1 /* CCL (Code Conversion Language) interpreter.
2 Copyright (C) 1995, 1997 Electrotechnical Laboratory, JAPAN.
3 Licensed to the Free Software Foundation.
4 Copyright (C) 2001, 2002 Free Software Foundation, Inc.
6 National Institute of Advanced Industrial Science and Technology (AIST)
7 Registration Number H13PRO009
9 This file is part of GNU Emacs.
11 GNU Emacs is free software; you can redistribute it and/or modify
12 it under the terms of the GNU General Public License as published by
13 the Free Software Foundation; either version 2, or (at your option)
16 GNU Emacs is distributed in the hope that it will be useful,
17 but WITHOUT ANY WARRANTY; without even the implied warranty of
18 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
19 GNU General Public License for more details.
21 You should have received a copy of the GNU General Public License
22 along with GNU Emacs; see the file COPYING. If not, write to
23 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
24 Boston, MA 02111-1307, USA. */
31 #include "character.h"
36 Lisp_Object Qccl
, Qcclp
;
38 /* This contains all code conversion map available to CCL. */
39 Lisp_Object Vcode_conversion_map_vector
;
41 /* Alist of fontname patterns vs corresponding CCL program. */
42 Lisp_Object Vfont_ccl_encoder_alist
;
44 /* This symbol is a property which assocates with ccl program vector.
45 Ex: (get 'ccl-big5-encoder 'ccl-program) returns ccl program vector. */
46 Lisp_Object Qccl_program
;
48 /* These symbols are properties which associate with code conversion
49 map and their ID respectively. */
50 Lisp_Object Qcode_conversion_map
;
51 Lisp_Object Qcode_conversion_map_id
;
53 /* Symbols of ccl program have this property, a value of the property
54 is an index for Vccl_protram_table. */
55 Lisp_Object Qccl_program_idx
;
57 /* Table of registered CCL programs. Each element is a vector of
58 NAME, CCL_PROG, RESOLVEDP, and UPDATEDP, where NAME (symbol) is the
59 name of the program, CCL_PROG (vector) is the compiled code of the
60 program, RESOLVEDP (t or nil) is the flag to tell if symbols in
61 CCL_PROG is already resolved to index numbers or not, UPDATEDP (t
62 or nil) is the flat to tell if the CCL program is updated after it
64 Lisp_Object Vccl_program_table
;
66 /* Vector of registered hash tables for translation. */
67 Lisp_Object Vtranslation_hash_table_vector
;
69 /* Return a hash table of id number ID. */
70 #define GET_HASH_TABLE(id) \
71 (XHASH_TABLE (XCDR(XVECTOR(Vtranslation_hash_table_vector)->contents[(id)])))
73 extern int charset_unicode
;
75 /* CCL (Code Conversion Language) is a simple language which has
76 operations on one input buffer, one output buffer, and 7 registers.
77 The syntax of CCL is described in `ccl.el'. Emacs Lisp function
78 `ccl-compile' compiles a CCL program and produces a CCL code which
79 is a vector of integers. The structure of this vector is as
80 follows: The 1st element: buffer-magnification, a factor for the
81 size of output buffer compared with the size of input buffer. The
82 2nd element: address of CCL code to be executed when encountered
83 with end of input stream. The 3rd and the remaining elements: CCL
86 /* Header of CCL compiled code */
87 #define CCL_HEADER_BUF_MAG 0
88 #define CCL_HEADER_EOF 1
89 #define CCL_HEADER_MAIN 2
91 /* CCL code is a sequence of 28-bit non-negative integers (i.e. the
92 MSB is always 0), each contains CCL command and/or arguments in the
95 |----------------- integer (28-bit) ------------------|
96 |------- 17-bit ------|- 3-bit --|- 3-bit --|- 5-bit -|
97 |--constant argument--|-register-|-register-|-command-|
98 ccccccccccccccccc RRR rrr XXXXX
100 |------- relative address -------|-register-|-command-|
101 cccccccccccccccccccc rrr XXXXX
103 |------------- constant or other args ----------------|
104 cccccccccccccccccccccccccccc
106 where, `cc...c' is a non-negative integer indicating constant value
107 (the left most `c' is always 0) or an absolute jump address, `RRR'
108 and `rrr' are CCL register number, `XXXXX' is one of the following
113 Each comment fields shows one or more lines for command syntax and
114 the following lines for semantics of the command. In semantics, IC
115 stands for Instruction Counter. */
117 #define CCL_SetRegister 0x00 /* Set register a register value:
118 1:00000000000000000RRRrrrXXXXX
119 ------------------------------
123 #define CCL_SetShortConst 0x01 /* Set register a short constant value:
124 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
125 ------------------------------
126 reg[rrr] = CCCCCCCCCCCCCCCCCCC;
129 #define CCL_SetConst 0x02 /* Set register a constant value:
130 1:00000000000000000000rrrXXXXX
132 ------------------------------
137 #define CCL_SetArray 0x03 /* Set register an element of array:
138 1:CCCCCCCCCCCCCCCCCRRRrrrXXXXX
142 ------------------------------
143 if (0 <= reg[RRR] < CC..C)
144 reg[rrr] = ELEMENT[reg[RRR]];
148 #define CCL_Jump 0x04 /* Jump:
149 1:A--D--D--R--E--S--S-000XXXXX
150 ------------------------------
154 /* Note: If CC..C is greater than 0, the second code is omitted. */
156 #define CCL_JumpCond 0x05 /* Jump conditional:
157 1:A--D--D--R--E--S--S-rrrXXXXX
158 ------------------------------
164 #define CCL_WriteRegisterJump 0x06 /* Write register and jump:
165 1:A--D--D--R--E--S--S-rrrXXXXX
166 ------------------------------
171 #define CCL_WriteRegisterReadJump 0x07 /* Write register, read, and jump:
172 1:A--D--D--R--E--S--S-rrrXXXXX
173 2:A--D--D--R--E--S--S-rrrYYYYY
174 -----------------------------
180 /* Note: If read is suspended, the resumed execution starts from the
181 second code (YYYYY == CCL_ReadJump). */
183 #define CCL_WriteConstJump 0x08 /* Write constant and jump:
184 1:A--D--D--R--E--S--S-000XXXXX
186 ------------------------------
191 #define CCL_WriteConstReadJump 0x09 /* Write constant, read, and jump:
192 1:A--D--D--R--E--S--S-rrrXXXXX
194 3:A--D--D--R--E--S--S-rrrYYYYY
195 -----------------------------
201 /* Note: If read is suspended, the resumed execution starts from the
202 second code (YYYYY == CCL_ReadJump). */
204 #define CCL_WriteStringJump 0x0A /* Write string and jump:
205 1:A--D--D--R--E--S--S-000XXXXX
207 3:0000STRIN[0]STRIN[1]STRIN[2]
209 ------------------------------
210 write_string (STRING, LENGTH);
214 #define CCL_WriteArrayReadJump 0x0B /* Write an array element, read, and jump:
215 1:A--D--D--R--E--S--S-rrrXXXXX
220 N:A--D--D--R--E--S--S-rrrYYYYY
221 ------------------------------
222 if (0 <= reg[rrr] < LENGTH)
223 write (ELEMENT[reg[rrr]]);
224 IC += LENGTH + 2; (... pointing at N+1)
228 /* Note: If read is suspended, the resumed execution starts from the
229 Nth code (YYYYY == CCL_ReadJump). */
231 #define CCL_ReadJump 0x0C /* Read and jump:
232 1:A--D--D--R--E--S--S-rrrYYYYY
233 -----------------------------
238 #define CCL_Branch 0x0D /* Jump by branch table:
239 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
240 2:A--D--D--R--E-S-S[0]000XXXXX
241 3:A--D--D--R--E-S-S[1]000XXXXX
243 ------------------------------
244 if (0 <= reg[rrr] < CC..C)
245 IC += ADDRESS[reg[rrr]];
247 IC += ADDRESS[CC..C];
250 #define CCL_ReadRegister 0x0E /* Read bytes into registers:
251 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
252 2:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
254 ------------------------------
259 #define CCL_WriteExprConst 0x0F /* write result of expression:
260 1:00000OPERATION000RRR000XXXXX
262 ------------------------------
263 write (reg[RRR] OPERATION CONSTANT);
267 /* Note: If the Nth read is suspended, the resumed execution starts
