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, and RESOLVEDP where NAME (symbol) is the name of
59 the program, CCL_PROG (vector) is the compiled code of the program,
60 RESOLVEDP (t or nil) is the flag to tell if symbols in CCL_PROG is
61 already resolved to index numbers or not. */
62 Lisp_Object Vccl_program_table
;
64 /* Vector of registered hash tables for translation. */
65 Lisp_Object Vtranslation_hash_table_vector
;
67 /* Return a hash table of id number ID. */
68 #define GET_HASH_TABLE(id) \
69 (XHASH_TABLE (XCDR(XVECTOR(Vtranslation_hash_table_vector)->contents[(id)])))
71 extern int charset_unicode
;
73 /* CCL (Code Conversion Language) is a simple language which has
74 operations on one input buffer, one output buffer, and 7 registers.
75 The syntax of CCL is described in `ccl.el'. Emacs Lisp function
76 `ccl-compile' compiles a CCL program and produces a CCL code which
77 is a vector of integers. The structure of this vector is as
78 follows: The 1st element: buffer-magnification, a factor for the
79 size of output buffer compared with the size of input buffer. The
80 2nd element: address of CCL code to be executed when encountered
81 with end of input stream. The 3rd and the remaining elements: CCL
84 /* Header of CCL compiled code */
85 #define CCL_HEADER_BUF_MAG 0
86 #define CCL_HEADER_EOF 1
87 #define CCL_HEADER_MAIN 2
89 /* CCL code is a sequence of 28-bit non-negative integers (i.e. the
90 MSB is always 0), each contains CCL command and/or arguments in the
93 |----------------- integer (28-bit) ------------------|
94 |------- 17-bit ------|- 3-bit --|- 3-bit --|- 5-bit -|
95 |--constant argument--|-register-|-register-|-command-|
96 ccccccccccccccccc RRR rrr XXXXX
98 |------- relative address -------|-register-|-command-|
99 cccccccccccccccccccc rrr XXXXX
101 |------------- constant or other args ----------------|
102 cccccccccccccccccccccccccccc
104 where, `cc...c' is a non-negative integer indicating constant value
105 (the left most `c' is always 0) or an absolute jump address, `RRR'
106 and `rrr' are CCL register number, `XXXXX' is one of the following
111 Each comment fields shows one or more lines for command syntax and
112 the following lines for semantics of the command. In semantics, IC
113 stands for Instruction Counter. */
115 #define CCL_SetRegister 0x00 /* Set register a register value:
116 1:00000000000000000RRRrrrXXXXX
117 ------------------------------
121 #define CCL_SetShortConst 0x01 /* Set register a short constant value:
122 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
123 ------------------------------
124 reg[rrr] = CCCCCCCCCCCCCCCCCCC;
127 #define CCL_SetConst 0x02 /* Set register a constant value:
128 1:00000000000000000000rrrXXXXX
130 ------------------------------
135 #define CCL_SetArray 0x03 /* Set register an element of array:
136 1:CCCCCCCCCCCCCCCCCRRRrrrXXXXX
140 ------------------------------
141 if (0 <= reg[RRR] < CC..C)
142 reg[rrr] = ELEMENT[reg[RRR]];
146 #define CCL_Jump 0x04 /* Jump:
147 1:A--D--D--R--E--S--S-000XXXXX
148 ------------------------------
152 /* Note: If CC..C is greater than 0, the second code is omitted. */
154 #define CCL_JumpCond 0x05 /* Jump conditional:
155 1:A--D--D--R--E--S--S-rrrXXXXX
156 ------------------------------
162 #define CCL_WriteRegisterJump 0x06 /* Write register and jump:
163 1:A--D--D--R--E--S--S-rrrXXXXX
164 ------------------------------
169 #define CCL_WriteRegisterReadJump 0x07 /* Write register, read, and jump:
170 1:A--D--D--R--E--S--S-rrrXXXXX
171 2:A--D--D--R--E--S--S-rrrYYYYY
172 -----------------------------
178 /* Note: If read is suspended, the resumed execution starts from the
179 second code (YYYYY == CCL_ReadJump). */
181 #define CCL_WriteConstJump 0x08 /* Write constant and jump:
182 1:A--D--D--R--E--S--S-000XXXXX
184 ------------------------------
189 #define CCL_WriteConstReadJump 0x09 /* Write constant, read, and jump:
190 1:A--D--D--R--E--S--S-rrrXXXXX
192 3:A--D--D--R--E--S--S-rrrYYYYY
193 -----------------------------
199 /* Note: If read is suspended, the resumed execution starts from the
200 second code (YYYYY == CCL_ReadJump). */
202 #define CCL_WriteStringJump 0x0A /* Write string and jump:
203 1:A--D--D--R--E--S--S-000XXXXX
205 3:0000STRIN[0]STRIN[1]STRIN[2]
207 ------------------------------
208 write_string (STRING, LENGTH);
212 #define CCL_WriteArrayReadJump 0x0B /* Write an array element, read, and jump:
213 1:A--D--D--R--E--S--S-rrrXXXXX
218 N:A--D--D--R--E--S--S-rrrYYYYY
219 ------------------------------
220 if (0 <= reg[rrr] < LENGTH)
221 write (ELEMENT[reg[rrr]]);
222 IC += LENGTH + 2; (... pointing at N+1)
226 /* Note: If read is suspended, the resumed execution starts from the
227 Nth code (YYYYY == CCL_ReadJump). */
229 #define CCL_ReadJump 0x0C /* Read and jump:
230 1:A--D--D--R--E--S--S-rrrYYYYY
231 -----------------------------
236 #define CCL_Branch 0x0D /* Jump by branch table:
237 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
238 2:A--D--D--R--E-S-S[0]000XXXXX
239 3:A--D--D--R--E-S-S[1]000XXXXX
241 ------------------------------
242 if (0 <= reg[rrr] < CC..C)
243 IC += ADDRESS[reg[rrr]];
245 IC += ADDRESS[CC..C];
248 #define CCL_ReadRegister 0x0E /* Read bytes into registers:
249 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
