif /* not emacs */
40
41 /* This contains all code conversion map available to CCL. */
42 Lisp_Object Vcode_conversion_map_vector;
43
44 /* Alist of fontname patterns vs corresponding CCL program. */
45 Lisp_Object Vfont_ccl_encoder_alist;
46
47 /* This symbol is a property which assocates with ccl program vector.
48 Ex: (get 'ccl-big5-encoder 'ccl-program) returns ccl program vector. */
49 Lisp_Object Qccl_program;
50
51 /* These symbols are properties which associate with code conversion
52 map and their ID respectively. */
53 Lisp_Object Qcode_conversion_map;
54 Lisp_Object Qcode_conversion_map_id;
55
56 /* Symbols of ccl program have this property, a value of the property
57 is an index for Vccl_protram_table. */
58 Lisp_Object Qccl_program_idx;
59
60 /* Table of registered CCL programs. Each element is a vector of
61 NAME, CCL_PROG, and RESOLVEDP where NAME (symbol) is the name of
62 the program, CCL_PROG (vector) is the compiled code of the program,
63 RESOLVEDP (t or nil) is the flag to tell if symbols in CCL_PROG is
64 already resolved to index numbers or not. */
65 Lisp_Object Vccl_program_table;
66
67 /* CCL (Code Conversion Language) is a simple language which has
68 operations on one input buffer, one output buffer, and 7 registers.
69 The syntax of CCL is described in `ccl.el'. Emacs Lisp function
70 `ccl-compile' compiles a CCL program and produces a CCL code which
71 is a vector of integers. The structure of this vector is as
72 follows: The 1st element: buffer-magnification, a factor for the
73 size of output buffer compared with the size of input buffer. The
74 2nd element: address of CCL code to be executed when encountered
75 with end of input stream. The 3rd and the remaining elements: CCL
76 codes. */
77
78 /* Header of CCL compiled code */
79 #define CCL_HEADER_BUF_MAG 0
80 #define CCL_HEADER_EOF 1
81 #define CCL_HEADER_MAIN 2
82
83 /* CCL code is a sequence of 28-bit non-negative integers (i.e. the
84 MSB is always 0), each contains CCL command and/or arguments in the
85 following format:
86
87 |----------------- integer (28-bit) ------------------|
88 |------- 17-bit ------|- 3-bit --|- 3-bit --|- 5-bit -|
89 |--constant argument--|-register-|-register-|-command-|
90 ccccccccccccccccc RRR rrr XXXXX
91 or
92 |------- relative address -------|-register-|-command-|
93 cccccccccccccccccccc rrr XXXXX
94 or
95 |------------- constant or other args ----------------|
96 cccccccccccccccccccccccccccc
97
98 where, `cc...c' is a non-negative integer indicating constant value
99 (the left most `c' is always 0) or an absolute jump address, `RRR'
100 and `rrr' are CCL register number, `XXXXX' is one of the following
101 CCL commands. */
102
103 /* CCL commands
104
105 Each comment fields shows one or more lines for command syntax and
106 the following lines for semantics of the command. In semantics, IC
107 stands for Instruction Counter. */
108
109 #define CCL_SetRegister 0x00 /* Set register a register value:
110 1:00000000000000000RRRrrrXXXXX
111 ------------------------------
112 reg[rrr] = reg[RRR];
113 */
114
115 #define CCL_SetShortConst 0x01 /* Set register a short constant value:
116 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
117 ------------------------------
118 reg[rrr] = CCCCCCCCCCCCCCCCCCC;
119 */
120
121 #define CCL_SetConst 0x02 /* Set register a constant value:
122 1:00000000000000000000rrrXXXXX
123 2:CONSTANT
124 ------------------------------
125 reg[rrr] = CONSTANT;
126 IC++;
127 */
128
129 #define CCL_SetArray 0x03 /* Set register an element of array:
130 1:CCCCCCCCCCCCCCCCCRRRrrrXXXXX
131 2:ELEMENT[0]
132 3:ELEMENT[1]
133 ...
134 ------------------------------
135 if (0 <= reg[RRR] < CC..C)
136 reg[rrr] = ELEMENT[reg[RRR]];
137 IC += CC..C;
138 */
139
140 #define CCL_Jump 0x04 /* Jump:
141 1:A--D--D--R--E--S--S-000XXXXX
142 ------------------------------
143 IC += ADDRESS;
144 */
145
146 /* Note: If CC..C is greater than 0, the second code is omitted. */
147
148 #define CCL_JumpCond 0x05 /* Jump conditional:
149 1:A--D--D--R--E--S--S-rrrXXXXX
150 ------------------------------
151 if (!reg[rrr])
152 IC += ADDRESS;
153 */
154
155
156 #define CCL_WriteRegisterJump 0x06 /* Write register and jump:
157 1:A--D--D--R--E--S--S-rrrXXXXX
158 ------------------------------
159 write (reg[rrr]);
160 IC += ADDRESS;
161 */
162
163 #define CCL_WriteRegisterReadJump 0x07 /* Write register, read, and jump:
164 1:A--D--D--R--E--S--S-rrrXXXXX
165 2:A--D--D--R--E--S--S-rrrYYYYY
166 -----------------------------
167 write (reg[rrr]);
168 IC++;
169 read (reg[rrr]);
170 IC += ADDRESS;
171 */
172 /* Note: If read is suspended, the resumed execution starts from the
173 second code (YYYYY == CCL_ReadJump). */
174
175 #define CCL_WriteConstJump 0x08 /* Write constant and jump:
176 1:A--D--D--R--E--S--S-000XXXXX
177 2:CONST
178 ------------------------------
179 write (CONST);
180 IC += ADDRESS;
181 */
182
183 #define CCL_WriteConstReadJump 0x09 /* Write constant, read, and jump:
184 1:A--D--D--R--E--S--S-rrrXXXXX
185 2:CONST
186 3:A--D--D--R--E--S--S-rrrYYYYY
187 -----------------------------
188 write (CONST);
189 IC += 2;
190 read (reg[rrr]);
191 IC += ADDRESS;
192 */
193 /* Note: If read is suspended, the resumed execution starts from the
194 second code (YYYYY == CCL_ReadJump). */
195
196 #define CCL_WriteStringJump 0x0A /* Write string and jump:
197 1:A--D--D--R--E--S--S-000XXXXX
198 2:LENGTH
199 3:0000STRIN[0]STRIN[1]STRIN[2]
200 ...
201 ------------------------------
202 write_string (STRING, LENGTH);
203 IC += ADDRESS;
204 */
205
206 #define CCL_WriteArrayReadJump 0x0B /* Write an array element, read, and jump:
207 1:A--D--D--R--E--S--S-rrrXXXXX
208 2:LENGTH
209 3:ELEMENET[0]
210 4:ELEMENET[1]
211 ...
212 N:A--D--D--R--E--S--S-rrrYYYYY
213 ------------------------------
214 if (0 <= reg[rrr] < LENGTH)
215 write (ELEMENT[reg[rrr]]);
216 IC += LENGTH + 2; (... pointing at N+1)
217 read (reg[rrr]);
218 IC += ADDRESS;
219 */
220 /* Note: If read is suspended, the resumed execution starts from the
221 Nth code (YYYYY == CCL_ReadJump). */
222
223 #define CCL_ReadJump 0x0C /* Read and jump:
224 1:A--D--D--R--E--S--S-rrrYYYYY
225 -----------------------------
226 read (reg[rrr]);
227 IC += ADDRESS;
228 */
229
230 #define CCL_Branch 0x0D /* Jump by branch table:
231 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
232 2:A--D--D--R--E-S-S[0]000XXXXX
233 3:A--D--D--R--E-S-S[1]000XXXXX
234 ...
235 ------------------------------
236 if (0 <= reg[rrr] < CC..C)
237 IC += ADDRESS[reg[rrr]];
238 else
239 IC += ADDRESS[CC..C];
240 */
241
242 #define CCL_ReadRegister 0x0E /* Read bytes into registers:
243 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
244 2:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
245 ...
246 ------------------------------
247 while (CCC--)
248 read (reg[rrr]);
249 */
250
251 #define CCL_WriteExprConst 0x0F /* write result of expression:
252 1:00000OPERATION000RRR000XXXXX
253 2:CONSTANT
254 ------------------------------
255 write (reg[RRR] OPERATION CONSTANT);
256 IC++;
257 */
258
259 /* Note: If the Nth read is suspended, the resumed execution starts
260 from the Nth code. */
261
262 #define CCL_ReadBranch 0x10 /* Read one byte into a register,
263 and jump by branch table:
264 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
265 2:A--D--D--R--E-S-S[0]000XXXXX
266 3:A--D--D--R--E-S-S[1]000XXXXX
267 ...
268 ------------------------------
269 read (read[rrr]);
270 if (0 <= reg[rrr] < CC..C)
271 IC += ADDRESS[reg[rrr]];
272 else
273 IC += ADDRESS[CC..C];
274 */
275
276 #define CCL_WriteRegister 0x11 /* Write registers:
277 1:CCCCCCCCCCCCCCCCCCCrrrXXXXX
278 2:CCCCCCCCCCCCCCCCCCCrrrXXXXX
279 ...
280 ------------------------------
281 while (CCC--)
282 write (reg[rrr]);
283 ...
284 */
285
286 /* Note: If the Nth write is suspended, the resumed execution
287 starts from the Nth code. */
288
289 #define CCL_WriteExprRegister 0x12 /* Write result of expression
290 1:00000OPERATIONRrrRRR000XXXXX
291 ------------------------------
292 write (reg[RRR] OPERATION reg[Rrr]);
293 */
294
295 #define CCL_Call 0x13 /* Call the CCL program whose ID is
296 CC..C or cc..c.
297 1:CCCCCCCCCCCCCCCCCCCCFFFXXXXX
298 [2:00000000cccccccccccccccccccc]
299 ------------------------------
300 if (FFF)
301 call (cc..c)
302 IC++;
303 else
304 call (CC..C)
305 */
306
307 #define CCL_WriteConstString 0x14 /* Write a constant or a string:
308 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
309 [2:0000STRIN[0]STRIN[1]STRIN[2]]
310 [...]
311 -----------------------------
312 if (!rrr)
313 write (CC..C)
314 else
315 write_string (STRING, CC..C);
316 IC += (CC..C + 2) / 3;
317 */
318
319 #define CCL_WriteArray 0x15 /* Write an element of array:
320 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
321 2:ELEMENT[0]
322 3:ELEMENT[1]
323 ...