268 from the Nth code. */
270 #define CCL_ReadBranch 0x10 /* Read one byte into a register,
271 and jump by branch table:
272 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
273 2:A--D--D--R--E-S-S[0]000XXXXX
274 3:A--D--D--R--E-S-S[1]000XXXXX
276 ------------------------------
278 if (0 <= reg[rrr] < CC..C)
279 IC += ADDRESS[reg[rrr]];
281 IC += ADDRESS[CC..C];
284 #define CCL_WriteRegister 0x11 /* Write registers:
285 1:CCCCCCCCCCCCCCCCCCCrrrXXXXX
286 2:CCCCCCCCCCCCCCCCCCCrrrXXXXX
288 ------------------------------
294 /* Note: If the Nth write is suspended, the resumed execution
295 starts from the Nth code. */
297 #define CCL_WriteExprRegister 0x12 /* Write result of expression
298 1:00000OPERATIONRrrRRR000XXXXX
299 ------------------------------
300 write (reg[RRR] OPERATION reg[Rrr]);
303 #define CCL_Call 0x13 /* Call the CCL program whose ID is
305 1:CCCCCCCCCCCCCCCCCCCCFFFXXXXX
306 [2:00000000cccccccccccccccccccc]
307 ------------------------------
315 #define CCL_WriteConstString 0x14 /* Write a constant or a string:
316 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
317 [2:0000STRIN[0]STRIN[1]STRIN[2]]
319 -----------------------------
323 write_string (STRING, CC..C);
324 IC += (CC..C + 2) / 3;
327 #define CCL_WriteArray 0x15 /* Write an element of array:
328 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
332 ------------------------------
333 if (0 <= reg[rrr] < CC..C)
334 write (ELEMENT[reg[rrr]]);
338 #define CCL_End 0x16 /* Terminate:
339 1:00000000000000000000000XXXXX
340 ------------------------------
344 /* The following two codes execute an assignment arithmetic/logical
345 operation. The form of the operation is like REG OP= OPERAND. */
347 #define CCL_ExprSelfConst 0x17 /* REG OP= constant:
348 1:00000OPERATION000000rrrXXXXX
350 ------------------------------
351 reg[rrr] OPERATION= CONSTANT;
354 #define CCL_ExprSelfReg 0x18 /* REG1 OP= REG2:
355 1:00000OPERATION000RRRrrrXXXXX
356 ------------------------------
357 reg[rrr] OPERATION= reg[RRR];
360 /* The following codes execute an arithmetic/logical operation. The
361 form of the operation is like REG_X = REG_Y OP OPERAND2. */
363 #define CCL_SetExprConst 0x19 /* REG_X = REG_Y OP constant:
364 1:00000OPERATION000RRRrrrXXXXX
366 ------------------------------
367 reg[rrr] = reg[RRR] OPERATION CONSTANT;
371 #define CCL_SetExprReg 0x1A /* REG1 = REG2 OP REG3:
372 1:00000OPERATIONRrrRRRrrrXXXXX
373 ------------------------------
374 reg[rrr] = reg[RRR] OPERATION reg[Rrr];
377 #define CCL_JumpCondExprConst 0x1B /* Jump conditional according to
378 an operation on constant:
379 1:A--D--D--R--E--S--S-rrrXXXXX
382 -----------------------------
383 reg[7] = reg[rrr] OPERATION CONSTANT;
390 #define CCL_JumpCondExprReg 0x1C /* Jump conditional according to
391 an operation on register:
392 1:A--D--D--R--E--S--S-rrrXXXXX
395 -----------------------------
396 reg[7] = reg[rrr] OPERATION reg[RRR];
403 #define CCL_ReadJumpCondExprConst 0x1D /* Read and jump conditional according
404 to an operation on constant:
405 1:A--D--D--R--E--S--S-rrrXXXXX
408 -----------------------------
410 reg[7] = reg[rrr] OPERATION CONSTANT;
417 #define CCL_ReadJumpCondExprReg 0x1E /* Read and jump conditional according
418 to an operation on register:
419 1:A--D--D--R--E--S--S-rrrXXXXX
422 -----------------------------
424 reg[7] = reg[rrr] OPERATION reg[RRR];
431 #define CCL_Extension 0x1F /* Extended CCL code
432 1:ExtendedCOMMNDRrrRRRrrrXXXXX
435 ------------------------------
436 extended_command (rrr,RRR,Rrr,ARGS)
440 Here after, Extended CCL Instructions.
441 Bit length of extended command is 14.
442 Therefore, the instruction code range is 0..16384(0x3fff).
445 /* Read a multibyte characeter.
446 A code point is stored into reg[rrr]. A charset ID is stored into
449 #define CCL_ReadMultibyteChar2 0x00 /* Read Multibyte Character
450 1:ExtendedCOMMNDRrrRRRrrrXXXXX */
452 /* Write a multibyte character.
453 Write a character whose code point is reg[rrr] and the charset ID
456 #define CCL_WriteMultibyteChar2 0x01 /* Write Multibyte Character
457 1:ExtendedCOMMNDRrrRRRrrrXXXXX */
459 /* Translate a character whose code point is reg[rrr] and the charset
460 ID is reg[RRR] by a translation table whose ID is reg[Rrr].
462 A translated character is set in reg[rrr] (code point) and reg[RRR]
465 #define CCL_TranslateCharacter 0x02 /* Translate a multibyte character
466 1:ExtendedCOMMNDRrrRRRrrrXXXXX */
468 /* Translate a character whose code point is reg[rrr] and the charset
469 ID is reg[RRR] by a translation table whose ID is ARGUMENT.
471 A translated character is set in reg[rrr] (code point) and reg[RRR]
474 #define CCL_TranslateCharacterConstTbl 0x03 /* Translate a multibyte character
475 1:ExtendedCOMMNDRrrRRRrrrXXXXX
476 2:ARGUMENT(Translation Table ID)
479 /* Iterate looking up MAPs for reg[rrr] starting from the Nth (N =
480 reg[RRR]) MAP until some value is found.
482 Each MAP is a Lisp vector whose element is number, nil, t, or
484 If the element is nil, ignore the map and proceed to the next map.
485 If the element is t or lambda, finish without changing reg[rrr].
486 If the element is a number, set reg[rrr] to the number and finish.
488 Detail of the map structure is descibed in the comment for
489 CCL_MapMultiple below. */
491 #define CCL_IterateMultipleMap 0x10 /* Iterate multiple maps
492 1:ExtendedCOMMNDXXXRRRrrrXXXXX
499 /* Map the code in reg[rrr] by MAPs starting from the Nth (N =
502 MAPs are supplied in the succeeding CCL codes as follows:
504 When CCL program gives this nested structure of map to this command:
507 (MAP-ID121 MAP-ID122 MAP-ID123)
510 (MAP-ID211 (MAP-ID2111) MAP-ID212)
512 the compiled CCL codes has this sequence:
513 CCL_MapMultiple (CCL code of this command)
514 16 (total number of MAPs and SEPARATORs)
532 A value of each SEPARATOR follows this rule:
533 MAP-SET := SEPARATOR [(MAP-ID | MAP-SET)]+
534 SEPARATOR := -(number of MAP-IDs and SEPARATORs in the MAP-SET)
536 (*)....Nest level of MAP-SET must not be over than MAX_MAP_SET_LEVEL.
538 When some map fails to map (i.e. it doesn't have a value for
539 reg[rrr]), the mapping is treated as identity.
541 The mapping is iterated for all maps in each map set (set of maps
542 separated by SEPARATOR) except in the case that lambda is
543 encountered. More precisely, the mapping proceeds as below:
545 At first, VAL0 is set to reg[rrr], and it is translated by the
546 first map to VAL1. Then, VAL1 is translated by the next map to
547 VAL2. This mapping is iterated until the last map is used. The
548 result of the mapping is the last value of VAL?. When the mapping
549 process reached to the end of the map set, it moves to the next
550 map set. If the next does not exit, the mapping process terminates,
551 and regard the last value as a result.
553 But, when VALm is mapped to VALn and VALn is not a number, the
554 mapping proceed as below:
556 If VALn is nil, the lastest map is ignored and the mapping of VALm
557 proceed to the next map.
559 In VALn is t, VALm is reverted to reg[rrr] and the mapping of VALm
560 proceed to the next map.
562 If VALn is lambda, move to the next map set like reaching to the
563 end of the current map set.
565 If VALn is a symbol, call the CCL program refered by it.
566 Then, use reg[rrr] as a mapped value except for -1, -2 and -3.
567 Such special values are regarded as nil, t, and lambda respectively.
569 Each map is a Lisp vector of the following format (a) or (b):
570 (a)......[STARTPOINT VAL1 VAL2 ...]