250 2:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
252 ------------------------------
257 #define CCL_WriteExprConst 0x0F /* write result of expression:
258 1:00000OPERATION000RRR000XXXXX
260 ------------------------------
261 write (reg[RRR] OPERATION CONSTANT);
265 /* Note: If the Nth read is suspended, the resumed execution starts
266 from the Nth code. */
268 #define CCL_ReadBranch 0x10 /* Read one byte into a register,
269 and jump by branch table:
270 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
271 2:A--D--D--R--E-S-S[0]000XXXXX
272 3:A--D--D--R--E-S-S[1]000XXXXX
274 ------------------------------
276 if (0 <= reg[rrr] < CC..C)
277 IC += ADDRESS[reg[rrr]];
279 IC += ADDRESS[CC..C];
282 #define CCL_WriteRegister 0x11 /* Write registers:
283 1:CCCCCCCCCCCCCCCCCCCrrrXXXXX
284 2:CCCCCCCCCCCCCCCCCCCrrrXXXXX
286 ------------------------------
292 /* Note: If the Nth write is suspended, the resumed execution
293 starts from the Nth code. */
295 #define CCL_WriteExprRegister 0x12 /* Write result of expression
296 1:00000OPERATIONRrrRRR000XXXXX
297 ------------------------------
298 write (reg[RRR] OPERATION reg[Rrr]);
301 #define CCL_Call 0x13 /* Call the CCL program whose ID is
303 1:CCCCCCCCCCCCCCCCCCCCFFFXXXXX
304 [2:00000000cccccccccccccccccccc]
305 ------------------------------
313 #define CCL_WriteConstString 0x14 /* Write a constant or a string:
314 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
315 [2:0000STRIN[0]STRIN[1]STRIN[2]]
317 -----------------------------
321 write_string (STRING, CC..C);
322 IC += (CC..C + 2) / 3;
325 #define CCL_WriteArray 0x15 /* Write an element of array:
326 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
330 ------------------------------
331 if (0 <= reg[rrr] < CC..C)
332 write (ELEMENT[reg[rrr]]);
336 #define CCL_End 0x16 /* Terminate:
337 1:00000000000000000000000XXXXX
338 ------------------------------
342 /* The following two codes execute an assignment arithmetic/logical
343 operation. The form of the operation is like REG OP= OPERAND. */
345 #define CCL_ExprSelfConst 0x17 /* REG OP= constant:
346 1:00000OPERATION000000rrrXXXXX
348 ------------------------------
349 reg[rrr] OPERATION= CONSTANT;
352 #define CCL_ExprSelfReg 0x18 /* REG1 OP= REG2:
353 1:00000OPERATION000RRRrrrXXXXX
354 ------------------------------
355 reg[rrr] OPERATION= reg[RRR];
358 /* The following codes execute an arithmetic/logical operation. The
359 form of the operation is like REG_X = REG_Y OP OPERAND2. */
361 #define CCL_SetExprConst 0x19 /* REG_X = REG_Y OP constant:
362 1:00000OPERATION000RRRrrrXXXXX
364 ------------------------------
365 reg[rrr] = reg[RRR] OPERATION CONSTANT;
369 #define CCL_SetExprReg 0x1A /* REG1 = REG2 OP REG3:
370 1:00000OPERATIONRrrRRRrrrXXXXX
371 ------------------------------
372 reg[rrr] = reg[RRR] OPERATION reg[Rrr];
375 #define CCL_JumpCondExprConst 0x1B /* Jump conditional according to
376 an operation on constant:
377 1:A--D--D--R--E--S--S-rrrXXXXX
380 -----------------------------
381 reg[7] = reg[rrr] OPERATION CONSTANT;
388 #define CCL_JumpCondExprReg 0x1C /* Jump conditional according to
389 an operation on register:
390 1:A--D--D--R--E--S--S-rrrXXXXX
393 -----------------------------
394 reg[7] = reg[rrr] OPERATION reg[RRR];
401 #define CCL_ReadJumpCondExprConst 0x1D /* Read and jump conditional according
402 to an operation on constant:
403 1:A--D--D--R--E--S--S-rrrXXXXX
406 -----------------------------
408 reg[7] = reg[rrr] OPERATION CONSTANT;
415 #define CCL_ReadJumpCondExprReg 0x1E /* Read and jump conditional according
416 to an operation on register:
417 1:A--D--D--R--E--S--S-rrrXXXXX
420 -----------------------------
422 reg[7] = reg[rrr] OPERATION reg[RRR];
429 #define CCL_Extension 0x1F /* Extended CCL code
430 1:ExtendedCOMMNDRrrRRRrrrXXXXX
433 ------------------------------
434 extended_command (rrr,RRR,Rrr,ARGS)
438 Here after, Extended CCL Instructions.
439 Bit length of extended command is 14.
440 Therefore, the instruction code range is 0..16384(0x3fff).
443 /* Read a multibyte characeter.
444 A code point is stored into reg[rrr]. A charset ID is stored into
447 #define CCL_ReadMultibyteChar2 0x00 /* Read Multibyte Character
448 1:ExtendedCOMMNDRrrRRRrrrXXXXX */
450 /* Write a multibyte character.
451 Write a character whose code point is reg[rrr] and the charset ID
454 #define CCL_WriteMultibyteChar2 0x01 /* Write Multibyte Character
455 1:ExtendedCOMMNDRrrRRRrrrXXXXX */
457 /* Translate a character whose code point is reg[rrr] and the charset
458 ID is reg[RRR] by a translation table whose ID is reg[Rrr].
460 A translated character is set in reg[rrr] (code point) and reg[RRR]
463 #define CCL_TranslateCharacter 0x02 /* Translate a multibyte character
464 1:ExtendedCOMMNDRrrRRRrrrXXXXX */
466 /* Translate a character whose code point is reg[rrr] and the charset
467 ID is reg[RRR] by a translation table whose ID is ARGUMENT.
469 A translated character is set in reg[rrr] (code point) and reg[RRR]
472 #define CCL_TranslateCharacterConstTbl 0x03 /* Translate a multibyte character
473 1:ExtendedCOMMNDRrrRRRrrrXXXXX
474 2:ARGUMENT(Translation Table ID)
477 /* Iterate looking up MAPs for reg[rrr] starting from the Nth (N =
478 reg[RRR]) MAP until some value is found.