324 ------------------------------
325 if (0 <= reg[rrr] < CC..C)
326 write (ELEMENT[reg[rrr]]);
327 IC += CC..C;
328 */
329
330 #define CCL_End 0x16 /* Terminate:
331 1:00000000000000000000000XXXXX
332 ------------------------------
333 terminate ();
334 */
335
336 /* The following two codes execute an assignment arithmetic/logical
337 operation. The form of the operation is like REG OP= OPERAND. */
338
339 #define CCL_ExprSelfConst 0x17 /* REG OP= constant:
340 1:00000OPERATION000000rrrXXXXX
341 2:CONSTANT
342 ------------------------------
343 reg[rrr] OPERATION= CONSTANT;
344 */
345
346 #define CCL_ExprSelfReg 0x18 /* REG1 OP= REG2:
347 1:00000OPERATION000RRRrrrXXXXX
348 ------------------------------
349 reg[rrr] OPERATION= reg[RRR];
350 */
351
352 /* The following codes execute an arithmetic/logical operation. The
353 form of the operation is like REG_X = REG_Y OP OPERAND2. */
354
355 #define CCL_SetExprConst 0x19 /* REG_X = REG_Y OP constant:
356 1:00000OPERATION000RRRrrrXXXXX
357 2:CONSTANT
358 ------------------------------
359 reg[rrr] = reg[RRR] OPERATION CONSTANT;
360 IC++;
361 */
362
363 #define CCL_SetExprReg 0x1A /* REG1 = REG2 OP REG3:
364 1:00000OPERATIONRrrRRRrrrXXXXX
365 ------------------------------
366 reg[rrr] = reg[RRR] OPERATION reg[Rrr];
367 */
368
369 #define CCL_JumpCondExprConst 0x1B /* Jump conditional according to
370 an operation on constant:
371 1:A--D--D--R--E--S--S-rrrXXXXX
372 2:OPERATION
373 3:CONSTANT
374 -----------------------------
375 reg[7] = reg[rrr] OPERATION CONSTANT;
376 if (!(reg[7]))
377 IC += ADDRESS;
378 else
379 IC += 2
380 */
381
382 #define CCL_JumpCondExprReg 0x1C /* Jump conditional according to
383 an operation on register:
384 1:A--D--D--R--E--S--S-rrrXXXXX
385 2:OPERATION
386 3:RRR
387 -----------------------------
388 reg[7] = reg[rrr] OPERATION reg[RRR];
389 if (!reg[7])
390 IC += ADDRESS;
391 else
392 IC += 2;
393 */
394
395 #define CCL_ReadJumpCondExprConst 0x1D /* Read and jump conditional according
396 to an operation on constant:
397 1:A--D--D--R--E--S--S-rrrXXXXX
398 2:OPERATION
399 3:CONSTANT
400 -----------------------------
401 read (reg[rrr]);
402 reg[7] = reg[rrr] OPERATION CONSTANT;
403 if (!reg[7])
404 IC += ADDRESS;
405 else
406 IC += 2;
407 */
408
409 #define CCL_ReadJumpCondExprReg 0x1E /* Read and jump conditional according
410 to an operation on register:
411 1:A--D--D--R--E--S--S-rrrXXXXX
412 2:OPERATION
413 3:RRR
414 -----------------------------
415 read (reg[rrr]);
416 reg[7] = reg[rrr] OPERATION reg[RRR];
417 if (!reg[7])
418 IC += ADDRESS;
419 else
420 IC += 2;
421 */
422
423 #define CCL_Extension 0x1F /* Extended CCL code
424 1:ExtendedCOMMNDRrrRRRrrrXXXXX
425 2:ARGUEMENT
426 3:...
427 ------------------------------
428 extended_command (rrr,RRR,Rrr,ARGS)
429 */
430
431 /*
432 Here after, Extended CCL Instructions.
433 Bit length of extended command is 14.
434 Therefore, the instruction code range is 0..16384(0x3fff).
435 */
436
437 /* Read a multibyte characeter.
438 A code point is stored into reg[rrr]. A charset ID is stored into
439 reg[RRR]. */
440
441 #define CCL_ReadMultibyteChar2 0x00 /* Read Multibyte Character
442 1:ExtendedCOMMNDRrrRRRrrrXXXXX */
443
444 /* Write a multibyte character.
445 Write a character whose code point is reg[rrr] and the charset ID
446 is reg[RRR]. */
447
448 #define CCL_WriteMultibyteChar2 0x01 /* Write Multibyte Character
449 1:ExtendedCOMMNDRrrRRRrrrXXXXX */
450
451 /* Translate a character whose code point is reg[rrr] and the charset
452 ID is reg[RRR] by a translation table whose ID is reg[Rrr].
453
454 A translated character is set in reg[rrr] (code point) and reg[RRR]
455 (charset ID). */
456
457 #define CCL_TranslateCharacter 0x02 /* Translate a multibyte character
458 1:ExtendedCOMMNDRrrRRRrrrXXXXX */
459
460 /* Translate a character whose code point is reg[rrr] and the charset
461 ID is reg[RRR] by a translation table whose ID is ARGUMENT.
462
463 A translated character is set in reg[rrr] (code point) and reg[RRR]
464 (charset ID). */
465
466 #define CCL_TranslateCharacterConstTbl 0x03 /* Translate a multibyte character
467 1:ExtendedCOMMNDRrrRRRrrrXXXXX
468 2:ARGUMENT(Translation Table ID)
469 */
470
471 /* Iterate looking up MAPs for reg[rrr] starting from the Nth (N =
472 reg[RRR]) MAP until some value is found.
473
474 Each MAP is a Lisp vector whose element is number, nil, t, or
475 lambda.
476 If the element is nil, ignore the map and proceed to the next map.
477 If the element is t or lambda, finish without changing reg[rrr].
478 If the element is a number, set reg[rrr] to the number and finish.
479
480 Detail of the map structure is descibed in the comment for
481 CCL_MapMultiple below. */
482
483 #define CCL_IterateMultipleMap 0x10 /* Iterate multiple maps
484 1:ExtendedCOMMNDXXXRRRrrrXXXXX
485 2:NUMBER of MAPs
486 3:MAP-ID1
487 4:MAP-ID2
488 ...
489 */
490
491 /* Map the code in reg[rrr] by MAPs starting from the Nth (N =
492 reg[RRR]) map.
493
494 MAPs are supplied in the succeeding CCL codes as follows:
495
496 When CCL program gives this nested structure of map to this command:
497 ((MAP-ID11
498 MAP-ID12
499 (MAP-ID121 MAP-ID122 MAP-ID123)
500 MAP-ID13)
501 (MAP-ID21
502 (MAP-ID211 (MAP-ID2111) MAP-ID212)
503 MAP-ID22)),
504 the compiled CCL codes has this sequence:
505 CCL_MapMultiple (CCL code of this command)
506 16 (total number of MAPs and SEPARATORs)
507 -7 (1st SEPARATOR)
508 MAP-ID11
509 MAP-ID12
510 -3 (2nd SEPARATOR)
511 MAP-ID121
512 MAP-ID122
513 MAP-ID123
514 MAP-ID13
515 -7 (3rd SEPARATOR)
516 MAP-ID21
517 -4 (4th SEPARATOR)
518 MAP-ID211
519 -1 (5th SEPARATOR)
520 MAP_ID2111
521 MAP-ID212
522 MAP-ID22
523
524 A value of each SEPARATOR follows this rule:
525 MAP-SET := SEPARATOR [(MAP-ID | MAP-SET)]+
526 SEPARATOR := -(number of MAP-IDs and SEPARATORs in the MAP-SET)
527
528 (*)....Nest level of MAP-SET must not be over than MAX_MAP_SET_LEVEL.
529
530 When some map fails to map (i.e. it doesn't have a value for
531 reg[rrr]), the mapping is treated as identity.
532
533 The mapping is iterated for all maps in each map set (set of maps
534 separated by SEPARATOR) except in the case that lambda is
535 encountered. More precisely, the mapping proceeds as below:
536
537 At first, VAL0 is set to reg[rrr], and it is translated by the
538 first map to VAL1. Then, VAL1 is translated by the next map to
539 VAL2. This mapping is iterated until the last map is used. The
540 result of the mapping is the last value of VAL?. When the mapping
541 process reached to the end of the map set, it moves to the next
542 map set. If the next does not exit, the mapping process terminates,
543 and regard the last value as a result.
544
545 But, when VALm is mapped to VALn and VALn is not a number, the
546 mapping proceed as below:
547
548 If VALn is nil, the lastest map is ignored and the mapping of VALm
549 proceed to the next map.
550
551 In VALn is t, VALm is reverted to reg[rrr] and the mapping of VALm
552 proceed to the next map.
553
554 If VALn is lambda, move to the next map set like reaching to the
555 end of the current map set.
556
557 If VALn is a symbol, call the CCL program refered by it.
558 Then, use reg[rrr] as a mapped value except for -1, -2 and -3.
559 Such special values are regarded as nil, t, and lambda respectively.
560
561 Each map is a Lisp vector of the following format (a) or (b):
562 (a)......[STARTPOINT VAL1 VAL2 ...]
563 (b)......[t VAL STARTPOINT ENDPOINT],
564 where
565 STARTPOINT is an offset to be used for indexing a map,
566 ENDPOINT is a maximum index number of a map,
567 VAL and VALn is a number, nil, t, or lambda.
568
569 Valid index range of a map of type (a) is:
570 STARTPOINT <= index < STARTPOINT + map_size - 1
571 Valid index range of a map of type (b) is:
572 STARTPOINT <= index < ENDPOINT */
573
574 #define CCL_MapMultiple 0x11 /* Mapping by multiple code conversion maps
575 1:ExtendedCOMMNDXXXRRRrrrXXXXX
576 2:N-2
577 3:SEPARATOR_1 (< 0)
578 4:MAP-ID_1
579 5:MAP-ID_2
580 ...
581 M:SEPARATOR_x (< 0)
582 M+1:MAP-ID_y
583 ...