571 (b)......[t VAL STARTPOINT ENDPOINT],
573 STARTPOINT is an offset to be used for indexing a map,
574 ENDPOINT is a maximum index number of a map,
575 VAL and VALn is a number, nil, t, or lambda.
577 Valid index range of a map of type (a) is:
578 STARTPOINT <= index < STARTPOINT + map_size - 1
579 Valid index range of a map of type (b) is:
580 STARTPOINT <= index < ENDPOINT */
582 #define CCL_MapMultiple 0x11 /* Mapping by multiple code conversion maps
583 1:ExtendedCOMMNDXXXRRRrrrXXXXX
595 #define MAX_MAP_SET_LEVEL 30
603 static tr_stack mapping_stack
[MAX_MAP_SET_LEVEL
];
604 static tr_stack
*mapping_stack_pointer
;
606 /* If this variable is non-zero, it indicates the stack_idx
607 of immediately called by CCL_MapMultiple. */
608 static int stack_idx_of_map_multiple
;
610 #define PUSH_MAPPING_STACK(restlen, orig) \
613 mapping_stack_pointer->rest_length = (restlen); \
614 mapping_stack_pointer->orig_val = (orig); \
615 mapping_stack_pointer++; \
619 #define POP_MAPPING_STACK(restlen, orig) \
622 mapping_stack_pointer--; \
623 (restlen) = mapping_stack_pointer->rest_length; \
624 (orig) = mapping_stack_pointer->orig_val; \
628 #define CCL_CALL_FOR_MAP_INSTRUCTION(symbol, ret_ic) \
631 struct ccl_program called_ccl; \
632 if (stack_idx >= 256 \
633 || (setup_ccl_program (&called_ccl, (symbol)) != 0)) \
637 ccl_prog = ccl_prog_stack_struct[0].ccl_prog; \
638 ic = ccl_prog_stack_struct[0].ic; \
639 eof_ic = ccl_prog_stack_struct[0].eof_ic; \
643 ccl_prog_stack_struct[stack_idx].ccl_prog = ccl_prog; \
644 ccl_prog_stack_struct[stack_idx].ic = (ret_ic); \
645 ccl_prog_stack_struct[stack_idx].eof_ic = eof_ic; \
647 ccl_prog = called_ccl.prog; \
648 ic = CCL_HEADER_MAIN; \
649 eof_ic = XFASTINT (ccl_prog[CCL_HEADER_EOF]); \
654 #define CCL_MapSingle 0x12 /* Map by single code conversion map
655 1:ExtendedCOMMNDXXXRRRrrrXXXXX
657 ------------------------------
658 Map reg[rrr] by MAP-ID.
659 If some valid mapping is found,
660 set reg[rrr] to the result,
665 #define CCL_LookupIntConstTbl 0x13 /* Lookup multibyte character by
666 integer key. Afterwards R7 set
667 to 1 iff lookup succeeded.
668 1:ExtendedCOMMNDRrrRRRXXXXXXXX
669 2:ARGUMENT(Hash table ID) */
671 #define CCL_LookupCharConstTbl 0x14 /* Lookup integer by multibyte
672 character key. Afterwards R7 set
673 to 1 iff lookup succeeded.
674 1:ExtendedCOMMNDRrrRRRrrrXXXXX
675 2:ARGUMENT(Hash table ID) */
677 /* CCL arithmetic/logical operators. */
678 #define CCL_PLUS 0x00 /* X = Y + Z */
679 #define CCL_MINUS 0x01 /* X = Y - Z */
680 #define CCL_MUL 0x02 /* X = Y * Z */
681 #define CCL_DIV 0x03 /* X = Y / Z */
682 #define CCL_MOD 0x04 /* X = Y % Z */
683 #define CCL_AND 0x05 /* X = Y & Z */
684 #define CCL_OR 0x06 /* X = Y | Z */
685 #define CCL_XOR 0x07 /* X = Y ^ Z */
686 #define CCL_LSH 0x08 /* X = Y << Z */
687 #define CCL_RSH 0x09 /* X = Y >> Z */
688 #define CCL_LSH8 0x0A /* X = (Y << 8) | Z */
689 #define CCL_RSH8 0x0B /* X = Y >> 8, r[7] = Y & 0xFF */
690 #define CCL_DIVMOD 0x0C /* X = Y / Z, r[7] = Y % Z */
691 #define CCL_LS 0x10 /* X = (X < Y) */
692 #define CCL_GT 0x11 /* X = (X > Y) */
693 #define CCL_EQ 0x12 /* X = (X == Y) */
694 #define CCL_LE 0x13 /* X = (X <= Y) */
695 #define CCL_GE 0x14 /* X = (X >= Y) */
696 #define CCL_NE 0x15 /* X = (X != Y) */
698 #define CCL_DECODE_SJIS 0x16 /* X = HIGHER_BYTE (DE-SJIS (Y, Z))
699 r[7] = LOWER_BYTE (DE-SJIS (Y, Z)) */
700 #define CCL_ENCODE_SJIS 0x17 /* X = HIGHER_BYTE (SJIS (Y, Z))
701 r[7] = LOWER_BYTE (SJIS (Y, Z) */
703 /* Terminate CCL program successfully. */
704 #define CCL_SUCCESS \
707 ccl->status = CCL_STAT_SUCCESS; \
712 /* Suspend CCL program because of reading from empty input buffer or
713 writing to full output buffer. When this program is resumed, the
714 same I/O command is executed. */
715 #define CCL_SUSPEND(stat) \
719 ccl->status = stat; \
724 /* Terminate CCL program because of invalid command. Should not occur
725 in the normal case. */
728 #define CCL_INVALID_CMD \
731 ccl->status = CCL_STAT_INVALID_CMD; \
732 goto ccl_error_handler; \
738 #define CCL_INVALID_CMD \
741 ccl_debug_hook (this_ic); \
742 ccl->status = CCL_STAT_INVALID_CMD; \
743 goto ccl_error_handler; \
749 /* Encode one character CH to multibyte form and write to the current
750 output buffer. If CH is less than 256, CH is written as is. */
751 #define CCL_WRITE_CHAR(ch) \
755 else if (dst < dst_end) \
758 CCL_SUSPEND (CCL_STAT_SUSPEND_BY_DST); \
761 /* Write a string at ccl_prog[IC] of length LEN to the current output
763 #define CCL_WRITE_STRING(len) \
768 else if (dst + len <= dst_end) \
769 for (i = 0; i < len; i++) \
770 *dst++ = ((XFASTINT (ccl_prog[ic + (i / 3)])) \
771 >> ((2 - (i % 3)) * 8)) & 0xFF; \
773 CCL_SUSPEND (CCL_STAT_SUSPEND_BY_DST); \
776 /* Read one byte from the current input buffer into Rth register. */
777 #define CCL_READ_CHAR(r) \
781 else if (src < src_end) \
783 else if (ccl->last_block) \
790 CCL_SUSPEND (CCL_STAT_SUSPEND_BY_SRC); \
793 /* Decode CODE by a charset whose id is ID. If ID is 0, return CODE
794 as is for backward compatibility. Assume that we can use the
795 variable `charset'. */
797 #define CCL_DECODE_CHAR(id, code) \
798 ((id) == 0 ? (code) \
799 : (charset = CHARSET_FROM_ID ((id)), DECODE_CHAR (charset, (code))))
801 /* Encode character C by some of charsets in CHARSET_LIST. Set ID to
802 the id of the used charset, ENCODED to the resulf of encoding.
803 Assume that we can use the variable `charset'. */
805 #define CCL_ENCODE_CHAR(c, charset_list, id, encoded) \
809 charset = char_charset ((c), (charset_list), &code); \
810 if (! charset && ! NILP (charset_list)) \
811 charset = char_charset ((c), Qnil, &code); \
814 (id) = CHARSET_ID (charset); \
819 /* Execute CCL code on characters at SOURCE (length SRC_SIZE). The
820 resulting text goes to a place pointed by DESTINATION, the length
821 of which should not exceed DST_SIZE. As a side effect, how many
822 characters are consumed and produced are recorded in CCL->consumed
823 and CCL->produced, and the contents of CCL registers are updated.