480 Each MAP is a Lisp vector whose element is number, nil, t, or
482 If the element is nil, ignore the map and proceed to the next map.
483 If the element is t or lambda, finish without changing reg[rrr].
484 If the element is a number, set reg[rrr] to the number and finish.
486 Detail of the map structure is descibed in the comment for
487 CCL_MapMultiple below. */
489 #define CCL_IterateMultipleMap 0x10 /* Iterate multiple maps
490 1:ExtendedCOMMNDXXXRRRrrrXXXXX
497 /* Map the code in reg[rrr] by MAPs starting from the Nth (N =
500 MAPs are supplied in the succeeding CCL codes as follows:
502 When CCL program gives this nested structure of map to this command:
505 (MAP-ID121 MAP-ID122 MAP-ID123)
508 (MAP-ID211 (MAP-ID2111) MAP-ID212)
510 the compiled CCL codes has this sequence:
511 CCL_MapMultiple (CCL code of this command)
512 16 (total number of MAPs and SEPARATORs)
530 A value of each SEPARATOR follows this rule:
531 MAP-SET := SEPARATOR [(MAP-ID | MAP-SET)]+
532 SEPARATOR := -(number of MAP-IDs and SEPARATORs in the MAP-SET)
534 (*)....Nest level of MAP-SET must not be over than MAX_MAP_SET_LEVEL.
536 When some map fails to map (i.e. it doesn't have a value for
537 reg[rrr]), the mapping is treated as identity.
539 The mapping is iterated for all maps in each map set (set of maps
540 separated by SEPARATOR) except in the case that lambda is
541 encountered. More precisely, the mapping proceeds as below:
543 At first, VAL0 is set to reg[rrr], and it is translated by the
544 first map to VAL1. Then, VAL1 is translated by the next map to
545 VAL2. This mapping is iterated until the last map is used. The
546 result of the mapping is the last value of VAL?. When the mapping
547 process reached to the end of the map set, it moves to the next
548 map set. If the next does not exit, the mapping process terminates,
549 and regard the last value as a result.
551 But, when VALm is mapped to VALn and VALn is not a number, the
552 mapping proceed as below:
554 If VALn is nil, the lastest map is ignored and the mapping of VALm
555 proceed to the next map.
557 In VALn is t, VALm is reverted to reg[rrr] and the mapping of VALm
558 proceed to the next map.
560 If VALn is lambda, move to the next map set like reaching to the
561 end of the current map set.
563 If VALn is a symbol, call the CCL program refered by it.
564 Then, use reg[rrr] as a mapped value except for -1, -2 and -3.
565 Such special values are regarded as nil, t, and lambda respectively.
567 Each map is a Lisp vector of the following format (a) or (b):
568 (a)......[STARTPOINT VAL1 VAL2 ...]
569 (b)......[t VAL STARTPOINT ENDPOINT],
571 STARTPOINT is an offset to be used for indexing a map,
572 ENDPOINT is a maximum index number of a map,
573 VAL and VALn is a number, nil, t, or lambda.
575 Valid index range of a map of type (a) is:
576 STARTPOINT <= index < STARTPOINT + map_size - 1
577 Valid index range of a map of type (b) is:
578 STARTPOINT <= index < ENDPOINT */
580 #define CCL_MapMultiple 0x11 /* Mapping by multiple code conversion maps
581 1:ExtendedCOMMNDXXXRRRrrrXXXXX
593 #define MAX_MAP_SET_LEVEL 30
601 static tr_stack mapping_stack
[MAX_MAP_SET_LEVEL
];
602 static tr_stack
*mapping_stack_pointer
;
604 /* If this variable is non-zero, it indicates the stack_idx
605 of immediately called by CCL_MapMultiple. */
606 static int stack_idx_of_map_multiple
;
608 #define PUSH_MAPPING_STACK(restlen, orig) \
611 mapping_stack_pointer->rest_length = (restlen); \
612 mapping_stack_pointer->orig_val = (orig); \
613 mapping_stack_pointer++; \
617 #define POP_MAPPING_STACK(restlen, orig) \
620 mapping_stack_pointer--; \
621 (restlen) = mapping_stack_pointer->rest_length; \
622 (orig) = mapping_stack_pointer->orig_val; \
626 #define CCL_CALL_FOR_MAP_INSTRUCTION(symbol, ret_ic) \
629 struct ccl_program called_ccl; \
630 if (stack_idx >= 256 \
631 || (setup_ccl_program (&called_ccl, (symbol)) != 0)) \
635 ccl_prog = ccl_prog_stack_struct[0].ccl_prog; \
636 ic = ccl_prog_stack_struct[0].ic; \
637 eof_ic = ccl_prog_stack_struct[0].eof_ic; \
641 ccl_prog_stack_struct[stack_idx].ccl_prog = ccl_prog; \
642 ccl_prog_stack_struct[stack_idx].ic = (ret_ic); \
643 ccl_prog_stack_struct[stack_idx].eof_ic = eof_ic; \
645 ccl_prog = called_ccl.prog; \
646 ic = CCL_HEADER_MAIN; \
647 eof_ic = XFASTINT (ccl_prog[CCL_HEADER_EOF]); \
652 #define CCL_MapSingle 0x12 /* Map by single code conversion map
653 1:ExtendedCOMMNDXXXRRRrrrXXXXX
655 ------------------------------
656 Map reg[rrr] by MAP-ID.
657 If some valid mapping is found,
658 set reg[rrr] to the result,
663 #define CCL_LookupIntConstTbl 0x13 /* Lookup multibyte character by
664 integer key. Afterwards R7 set
665 to 1 iff lookup succeeded.
666 1:ExtendedCOMMNDRrrRRRXXXXXXXX
667 2:ARGUMENT(Hash table ID) */
669 #define CCL_LookupCharConstTbl 0x14 /* Lookup integer by multibyte
670 character key. Afterwards R7 set
671 to 1 iff lookup succeeded.