584 N:SEPARATOR_z (< 0)
585 */
586
587 #define MAX_MAP_SET_LEVEL 30
588
589 typedef struct
590 {
591 int rest_length;
592 int orig_val;
593 } tr_stack;
594
595 static tr_stack mapping_stack[MAX_MAP_SET_LEVEL];
596 static tr_stack *mapping_stack_pointer;
597
598 /* If this variable is non-zero, it indicates the stack_idx
599 of immediately called by CCL_MapMultiple. */
600 static int stack_idx_of_map_multiple;
601
602 #define PUSH_MAPPING_STACK(restlen, orig) \
603 do { \
604 mapping_stack_pointer->rest_length = (restlen); \
605 mapping_stack_pointer->orig_val = (orig); \
606 mapping_stack_pointer++; \
607 } while (0)
608
609 #define POP_MAPPING_STACK(restlen, orig) \
610 do { \
611 mapping_stack_pointer--; \
612 (restlen) = mapping_stack_pointer->rest_length; \
613 (orig) = mapping_stack_pointer->orig_val; \
614 } while (0)
615
616 #define CCL_CALL_FOR_MAP_INSTRUCTION(symbol, ret_ic) \
617 if (1) \
618 { \
619 struct ccl_program called_ccl; \
620 if (stack_idx >= 256 \
621 || (setup_ccl_program (&called_ccl, (symbol)) != 0)) \
622 { \
623 if (stack_idx > 0) \
624 { \
625 ccl_prog = ccl_prog_stack_struct[0].ccl_prog; \
626 ic = ccl_prog_stack_struct[0].ic; \
627 } \
628 CCL_INVALID_CMD; \
629 } \
630 ccl_prog_stack_struct[stack_idx].ccl_prog = ccl_prog; \
631 ccl_prog_stack_struct[stack_idx].ic = (ret_ic); \
632 stack_idx++; \
633 ccl_prog = called_ccl.prog; \
634 ic = CCL_HEADER_MAIN; \
635 goto ccl_repeat; \
636 } \
637 else
638
639 #define CCL_MapSingle 0x12 /* Map by single code conversion map
640 1:ExtendedCOMMNDXXXRRRrrrXXXXX
641 2:MAP-ID
642 ------------------------------
643 Map reg[rrr] by MAP-ID.
644 If some valid mapping is found,
645 set reg[rrr] to the result,
646 else
647 set reg[RRR] to -1.
648 */
649
650 /* CCL arithmetic/logical operators. */
651 #define CCL_PLUS 0x00 /* X = Y + Z */
652 #define CCL_MINUS 0x01 /* X = Y - Z */
653 #define CCL_MUL 0x02 /* X = Y * Z */
654 #define CCL_DIV 0x03 /* X = Y / Z */
655 #define CCL_MOD 0x04 /* X = Y % Z */
656 #define CCL_AND 0x05 /* X = Y & Z */
657 #define CCL_OR 0x06 /* X = Y | Z */
658 #define CCL_XOR 0x07 /* X = Y ^ Z */
659 #define CCL_LSH 0x08 /* X = Y << Z */
660 #define CCL_RSH 0x09 /* X = Y >> Z */
661 #define CCL_LSH8 0x0A /* X = (Y << 8) | Z */
662 #define CCL_RSH8 0x0B /* X = Y >> 8, r[7] = Y & 0xFF */
663 #define CCL_DIVMOD 0x0C /* X = Y / Z, r[7] = Y % Z */
664 #define CCL_LS 0x10 /* X = (X < Y) */
665 #define CCL_GT 0x11 /* X = (X > Y) */
666 #define CCL_EQ 0x12 /* X = (X == Y) */
667 #define CCL_LE 0x13 /* X = (X <= Y) */
668 #define CCL_GE 0x14 /* X = (X >= Y) */
669 #define CCL_NE 0x15 /* X = (X != Y) */
670
671 #define CCL_DECODE_SJIS 0x16 /* X = HIGHER_BYTE (DE-SJIS (Y, Z))
672 r[7] = LOWER_BYTE (DE-SJIS (Y, Z)) */
673 #define CCL_ENCODE_SJIS 0x17 /* X = HIGHER_BYTE (SJIS (Y, Z))
674 r[7] = LOWER_BYTE (SJIS (Y, Z) */
675
676 /* Terminate CCL program successfully. */
677 #define CCL_SUCCESS \
678 if (1) \
679 { \
680 ccl->status = CCL_STAT_SUCCESS; \
681 goto ccl_finish; \
682 } \
683 else
684
685 /* Suspend CCL program because of reading from empty input buffer or
686 writing to full output buffer. When this program is resumed, the
687 same I/O command is executed. */
688 #define CCL_SUSPEND(stat) \
689 if (1) \
690 { \
691 ic--; \
692 ccl->status = stat; \
693 goto ccl_finish; \
694 } \
695 else
696
697 /* Terminate CCL program because of invalid command. Should not occur
698 in the normal case. */
699 #define CCL_INVALID_CMD \
700 if (1) \
701 { \
702 ccl->status = CCL_STAT_INVALID_CMD; \
703 goto ccl_error_handler; \
704 } \
705 else
706
707 /* Encode one character CH to multibyte form and write to the current
708 output buffer. If CH is less than 256, CH is written as is. */
709 #define CCL_WRITE_CHAR(ch) \
710 do { \
711 int bytes = SINGLE_BYTE_CHAR_P (ch) ? 1: CHAR_BYTES (ch); \
712 if (!dst) \
713 CCL_INVALID_CMD; \
714 else if (dst + bytes + extra_bytes < (dst_bytes ? dst_end : src)) \
715 { \
716 if (bytes == 1) \
717 { \
718 *dst++ = (ch); \
719 if ((ch) >= 0x80 && (ch) < 0xA0) \
720 /* We may have to convert this eight-bit char to \
721 multibyte form later. */ \
722 extra_bytes++; \
723 } \
724 else if (CHAR_VALID_P (ch, 0)) \
725 dst += CHAR_STRING (ch, dst); \
726 else \
727 CCL_INVALID_CMD; \
728 } \
729 else \
730 CCL_SUSPEND (CCL_STAT_SUSPEND_BY_DST); \
731 } while (0)
732
733 /* Write a string at ccl_prog[IC] of length LEN to the current output
734 buffer. */
735 #define CCL_WRITE_STRING(len) \
736 do { \
737 if (!dst) \
738 CCL_INVALID_CMD; \
739 else if (dst + len <= (dst_bytes ? dst_end : src)) \
740 for (i = 0; i < len; i++) \
741 *dst++ = ((XFASTINT (ccl_prog[ic + (i / 3)])) \
742 >> ((2 - (i % 3)) * 8)) & 0xFF; \
743 else \
744 CCL_SUSPEND (CCL_STAT_SUSPEND_BY_DST); \
745 } while (0)
746
747 /* Read one byte from the current input buffer into REGth register. */
748 #define CCL_READ_CHAR(REG) \
749 do { \
750 if (!src) \
751 CCL_INVALID_CMD; \
752 else if (src < src_end) \
753 { \
754 REG = *src++; \
755 if (REG == '\n' \
756 && ccl->eol_type != CODING_EOL_LF) \
757 { \
758 /* We are encoding. */ \
759 if (ccl->eol_type == CODING_EOL_CRLF) \
760 { \
761 if (ccl->cr_consumed) \
762 ccl->cr_consumed = 0; \
763 else \
764 { \
765 ccl->cr_consumed = 1; \
766 REG = '\r'; \
767 src--; \
768 } \
769 } \
770 else \
771 REG = '\r'; \
772 } \
773 if (REG == LEADING_CODE_8_BIT_CONTROL \
774 && ccl->multibyte) \
775 REG = *src++ - 0x20; \
776 } \
777 else if (ccl->last_block) \
778 { \
779 ic = ccl->eof_ic; \
780 goto ccl_repeat; \
781 } \
782 else \
783 CCL_SUSPEND (CCL_STAT_SUSPEND_BY_SRC); \
784 } while (0)
785
786
787 /* Set C to the character code made from CHARSET and CODE. This is
788 like MAKE_CHAR but check the validity of CHARSET and CODE. If they
789 are not valid, set C to (CODE & 0xFF) because that is usually the
790 case that CCL_ReadMultibyteChar2 read an invalid code and it set
791 CODE to that invalid byte. */
792
793 #define CCL_MAKE_CHAR(charset, code, c) \
794 do { \
795 if (charset == CHARSET_ASCII) \
796 c = code & 0xFF; \
797 else if (CHARSET_DEFINED_P (charset) \
798 && (code & 0x7F) >= 32 \
799 && (code < 256 || ((code >> 7) & 0x7F) >= 32)) \
800 { \
801 int c1 = code & 0x7F, c2 = 0; \
802 \
803 if (code >= 256) \
804 c2 = c1, c1 = (code >> 7) & 0x7F; \
805 c = MAKE_CHAR (charset, c1, c2); \
806 } \
807 else \
808 c = code & 0xFF; \
809 } while (0)
810
811
812 /* Execute CCL code on SRC_BYTES length text at SOURCE. The resulting
813 text goes to a place pointed by DESTINATION, the length of which
814 should not exceed DST_BYTES. The bytes actually processed is
815 returned as *CONSUMED. The return value is the length of the
816 resulting text. As a side effect, the contents of CCL registers
817 are updated. If SOURCE or DESTINATION is NULL, only operations on
818 registers are permitted. */
819
820 #ifdef CCL_DEBUG
821 #define CCL_DEBUG_BACKTRACE_LEN 256
822 int ccl_backtrace_table[CCL_BACKTRACE_TABLE];
823 int ccl_backtrace_idx;
824 #endif
825
826 struct ccl_prog_stack
827 {
828 Lisp_Object *ccl_prog; /* Pointer to an array of CCL code. */
829 int ic; /* Instruction Counter. */
830 };
831
832 /* For the moment, we only support depth 256 of stack. */
833 static struct ccl_prog_stack ccl_prog_stack_struct[256];
834
835 int
836 ccl_driver (ccl, source, destination, src_bytes, dst_bytes, consumed)
837 struct ccl_program *ccl;
838 unsigned char *source, *destination;
839 int src_bytes, dst_bytes;
840 int *consumed;
841 {
842 register int *reg = ccl->reg;
843 register int ic = ccl->ic;
844 register int code, field1, field2;
845 register Lisp_Object *ccl_prog = ccl->prog;
846 unsigned char *src = source, *src_end = src + src_bytes;
847 unsigned char *dst = destination, *dst_end = dst + dst_bytes;
848 int jump_address;
849 int i, j, op;
850 int stack_idx = ccl->stack_idx;
851 /* Instruction counter of the current CCL code. */
852 int this_ic;
853 /* CCL_WRITE_CHAR will produce 8-bit code of range 0x80..0x9F. But,
854 each of them will be converted to multibyte form of 2-byte
855 sequence. For that conversion, we remember how many more bytes