824 If SOURCE or DESTINATION is NULL, only operations on registers are
828 #define CCL_DEBUG_BACKTRACE_LEN 256
829 int ccl_backtrace_table
[CCL_DEBUG_BACKTRACE_LEN
];
830 int ccl_backtrace_idx
;
833 ccl_debug_hook (int ic
)
840 struct ccl_prog_stack
842 Lisp_Object
*ccl_prog
; /* Pointer to an array of CCL code. */
843 int ic
; /* Instruction Counter. */
844 int eof_ic
; /* Instruction Counter to jump on EOF. */
847 /* For the moment, we only support depth 256 of stack. */
848 static struct ccl_prog_stack ccl_prog_stack_struct
[256];
851 ccl_driver (ccl
, source
, destination
, src_size
, dst_size
, charset_list
)
852 struct ccl_program
*ccl
;
853 int *source
, *destination
;
854 int src_size
, dst_size
;
855 Lisp_Object charset_list
;
857 register int *reg
= ccl
->reg
;
858 register int ic
= ccl
->ic
;
859 register int code
= 0, field1
, field2
;
860 register Lisp_Object
*ccl_prog
= ccl
->prog
;
861 int *src
= source
, *src_end
= src
+ src_size
;
862 int *dst
= destination
, *dst_end
= dst
+ dst_size
;
865 int stack_idx
= ccl
->stack_idx
;
866 /* Instruction counter of the current CCL code. */
868 struct charset
*charset
;
869 int eof_ic
= ccl
->eof_ic
;
873 ic
= CCL_HEADER_MAIN
;
875 if (ccl
->buf_magnification
== 0) /* We can't read/produce any bytes. */
878 /* Set mapping stack pointer. */
879 mapping_stack_pointer
= mapping_stack
;
882 ccl_backtrace_idx
= 0;
889 ccl_backtrace_table
[ccl_backtrace_idx
++] = ic
;
890 if (ccl_backtrace_idx
>= CCL_DEBUG_BACKTRACE_LEN
)
891 ccl_backtrace_idx
= 0;
892 ccl_backtrace_table
[ccl_backtrace_idx
] = 0;
895 if (!NILP (Vquit_flag
) && NILP (Vinhibit_quit
))
897 /* We can't just signal Qquit, instead break the loop as if
898 the whole data is processed. Don't reset Vquit_flag, it
899 must be handled later at a safer place. */
901 src
= source
+ src_size
;
902 ccl
->status
= CCL_STAT_QUIT
;
907 code
= XINT (ccl_prog
[ic
]); ic
++;
909 field2
= (code
& 0xFF) >> 5;
912 #define RRR (field1 & 7)
913 #define Rrr ((field1 >> 3) & 7)
915 #define EXCMD (field1 >> 6)
919 case CCL_SetRegister
: /* 00000000000000000RRRrrrXXXXX */
923 case CCL_SetShortConst
: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
927 case CCL_SetConst
: /* 00000000000000000000rrrXXXXX */
928 reg
[rrr
] = XINT (ccl_prog
[ic
]);
932 case CCL_SetArray
: /* CCCCCCCCCCCCCCCCCCCCRRRrrrXXXXX */
935 if ((unsigned int) i
< j
)
936 reg
[rrr
] = XINT (ccl_prog
[ic
+ i
]);
940 case CCL_Jump
: /* A--D--D--R--E--S--S-000XXXXX */
944 case CCL_JumpCond
: /* A--D--D--R--E--S--S-rrrXXXXX */
949 case CCL_WriteRegisterJump
: /* A--D--D--R--E--S--S-rrrXXXXX */
955 case CCL_WriteRegisterReadJump
: /* A--D--D--R--E--S--S-rrrXXXXX */
959 CCL_READ_CHAR (reg
[rrr
]);
963 case CCL_WriteConstJump
: /* A--D--D--R--E--S--S-000XXXXX */
964 i
= XINT (ccl_prog
[ic
]);
969 case CCL_WriteConstReadJump
: /* A--D--D--R--E--S--S-rrrXXXXX */
970 i
= XINT (ccl_prog
[ic
]);
973 CCL_READ_CHAR (reg
[rrr
]);
977 case CCL_WriteStringJump
: /* A--D--D--R--E--S--S-000XXXXX */
978 j
= XINT (ccl_prog
[ic
]);
980 CCL_WRITE_STRING (j
);
984 case CCL_WriteArrayReadJump
: /* A--D--D--R--E--S--S-rrrXXXXX */
986 j
= XINT (ccl_prog
[ic
]);
987 if ((unsigned int) i
< j
)
989 i
= XINT (ccl_prog
[ic
+ 1 + i
]);
993 CCL_READ_CHAR (reg
[rrr
]);
994 ic
+= ADDR
- (j
+ 2);
997 case CCL_ReadJump
: /* A--D--D--R--E--S--S-rrrYYYYY */
998 CCL_READ_CHAR (reg
[rrr
]);
1002 case CCL_ReadBranch
: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
1003 CCL_READ_CHAR (reg
[rrr
]);
1004 /* fall through ... */
1005 case CCL_Branch
: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
1006 if ((unsigned int) reg
[rrr
] < field1
)
1007 ic
+= XINT (ccl_prog
[ic
+ reg
[rrr
]]);
1009 ic
+= XINT (ccl_prog
[ic
+ field1
]);
1012 case CCL_ReadRegister
: /* CCCCCCCCCCCCCCCCCCCCrrXXXXX */
1015 CCL_READ_CHAR (reg
[rrr
]);
1017 code
= XINT (ccl_prog
[ic
]); ic
++;
1019 field2
= (code
& 0xFF) >> 5;
1023 case CCL_WriteExprConst
: /* 1:00000OPERATION000RRR000XXXXX */
1026 j
= XINT (ccl_prog
[ic
]);
1028 jump_address
= ic
+ 1;
1031 case CCL_WriteRegister
: /* CCCCCCCCCCCCCCCCCCCrrrXXXXX */
1037 code
= XINT (ccl_prog
[ic
]); ic
++;
1039 field2
= (code
& 0xFF) >> 5;
1043 case CCL_WriteExprRegister
: /* 1:00000OPERATIONRrrRRR000XXXXX */
1051 case CCL_Call
: /* 1:CCCCCCCCCCCCCCCCCCCCFFFXXXXX */
1056 /* If FFF is nonzero, the CCL program ID is in the
1060 prog_id
= XINT (ccl_prog
[ic
]);
1066 if (stack_idx
>= 256
1068 || prog_id
>= ASIZE (Vccl_program_table
)
1069 || (slot
= AREF (Vccl_program_table
, prog_id
), !VECTORP (slot
))
1070 || !VECTORP (AREF (slot
, 1)))
1074 ccl_prog
= ccl_prog_stack_struct
[0].ccl_prog
;
1075 ic
= ccl_prog_stack_struct
[0].ic
;
1076 eof_ic
= ccl_prog_stack_struct
[0].eof_ic
;
1081 ccl_prog_stack_struct
[stack_idx
].ccl_prog
= ccl_prog
;
1082 ccl_prog_stack_struct
[stack_idx
].ic
= ic
;
1083 ccl_prog_stack_struct
[stack_idx
].eof_ic
= eof_ic
;
1085 ccl_prog
= XVECTOR (AREF (slot
, 1))->contents
;
1086 ic
= CCL_HEADER_MAIN
;
1087 eof_ic
= XFASTINT (ccl_prog
[CCL_HEADER_EOF
]);
1091 case CCL_WriteConstString
: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
1093 CCL_WRITE_CHAR (field1
);
1096 CCL_WRITE_STRING (field1
);
1097 ic
+= (field1
+ 2) / 3;
1101 case CCL_WriteArray
: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
1103 if ((unsigned int) i
< field1
)
1105 j
= XINT (ccl_prog
[ic
+ i
]);
1111 case CCL_End
: /* 0000000000000000000000XXXXX */
1115 ccl_prog
= ccl_prog_stack_struct
[stack_idx
].ccl_prog
;
1116 ic
= ccl_prog_stack_struct
[stack_idx
].ic
;
1117 eof_ic
= ccl_prog_stack_struct
[stack_idx
].eof_ic
;
1124 /* ccl->ic should points to this command code again to
1125 suppress further processing. */
1129 case CCL_ExprSelfConst
: /* 00000OPERATION000000rrrXXXXX */
1130 i
= XINT (ccl_prog
[ic
]);
1135 case CCL_ExprSelfReg
: /* 00000OPERATION000RRRrrrXXXXX */
1142 case CCL_PLUS
: reg
[rrr
] += i
; break;
1143 case CCL_MINUS
: reg
[rrr
] -= i
; break;
1144 case CCL_MUL
: reg
[rrr
] *= i
; break;
1145 case CCL_DIV
: reg
[rrr
] /= i
; break;
1146 case CCL_MOD
: reg
[rrr
] %= i
; break;
1147 case CCL_AND
: reg
[rrr
] &= i
; break;
1148 case CCL_OR
: reg
[rrr
] |= i
; break;
1149 case CCL_XOR
: reg
[rrr
] ^= i
; break;
1150 case CCL_LSH
: reg
[rrr
] <<= i
; break;
1151 case CCL_RSH
: reg
[rrr
] >>= i
; break;
1152 case CCL_LSH8
: reg
[rrr
] <<= 8; reg
[rrr
] |= i
; break;
1153 case CCL_RSH8
: reg
[7] = reg
[rrr
] & 0xFF; reg
[rrr
] >>= 8; break;