672 1:ExtendedCOMMNDRrrRRRrrrXXXXX
673 2:ARGUMENT(Hash table ID) */
675 /* CCL arithmetic/logical operators. */
676 #define CCL_PLUS 0x00 /* X = Y + Z */
677 #define CCL_MINUS 0x01 /* X = Y - Z */
678 #define CCL_MUL 0x02 /* X = Y * Z */
679 #define CCL_DIV 0x03 /* X = Y / Z */
680 #define CCL_MOD 0x04 /* X = Y % Z */
681 #define CCL_AND 0x05 /* X = Y & Z */
682 #define CCL_OR 0x06 /* X = Y | Z */
683 #define CCL_XOR 0x07 /* X = Y ^ Z */
684 #define CCL_LSH 0x08 /* X = Y << Z */
685 #define CCL_RSH 0x09 /* X = Y >> Z */
686 #define CCL_LSH8 0x0A /* X = (Y << 8) | Z */
687 #define CCL_RSH8 0x0B /* X = Y >> 8, r[7] = Y & 0xFF */
688 #define CCL_DIVMOD 0x0C /* X = Y / Z, r[7] = Y % Z */
689 #define CCL_LS 0x10 /* X = (X < Y) */
690 #define CCL_GT 0x11 /* X = (X > Y) */
691 #define CCL_EQ 0x12 /* X = (X == Y) */
692 #define CCL_LE 0x13 /* X = (X <= Y) */
693 #define CCL_GE 0x14 /* X = (X >= Y) */
694 #define CCL_NE 0x15 /* X = (X != Y) */
696 #define CCL_DECODE_SJIS 0x16 /* X = HIGHER_BYTE (DE-SJIS (Y, Z))
697 r[7] = LOWER_BYTE (DE-SJIS (Y, Z)) */
698 #define CCL_ENCODE_SJIS 0x17 /* X = HIGHER_BYTE (SJIS (Y, Z))
699 r[7] = LOWER_BYTE (SJIS (Y, Z) */
701 /* Terminate CCL program successfully. */
702 #define CCL_SUCCESS \
705 ccl->status = CCL_STAT_SUCCESS; \
710 /* Suspend CCL program because of reading from empty input buffer or
711 writing to full output buffer. When this program is resumed, the
712 same I/O command is executed. */
713 #define CCL_SUSPEND(stat) \
717 ccl->status = stat; \
722 /* Terminate CCL program because of invalid command. Should not occur
723 in the normal case. */
726 #define CCL_INVALID_CMD \
729 ccl->status = CCL_STAT_INVALID_CMD; \
730 goto ccl_error_handler; \
736 #define CCL_INVALID_CMD \
739 ccl_debug_hook (this_ic); \
740 ccl->status = CCL_STAT_INVALID_CMD; \
741 goto ccl_error_handler; \
747 /* Encode one character CH to multibyte form and write to the current
748 output buffer. If CH is less than 256, CH is written as is. */
749 #define CCL_WRITE_CHAR(ch) \
753 else if (dst < dst_end) \
756 CCL_SUSPEND (CCL_STAT_SUSPEND_BY_DST); \
759 /* Write a string at ccl_prog[IC] of length LEN to the current output
761 #define CCL_WRITE_STRING(len) \
766 else if (dst + len <= dst_end) \
767 for (i = 0; i < len; i++) \
768 *dst++ = ((XFASTINT (ccl_prog[ic + (i / 3)])) \
769 >> ((2 - (i % 3)) * 8)) & 0xFF; \
771 CCL_SUSPEND (CCL_STAT_SUSPEND_BY_DST); \
774 /* Read one byte from the current input buffer into Rth register. */
775 #define CCL_READ_CHAR(r) \
779 else if (src < src_end) \
781 else if (ccl->last_block) \
788 CCL_SUSPEND (CCL_STAT_SUSPEND_BY_SRC); \
791 /* Decode CODE by a charset whose id is ID. If ID is 0, return CODE
792 as is for backward compatibility. Assume that we can use the
793 variable `charset'. */
795 #define CCL_DECODE_CHAR(id, code) \
796 ((id) == 0 ? (code) \
797 : (charset = CHARSET_FROM_ID ((id)), DECODE_CHAR (charset, (code))))
799 /* Encode character C by some of charsets in CHARSET_LIST. Set ID to
800 the id of the used charset, ENCODED to the resulf of encoding.
801 Assume that we can use the variable `charset'. */
803 #define CCL_ENCODE_CHAR(c, charset_list, id, encoded) \
807 charset = char_charset ((c), (charset_list), &code); \
808 if (! charset && ! NILP (charset_list)) \
809 charset = char_charset ((c), Qnil, &code); \
812 (id) = CHARSET_ID (charset); \
817 /* Execute CCL code on characters at SOURCE (length SRC_SIZE). The
818 resulting text goes to a place pointed by DESTINATION, the length
819 of which should not exceed DST_SIZE. As a side effect, how many
820 characters are consumed and produced are recorded in CCL->consumed
821 and CCL->produced, and the contents of CCL registers are updated.
822 If SOURCE or DESTINATION is NULL, only operations on registers are
826 #define CCL_DEBUG_BACKTRACE_LEN 256
827 int ccl_backtrace_table
[CCL_DEBUG_BACKTRACE_LEN
];
828 int ccl_backtrace_idx
;
831 ccl_debug_hook (int ic
)
838 struct ccl_prog_stack
840 Lisp_Object
*ccl_prog
; /* Pointer to an array of CCL code. */
841 int ic
; /* Instruction Counter. */
842 int eof_ic
; /* Instruction Counter to jump on EOF. */
845 /* For the moment, we only support depth 256 of stack. */
846 static struct ccl_prog_stack ccl_prog_stack_struct
[256];
849 ccl_driver (ccl
, source
, destination
, src_size
, dst_size
, charset_list
)
850 struct ccl_program
*ccl
;
851 int *source
, *destination
;
852 int src_size
, dst_size
;
853 Lisp_Object charset_list
;
855 register int *reg
= ccl
->reg
;
856 register int ic
= ccl
->ic
;
857 register int code
= 0, field1
, field2
;
858 register Lisp_Object
*ccl_prog
= ccl
->prog
;
859 int *src
= source