856 we must keep in DESTINATION in this variable. */
857 int extra_bytes = 0;
858
859 if (ic >= ccl->eof_ic)
860 ic = CCL_HEADER_MAIN;
861
862 if (ccl->buf_magnification ==0) /* We can't produce any bytes. */
863 dst = NULL;
864
865 /* Set mapping stack pointer. */
866 mapping_stack_pointer = mapping_stack;
867
868 #ifdef CCL_DEBUG
869 ccl_backtrace_idx = 0;
870 #endif
871
872 for (;;)
873 {
874 ccl_repeat:
875 #ifdef CCL_DEBUG
876 ccl_backtrace_table[ccl_backtrace_idx++] = ic;
877 if (ccl_backtrace_idx >= CCL_DEBUG_BACKTRACE_LEN)
878 ccl_backtrace_idx = 0;
879 ccl_backtrace_table[ccl_backtrace_idx] = 0;
880 #endif
881
882 if (!NILP (Vquit_flag) && NILP (Vinhibit_quit))
883 {
884 /* We can't just signal Qquit, instead break the loop as if
885 the whole data is processed. Don't reset Vquit_flag, it
886 must be handled later at a safer place. */
887 if (consumed)
888 src = source + src_bytes;
889 ccl->status = CCL_STAT_QUIT;
890 break;
891 }
892
893 this_ic = ic;
894 code = XINT (ccl_prog[ic]); ic++;
895 field1 = code >> 8;
896 field2 = (code & 0xFF) >> 5;
897
898 #define rrr field2
899 #define RRR (field1 & 7)
900 #define Rrr ((field1 >> 3) & 7)
901 #define ADDR field1
902 #define EXCMD (field1 >> 6)
903
904 switch (code & 0x1F)
905 {
906 case CCL_SetRegister: /* 00000000000000000RRRrrrXXXXX */
907 reg[rrr] = reg[RRR];
908 break;
909
910 case CCL_SetShortConst: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
911 reg[rrr] = field1;
912 break;
913
914 case CCL_SetConst: /* 00000000000000000000rrrXXXXX */
915 reg[rrr] = XINT (ccl_prog[ic]);
916 ic++;
917 break;
918
919 case CCL_SetArray: /* CCCCCCCCCCCCCCCCCCCCRRRrrrXXXXX */
920 i = reg[RRR];
921 j = field1 >> 3;
922 if ((unsigned int) i < j)
923 reg[rrr] = XINT (ccl_prog[ic + i]);
924 ic += j;
925 break;
926
927 case CCL_Jump: /* A--D--D--R--E--S--S-000XXXXX */
928 ic += ADDR;
929 break;
930
931 case CCL_JumpCond: /* A--D--D--R--E--S--S-rrrXXXXX */
932 if (!reg[rrr])
933 ic += ADDR;
934 break;
935
936 case CCL_WriteRegisterJump: /* A--D--D--R--E--S--S-rrrXXXXX */
937 i = reg[rrr];
938 CCL_WRITE_CHAR (i);
939 ic += ADDR;
940 break;
941
942 case CCL_WriteRegisterReadJump: /* A--D--D--R--E--S--S-rrrXXXXX */
943 i = reg[rrr];
944 CCL_WRITE_CHAR (i);
945 ic++;
946 CCL_READ_CHAR (reg[rrr]);
947 ic += ADDR - 1;
948 break;
949
950 case CCL_WriteConstJump: /* A--D--D--R--E--S--S-000XXXXX */
951 i = XINT (ccl_prog[ic]);
952 CCL_WRITE_CHAR (i);
953 ic += ADDR;
954 break;
955
956 case CCL_WriteConstReadJump: /* A--D--D--R--E--S--S-rrrXXXXX */
957 i = XINT (ccl_prog[ic]);
958 CCL_WRITE_CHAR (i);
959 ic++;
960 CCL_READ_CHAR (reg[rrr]);
961 ic += ADDR - 1;
962 break;
963
964 case CCL_WriteStringJump: /* A--D--D--R--E--S--S-000XXXXX */
965 j = XINT (ccl_prog[ic]);
966 ic++;
967 CCL_WRITE_STRING (j);
968 ic += ADDR - 1;
969 break;
970
971 case CCL_WriteArrayReadJump: /* A--D--D--R--E--S--S-rrrXXXXX */
972 i = reg[rrr];
973 j = XINT (ccl_prog[ic]);
974 if ((unsigned int) i < j)
975 {
976 i = XINT (ccl_prog[ic + 1 + i]);
977 CCL_WRITE_CHAR (i);
978 }
979 ic += j + 2;
980 CCL_READ_CHAR (reg[rrr]);
981 ic += ADDR - (j + 2);
982 break;
983
984 case CCL_ReadJump: /* A--D--D--R--E--S--S-rrrYYYYY */
985 CCL_READ_CHAR (reg[rrr]);
986 ic += ADDR;
987 break;
988
989 case CCL_ReadBranch: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
990 CCL_READ_CHAR (reg[rrr]);
991 /* fall through ... */
992 case CCL_Branch: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
993 if ((unsigned int) reg[rrr] < field1)
994 ic += XINT (ccl_prog[ic + reg[rrr]]);
995 else
996 ic += XINT (ccl_prog[ic + field1]);
997 break;
998
999 case CCL_ReadRegister: /* CCCCCCCCCCCCCCCCCCCCrrXXXXX */
1000 while (1)
1001 {
1002 CCL_READ_CHAR (reg[rrr]);
1003 if (!field1) break;
1004 code = XINT (ccl_prog[ic]); ic++;
1005 field1 = code >> 8;
1006 field2 = (code & 0xFF) >> 5;
1007 }
1008 break;
1009
1010 case CCL_WriteExprConst: /* 1:00000OPERATION000RRR000XXXXX */
1011 rrr = 7;
1012 i = reg[RRR];
1013 j = XINT (ccl_prog[ic]);
1014 op = field1 >> 6;
1015 jump_address = ic + 1;
1016 goto ccl_set_expr;
1017
1018 case CCL_WriteRegister: /* CCCCCCCCCCCCCCCCCCCrrrXXXXX */
1019 while (1)
1020 {
1021 i = reg[rrr];
1022 CCL_WRITE_CHAR (i);
1023 if (!field1) break;
1024 code = XINT (ccl_prog[ic]); ic++;
1025 field1 = code >> 8;
1026 field2 = (code & 0xFF) >> 5;
1027 }
1028 break;
1029
1030 case CCL_WriteExprRegister: /* 1:00000OPERATIONRrrRRR000XXXXX */
1031 rrr = 7;
1032 i = reg[RRR];
1033 j = reg[Rrr];
1034 op = field1 >> 6;
1035 jump_address = ic;
1036 goto ccl_set_expr;
1037
1038 case CCL_Call: /* 1:CCCCCCCCCCCCCCCCCCCCFFFXXXXX */
1039 {
1040 Lisp_Object slot;
1041 int prog_id;
1042
1043 /* If FFF is nonzero, the CCL program ID is in the
1044 following code. */
1045 if (rrr)
1046 {
1047 prog_id = XINT (ccl_prog[ic]);
1048 ic++;
1049 }
1050 else
1051 prog_id = field1;
1052
1053 if (stack_idx >= 256
1054 || prog_id < 0
1055 || prog_id >= XVECTOR (Vccl_program_table)->size
1056 || (slot = XVECTOR (Vccl_program_table)->contents[prog_id],
1057 !VECTORP (slot))
1058 || !VECTORP (XVECTOR (slot)->contents[1]))
1059 {
1060 if (stack_idx > 0)
1061 {
1062 ccl_prog = ccl_prog_stack_struct[0].ccl_prog;
1063 ic = ccl_prog_stack_struct[0].ic;
1064 }
1065 CCL_INVALID_CMD;
1066 }
1067
1068 ccl_prog_stack_struct[stack_idx].ccl_prog = ccl_prog;
1069 ccl_prog_stack_struct[stack_idx].ic = ic;
1070 stack_idx++;
1071 ccl_prog = XVECTOR (XVECTOR (slot)->contents[1])->contents;
1072 ic = CCL_HEADER_MAIN;
1073 }
1074 break;
1075
1076 case CCL_WriteConstString: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
1077 if (!rrr)
1078 CCL_WRITE_CHAR (field1);
1079 else
1080 {
1081 CCL_WRITE_STRING (field1);
1082 ic += (field1 + 2) / 3;
1083 }
1084 break;
1085
1086 case CCL_WriteArray: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
1087 i = reg[rrr];
1088 if ((unsigned int) i < field1)
1089 {
1090 j = XINT (ccl_prog[ic + i]);
1091 CCL_WRITE_CHAR (j);
1092 }
1093 ic += field1;
1094 break;
1095
1096 case CCL_End: /* 0000000000000000000000XXXXX */
1097 if (stack_idx > 0)
1098 {
1099 stack_idx--;
1100 ccl_prog = ccl_prog_stack_struct[stack_idx].ccl_prog;
1101 ic = ccl_prog_stack_struct[stack_idx].ic;
1102 break;
1103 }
1104 if (src)
1105 src = src_end;
1106 /* ccl->ic should points to this command code again to
1107 suppress further processing. */
1108 ic--;
1109 CCL_SUCCESS;
1110
1111 case CCL_ExprSelfConst: /* 00000OPERATION000000rrrXXXXX */
1112 i = XINT (ccl_prog[ic]);
1113 ic++;
1114 op = field1 >> 6;
1115 goto ccl_expr_self;
1116
1117 case CCL_ExprSelfReg: /* 00000OPERATION000RRRrrrXXXXX */
1118 i = reg[RRR];
1119 op = field1 >> 6;
1120
1121 ccl_expr_self:
1122 switch (op)
1123 {
1124 case CCL_PLUS: reg[rrr] += i; break;
1125 case CCL_MINUS: reg[rrr] -= i; break;
1126 case CCL_MUL: reg[rrr] *= i; break;
1127 case CCL_DIV: reg[rrr] /= i; break;
1128 case CCL_MOD: reg[rrr] %= i; break;
1129 case CCL_AND: reg[rrr] &= i; break;
1130 case CCL_OR: reg[rrr] |= i; break;
1131 case CCL_XOR: reg[rrr] ^= i; break;
1132 case CCL_LSH: reg[rrr] <<= i; break;
1133 case CCL_RSH: reg[rrr] >>= i; break;
1134 case CCL_LSH8: reg[rrr] <<= 8; reg[rrr] |= i; break;
1135 case CCL_RSH8: reg[7] = reg[rrr] & 0xFF; reg[rrr] >>= 8; break;
1136 case CCL_DIVMOD: reg[7] = reg[rrr] % i; reg[rrr] /= i; break;
1137 case CCL_LS: reg[rrr] = reg[rrr] < i; break;
1138 case CCL_GT: reg[rrr] = reg[rrr] > i; break;
1139 case CCL_EQ: reg[rrr] = reg[rrr] == i; break;
1140 case CCL_LE: reg[rrr] = reg[rrr] <= i; break;
1141 case CCL_GE: reg[rrr] = reg[rrr] >= i; break;
1142 case CCL_NE: reg[rrr] = reg[rrr] != i; break;
1143 default: CCL_INVALID_CMD;
1144 }
1145 break;
1146
1147 case CCL_SetExprConst: /* 00000OPERATION000RRRrrrXXXXX */
1148 i = reg[RRR];
1149 j = XINT (ccl_prog[ic]);
1150 op = field1 >> 6;
1151 jump_address = ++ic;
1152 goto ccl_set_expr;
1153
1154 case CCL_SetExprReg: /* 00000OPERATIONRrrRRRrrrXXXXX */
1155 i = reg[RRR];
1156 j = reg[Rrr];
1157 op = field1 >> 6;
1158 jump_address = ic;
1159 goto ccl_set_expr;
1160
1161 case CCL_ReadJumpCondExprConst: /* A--D--D--R--E--S--S-rrrXXXXX */