1154 case CCL_DIVMOD
: reg
[7] = reg
[rrr
] % i
; reg
[rrr
] /= i
; break;
1155 case CCL_LS
: reg
[rrr
] = reg
[rrr
] < i
; break;
1156 case CCL_GT
: reg
[rrr
] = reg
[rrr
] > i
; break;
1157 case CCL_EQ
: reg
[rrr
] = reg
[rrr
] == i
; break;
1158 case CCL_LE
: reg
[rrr
] = reg
[rrr
] <= i
; break;
1159 case CCL_GE
: reg
[rrr
] = reg
[rrr
] >= i
; break;
1160 case CCL_NE
: reg
[rrr
] = reg
[rrr
] != i
; break;
1161 default: CCL_INVALID_CMD
;
1165 case CCL_SetExprConst
: /* 00000OPERATION000RRRrrrXXXXX */
1167 j
= XINT (ccl_prog
[ic
]);
1169 jump_address
= ++ic
;
1172 case CCL_SetExprReg
: /* 00000OPERATIONRrrRRRrrrXXXXX */
1179 case CCL_ReadJumpCondExprConst
: /* A--D--D--R--E--S--S-rrrXXXXX */
1180 CCL_READ_CHAR (reg
[rrr
]);
1181 case CCL_JumpCondExprConst
: /* A--D--D--R--E--S--S-rrrXXXXX */
1183 op
= XINT (ccl_prog
[ic
]);
1184 jump_address
= ic
++ + ADDR
;
1185 j
= XINT (ccl_prog
[ic
]);
1190 case CCL_ReadJumpCondExprReg
: /* A--D--D--R--E--S--S-rrrXXXXX */
1191 CCL_READ_CHAR (reg
[rrr
]);
1192 case CCL_JumpCondExprReg
:
1194 op
= XINT (ccl_prog
[ic
]);
1195 jump_address
= ic
++ + ADDR
;
1196 j
= reg
[XINT (ccl_prog
[ic
])];
1203 case CCL_PLUS
: reg
[rrr
] = i
+ j
; break;
1204 case CCL_MINUS
: reg
[rrr
] = i
- j
; break;
1205 case CCL_MUL
: reg
[rrr
] = i
* j
; break;
1206 case CCL_DIV
: reg
[rrr
] = i
/ j
; break;
1207 case CCL_MOD
: reg
[rrr
] = i
% j
; break;
1208 case CCL_AND
: reg
[rrr
] = i
& j
; break;
1209 case CCL_OR
: reg
[rrr
] = i
| j
; break;
1210 case CCL_XOR
: reg
[rrr
] = i
^ j
;; break;
1211 case CCL_LSH
: reg
[rrr
] = i
<< j
; break;
1212 case CCL_RSH
: reg
[rrr
] = i
>> j
; break;
1213 case CCL_LSH8
: reg
[rrr
] = (i
<< 8) | j
; break;
1214 case CCL_RSH8
: reg
[rrr
] = i
>> 8; reg
[7] = i
& 0xFF; break;
1215 case CCL_DIVMOD
: reg
[rrr
] = i
/ j
; reg
[7] = i
% j
; break;
1216 case CCL_LS
: reg
[rrr
] = i
< j
; break;
1217 case CCL_GT
: reg
[rrr
] = i
> j
; break;
1218 case CCL_EQ
: reg
[rrr
] = i
== j
; break;
1219 case CCL_LE
: reg
[rrr
] = i
<= j
; break;
1220 case CCL_GE
: reg
[rrr
] = i
>= j
; break;
1221 case CCL_NE
: reg
[rrr
] = i
!= j
; break;
1222 case CCL_DECODE_SJIS
:
1230 case CCL_ENCODE_SJIS
:
1238 default: CCL_INVALID_CMD
;
1241 if (code
== CCL_WriteExprConst
|| code
== CCL_WriteExprRegister
)
1254 case CCL_ReadMultibyteChar2
:
1258 CCL_ENCODE_CHAR (i
, charset_list
, reg
[RRR
], reg
[rrr
]);
1261 case CCL_WriteMultibyteChar2
:
1264 i
= CCL_DECODE_CHAR (reg
[RRR
], reg
[rrr
]);
1268 case CCL_TranslateCharacter
:
1269 i
= CCL_DECODE_CHAR (reg
[RRR
], reg
[rrr
]);
1270 op
= translate_char (GET_TRANSLATION_TABLE (reg
[Rrr
]), i
);
1271 CCL_ENCODE_CHAR (op
, charset_list
, reg
[RRR
], reg
[rrr
]);
1274 case CCL_TranslateCharacterConstTbl
:
1275 op
= XINT (ccl_prog
[ic
]); /* table */
1277 i
= CCL_DECODE_CHAR (reg
[RRR
], reg
[rrr
]);
1278 op
= translate_char (GET_TRANSLATION_TABLE (op
), i
);
1279 CCL_ENCODE_CHAR (op
, charset_list
, reg
[RRR
], reg
[rrr
]);
1282 case CCL_LookupIntConstTbl
:
1283 op
= XINT (ccl_prog
[ic
]); /* table */
1286 struct Lisp_Hash_Table
*h
= GET_HASH_TABLE (op
);
1288 op
= hash_lookup (h
, make_number (reg
[RRR
]), NULL
);
1292 opl
= HASH_VALUE (h
, op
);
1293 if (! CHARACTERP (opl
))
1295 reg
[RRR
] = charset_unicode
;
1297 reg
[7] = 1; /* r7 true for success */
1304 case CCL_LookupCharConstTbl
:
1305 op
= XINT (ccl_prog
[ic
]); /* table */
1307 i
= CCL_DECODE_CHAR (reg
[RRR
], reg
[rrr
]);
1309 struct Lisp_Hash_Table
*h
= GET_HASH_TABLE (op
);
1311 op
= hash_lookup (h
, make_number (i
), NULL
);
1315 opl
= HASH_VALUE (h
, op
);
1316 if (!INTEGERP (opl
))
1318 reg
[RRR
] = XINT (opl
);
1319 reg
[7] = 1; /* r7 true for success */
1326 case CCL_IterateMultipleMap
:
1328 Lisp_Object map
, content
, attrib
, value
;
1329 int point
, size
, fin_ic
;
1331 j
= XINT (ccl_prog
[ic
++]); /* number of maps. */
1334 if ((j
> reg
[RRR
]) && (j
>= 0))
1349 size
= ASIZE (Vcode_conversion_map_vector
);
1350 point
= XINT (ccl_prog
[ic
++]);
1351 if (point
>= size
) continue;
1352 map
= AREF (Vcode_conversion_map_vector
, point
);
1354 /* Check map varidity. */
1355 if (!CONSP (map
)) continue;
1357 if (!VECTORP (map
)) continue;
1359 if (size
<= 1) continue;
1361 content
= AREF (map
, 0);
1364 [STARTPOINT VAL1 VAL2 ...] or
1365 [t ELELMENT STARTPOINT ENDPOINT] */
1366 if (NUMBERP (content
))
1368 point
= XUINT (content
);
1369 point
= op
- point
+ 1;
1370 if (!((point
>= 1) && (point
< size
))) continue;
1371 content
= AREF (map
, point
);
1373 else if (EQ (content
, Qt
))
1375 if (size
!= 4) continue;
1376 if ((op
>= XUINT (AREF (map
, 2)))
1377 && (op
< XUINT (AREF (map
, 3))))
1378 content
= AREF (map
, 1);
1387 else if (NUMBERP (content
))
1390 reg
[rrr
] = XINT(content
);
1393 else if (EQ (content
, Qt
) || EQ (content
, Qlambda
))
1398 else if (CONSP (content
))
1400 attrib
= XCAR (content
);
1401 value
= XCDR (content
);
1402 if (!NUMBERP (attrib
) || !NUMBERP (value
))
1405 reg
[rrr
] = XUINT (value
);
1408 else if (SYMBOLP (content
))
1409 CCL_CALL_FOR_MAP_INSTRUCTION (content
, fin_ic
);
1419 case CCL_MapMultiple
:
1421 Lisp_Object map
, content
, attrib
, value
;
1422 int point
, size
, map_vector_size
;
1423 int map_set_rest_length
, fin_ic
;
1424 int current_ic
= this_ic
;
1426 /* inhibit recursive call on MapMultiple. */
1427 if (stack_idx_of_map_multiple
> 0)
1429 if (stack_idx_of_map_multiple
<= stack_idx
)
1431 stack_idx_of_map_multiple
= 0;
1432 mapping_stack_pointer
= mapping_stack
;
1437 mapping_stack_pointer
= mapping_stack
;
1438 stack_idx_of_map_multiple
= 0;
1440 map_set_rest_length
=
1441 XINT (ccl_prog
[ic
++]); /* number of maps and separators. */
1442 fin_ic
= ic
+ map_set_rest_length
;
1445 if ((map_set_rest_length
> reg
[RRR
]) && (reg
[RRR
] >= 0))
1449 map_set_rest_length
-= i
;
1455 mapping_stack_pointer
= mapping_stack
;
1459 if (mapping_stack_pointer
<= (mapping_stack
+ 1))
1461 /* Set up initial state. */
1462 mapping_stack_pointer
= mapping_stack
;
1463 PUSH_MAPPING_STACK (0, op
);
1468 /* Recover after calling other ccl program. */
1471 POP_MAPPING_STACK (map_set_rest_length
, orig_op
);
1472 POP_MAPPING_STACK (map_set_rest_length
, reg
[rrr
]);
1476 /* Regard it as Qnil. */
1480 map_set_rest_length
--;
1483 /* Regard it as Qt. */
1487 map_set_rest_length
--;
1490 /* Regard it as Qlambda. */
1492 i
+= map_set_rest_length
;
1493 ic
+= map_set_rest_length
;
1494 map_set_rest_length
= 0;
1497 /* Regard it as normal mapping. */
1498 i