, *src_end
= src
+ src_size
;
860 int *dst
= destination
, *dst_end
= dst
+ dst_size
;
863 int stack_idx
= ccl
->stack_idx
;
864 /* Instruction counter of the current CCL code. */
866 struct charset
*charset
;
867 int eof_ic
= ccl
->eof_ic
;
871 ic
= CCL_HEADER_MAIN
;
873 if (ccl
->buf_magnification
== 0) /* We can't read/produce any bytes. */
876 /* Set mapping stack pointer. */
877 mapping_stack_pointer
= mapping_stack
;
880 ccl_backtrace_idx
= 0;
887 ccl_backtrace_table
[ccl_backtrace_idx
++] = ic
;
888 if (ccl_backtrace_idx
>= CCL_DEBUG_BACKTRACE_LEN
)
889 ccl_backtrace_idx
= 0;
890 ccl_backtrace_table
[ccl_backtrace_idx
] = 0;
893 if (!NILP (Vquit_flag
) && NILP (Vinhibit_quit
))
895 /* We can't just signal Qquit, instead break the loop as if
896 the whole data is processed. Don't reset Vquit_flag, it
897 must be handled later at a safer place. */
899 src
= source
+ src_size
;
900 ccl
->status
= CCL_STAT_QUIT
;
905 code
= XINT (ccl_prog
[ic
]); ic
++;
907 field2
= (code
& 0xFF) >> 5;
910 #define RRR (field1 & 7)
911 #define Rrr ((field1 >> 3) & 7)
913 #define EXCMD (field1 >> 6)
917 case CCL_SetRegister
: /* 00000000000000000RRRrrrXXXXX */
921 case CCL_SetShortConst
: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
925 case CCL_SetConst
: /* 00000000000000000000rrrXXXXX */
926 reg
[rrr
] = XINT (ccl_prog
[ic
]);
930 case CCL_SetArray
: /* CCCCCCCCCCCCCCCCCCCCRRRrrrXXXXX */
933 if ((unsigned int) i
< j
)
934 reg
[rrr
] = XINT (ccl_prog
[ic
+ i
]);
938 case CCL_Jump
: /* A--D--D--R--E--S--S-000XXXXX */
942 case CCL_JumpCond
: /* A--D--D--R--E--S--S-rrrXXXXX */
947 case CCL_WriteRegisterJump
: /* A--D--D--R--E--S--S-rrrXXXXX */
953 case CCL_WriteRegisterReadJump
: /* A--D--D--R--E--S--S-rrrXXXXX */
957 CCL_READ_CHAR (reg
[rrr
]);
961 case CCL_WriteConstJump
: /* A--D--D--R--E--S--S-000XXXXX */
962 i
= XINT (ccl_prog
[ic
]);
967 case CCL_WriteConstReadJump
: /* A--D--D--R--E--S--S-rrrXXXXX */
968 i
= XINT (ccl_prog
[ic
]);
971 CCL_READ_CHAR (reg
[rrr
]);
975 case CCL_WriteStringJump
: /* A--D--D--R--E--S--S-000XXXXX */
976 j
= XINT (ccl_prog
[ic
]);
978 CCL_WRITE_STRING (j
);
982 case CCL_WriteArrayReadJump
: /* A--D--D--R--E--S--S-rrrXXXXX */
984 j
= XINT (ccl_prog
[ic
]);
985 if ((unsigned int) i
< j
)
987 i
= XINT (ccl_prog
[ic
+ 1 + i
]);
991 CCL_READ_CHAR (reg
[rrr
]);
992 ic
+= ADDR
- (j
+ 2);
995 case CCL_ReadJump
: /* A--D--D--R--E--S--S-rrrYYYYY */
996 CCL_READ_CHAR (reg
[rrr
]);
1000 case CCL_ReadBranch
: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
1001 CCL_READ_CHAR (reg
[rrr
]);
1002 /* fall through ... */
1003 case CCL_Branch
: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
1004 if ((unsigned int) reg
[rrr
] < field1
)
1005 ic
+= XINT (ccl_prog
[ic
+ reg
[rrr
]]);
1007 ic
+= XINT (ccl_prog
[ic
+ field1
]);
1010 case CCL_ReadRegister
: /* CCCCCCCCCCCCCCCCCCCCrrXXXXX */
1013 CCL_READ_CHAR (reg
[rrr
]);
1015 code
= XINT (ccl_prog
[ic
]); ic
++;
1017 field2
= (code
& 0xFF) >> 5;
1021 case CCL_WriteExprConst
: /* 1:00000OPERATION000RRR000XXXXX */
1024 j
= XINT (ccl_prog
[ic
]);
1026 jump_address
= ic
+ 1;
1029 case CCL_WriteRegister
: /* CCCCCCCCCCCCCCCCCCCrrrXXXXX */
1035 code
= XINT (ccl_prog
[ic
]); ic
++;
1037 field2
= (code
& 0xFF) >> 5;
1041 case CCL_WriteExprRegister
: /* 1:00000OPERATIONRrrRRR000XXXXX */
1049 case CCL_Call
: /* 1:CCCCCCCCCCCCCCCCCCCCFFFXXXXX */
1054 /* If FFF is nonzero, the CCL program ID is in the
1058 prog_id
= XINT (ccl_prog
[ic
]);
1064 if (stack_idx
>= 256
1066 || prog_id
>= ASIZE (Vccl_program_table
)
1067 || (slot
= AREF (Vccl_program_table
, prog_id
), !VECTORP (slot
))
1068 || !VECTORP (AREF (slot
, 1)))
1072 ccl_prog
= ccl_prog_stack_struct
[0].ccl_prog
;
1073 ic
= ccl_prog_stack_struct
[0].ic
;
1074 eof_ic
= ccl_prog_stack_struct
[0].eof_ic
;
1079 ccl_prog_stack_struct
[stack_idx
].ccl_prog
= ccl_prog
;
1080 ccl_prog_stack_struct
[stack_idx
].ic
= ic
;
1081 ccl_prog_stack_struct
[stack_idx
].eof_ic
= eof_ic
;
1083 ccl_prog
= XVECTOR (AREF (slot
, 1))->contents
;
1084 ic
= CCL_HEADER_MAIN
;
1085 eof_ic
= XFASTINT (ccl_prog
[CCL_HEADER_EOF
]);
1089 case CCL_WriteConstString
: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
1091 CCL_WRITE_CHAR (field1
);
1094 CCL_WRITE_STRING (field1
);
1095 ic
+= (field1
+ 2) / 3;
1099 case CCL_WriteArray
: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
1101 if ((unsigned int) i
< field1
)
1103 j
= XINT (ccl_prog
[ic
+ i
]);
1109 case CCL_End
: /* 0000000000000000000000XXXXX */
1113 ccl_prog
= ccl_prog_stack_struct
[stack_idx
].ccl_prog
;
1114 ic
= ccl_prog_stack_struct
[stack_idx
].ic
;
1115 eof_ic
= ccl_prog_stack_struct
[stack_idx
].eof_ic
;