1162 CCL_READ_CHAR (reg[rrr]);
1163 case CCL_JumpCondExprConst: /* A--D--D--R--E--S--S-rrrXXXXX */
1164 i = reg[rrr];
1165 op = XINT (ccl_prog[ic]);
1166 jump_address = ic++ + ADDR;
1167 j = XINT (ccl_prog[ic]);
1168 ic++;
1169 rrr = 7;
1170 goto ccl_set_expr;
1171
1172 case CCL_ReadJumpCondExprReg: /* A--D--D--R--E--S--S-rrrXXXXX */
1173 CCL_READ_CHAR (reg[rrr]);
1174 case CCL_JumpCondExprReg:
1175 i = reg[rrr];
1176 op = XINT (ccl_prog[ic]);
1177 jump_address = ic++ + ADDR;
1178 j = reg[XINT (ccl_prog[ic])];
1179 ic++;
1180 rrr = 7;
1181
1182 ccl_set_expr:
1183 switch (op)
1184 {
1185 case CCL_PLUS: reg[rrr] = i + j; break;
1186 case CCL_MINUS: reg[rrr] = i - j; break;
1187 case CCL_MUL: reg[rrr] = i * j; break;
1188 case CCL_DIV: reg[rrr] = i / j; break;
1189 case CCL_MOD: reg[rrr] = i % j; break;
1190 case CCL_AND: reg[rrr] = i & j; break;
1191 case CCL_OR: reg[rrr] = i | j; break;
1192 case CCL_XOR: reg[rrr] = i ^ j;; break;
1193 case CCL_LSH: reg[rrr] = i << j; break;
1194 case CCL_RSH: reg[rrr] = i >> j; break;
1195 case CCL_LSH8: reg[rrr] = (i << 8) | j; break;
1196 case CCL_RSH8: reg[rrr] = i >> 8; reg[7] = i & 0xFF; break;
1197 case CCL_DIVMOD: reg[rrr] = i / j; reg[7] = i % j; break;
1198 case CCL_LS: reg[rrr] = i < j; break;
1199 case CCL_GT: reg[rrr] = i > j; break;
1200 case CCL_EQ: reg[rrr] = i == j; break;
1201 case CCL_LE: reg[rrr] = i <= j; break;
1202 case CCL_GE: reg[rrr] = i >= j; break;
1203 case CCL_NE: reg[rrr] = i != j; break;
1204 case CCL_DECODE_SJIS: DECODE_SJIS (i, j, reg[rrr], reg[7]); break;
1205 case CCL_ENCODE_SJIS: ENCODE_SJIS (i, j, reg[rrr], reg[7]); break;
1206 default: CCL_INVALID_CMD;
1207 }
1208 code &= 0x1F;
1209 if (code == CCL_WriteExprConst || code == CCL_WriteExprRegister)
1210 {
1211 i = reg[rrr];
1212 CCL_WRITE_CHAR (i);
1213 ic = jump_address;
1214 }
1215 else if (!reg[rrr])
1216 ic = jump_address;
1217 break;
1218
1219 case CCL_Extension:
1220 switch (EXCMD)
1221 {
1222 case CCL_ReadMultibyteChar2:
1223 if (!src)
1224 CCL_INVALID_CMD;
1225
1226 if (src >= src_end)
1227 {
1228 src++;
1229 goto ccl_read_multibyte_character_suspend;
1230 }
1231
1232 i = *src++;
1233 if (i == '\n' && ccl->eol_type != CODING_EOL_LF)
1234 {
1235 /* We are encoding. */
1236 if (ccl->eol_type == CODING_EOL_CRLF)
1237 {
1238 if (ccl->cr_consumed)
1239 ccl->cr_consumed = 0;
1240 else
1241 {
1242 ccl->cr_consumed = 1;
1243 i = '\r';
1244 src--;
1245 }
1246 }
1247 else
1248 i = '\r';
1249 reg[rrr] = i;
1250 reg[RRR] = CHARSET_ASCII;
1251 }
1252 else if (i < 0x80)
1253 {
1254 /* ASCII */
1255 reg[rrr] = i;
1256 reg[RRR] = CHARSET_ASCII;
1257 }
1258 else if (i <= MAX_CHARSET_OFFICIAL_DIMENSION1)
1259 {
1260 if (src >= src_end)
1261 goto ccl_read_multibyte_character_suspend;
1262 reg[RRR] = i;
1263 reg[rrr] = (*src++ & 0x7F);
1264 }
1265 else if (i <= MAX_CHARSET_OFFICIAL_DIMENSION2)
1266 {
1267 if ((src + 1) >= src_end)
1268 goto ccl_read_multibyte_character_suspend;
1269 reg[RRR] = i;
1270 i = (*src++ & 0x7F);
1271 reg[rrr] = ((i << 7) | (*src & 0x7F));
1272 src++;
1273 }
1274 else if ((i == LEADING_CODE_PRIVATE_11)
1275 || (i == LEADING_CODE_PRIVATE_12))
1276 {
1277 if ((src + 1) >= src_end)
1278 goto ccl_read_multibyte_character_suspend;
1279 reg[RRR] = *src++;
1280 reg[rrr] = (*src++ & 0x7F);
1281 }
1282 else if ((i == LEADING_CODE_PRIVATE_21)
1283 || (i == LEADING_CODE_PRIVATE_22))
1284 {
1285 if ((src + 2) >= src_end)
1286 goto ccl_read_multibyte_character_suspend;
1287 reg[RRR] = *src++;
1288 i = (*src++ & 0x7F);
1289 reg[rrr] = ((i << 7) | (*src & 0x7F));
1290 src++;
1291 }
1292 else if (i == LEADING_CODE_8_BIT_CONTROL)
1293 {
1294 if (src >= src_end)
1295 goto ccl_read_multibyte_character_suspend;
1296 reg[RRR] = CHARSET_8_BIT_CONTROL;
1297 reg[rrr] = (*src++ - 0x20);
1298 }
1299 else if (i >= 0xA0)
1300 {
1301 reg[RRR] = CHARSET_8_BIT_GRAPHIC;
1302 reg[rrr] = i;
1303 }
1304 else
1305 {
1306 /* INVALID CODE. Return a single byte character. */
1307 reg[RRR] = CHARSET_ASCII;
1308 reg[rrr] = i;
1309 }
1310 break;
1311
1312 ccl_read_multibyte_character_suspend:
1313 src--;
1314 if (ccl->last_block)
1315 {
1316 ic = ccl->eof_ic;
1317 goto ccl_repeat;
1318 }
1319 else
1320 CCL_SUSPEND (CCL_STAT_SUSPEND_BY_SRC);
1321
1322 break;
1323
1324 case CCL_WriteMultibyteChar2:
1325 i = reg[RRR]; /* charset */
1326 if (i == CHARSET_ASCII
1327 || i == CHARSET_8_BIT_CONTROL
1328 || i == CHARSET_8_BIT_GRAPHIC)
1329 i = reg[rrr] & 0xFF;
1330 else if (CHARSET_DIMENSION (i) == 1)
1331 i = ((i - 0x70) << 7) | (reg[rrr] & 0x7F);
1332 else if (i < MIN_CHARSET_PRIVATE_DIMENSION2)
1333 i = ((i - 0x8F) << 14) | reg[rrr];
1334 else
1335 i = ((i - 0xE0) << 14) | reg[rrr];
1336
1337 CCL_WRITE_CHAR (i);
1338
1339 break;
1340
1341 case CCL_TranslateCharacter:
1342 CCL_MAKE_CHAR (reg[RRR], reg[rrr], i);
1343 op = translate_char (GET_TRANSLATION_TABLE (reg[Rrr]),
1344 i, -1, 0, 0);
1345 SPLIT_CHAR (op, reg[RRR], i, j);
1346 if (j != -1)
1347 i = (i << 7) | j;
1348
1349 reg[rrr] = i;
1350 break;
1351
1352 case CCL_TranslateCharacterConstTbl:
1353 op = XINT (ccl_prog[ic]); /* table */
1354 ic++;
1355 CCL_MAKE_CHAR (reg[RRR], reg[rrr], i);
1356 op = translate_char (GET_TRANSLATION_TABLE (op), i, -1, 0, 0);
1357 SPLIT_CHAR (op, reg[RRR], i, j);
1358 if (j != -1)
1359 i = (i << 7) | j;
1360
1361 reg[rrr] = i;
1362 break;
1363
1364 case CCL_IterateMultipleMap:
1365 {
1366 Lisp_Object map, content, attrib, value;
1367 int point, size, fin_ic;
1368
1369 j = XINT (ccl_prog[ic++]); /* number of maps. */
1370 fin_ic = ic + j;
1371 op = reg[rrr];
1372 if ((j > reg[RRR]) && (j >= 0))
1373 {
1374 ic += reg[RRR];
1375 i = reg[RRR];
1376 }
1377 else
1378 {
1379 reg[RRR] = -1;
1380 ic = fin_ic;
1381 break;
1382 }
1383
1384 for (;i < j;i++)
1385 {
1386
1387 size = XVECTOR (Vcode_conversion_map_vector)->size;
1388 point = XINT (ccl_prog[ic++]);
1389 if (point >= size) continue;
1390 map =
1391 XVECTOR (Vcode_conversion_map_vector)->contents[point];
1392
1393 /* Check map varidity. */
1394 if (!CONSP (map)) continue;
1395 map = XCDR (map);
1396 if (!VECTORP (map)) continue;
1397 size = XVECTOR (map)->size;
1398 if (size <= 1) continue;
1399
1400 content = XVECTOR (map)->contents[0];
1401
1402 /* check map type,
1403 [STARTPOINT VAL1 VAL2 ...] or
1404 [t ELELMENT STARTPOINT ENDPOINT] */
1405 if (NUMBERP (content))
1406 {
1407 point = XUINT (content);
1408 point = op - point + 1;
1409 if (!((point >= 1) && (point < size))) continue;
1410 content = XVECTOR (map)->contents[point];
1411 }
1412 else if (EQ (content, Qt))
1413 {
1414 if (size != 4) continue;
1415 if ((op >= XUINT (XVECTOR (map)->contents[2]))
1416 && (op < XUINT (XVECTOR (map)->contents[3])))
1417 content = XVECTOR (map)->contents[1];
1418 else
1419 continue;
1420 }
1421 else
1422 continue;
1423
1424 if (NILP (content))
1425 continue;
1426 else if (NUMBERP (content))
1427 {
1428 reg[RRR] = i;
1429 reg[rrr] = XINT(content);
1430 break;
1431 }
1432 else if (EQ (content, Qt) || EQ (content, Qlambda))
1433 {
1434 reg[RRR] = i;
1435 break;
1436 }
1437 else if (CONSP (content))
1438 {
1439 attrib = XCAR (content);
1440 value = XCDR (content);
1441 if (!NUMBERP (attrib) || !NUMBERP (value))
1442 continue;
1443 reg[RRR] = i;
1444 reg[rrr] = XUINT (value);
1445 break;
1446 }
1447 else if (SYMBOLP (content))
1448 CCL_CALL_FOR_MAP_INSTRUCTION (content, fin_ic);
1449 else
1450 CCL_INVALID_CMD;
1451 }
1452 if (i == j)
1453 reg[RRR] = -1;
1454 ic = fin_ic;
1455 }
1456 break;
1457
1458 case CCL_MapMultiple:
1459 {
1460 Lisp_Object map, content, attrib, value;
1461 int point, size, map_vector_size;
1462 int map_set_rest_length, fin_ic;
1463 int current_ic = this_ic;
1464
1465 /* inhibit recursive call on MapMultiple. */
1466 if (stack_idx_of_map_multiple > 0)
1467 {
1468 if (stack_idx_of_map_multiple <= stack_idx)
1469 {
1470 stack_idx_of_map_multiple = 0;
1471 mapping_stack_pointer = mapping_stack;
1472 CCL_INVALID_CMD;
1473 }
1474 }
1475 else
1476 mapping_stack_pointer = mapping_stack;
1477 stack_idx_of_map_multiple = 0;
1478
1479 map_set_rest_length =