+= map_set_rest_length
;
1499 ic
+= map_set_rest_length
;
1500 POP_MAPPING_STACK (map_set_rest_length
, reg
[rrr
]);
1504 map_vector_size
= ASIZE (Vcode_conversion_map_vector
);
1507 for (;map_set_rest_length
> 0;i
++, ic
++, map_set_rest_length
--)
1509 point
= XINT(ccl_prog
[ic
]);
1512 /* +1 is for including separator. */
1514 if (mapping_stack_pointer
1515 >= &mapping_stack
[MAX_MAP_SET_LEVEL
])
1517 PUSH_MAPPING_STACK (map_set_rest_length
- point
,
1519 map_set_rest_length
= point
;
1524 if (point
>= map_vector_size
) continue;
1525 map
= AREF (Vcode_conversion_map_vector
, point
);
1527 /* Check map varidity. */
1528 if (!CONSP (map
)) continue;
1530 if (!VECTORP (map
)) continue;
1532 if (size
<= 1) continue;
1534 content
= AREF (map
, 0);
1537 [STARTPOINT VAL1 VAL2 ...] or
1538 [t ELEMENT STARTPOINT ENDPOINT] */
1539 if (NUMBERP (content
))
1541 point
= XUINT (content
);
1542 point
= op
- point
+ 1;
1543 if (!((point
>= 1) && (point
< size
))) continue;
1544 content
= AREF (map
, point
);
1546 else if (EQ (content
, Qt
))
1548 if (size
!= 4) continue;
1549 if ((op
>= XUINT (AREF (map
, 2))) &&
1550 (op
< XUINT (AREF (map
, 3))))
1551 content
= AREF (map
, 1);
1562 if (NUMBERP (content
))
1564 op
= XINT (content
);
1565 i
+= map_set_rest_length
- 1;
1566 ic
+= map_set_rest_length
- 1;
1567 POP_MAPPING_STACK (map_set_rest_length
, reg
[rrr
]);
1568 map_set_rest_length
++;
1570 else if (CONSP (content
))
1572 attrib
= XCAR (content
);
1573 value
= XCDR (content
);
1574 if (!NUMBERP (attrib
) || !NUMBERP (value
))
1577 i
+= map_set_rest_length
- 1;
1578 ic
+= map_set_rest_length
- 1;
1579 POP_MAPPING_STACK (map_set_rest_length
, reg
[rrr
]);
1580 map_set_rest_length
++;
1582 else if (EQ (content
, Qt
))
1586 else if (EQ (content
, Qlambda
))
1588 i
+= map_set_rest_length
;
1589 ic
+= map_set_rest_length
;
1592 else if (SYMBOLP (content
))
1594 if (mapping_stack_pointer
1595 >= &mapping_stack
[MAX_MAP_SET_LEVEL
])
1597 PUSH_MAPPING_STACK (map_set_rest_length
, reg
[rrr
]);
1598 PUSH_MAPPING_STACK (map_set_rest_length
, op
);
1599 stack_idx_of_map_multiple
= stack_idx
+ 1;
1600 CCL_CALL_FOR_MAP_INSTRUCTION (content
, current_ic
);
1605 if (mapping_stack_pointer
<= (mapping_stack
+ 1))
1607 POP_MAPPING_STACK (map_set_rest_length
, reg
[rrr
]);
1608 i
+= map_set_rest_length
;
1609 ic
+= map_set_rest_length
;
1610 POP_MAPPING_STACK (map_set_rest_length
, reg
[rrr
]);
1620 Lisp_Object map
, attrib
, value
, content
;
1622 j
= XINT (ccl_prog
[ic
++]); /* map_id */
1624 if (j
>= ASIZE (Vcode_conversion_map_vector
))
1629 map
= AREF (Vcode_conversion_map_vector
, j
);
1642 point
= XUINT (AREF (map
, 0));
1643 point
= op
- point
+ 1;
1646 (!((point
>= 1) && (point
< size
))))
1651 content
= AREF (map
, point
);
1654 else if (NUMBERP (content
))
1655 reg
[rrr
] = XINT (content
);
1656 else if (EQ (content
, Qt
));
1657 else if (CONSP (content
))
1659 attrib
= XCAR (content
);
1660 value
= XCDR (content
);
1661 if (!NUMBERP (attrib
) || !NUMBERP (value
))
1663 reg
[rrr
] = XUINT(value
);
1666 else if (SYMBOLP (content
))
1667 CCL_CALL_FOR_MAP_INSTRUCTION (content
, ic
);
1685 /* The suppress_error member is set when e.g. a CCL-based coding
1686 system is used for terminal output. */
1687 if (!ccl
->suppress_error
&& destination
)
1689 /* We can insert an error message only if DESTINATION is
1690 specified and we still have a room to store the message
1698 switch (ccl
->status
)
1700 case CCL_STAT_INVALID_CMD
:
1701 sprintf(msg
, "\nCCL: Invalid command %x (ccl_code = %x) at %d.",
1702 code
& 0x1F, code
, this_ic
);
1705 int i
= ccl_backtrace_idx
- 1;
1708 msglen
= strlen (msg
);
1709 if (dst
+ msglen
<= (dst_bytes
? dst_end
: src
))
1711 bcopy (msg
, dst
, msglen
);
1715 for (j
= 0; j
< CCL_DEBUG_BACKTRACE_LEN
; j
++, i
--)
1717 if (i
< 0) i
= CCL_DEBUG_BACKTRACE_LEN
- 1;
1718 if (ccl_backtrace_table
[i
] == 0)
1720 sprintf(msg
, " %d", ccl_backtrace_table
[i
]);
1721 msglen
= strlen (msg
);
1722 if (dst
+ msglen
> (dst_bytes
? dst_end
: src
))
1724 bcopy (msg
, dst
, msglen
);
1733 sprintf(msg
, "\nCCL: Quited.");
1737 sprintf(msg
, "\nCCL: Unknown error type (%d)", ccl
->status
);
1740 msglen
= strlen (msg
);
1741 if (dst
+ msglen
<= dst_end
)
1743 for (i
= 0; i
< msglen
; i
++)
1747 if (ccl
->status
== CCL_STAT_INVALID_CMD
)
1749 #if 0 /* If the remaining bytes contain 0x80..0x9F, copying them
1750 results in an invalid multibyte sequence. */
1752 /* Copy the remaining source data. */
1753 int i
= src_end
- src
;
1754 if (dst_bytes
&& (dst_end
- dst
) < i
)
1756 bcopy (src
, dst
, i
);
1760 /* Signal that we've consumed everything. */
1768 ccl
->stack_idx
= stack_idx
;
1769 ccl
->prog
= ccl_prog
;
1770 ccl
->consumed
= src
- source
;
1771 ccl
->produced
= dst
- destination
;
1774 /* Resolve symbols in the specified CCL code (Lisp vector). This
1775 function converts symbols of code conversion maps and character
1776 translation tables embeded in the CCL code into their ID numbers.
1778 The return value is a vector (CCL itself or a new vector in which
1779 all symbols are resolved), Qt if resolving of some symbol failed,
1780 or nil if CCL contains invalid data. */
1783 resolve_symbol_ccl_program (ccl
)
1786 int i
, veclen
, unresolved
= 0;
1787 Lisp_Object result
, contents
, val
;
1790 veclen
= ASIZE (result
);
1792 for (i
= 0; i
< veclen
; i
++)
1794 contents
= AREF (result
, i
);
1795 if (INTEGERP (contents
))
1797 else if (CONSP (contents
)
1798 && SYMBOLP (XCAR (contents
))
1799 && SYMBOLP (XCDR (contents
)))
1801 /* This is the new style for embedding symbols. The form is
1802 (SYMBOL . PROPERTY). (get SYMBOL PROPERTY) should give
1805 if (EQ (result
, ccl
))
1806 result
= Fcopy_sequence (ccl
);
1808 val
= Fget (XCAR (contents
), XCDR (contents
));
1810 AREF (result
, i
) = val
;
1815 else if (SYMBOLP (contents
))
1817 /* This is the old style for embedding symbols. This style
1818 may lead to a bug if, for instance, a translation table
1819 and a code conversion map have the same name. */
1820 if (EQ (result
, ccl
))
1821 result
= Fcopy_sequence (ccl
);
1823 val
= Fget (contents
, Qtranslation_table_id
);
1825 AREF (result
, i
) = val
;
1828 val
= Fget (contents
, Qcode_conversion_map_id
);
1830 AREF (result
, i
) = val
;
1833 val
= Fget (contents
, Qccl_program_idx
);
1835 AREF (result
, i
) = val
;
1845 return (unresolved
? Qt
: result
);
1848 /* Return the compiled code (vector) of CCL program CCL_PROG.