1122 /* ccl->ic should points to this command code again to
1123 suppress further processing. */
1127 case CCL_ExprSelfConst
: /* 00000OPERATION000000rrrXXXXX */
1128 i
= XINT (ccl_prog
[ic
]);
1133 case CCL_ExprSelfReg
: /* 00000OPERATION000RRRrrrXXXXX */
1140 case CCL_PLUS
: reg
[rrr
] += i
; break;
1141 case CCL_MINUS
: reg
[rrr
] -= i
; break;
1142 case CCL_MUL
: reg
[rrr
] *= i
; break;
1143 case CCL_DIV
: reg
[rrr
] /= i
; break;
1144 case CCL_MOD
: reg
[rrr
] %= i
; break;
1145 case CCL_AND
: reg
[rrr
] &= i
; break;
1146 case CCL_OR
: reg
[rrr
] |= i
; break;
1147 case CCL_XOR
: reg
[rrr
] ^= i
; break;
1148 case CCL_LSH
: reg
[rrr
] <<= i
; break;
1149 case CCL_RSH
: reg
[rrr
] >>= i
; break;
1150 case CCL_LSH8
: reg
[rrr
] <<= 8; reg
[rrr
] |= i
; break;
1151 case CCL_RSH8
: reg
[7] = reg
[rrr
] & 0xFF; reg
[rrr
] >>= 8; break;
1152 case CCL_DIVMOD
: reg
[7] = reg
[rrr
] % i
; reg
[rrr
] /= i
; break;
1153 case CCL_LS
: reg
[rrr
] = reg
[rrr
] < i
; break;
1154 case CCL_GT
: reg
[rrr
] = reg
[rrr
] > i
; break;
1155 case CCL_EQ
: reg
[rrr
] = reg
[rrr
] == i
; break;
1156 case CCL_LE
: reg
[rrr
] = reg
[rrr
] <= i
; break;
1157 case CCL_GE
: reg
[rrr
] = reg
[rrr
] >= i
; break;
1158 case CCL_NE
: reg
[rrr
] = reg
[rrr
] != i
; break;
1159 default: CCL_INVALID_CMD
;
1163 case CCL_SetExprConst
: /* 00000OPERATION000RRRrrrXXXXX */
1165 j
= XINT (ccl_prog
[ic
]);
1167 jump_address
= ++ic
;
1170 case CCL_SetExprReg
: /* 00000OPERATIONRrrRRRrrrXXXXX */
1177 case CCL_ReadJumpCondExprConst
: /* A--D--D--R--E--S--S-rrrXXXXX */
1178 CCL_READ_CHAR (reg
[rrr
]);
1179 case CCL_JumpCondExprConst
: /* A--D--D--R--E--S--S-rrrXXXXX */
1181 op
= XINT (ccl_prog
[ic
]);
1182 jump_address
= ic
++ + ADDR
;
1183 j
= XINT (ccl_prog
[ic
]);
1188 case CCL_ReadJumpCondExprReg
: /* A--D--D--R--E--S--S-rrrXXXXX */
1189 CCL_READ_CHAR (reg
[rrr
]);
1190 case CCL_JumpCondExprReg
:
1192 op
= XINT (ccl_prog
[ic
]);
1193 jump_address
= ic
++ + ADDR
;
1194 j
= reg
[XINT (ccl_prog
[ic
])];
1201 case CCL_PLUS
: reg
[rrr
] = i
+ j
; break;
1202 case CCL_MINUS
: reg
[rrr
] = i
- j
; break;
1203 case CCL_MUL
: reg
[rrr
] = i
* j
; break;
1204 case CCL_DIV
: reg
[rrr
] = i
/ j
; break;
1205 case CCL_MOD
: reg
[rrr
] = i
% j
; break;
1206 case CCL_AND
: reg
[rrr
] = i
& j
; break;
1207 case CCL_OR
: reg
[rrr
] = i
| j
; break;
1208 case CCL_XOR
: reg
[rrr
] = i
^ j
;; break;
1209 case CCL_LSH
: reg
[rrr
] = i
<< j
; break;
1210 case CCL_RSH
: reg
[rrr
] = i
>> j
; break;
1211 case CCL_LSH8
: reg
[rrr
] = (i
<< 8) | j
; break;
1212 case CCL_RSH8
: reg
[rrr
] = i
>> 8; reg
[7] = i
& 0xFF; break;
1213 case CCL_DIVMOD
: reg
[rrr
] = i
/ j
; reg
[7] = i
% j
; break;
1214 case CCL_LS
: reg
[rrr
] = i
< j
; break;
1215 case CCL_GT
: reg
[rrr
] = i
> j
; break;
1216 case CCL_EQ
: reg
[rrr
] = i
== j
; break;
1217 case CCL_LE
: reg
[rrr
] = i
<= j
; break;
1218 case CCL_GE
: reg
[rrr
] = i
>= j
; break;
1219 case CCL_NE
: reg
[rrr
] = i
!= j
; break;
1220 case CCL_DECODE_SJIS
:
1228 case CCL_ENCODE_SJIS
:
1236 default: CCL_INVALID_CMD
;
1239 if (code
== CCL_WriteExprConst
|| code
== CCL_WriteExprRegister
)
1252 case CCL_ReadMultibyteChar2
:
1256 CCL_ENCODE_CHAR (i
, charset_list
, reg
[RRR
], reg
[rrr
]);
1259 case CCL_WriteMultibyteChar2
:
1262 i
= CCL_DECODE_CHAR (reg
[RRR
], reg
[rrr
]);
1266 case CCL_TranslateCharacter
:
1267 i
= CCL_DECODE_CHAR (reg
[RRR
], reg
[rrr
]);
1268 op
= translate_char (GET_TRANSLATION_TABLE (reg
[Rrr
]), i
);
1269 CCL_ENCODE_CHAR (op
, charset_list
, reg
[RRR
], reg
[rrr
]);
1272 case CCL_TranslateCharacterConstTbl
:
1273 op
= XINT (ccl_prog
[ic
]); /* table */
1275 i
= CCL_DECODE_CHAR (reg
[RRR
], reg
[rrr
]);
1276 op
= translate_char (GET_TRANSLATION_TABLE (op
), i
);
1277 CCL_ENCODE_CHAR (op
, charset_list
, reg
[RRR
], reg
[rrr
]);
1280 case CCL_LookupIntConstTbl
:
1281 op
= XINT (ccl_prog
[ic
]); /* table */
1284 struct Lisp_Hash_Table
*h
= GET_HASH_TABLE (op
);
1286 op
= hash_lookup (h
, make_number (reg
[RRR
]), NULL
);
1290 opl
= HASH_VALUE (h
, op
);
1291 if (! CHARACTERP (opl
))
1293 reg
[RRR
] = charset_unicode
;
1295 reg
[7] = 1; /* r7 true for success */
1302 case CCL_LookupCharConstTbl
:
1303 op
= XINT (ccl_prog
[ic
]); /* table */
1305 i
= CCL_DECODE_CHAR (reg
[RRR
], reg
[rrr
]);
1307 struct Lisp_Hash_Table
*h
= GET_HASH_TABLE (op
);
1309 op
= hash_lookup (h
, make_number (i
), NULL
);
1313 opl
= HASH_VALUE (h
, op
);
1314 if (!INTEGERP (opl
))
1316 reg
[RRR
] = XINT (opl
);
1317 reg
[7] = 1; /* r7 true for success */
1324 case CCL_IterateMultipleMap
:
1326 Lisp_Object map
, content
, attrib
, value
;
1327 int point
, size
, fin_ic
;
1329 j
= XINT (ccl_prog
[ic
++]); /* number of maps. */