1480 XINT (ccl_prog[ic++]); /* number of maps and separators. */
1481 fin_ic = ic + map_set_rest_length;
1482 op = reg[rrr];
1483
1484 if ((map_set_rest_length > reg[RRR]) && (reg[RRR] >= 0))
1485 {
1486 ic += reg[RRR];
1487 i = reg[RRR];
1488 map_set_rest_length -= i;
1489 }
1490 else
1491 {
1492 ic = fin_ic;
1493 reg[RRR] = -1;
1494 mapping_stack_pointer = mapping_stack;
1495 break;
1496 }
1497
1498 if (mapping_stack_pointer <= (mapping_stack + 1))
1499 {
1500 /* Set up initial state. */
1501 mapping_stack_pointer = mapping_stack;
1502 PUSH_MAPPING_STACK (0, op);
1503 reg[RRR] = -1;
1504 }
1505 else
1506 {
1507 /* Recover after calling other ccl program. */
1508 int orig_op;
1509
1510 POP_MAPPING_STACK (map_set_rest_length, orig_op);
1511 POP_MAPPING_STACK (map_set_rest_length, reg[rrr]);
1512 switch (op)
1513 {
1514 case -1:
1515 /* Regard it as Qnil. */
1516 op = orig_op;
1517 i++;
1518 ic++;
1519 map_set_rest_length--;
1520 break;
1521 case -2:
1522 /* Regard it as Qt. */
1523 op = reg[rrr];
1524 i++;
1525 ic++;
1526 map_set_rest_length--;
1527 break;
1528 case -3:
1529 /* Regard it as Qlambda. */
1530 op = orig_op;
1531 i += map_set_rest_length;
1532 ic += map_set_rest_length;
1533 map_set_rest_length = 0;
1534 break;
1535 default:
1536 /* Regard it as normal mapping. */
1537 i += map_set_rest_length;
1538 ic += map_set_rest_length;
1539 POP_MAPPING_STACK (map_set_rest_length, reg[rrr]);
1540 break;
1541 }
1542 }
1543 map_vector_size = XVECTOR (Vcode_conversion_map_vector)->size;
1544
1545 do {
1546 for (;map_set_rest_length > 0;i++, ic++, map_set_rest_length--)
1547 {
1548 point = XINT(ccl_prog[ic]);
1549 if (point < 0)
1550 {
1551 /* +1 is for including separator. */
1552 point = -point + 1;
1553 if (mapping_stack_pointer
1554 >= &mapping_stack[MAX_MAP_SET_LEVEL])
1555 CCL_INVALID_CMD;
1556 PUSH_MAPPING_STACK (map_set_rest_length - point,
1557 reg[rrr]);
1558 map_set_rest_length = point;
1559 reg[rrr] = op;
1560 continue;
1561 }
1562
1563 if (point >= map_vector_size) continue;
1564 map = (XVECTOR (Vcode_conversion_map_vector)
1565 ->contents[point]);
1566
1567 /* Check map varidity. */
1568 if (!CONSP (map)) continue;
1569 map = XCDR (map);
1570 if (!VECTORP (map)) continue;
1571 size = XVECTOR (map)->size;
1572 if (size <= 1) continue;
1573
1574 content = XVECTOR (map)->contents[0];
1575
1576 /* check map type,
1577 [STARTPOINT VAL1 VAL2 ...] or
1578 [t ELEMENT STARTPOINT ENDPOINT] */
1579 if (NUMBERP (content))
1580 {
1581 point = XUINT (content);
1582 point = op - point + 1;
1583 if (!((point >= 1) && (point < size))) continue;
1584 content = XVECTOR (map)->contents[point];
1585 }
1586 else if (EQ (content, Qt))
1587 {
1588 if (size != 4) continue;
1589 if ((op >= XUINT (XVECTOR (map)->contents[2])) &&
1590 (op < XUINT (XVECTOR (map)->contents[3])))
1591 content = XVECTOR (map)->contents[1];
1592 else
1593 continue;
1594 }
1595 else
1596 continue;
1597
1598 if (NILP (content))
1599 continue;
1600
1601 reg[RRR] = i;
1602 if (NUMBERP (content))
1603 {
1604 op = XINT (content);
1605 i += map_set_rest_length - 1;
1606 ic += map_set_rest_length - 1;
1607 POP_MAPPING_STACK (map_set_rest_length, reg[rrr]);
1608 map_set_rest_length++;
1609 }
1610 else if (CONSP (content))
1611 {
1612 attrib = XCAR (content);
1613 value = XCDR (content);
1614 if (!NUMBERP (attrib) || !NUMBERP (value))
1615 continue;
1616 op = XUINT (value);
1617 i += map_set_rest_length - 1;
1618 ic += map_set_rest_length - 1;
1619 POP_MAPPING_STACK (map_set_rest_length, reg[rrr]);
1620 map_set_rest_length++;
1621 }
1622 else if (EQ (content, Qt))
1623 {
1624 op = reg[rrr];
1625 }
1626 else if (EQ (content, Qlambda))
1627 {
1628 i += map_set_rest_length;
1629 ic += map_set_rest_length;
1630 break;
1631 }
1632 else if (SYMBOLP (content))
1633 {
1634 if (mapping_stack_pointer
1635 >= &mapping_stack[MAX_MAP_SET_LEVEL])
1636 CCL_INVALID_CMD;
1637 PUSH_MAPPING_STACK (map_set_rest_length, reg[rrr]);
1638 PUSH_MAPPING_STACK (map_set_rest_length, op);
1639 stack_idx_of_map_multiple = stack_idx + 1;
1640 CCL_CALL_FOR_MAP_INSTRUCTION (content, current_ic);
1641 }
1642 else
1643 CCL_INVALID_CMD;
1644 }
1645 if (mapping_stack_pointer <= (mapping_stack + 1))
1646 break;
1647 POP_MAPPING_STACK (map_set_rest_length, reg[rrr]);
1648 i += map_set_rest_length;
1649 ic += map_set_rest_length;
1650 POP_MAPPING_STACK (map_set_rest_length, reg[rrr]);
1651 } while (1);
1652
1653 ic = fin_ic;
1654 }
1655 reg[rrr] = op;
1656 break;
1657
1658 case CCL_MapSingle:
1659 {
1660 Lisp_Object map, attrib, value, content;
1661 int size, point;
1662 j = XINT (ccl_prog[ic++]); /* map_id */
1663 op = reg[rrr];
1664 if (j >= XVECTOR (Vcode_conversion_map_vector)->size)
1665 {
1666 reg[RRR] = -1;
1667 break;
1668 }
1669 map = XVECTOR (Vcode_conversion_map_vector)->contents[j];
1670 if (!CONSP (map))
1671 {
1672 reg[RRR] = -1;
1673 break;
1674 }
1675 map = XCDR (map);
1676 if (!VECTORP (map))
1677 {
1678 reg[RRR] = -1;
1679 break;
1680 }
1681 size = XVECTOR (map)->size;
1682 point = XUINT (XVECTOR (map)->contents[0]);
1683 point = op - point + 1;
1684 reg[RRR] = 0;
1685 if ((size <= 1) ||
1686 (!((point >= 1) && (point < size))))
1687 reg[RRR] = -1;
1688 else
1689 {
1690 reg[RRR] = 0;
1691 content = XVECTOR (map)->contents[point];
1692 if (NILP (content))
1693 reg[RRR] = -1;
1694 else if (NUMBERP (content))
1695 reg[rrr] = XINT (content);
1696 else if (EQ (content, Qt));
1697 else if (CONSP (content))
1698 {
1699 attrib = XCAR (content);
1700 value = XCDR (content);
1701 if (!NUMBERP (attrib) || !NUMBERP (value))
1702 continue;
1703 reg[rrr] = XUINT(value);
1704 break;
1705 }
1706 else if (SYMBOLP (content))
1707 CCL_CALL_FOR_MAP_INSTRUCTION (content, ic);
1708 else
1709 reg[RRR] = -1;
1710 }
1711 }
1712 break;
1713
1714 default:
1715 CCL_INVALID_CMD;
1716 }
1717 break;
1718
1719 default:
1720 CCL_INVALID_CMD;
1721 }
1722 }
1723
1724 ccl_error_handler:
1725 /* The suppress_error member is set when e.g. a CCL-based coding
1726 system is used for terminal output. */
1727 if (!ccl->suppress_error && destination)
1728 {
1729 /* We can insert an error message only if DESTINATION is
1730 specified and we still have a room to store the message
1731 there. */
1732 char msg[256];
1733 int msglen;
1734
1735 if (!dst)
1736 dst = destination;
1737
1738 switch (ccl->status)
1739 {
1740 case CCL_STAT_INVALID_CMD:
1741 sprintf(msg, "\nCCL: Invalid command %x (ccl_code = %x) at %d.",
1742 code & 0x1F, code, this_ic);
1743 #ifdef CCL_DEBUG
1744 {
1745 int i = ccl_backtrace_idx - 1;
1746 int j;
1747
1748 msglen = strlen (msg);
1749 if (dst + msglen <= (dst_bytes ? dst_end : src))
1750 {
1751 bcopy (msg, dst, msglen);
1752 dst += msglen;
1753 }
1754
1755 for (j = 0; j < CCL_DEBUG_BACKTRACE_LEN; j++, i--)
1756 {
1757 if (i < 0) i = CCL_DEBUG_BACKTRACE_LEN - 1;
1758 if (ccl_backtrace_table[i] == 0)
1759 break;
1760 sprintf(msg, " %d", ccl_backtrace_table[i]);
1761 msglen = strlen (msg);
1762 if (dst + msglen > (dst_bytes ? dst_end : src))
1763 break;
1764 bcopy (msg, dst, msglen);
1765 dst += msglen;
1766 }
1767 goto ccl_finish;
1768 }
1769 #endif
1770 break;
1771
1772 case CCL_STAT_QUIT:
1773 sprintf(msg, "\nCCL: Quited.");
1774 break;
1775
1776 default:
1777 sprintf(msg, "\nCCL: Unknown error type (%d).", ccl->status);
1778 }
1779
1780 msglen = strlen (msg);
1781 if (dst + msglen <= (dst_bytes ? dst_end : src))
1782 {
1783 bcopy (msg, dst, msglen);
1784 dst += msglen;
1785 }
1786 if (ccl->status == CCL_STAT_INVALID_CMD)
1787 {
1788 /* Copy the remaining source data. */
1789 int i = src_end - src;
1790 if (dst_bytes && (dst_end - dst) < i)
1791 i = dst_end - dst;
1792 bcopy (src, dst, i);
1793 src += i;
1794 dst += i;
1795 }
1796 }
1797
1798 ccl_finish:
1799 ccl->ic = ic;
1800 ccl->stack_idx = stack_idx;
1801 ccl->prog = ccl_prog;
1802 if (consumed) *consumed = src - source;
1803 return (dst ? dst - destination : 0);
1804 }
1805
1806 /* Resolve symbols in the specified CCL code (Lisp vector). This
1807 function converts symbols of code conversion maps and character
1808 translation tables embeded in the CCL code into their ID numbers.