1849 CCL_PROG is a name (symbol) of the program or already compiled
1850 code. If necessary, resolve symbols in the compiled code to index
1851 numbers. If we failed to get the compiled code or to resolve
1852 symbols, return Qnil. */
1855 ccl_get_compiled_code (ccl_prog
, idx
)
1856 Lisp_Object ccl_prog
;
1859 Lisp_Object val
, slot
;
1861 if (VECTORP (ccl_prog
))
1863 val
= resolve_symbol_ccl_program (ccl_prog
);
1865 return (VECTORP (val
) ? val
: Qnil
);
1867 if (!SYMBOLP (ccl_prog
))
1870 val
= Fget (ccl_prog
, Qccl_program_idx
);
1872 || XINT (val
) >= ASIZE (Vccl_program_table
))
1874 slot
= AREF (Vccl_program_table
, XINT (val
));
1875 if (! VECTORP (slot
)
1876 || ASIZE (slot
) != 4
1877 || ! VECTORP (AREF (slot
, 1)))
1880 if (NILP (AREF (slot
, 2)))
1882 val
= resolve_symbol_ccl_program (AREF (slot
, 1));
1883 if (! VECTORP (val
))
1885 AREF (slot
, 1) = val
;
1886 AREF (slot
, 2) = Qt
;
1888 return AREF (slot
, 1);
1891 /* Setup fields of the structure pointed by CCL appropriately for the
1892 execution of CCL program CCL_PROG. CCL_PROG is the name (symbol)
1893 of the CCL program or the already compiled code (vector).
1894 Return 0 if we succeed this setup, else return -1.
1896 If CCL_PROG is nil, we just reset the structure pointed by CCL. */
1898 setup_ccl_program (ccl
, ccl_prog
)
1899 struct ccl_program
*ccl
;
1900 Lisp_Object ccl_prog
;
1904 if (! NILP (ccl_prog
))
1906 struct Lisp_Vector
*vp
;
1908 ccl_prog
= ccl_get_compiled_code (ccl_prog
, &ccl
->idx
);
1909 if (! VECTORP (ccl_prog
))
1911 vp
= XVECTOR (ccl_prog
);
1912 ccl
->size
= vp
->size
;
1913 ccl
->prog
= vp
->contents
;
1914 ccl
->eof_ic
= XINT (vp
->contents
[CCL_HEADER_EOF
]);
1915 ccl
->buf_magnification
= XINT (vp
->contents
[CCL_HEADER_BUF_MAG
]);
1920 slot
= AREF (Vccl_program_table
, ccl
->idx
);
1921 ASET (slot
, 3, Qnil
);
1924 ccl
->ic
= CCL_HEADER_MAIN
;
1925 for (i
= 0; i
< 8; i
++)
1927 ccl
->last_block
= 0;
1928 ccl
->private_state
= 0;
1931 ccl
->suppress_error
= 0;
1932 ccl
->eight_bit_control
= 0;
1937 /* Check if CCL is updated or not. If not, re-setup members of CCL. */
1940 check_ccl_update (ccl
)
1941 struct ccl_program
*ccl
;
1943 struct Lisp_Vector
*vp
;
1944 Lisp_Object slot
, ccl_prog
;
1948 slot
= AREF (Vccl_program_table
, ccl
->idx
);
1949 if (NILP (AREF (slot
, 3)))
1951 ccl_prog
= ccl_get_compiled_code (AREF (slot
, 0), &ccl
->idx
);
1952 if (! VECTORP (ccl_prog
))
1954 ccl
->size
= ASIZE (ccl_prog
);
1955 ccl
->prog
= XVECTOR (ccl_prog
)->contents
;
1956 ccl
->eof_ic
= XINT (AREF (ccl_prog
, CCL_HEADER_EOF
));
1957 ccl
->buf_magnification
= XINT (AREF (ccl_prog
, CCL_HEADER_BUF_MAG
));
1958 ASET (slot
, 3, Qnil
);
1963 DEFUN ("ccl-program-p", Fccl_program_p
, Sccl_program_p
, 1, 1, 0,
1964 doc
: /* Return t if OBJECT is a CCL program name or a compiled CCL program code.
1965 See the documentation of `define-ccl-program' for the detail of CCL program. */)
1971 if (VECTORP (object
))
1973 val
= resolve_symbol_ccl_program (object
);
1974 return (VECTORP (val
) ? Qt
: Qnil
);
1976 if (!SYMBOLP (object
))
1979 val
= Fget (object
, Qccl_program_idx
);
1980 return ((! NATNUMP (val
)
1981 || XINT (val
) >= ASIZE (Vccl_program_table
))
1985 DEFUN ("ccl-execute", Fccl_execute
, Sccl_execute
, 2, 2, 0,
1986 doc
: /* Execute CCL-PROGRAM with registers initialized by REGISTERS.
1988 CCL-PROGRAM is a CCL program name (symbol)
1989 or compiled code generated by `ccl-compile' (for backward compatibility.
1990 In the latter case, the execution overhead is bigger than in the former).
1991 No I/O commands should appear in CCL-PROGRAM.
1993 REGISTERS is a vector of [R0 R1 ... R7] where RN is an initial value
1994 for the Nth register.
1996 As side effect, each element of REGISTERS holds the value of
1997 the corresponding register after the execution.
1999 See the documentation of `define-ccl-program' for a definition of CCL
2002 Lisp_Object ccl_prog
, reg
;
2004 struct ccl_program ccl
;
2007 if (setup_ccl_program (&ccl
, ccl_prog
) < 0)
2008 error ("Invalid CCL program");
2011 if (ASIZE (reg
) != 8)
2012 error ("Length of vector REGISTERS is not 8");
2014 for (i
= 0; i
< 8; i
++)
2015 ccl
.reg
[i
] = (INTEGERP (AREF (reg
, i
))
2016 ? XINT (AREF (reg
, i
))
2019 ccl_driver (&ccl
, NULL
, NULL
, 0, 0, Qnil
);
2021 if (ccl
.status
!= CCL_STAT_SUCCESS
)
2022 error ("Error in CCL program at %dth code", ccl
.ic
);
2024 for (i
= 0; i
< 8; i
++)
2025 XSETINT (AREF (reg
, i
), ccl
.reg
[i
]);
2029 DEFUN ("ccl-execute-on-string", Fccl_execute_on_string
, Sccl_execute_on_string
,
2031 doc
: /* Execute CCL-PROGRAM with initial STATUS on STRING.
2033 CCL-PROGRAM is a symbol registered by register-ccl-program,
2034 or a compiled code generated by `ccl-compile' (for backward compatibility,
2035 in this case, the execution is slower).
2037 Read buffer is set to STRING, and write buffer is allocated automatically.
2039 STATUS is a vector of [R0 R1 ... R7 IC], where
2040 R0..R7 are initial values of corresponding registers,
2041 IC is the instruction counter specifying from where to start the program.
2042 If R0..R7 are nil, they are initialized to 0.
2043 If IC is nil, it is initialized to head of the CCL program.
2045 If optional 4th arg CONTINUE is non-nil, keep IC on read operation
2046 when read buffer is exausted, else, IC is always set to the end of
2047 CCL-PROGRAM on exit.
2049 It returns the contents of write buffer as a string,
2050 and as side effect, STATUS is updated.
2051 If the optional 5th arg UNIBYTE-P is non-nil, the returned string
2052 is a unibyte string. By default it is a multibyte string.