1332 if ((j
> reg
[RRR
]) && (j
>= 0))
1347 size
= ASIZE (Vcode_conversion_map_vector
);
1348 point
= XINT (ccl_prog
[ic
++]);
1349 if (point
>= size
) continue;
1350 map
= AREF (Vcode_conversion_map_vector
, point
);
1352 /* Check map varidity. */
1353 if (!CONSP (map
)) continue;
1355 if (!VECTORP (map
)) continue;
1357 if (size
<= 1) continue;
1359 content
= AREF (map
, 0);
1362 [STARTPOINT VAL1 VAL2 ...] or
1363 [t ELELMENT STARTPOINT ENDPOINT] */
1364 if (NUMBERP (content
))
1366 point
= XUINT (content
);
1367 point
= op
- point
+ 1;
1368 if (!((point
>= 1) && (point
< size
))) continue;
1369 content
= AREF (map
, point
);
1371 else if (EQ (content
, Qt
))
1373 if (size
!= 4) continue;
1374 if ((op
>= XUINT (AREF (map
, 2)))
1375 && (op
< XUINT (AREF (map
, 3))))
1376 content
= AREF (map
, 1);
1385 else if (NUMBERP (content
))
1388 reg
[rrr
] = XINT(content
);
1391 else if (EQ (content
, Qt
) || EQ (content
, Qlambda
))
1396 else if (CONSP (content
))
1398 attrib
= XCAR (content
);
1399 value
= XCDR (content
);
1400 if (!NUMBERP (attrib
) || !NUMBERP (value
))
1403 reg
[rrr
] = XUINT (value
);
1406 else if (SYMBOLP (content
))
1407 CCL_CALL_FOR_MAP_INSTRUCTION (content
, fin_ic
);
1417 case CCL_MapMultiple
:
1419 Lisp_Object map
, content
, attrib
, value
;
1420 int point
, size
, map_vector_size
;
1421 int map_set_rest_length
, fin_ic
;
1422 int current_ic
= this_ic
;
1424 /* inhibit recursive call on MapMultiple. */
1425 if (stack_idx_of_map_multiple
> 0)
1427 if (stack_idx_of_map_multiple
<= stack_idx
)
1429 stack_idx_of_map_multiple
= 0;
1430 mapping_stack_pointer
= mapping_stack
;
1435 mapping_stack_pointer
= mapping_stack
;
1436 stack_idx_of_map_multiple
= 0;
1438 map_set_rest_length
=
1439 XINT (ccl_prog
[ic
++]); /* number of maps and separators. */
1440 fin_ic
= ic
+ map_set_rest_length
;
1443 if ((map_set_rest_length
> reg
[RRR
]) && (reg
[RRR
] >= 0))
1447 map_set_rest_length
-= i
;
1453 mapping_stack_pointer
= mapping_stack
;
1457 if (mapping_stack_pointer
<= (mapping_stack
+ 1))
1459 /* Set up initial state. */
1460 mapping_stack_pointer
= mapping_stack
;
1461 PUSH_MAPPING_STACK (0, op
);
1466 /* Recover after calling other ccl program. */
1469 POP_MAPPING_STACK (map_set_rest_length
, orig_op
);
1470 POP_MAPPING_STACK (map_set_rest_length
, reg
[rrr
]);
1474 /* Regard it as Qnil. */
1478 map_set_rest_length
--;
1481 /* Regard it as Qt. */
1485 map_set_rest_length
--;
1488 /* Regard it as Qlambda. */
1490 i
+= map_set_rest_length
;
1491 ic
+= map_set_rest_length
;
1492 map_set_rest_length
= 0;
1495 /* Regard it as normal mapping. */
1496 i
+= map_set_rest_length
;
1497 ic
+= map_set_rest_length
;
1498 POP_MAPPING_STACK (map_set_rest_length
, reg
[rrr
]);
1502 map_vector_size
= ASIZE (Vcode_conversion_map_vector
);
1505 for (;map_set_rest_length
> 0;i
++, ic
++, map_set_rest_length
--)
1507 point
= XINT(ccl_prog
[ic
]);
1510 /* +1 is for including separator. */
1512 if (mapping_stack_pointer
1513 >= &mapping_stack
[MAX_MAP_SET_LEVEL
])
1515 PUSH_MAPPING_STACK (map_set_rest_length
- point
,
1517 map_set_rest_length
= point
;
1522 if (point
>= map_vector_size
) continue;
1523 map
= AREF (Vcode_conversion_map_vector
, point
);
1525 /* Check map varidity. */
1526 if (!CONSP (map
)) continue;
1528 if (!VECTORP (map
)) continue;
1530 if (size
<= 1) continue;
1532 content
= AREF (map
, 0);
1535 [STARTPOINT VAL1 VAL2 ...] or
1536 [t ELEMENT STARTPOINT ENDPOINT] */
1537 if (NUMBERP (content
))
1539 point
= XUINT (content
);
1540 point
= op
- point
+ 1;
1541 if (!((point
>= 1) && (point
< size
))) continue;
1542 content
= AREF (map
, point
);
1544 else if (EQ (content
, Qt
))
1546 if (size
!= 4) continue;
1547 if ((op
>= XUINT (AREF (map
, 2))) &&
1548 (op
< XUINT (AREF (map
, 3))))
1549 content
= AREF (map
, 1);
1560 if (NUMBERP (content
))
1562 op
= XINT (content
);
1563 i
+= map_set_rest_length
- 1;
1564 ic
+= map_set_rest_length
- 1;
1565 POP_MAPPING_STACK (map_set_rest_length
, reg
[rrr
]);
1566 map_set_rest_length
++;
1568 else if (CONSP (content
))
1570 attrib
= XCAR (content
);
1571 value
= XCDR (content
);
1572 if (!NUMBERP (attrib
) || !NUMBERP (value
))
1575 i
+= map_set_rest_length
- 1;
1576 ic
+= map_set_rest_length
- 1;
1577 POP_MAPPING_STACK (map_set_rest_length
, reg
[rrr
]);
1578 map_set_rest_length
++;
1580 else if (EQ (content
, Qt
))
1584 else if (EQ (content
, Qlambda
))
1586 i
+= map_set_rest_length
;
1587 ic
+= map_set_rest_length
;
1590 else if (SYMBOLP (content
))
1592 if (mapping_stack_pointer
1593 >= &mapping_stack
[MAX_MAP_SET_LEVEL
])
1595 PUSH_MAPPING_STACK (map_set_rest_length
, reg
[rrr
]);
1596 PUSH_MAPPING_STACK (map_set_rest_length