1809
1810 The return value is a vector (CCL itself or a new vector in which
1811 all symbols are resolved), Qt if resolving of some symbol failed,
1812 or nil if CCL contains invalid data. */
1813
1814 static Lisp_Object
1815 resolve_symbol_ccl_program (ccl)
1816 Lisp_Object ccl;
1817 {
1818 int i, veclen, unresolved = 0;
1819 Lisp_Object result, contents, val;
1820
1821 result = ccl;
1822 veclen = XVECTOR (result)->size;
1823
1824 for (i = 0; i < veclen; i++)
1825 {
1826 contents = XVECTOR (result)->contents[i];
1827 if (INTEGERP (contents))
1828 continue;
1829 else if (CONSP (contents)
1830 && SYMBOLP (XCAR (contents))
1831 && SYMBOLP (XCDR (contents)))
1832 {
1833 /* This is the new style for embedding symbols. The form is
1834 (SYMBOL . PROPERTY). (get SYMBOL PROPERTY) should give
1835 an index number. */
1836
1837 if (EQ (result, ccl))
1838 result = Fcopy_sequence (ccl);
1839
1840 val = Fget (XCAR (contents), XCDR (contents));
1841 if (NATNUMP (val))
1842 XVECTOR (result)->contents[i] = val;
1843 else
1844 unresolved = 1;
1845 continue;
1846 }
1847 else if (SYMBOLP (contents))
1848 {
1849 /* This is the old style for embedding symbols. This style
1850 may lead to a bug if, for instance, a translation table
1851 and a code conversion map have the same name. */
1852 if (EQ (result, ccl))
1853 result = Fcopy_sequence (ccl);
1854
1855 val = Fget (contents, Qtranslation_table_id);
1856 if (NATNUMP (val))
1857 XVECTOR (result)->contents[i] = val;
1858 else
1859 {
1860 val = Fget (contents, Qcode_conversion_map_id);
1861 if (NATNUMP (val))
1862 XVECTOR (result)->contents[i] = val;
1863 else
1864 {
1865 val = Fget (contents, Qccl_program_idx);
1866 if (NATNUMP (val))
1867 XVECTOR (result)->contents[i] = val;
1868 else
1869 unresolved = 1;
1870 }
1871 }
1872 continue;
1873 }
1874 return Qnil;
1875 }
1876
1877 return (unresolved ? Qt : result);
1878 }
1879
1880 /* Return the compiled code (vector) of CCL program CCL_PROG.
1881 CCL_PROG is a name (symbol) of the program or already compiled
1882 code. If necessary, resolve symbols in the compiled code to index
1883 numbers. If we failed to get the compiled code or to resolve
1884 symbols, return Qnil. */
1885
1886 static Lisp_Object
1887 ccl_get_compiled_code (ccl_prog)
1888 Lisp_Object ccl_prog;
1889 {
1890 Lisp_Object val, slot;
1891
1892 if (VECTORP (ccl_prog))
1893 {
1894 val = resolve_symbol_ccl_program (ccl_prog);
1895 return (VECTORP (val) ? val : Qnil);
1896 }
1897 if (!SYMBOLP (ccl_prog))
1898 return Qnil;
1899
1900 val = Fget (ccl_prog, Qccl_program_idx);
1901 if (! NATNUMP (val)
1902 || XINT (val) >= XVECTOR (Vccl_program_table)->size)
1903 return Qnil;
1904 slot = XVECTOR (Vccl_program_table)->contents[XINT (val)];
1905 if (! VECTORP (slot)
1906 || XVECTOR (slot)->size != 3
1907 || ! VECTORP (XVECTOR (slot)->contents[1]))
1908 return Qnil;
1909 if (NILP (XVECTOR (slot)->contents[2]))
1910 {
1911 val = resolve_symbol_ccl_program (XVECTOR (slot)->contents[1]);
1912 if (! VECTORP (val))
1913 return Qnil;
1914 XVECTOR (slot)->contents[1] = val;
1915 XVECTOR (slot)->contents[2] = Qt;
1916 }
1917 return XVECTOR (slot)->contents[1];
1918 }
1919
1920 /* Setup fields of the structure pointed by CCL appropriately for the
1921 execution of CCL program CCL_PROG. CCL_PROG is the name (symbol)
1922 of the CCL program or the already compiled code (vector).
1923 Return 0 if we succeed this setup, else return -1.
1924
1925 If CCL_PROG is nil, we just reset the structure pointed by CCL. */
1926 int
1927 setup_ccl_program (ccl, ccl_prog)
1928 struct ccl_program *ccl;
1929 Lisp_Object ccl_prog;
1930 {
1931 int i;
1932
1933 if (! NILP (ccl_prog))
1934 {
1935 struct Lisp_Vector *vp;
1936
1937 ccl_prog = ccl_get_compiled_code (ccl_prog);
1938 if (! VECTORP (ccl_prog))
1939 return -1;
1940 vp = XVECTOR (ccl_prog);
1941 ccl->size = vp->size;
1942 ccl->prog = vp->contents;
1943 ccl->eof_ic = XINT (vp->contents[CCL_HEADER_EOF]);
1944 ccl->buf_magnification = XINT (vp->contents[CCL_HEADER_BUF_MAG]);
1945 }
1946 ccl->ic = CCL_HEADER_MAIN;
1947 for (i = 0; i < 8; i++)
1948 ccl->reg[i] = 0;
1949 ccl->last_block = 0;
1950 ccl->private_state = 0;
1951 ccl->status = 0;
1952 ccl->stack_idx = 0;
1953 ccl->eol_type = CODING_EOL_LF;
1954 ccl->suppress_error = 0;
1955 return 0;
1956 }
1957
1958 #ifdef emacs
1959
1960 DEFUN ("ccl-program-p", Fccl_program_p, Sccl_program_p, 1, 1, 0,
1961 "Return t if OBJECT is a CCL program name or a compiled CCL program code.\n\
1962 See the documentation of `define-ccl-program' for the detail of CCL program.")
1963 (object)
1964 Lisp_Object object;
1965 {
1966 Lisp_Object val;
1967
1968 if (VECTORP (object))
1969 {
1970 val = resolve_symbol_ccl_program (object);
1971 return (VECTORP (val) ? Qt : Qnil);
1972 }
1973 if (!SYMBOLP (object))
1974 return Qnil;
1975
1976 val = Fget (object, Qccl_program_idx);
1977 return ((! NATNUMP (val)
1978 || XINT (val) >= XVECTOR (Vccl_program_table)->size)
1979 ? Qnil : Qt);
1980 }
1981
1982 DEFUN ("ccl-execute", Fccl_execute, Sccl_execute, 2, 2, 0,
1983 "Execute CCL-PROGRAM with registers initialized by REGISTERS.\n\
1984 \n\
1985 CCL-PROGRAM is a CCL program name (symbol)\n\
1986 or a compiled code generated by `ccl-compile' (for backward compatibility,\n\
1987 in this case, the overhead of the execution is bigger than the former case).\n\
1988 No I/O commands should appear in CCL-PROGRAM.\n\
1989 \n\
1990 REGISTERS is a vector of [R0 R1 ... R7] where RN is an initial value\n\
1991 of Nth register.\n\
1992 \n\
1993 As side effect, each element of REGISTERS holds the value of\n\
1994 corresponding register after the execution.\n\
1995 \n\
1996 See the documentation of `define-ccl-program' for the detail of CCL program.")
1997 (ccl_prog, reg)
1998 Lisp_Object ccl_prog, reg;
1999 {
2000 struct ccl_program ccl;
2001 int i;
2002
2003 if (setup_ccl_program (&ccl, ccl_prog) < 0)
2004 error ("Invalid CCL program");
2005
2006 CHECK_VECTOR (reg, 1);
2007 if (XVECTOR (reg)->size != 8)
2008 error ("Length of vector REGISTERS is not 8");
2009
2010 for (i = 0; i < 8; i++)
2011 ccl.reg[i] = (INTEGERP (XVECTOR (reg)->contents[i])
2012 ? XINT (XVECTOR (reg)->contents[i])
2013 : 0);
2014
2015 ccl_driver (&ccl, (unsigned char *)0, (unsigned char *)0, 0, 0, (int *)0);
2016 QUIT;
2017 if (ccl.status != CCL_STAT_SUCCESS)
2018 error ("Error in CCL program at %dth code", ccl.ic);
2019
2020 for (i = 0; i < 8; i++)
2021 XSETINT (XVECTOR (reg)->contents[i], ccl.reg[i]);
2022 return Qnil;
2023 }
2024
2025 DEFUN ("ccl-execute-on-string", Fccl_execute_on_string, Sccl_execute_on_string,
2026 3, 5, 0,
2027 "Execute CCL-PROGRAM with initial STATUS on STRING.\n\
2028 \n\
2029 CCL-PROGRAM is a symbol registered by register-ccl-program,\n\
2030 or a compiled code generated by `ccl-compile' (for backward compatibility,\n\