2054 See the documentation of `define-ccl-program' for the detail of CCL program. */)
2055 (ccl_prog
, status
, str
, contin
, unibyte_p
)
2056 Lisp_Object ccl_prog
, status
, str
, contin
, unibyte_p
;
2059 struct ccl_program ccl
;
2062 unsigned char *outbuf
, *outp
;
2063 int str_chars
, str_bytes
;
2064 #define CCL_EXECUTE_BUF_SIZE 1024
2065 int source
[CCL_EXECUTE_BUF_SIZE
], destination
[CCL_EXECUTE_BUF_SIZE
];
2066 int consumed_chars
, consumed_bytes
, produced_chars
;
2068 if (setup_ccl_program (&ccl
, ccl_prog
) < 0)
2069 error ("Invalid CCL program");
2071 CHECK_VECTOR (status
);
2072 if (ASIZE (status
) != 9)
2073 error ("Length of vector STATUS is not 9");
2076 str_chars
= SCHARS (str
);
2077 str_bytes
= SBYTES (str
);
2079 for (i
= 0; i
< 8; i
++)
2081 if (NILP (AREF (status
, i
)))
2082 XSETINT (AREF (status
, i
), 0);
2083 if (INTEGERP (AREF (status
, i
)))
2084 ccl
.reg
[i
] = XINT (AREF (status
, i
));
2086 if (INTEGERP (AREF (status
, i
)))
2088 i
= XFASTINT (AREF (status
, 8));
2089 if (ccl
.ic
< i
&& i
< ccl
.size
)
2093 outbufsize
= (ccl
.buf_magnification
2094 ? str_bytes
* ccl
.buf_magnification
+ 256
2096 outp
= outbuf
= (unsigned char *) xmalloc (outbufsize
);
2098 consumed_chars
= consumed_bytes
= 0;
2102 const unsigned char *p
= SDATA (str
) + consumed_bytes
;
2103 const unsigned char *endp
= SDATA (str
) + str_bytes
;
2107 if (endp
- p
== str_chars
- consumed_chars
)
2108 while (i
< CCL_EXECUTE_BUF_SIZE
&& p
< endp
)
2111 while (i
< CCL_EXECUTE_BUF_SIZE
&& p
< endp
)
2112 source
[i
++] = STRING_CHAR_ADVANCE (p
);
2113 consumed_chars
+= i
;
2114 consumed_bytes
= p
- SDATA (str
);
2116 if (consumed_bytes
== str_bytes
)
2117 ccl
.last_block
= NILP (contin
);
2122 ccl_driver (&ccl
, src
, destination
, src_size
, CCL_EXECUTE_BUF_SIZE
,
2124 produced_chars
+= ccl
.produced
;
2125 if (NILP (unibyte_p
))
2127 if (outp
- outbuf
+ MAX_MULTIBYTE_LENGTH
* ccl
.produced
2130 int offset
= outp
- outbuf
;
2131 outbufsize
+= MAX_MULTIBYTE_LENGTH
* ccl
.produced
;
2132 outbuf
= (unsigned char *) xrealloc (outbuf
, outbufsize
);
2133 outp
= outbuf
+ offset
;
2135 for (i
= 0; i
< ccl
.produced
; i
++)
2136 CHAR_STRING_ADVANCE (destination
[i
], outp
);
2140 if (outp
- outbuf
+ ccl
.produced
> outbufsize
)
2142 int offset
= outp
- outbuf
;
2143 outbufsize
+= ccl
.produced
;
2144 outbuf
= (unsigned char *) xrealloc (outbuf
, outbufsize
);
2145 outp
= outbuf
+ offset
;
2147 for (i
= 0; i
< ccl
.produced
; i
++)
2148 *outp
++ = destination
[i
];
2150 src
+= ccl
.consumed
;
2151 src_size
-= ccl
.consumed
;
2152 if (ccl
.status
!= CCL_STAT_SUSPEND_BY_DST
)
2156 if (ccl
.status
!= CCL_STAT_SUSPEND_BY_SRC
)
2160 if (ccl
.status
!= CCL_STAT_SUCCESS
2161 && ccl
.status
!= CCL_STAT_SUSPEND_BY_SRC
)
2162 error ("Error in CCL program at %dth code", ccl
.ic
);
2164 for (i
= 0; i
< 8; i
++)
2165 XSET (XVECTOR (status
)->contents
[i
], Lisp_Int
, ccl
.reg
[i
]);
2166 XSETINT (XVECTOR (status
)->contents
[8], ccl
.ic
);
2168 if (NILP (unibyte_p
))
2169 val
= make_multibyte_string ((char *) outbuf
, produced_chars
,
2172 val
= make_unibyte_string ((char *) outbuf
, produced_chars
);
2178 DEFUN ("register-ccl-program", Fregister_ccl_program
, Sregister_ccl_program
,
2180 doc
: /* Register CCL program CCL_PROG as NAME in `ccl-program-table'.
2181 CCL_PROG should be a compiled CCL program (vector), or nil.
2182 If it is nil, just reserve NAME as a CCL program name.
2183 Return index number of the registered CCL program. */)
2185 Lisp_Object name
, ccl_prog
;
2187 int len
= ASIZE (Vccl_program_table
);
2189 Lisp_Object resolved
;
2191 CHECK_SYMBOL (name
);
2193 if (!NILP (ccl_prog
))
2195 CHECK_VECTOR (ccl_prog
);
2196 resolved
= resolve_symbol_ccl_program (ccl_prog
);
2197 if (NILP (resolved
))
2198 error ("Error in CCL program");
2199 if (VECTORP (resolved
))
2201 ccl_prog
= resolved
;
2208 for (idx
= 0; idx
< len
; idx
++)
2212 slot
= AREF (Vccl_program_table
, idx
);
2213 if (!VECTORP (slot
))
2214 /* This is the first unsed slot. Register NAME here. */
2217 if (EQ (name
, AREF (slot
, 0)))
2219 /* Update this slot. */
2220 ASET (slot
, 1, ccl_prog
);
2221 ASET (slot
, 2, resolved
);
2223 return make_number (idx
);
2229 /* Extend the table. */
2230 Lisp_Object new_table
;
2233 new_table
= Fmake_vector (make_number (len
* 2), Qnil
);
2234 for (j
= 0; j
< len
; j
++)
2235 ASET (new_table
, j
, AREF (Vccl_program_table
, j
));
2236 Vccl_program_table
= new_table
;
2242 elt
= Fmake_vector (make_number (4), Qnil
);
2243 ASET (elt
, 0, name
);
2244 ASET (elt
, 1, ccl_prog
);
2245 ASET (elt
, 2, resolved
);
2247 ASET (Vccl_program_table
, idx
, elt
);
2250 Fput (name
, Qccl_program_idx
, make_number (idx
));
2251 return make_number (idx
);
2254 /* Register code conversion map.
2255 A code conversion map consists of numbers, Qt, Qnil, and Qlambda.
2256 The first element is the start code point.
2257 The other elements are mapped numbers.
2258 Symbol t means to map to an original number before mapping.
2259 Symbol nil means that the corresponding element is empty.
2260 Symbol lambda means to terminate mapping here.
2263 DEFUN ("register-code-conversion-map", Fregister_code_conversion_map
,
2264 Sregister_code_conversion_map
,
2266 doc
: /* Register SYMBOL as code conversion map MAP.
2267 Return index number of the registered map. */)
2269 Lisp_Object symbol
, map
;
2271 int len
= ASIZE (Vcode_conversion_map_vector
);
2275 CHECK_SYMBOL (symbol
);
2278 for (i
= 0; i
< len
; i
++)
2280 Lisp_Object slot
= AREF (Vcode_conversion_map_vector
, i
);
2285 if (EQ (symbol
, XCAR (slot
)))
2287 index
= make_number (i
);
2288 XSETCDR (slot
, map
);
2289 Fput (symbol
, Qcode_conversion_map
, map
);
2290 Fput (symbol
, Qcode_conversion_map_id
, index
);
2297 Lisp_Object new_vector
= Fmake_vector (make_number (len
* 2), Qnil
);
2300 for (j
= 0; j
< len
; j
++)
2301 AREF (new_vector
, j
)
2302 = AREF (Vcode_conversion_map_vector
, j
);
2303 Vcode_conversion_map_vector
= new_vector
;
2306 index
= make_number (i
);
2307 Fput (symbol
, Qcode_conversion_map
, map
);
2308 Fput (symbol
, Qcode_conversion_map_id
, index
);
2309 AREF (Vcode_conversion_map_vector
, i
) = Fcons (symbol
, map
);
2317 staticpro (&Vccl_program_table
);
2318 Vccl_program_table
= Fmake_vector (make_number (32), Qnil
);
2320 Qccl
= intern ("ccl");
2323 Qcclp
= intern ("cclp");
2326 Qccl_program
= intern ("ccl-program");
2327 staticpro (&Qccl_program
);
2329 Qccl_program_idx
= intern ("ccl-program-idx");
2330 staticpro (&Qccl_program_idx
);
2332 Qcode_conversion_map
= intern ("code-conversion-map");
2333 staticpro (&Qcode_conversion_map
);
2335 Qcode_conversion_map_id
= intern ("code-conversion-map-id");
2336 staticpro (&Qcode_conversion_map_id
);
2338 DEFVAR_LISP ("code-conversion-map-vector", &Vcode_conversion_map_vector
,
2339 doc
: /* Vector of code conversion maps. */);
2340 Vcode_conversion_map_vector
= Fmake_vector (make_number (16), Qnil
);
2342 DEFVAR_LISP ("font-ccl-encoder-alist", &Vfont_ccl_encoder_alist
,
2343 doc
: /* Alist of fontname patterns vs corresponding CCL program.
2344 Each element looks like (REGEXP . CCL-CODE),
2345 where CCL-CODE is a compiled CCL program.
2346 When a font whose name matches REGEXP is used for displaying a character,
2347 CCL-CODE is executed to calculate the code point in the font
2348 from the charset number and position code(s) of the character which are set
2349 in CCL registers R0, R1, and R2 before the execution.
2350 The code point in the font is set in CCL registers R1 and R2
2351 when the execution terminated.
2352 If the font is single-byte font, the register R2 is not used. */);
2353 Vfont_ccl_encoder_alist
= Qnil
;
2355 DEFVAR_LISP ("translation-hash-table-vector", &Vtranslation_hash_table_vector
,
2356 doc
: /* Vector containing all translation hash tables ever defined.
2357 Comprises pairs (SYMBOL . TABLE) where SYMBOL and TABLE were set up by calls
2358 to `define-translation-hash-table'. The vector is indexed by the table id
2360 Vtranslation_hash_table_vector
= Qnil
;
2362 defsubr (&Sccl_program_p
);
2363 defsubr (&Sccl_execute
);
2364 defsubr (&Sccl_execute_on_string
);
2365 defsubr (&Sregister_ccl_program
);
2366 defsubr (&Sregister_code_conversion_map
);
2369 /* arch-tag: bb9a37be-68ce-4576-8d3d-15d750e4a860
2370 (do not change this comment) */