2031 in this case, the execution is slower).\n\
2032 \n\
2033 Read buffer is set to STRING, and write buffer is allocated automatically.\n\
2034 \n\
2035 STATUS is a vector of [R0 R1 ... R7 IC], where\n\
2036 R0..R7 are initial values of corresponding registers,\n\
2037 IC is the instruction counter specifying from where to start the program.\n\
2038 If R0..R7 are nil, they are initialized to 0.\n\
2039 If IC is nil, it is initialized to head of the CCL program.\n\
2040 \n\
2041 If optional 4th arg CONTINUE is non-nil, keep IC on read operation\n\
2042 when read buffer is exausted, else, IC is always set to the end of\n\
2043 CCL-PROGRAM on exit.\n\
2044 \n\
2045 It returns the contents of write buffer as a string,\n\
2046 and as side effect, STATUS is updated.\n\
2047 If the optional 5th arg UNIBYTE-P is non-nil, the returned string\n\
2048 is a unibyte string. By default it is a multibyte string.\n\
2049 \n\
2050 See the documentation of `define-ccl-program' for the detail of CCL program.")
2051 (ccl_prog, status, str, contin, unibyte_p)
2052 Lisp_Object ccl_prog, status, str, contin, unibyte_p;
2053 {
2054 Lisp_Object val;
2055 struct ccl_program ccl;
2056 int i, produced;
2057 int outbufsize;
2058 char *outbuf;
2059 struct gcpro gcpro1, gcpro2;
2060
2061 if (setup_ccl_program (&ccl, ccl_prog) < 0)
2062 error ("Invalid CCL program");
2063
2064 CHECK_VECTOR (status, 1);
2065 if (XVECTOR (status)->size != 9)
2066 error ("Length of vector STATUS is not 9");
2067 CHECK_STRING (str, 2);
2068
2069 GCPRO2 (status, str);
2070
2071 for (i = 0; i < 8; i++)
2072 {
2073 if (NILP (XVECTOR (status)->contents[i]))
2074 XSETINT (XVECTOR (status)->contents[i], 0);
2075 if (INTEGERP (XVECTOR (status)->contents[i]))
2076 ccl.reg[i] = XINT (XVECTOR (status)->contents[i]);
2077 }
2078 if (INTEGERP (XVECTOR (status)->contents[i]))
2079 {
2080 i = XFASTINT (XVECTOR (status)->contents[8]);
2081 if (ccl.ic < i && i < ccl.size)
2082 ccl.ic = i;
2083 }
2084 outbufsize = STRING_BYTES (XSTRING (str)) * ccl.buf_magnification + 256;
2085 outbuf = (char *) xmalloc (outbufsize);
2086 ccl.last_block = NILP (contin);
2087 ccl.multibyte = STRING_MULTIBYTE (str);
2088 produced = ccl_driver (&ccl, XSTRING (str)->data, outbuf,
2089 STRING_BYTES (XSTRING (str)), outbufsize, (int *) 0);
2090 for (i = 0; i < 8; i++)
2091 XSET (XVECTOR (status)->contents[i], Lisp_Int, ccl.reg[i]);
2092 XSETINT (XVECTOR (status)->contents[8], ccl.ic);
2093 UNGCPRO;
2094
2095 if (NILP (unibyte_p))
2096 {
2097 int nchars;
2098
2099 produced = str_as_multibyte (outbuf, outbufsize, produced, &nchars);
2100 val = make_multibyte_string (outbuf, nchars, produced);
2101 }
2102 else
2103 val = make_unibyte_string (outbuf, produced);
2104 xfree (outbuf);
2105 QUIT;
2106 if (ccl.status == CCL_STAT_SUSPEND_BY_DST)
2107 error ("Output buffer for the CCL programs overflow");
2108 if (ccl.status != CCL_STAT_SUCCESS
2109 && ccl.status != CCL_STAT_SUSPEND_BY_SRC)
2110 error ("Error in CCL program at %dth code", ccl.ic);
2111
2112 return val;
2113 }
2114
2115 DEFUN ("register-ccl-program", Fregister_ccl_program, Sregister_ccl_program,
2116 2, 2, 0,
2117 "Register CCL program CCL_PROG as NAME in `ccl-program-table'.\n\
2118 CCL_PROG should be a compiled CCL program (vector), or nil.\n\
2119 If it is nil, just reserve NAME as a CCL program name.\n\
2120 Return index number of the registered CCL program.")
2121 (name, ccl_prog)
2122 Lisp_Object name, ccl_prog;
2123 {
2124 int len = XVECTOR (Vccl_program_table)->size;
2125 int idx;
2126 Lisp_Object resolved;
2127
2128 CHECK_SYMBOL (name, 0);
2129 resolved = Qnil;
2130 if (!NILP (ccl_prog))
2131 {
2132 CHECK_VECTOR (ccl_prog, 1);
2133 resolved = resolve_symbol_ccl_program (ccl_prog);
2134 if (NILP (resolved))
2135 error ("Error in CCL program");
2136 if (VECTORP (resolved))
2137 {
2138 ccl_prog = resolved;
2139 resolved = Qt;
2140 }
2141 else
2142 resolved = Qnil;
2143 }
2144
2145 for (idx = 0; idx < len; idx++)
2146 {
2147 Lisp_Object slot;
2148
2149 slot = XVECTOR (Vccl_program_table)->contents[idx];
2150 if (!VECTORP (slot))
2151 /* This is the first unsed slot. Register NAME here. */
2152 break;
2153
2154 if (EQ (name, XVECTOR (slot)->contents[0]))
2155 {
2156 /* Update this slot. */
2157 XVECTOR (slot)->contents[1] = ccl_prog;
2158 XVECTOR (slot)->contents[2] = resolved;
2159 return make_number (idx);
2160 }
2161 }
2162
2163 if (idx == len)
2164 {
2165 /* Extend the table. */
2166 Lisp_Object new_table;
2167 int j;
2168
2169 new_table = Fmake_vector (make_number (len * 2), Qnil);
2170 for (j = 0; j < len; j++)
2171 XVECTOR (new_table)->contents[j]
2172 = XVECTOR (Vccl_program_table)->contents[j];
2173 Vccl_program_table = new_table;
2174 }
2175
2176 {
2177 Lisp_Object elt;
2178
2179 elt = Fmake_vector (make_number (3), Qnil);
2180 XVECTOR (elt)->contents[0] = name;
2181 XVECTOR (elt)->contents[1] = ccl_prog;
2182 XVECTOR (elt)->contents[2] = resolved;
2183 XVECTOR (Vccl_program_table)->contents[idx] = elt;
2184 }
2185
2186 Fput (name, Qccl_program_idx, make_number (idx));
2187 return make_number (idx);
2188 }
2189
2190 /* Register code conversion map.
2191 A code conversion map consists of numbers, Qt, Qnil, and Qlambda.
2192 The first element is start code point.
2193 The rest elements are mapped numbers.
2194 Symbol t means to map to an original number before mapping.
2195 Symbol nil means that the corresponding element is empty.
2196 Symbol lambda menas to terminate mapping here.
2197 */
2198
2199 DEFUN ("register-code-conversion-map", Fregister_code_conversion_map,
2200 Sregister_code_conversion_map,
2201 2, 2, 0,
2202 "Register SYMBOL as code conversion map MAP.\n\
2203 Return index number of the registered map.")
2204 (symbol, map)
2205 Lisp_Object symbol, map;
2206 {
2207 int len = XVECTOR (Vcode_conversion_map_vector)->size;
2208 int i;
2209 Lisp_Object index;
2210
2211 CHECK_SYMBOL (symbol, 0);
2212 CHECK_VECTOR (map, 1);
2213
2214 for (i = 0; i < len; i++)
2215 {
2216 Lisp_Object slot = XVECTOR (Vcode_conversion_map_vector)->contents[i];
2217
2218 if (!CONSP (slot))
2219 break;
2220
2221 if (EQ (symbol, XCAR (slot)))
2222 {
2223 index = make_number (i);
2224 XCDR (slot) = map;
2225 Fput (symbol, Qcode_conversion_map, map);
2226 Fput (symbol, Qcode_conversion_map_id, index);
2227 return index;
2228 }
2229 }
2230
2231 if (i == len)
2232 {
2233 Lisp_Object new_vector = Fmake_vector (make_number (len * 2), Qnil);
2234 int j;
2235
2236 for (j = 0; j < len; j++)
2237 XVECTOR (new_vector)->contents[j]
2238 = XVECTOR (Vcode_conversion_map_vector)->contents[j];
2239 Vcode_conversion_map_vector = new_vector;
2240 }
2241
2242 index = make_number (i);
2243 Fput (symbol, Qcode_conversion_map, map);
2244 Fput (symbol, Qcode_conversion_map_id, index);
2245 XVECTOR (Vcode_conversion_map_vector)->contents[i] = Fcons (symbol, map);
2246 return index;
2247 }
2248
2249
2250 void
2251 syms_of_ccl ()
2252 {
2253 staticpro (&Vccl_program_table);
2254 Vccl_program_table = Fmake_vector (make_number (32), Qnil);
2255
2256 Qccl_program = intern ("ccl-program");
2257 staticpro (&Qccl_program);
2258
2259 Qccl_program_idx = intern ("ccl-program-idx");
2260 staticpro (&Qccl_program_idx);
2261
2262 Qcode_conversion_map = intern ("code-conversion-map");
2263 staticpro (&Qcode_conversion_map);
2264
2265 Qcode_conversion_map_id = intern ("code-conversion-map-id");
2266 staticpro (&Qcode_conversion_map_id);
2267
2268 DEFVAR_LISP ("code-conversion-map-vector", &Vcode_conversion_map_vector,
2269 "Vector of code conversion maps.");
2270 Vcode_conversion_map_vector = Fmake_vector (make_number (16), Qnil);
2271
2272 DEFVAR_LISP ("font-ccl-encoder-alist", &Vfont_ccl_encoder_alist,
2273 "Alist of fontname patterns vs corresponding CCL program.\n\
2274 Each element looks like (REGEXP . CCL-CODE),\n\
2275 where CCL-CODE is a compiled CCL program.\n\
2276 When a font whose name matches REGEXP is used for displaying a character,\n\
2277 CCL-CODE is executed to calculate the code point in the font\n\
2278 from the charset number and position code(s) of the character which are set\n\
2279 in CCL registers R0, R1, and R2 before the execution.\n\
2280 The code point in the font is set in CCL registers R1 and R2\n\
2281 when the execution terminated.\n\
2282 If the font is single-byte font, the register R2 is not used.");
2283 Vfont_ccl_encoder_alist = Qnil;
2284
2285 defsubr (&Sccl_program_p);
2286 defsubr (&Sccl_execute);
2287 defsubr (&Sccl_execute_on_string);
2288 defsubr (&Sregister_ccl_program);
2289 defsubr (&Sregister_code_conversion_map);
2290 }
2291
2292 #endif /* emacs */