, op
);
1597 stack_idx_of_map_multiple
= stack_idx
+ 1;
1598 CCL_CALL_FOR_MAP_INSTRUCTION (content
, current_ic
);
1603 if (mapping_stack_pointer
<= (mapping_stack
+ 1))
1605 POP_MAPPING_STACK (map_set_rest_length
, reg
[rrr
]);
1606 i
+= map_set_rest_length
;
1607 ic
+= map_set_rest_length
;
1608 POP_MAPPING_STACK (map_set_rest_length
, reg
[rrr
]);
1618 Lisp_Object map
, attrib
, value
, content
;
1620 j
= XINT (ccl_prog
[ic
++]); /* map_id */
1622 if (j
>= ASIZE (Vcode_conversion_map_vector
))
1627 map
= AREF (Vcode_conversion_map_vector
, j
);
1640 point
= XUINT (AREF (map
, 0));
1641 point
= op
- point
+ 1;
1644 (!((point
>= 1) && (point
< size
))))
1649 content
= AREF (map
, point
);
1652 else if (NUMBERP (content
))
1653 reg
[rrr
] = XINT (content
);
1654 else if (EQ (content
, Qt
));
1655 else if (CONSP (content
))
1657 attrib
= XCAR (content
);
1658 value
= XCDR (content
);
1659 if (!NUMBERP (attrib
) || !NUMBERP (value
))
1661 reg
[rrr
] = XUINT(value
);
1664 else if (SYMBOLP (content
))
1665 CCL_CALL_FOR_MAP_INSTRUCTION (content
, ic
);
1683 /* The suppress_error member is set when e.g. a CCL-based coding
1684 system is used for terminal output. */
1685 if (!ccl
->suppress_error
&& destination
)
1687 /* We can insert an error message only if DESTINATION is
1688 specified and we still have a room to store the message
1696 switch (ccl
->status
)
1698 case CCL_STAT_INVALID_CMD
:
1699 sprintf(msg
, "\nCCL: Invalid command %x (ccl_code = %x) at %d.",
1700 code
& 0x1F, code
, this_ic
);
1703 int i
= ccl_backtrace_idx
- 1;
1706 msglen
= strlen (msg
);
1707 if (dst
+ msglen
<= (dst_bytes
? dst_end
: src
))
1709 bcopy (msg
, dst
, msglen
);
1713 for (j
= 0; j
< CCL_DEBUG_BACKTRACE_LEN
; j
++, i
--)
1715 if (i
< 0) i
= CCL_DEBUG_BACKTRACE_LEN
- 1;
1716 if (ccl_backtrace_table
[i
] == 0)
1718 sprintf(msg
, " %d", ccl_backtrace_table
[i
]);
1719 msglen
= strlen (msg
);
1720 if (dst
+ msglen
> (dst_bytes
? dst_end
: src
))
1722 bcopy (msg
, dst
, msglen
);
1731 sprintf(msg
, "\nCCL: Quited.");
1735 sprintf(msg
, "\nCCL: Unknown error type (%d)", ccl
->status
);
1738 msglen
= strlen (msg
);
1739 if (dst
+ msglen
<= dst_end
)
1741 for (i
= 0; i
< msglen
; i
++)
1745 if (ccl
->status
== CCL_STAT_INVALID_CMD
)
1747 #if 0 /* If the remaining bytes contain 0x80..0x9F, copying them
1748 results in an invalid multibyte sequence. */
1750 /* Copy the remaining source data. */
1751 int i
= src_end
- src
;
1752 if (dst_bytes
&& (dst_end
- dst
) < i
)
1754 bcopy (src
, dst
, i
);
1758 /* Signal that we've consumed everything. */
1766 ccl
->stack_idx
= stack_idx
;
1767 ccl
->prog
= ccl_prog
;
1768 ccl
->consumed
= src
- source
;
1769 ccl
->produced
= dst
- destination
;
1772 /* Resolve symbols in the specified CCL code (Lisp vector). This
1773 function converts symbols of code conversion maps and character
1774 translation tables embeded in the CCL code into their ID numbers.
1776 The return value is a vector (CCL itself or a new vector in which
1777 all symbols are resolved), Qt if resolving of some symbol failed,
1778 or nil if CCL contains invalid data. */
1781 resolve_symbol_ccl_program (ccl
)
1784 int i
, veclen
, unresolved
= 0;
1785 Lisp_Object result
, contents
, val
;
1788 veclen
= ASIZE (result
);
1790 for (i
= 0; i
< veclen
; i
++)
1792 contents
= AREF (result
, i
);
1793 if (INTEGERP (contents
))
1795 else if (CONSP (contents
)
1796 && SYMBOLP (XCAR (contents
))
1797 && SYMBOLP (XCDR (contents
)))
1799 /* This is the new style for embedding symbols. The form is
1800 (SYMBOL . PROPERTY). (get SYMBOL PROPERTY) should give
1803 if (EQ (result
, ccl
))
1804 result
= Fcopy_sequence (ccl
);
1806 val
= Fget (XCAR (contents
), XCDR (contents
));
1808 AREF (result
, i
) = val
;
1813 else if (SYMBOLP (contents
))
1815 /* This is the old style for embedding symbols. This style
1816 may lead to a bug if, for instance, a translation table
1817 and a code conversion map have the same name. */
1818 if (EQ (result
, ccl
))
1819 result
= Fcopy_sequence (ccl
);
1821 val
= Fget (contents
, Qtranslation_table_id
);
1823 AREF (result
, i
) = val
;
1826 val
= Fget (contents
, Qcode_conversion_map_id
);
1828 AREF (result
, i
) = val
;
1831 val
= Fget (contents
, Qccl_program_idx
);
1833 AREF (result
, i
) = val
;
1843 return (unresolved
? Qt
: result
);
1846 /* Return the compiled code (vector) of CCL program CCL_PROG.
1847 CCL_PROG is a name (symbol) of the program or already compiled
1848 code. If necessary, resolve symbols in the compiled code to index
1849 numbers. If we failed to get the compiled code or to resolve
1850 symbols, return Qnil. */
1853 ccl_get_compiled_code (ccl_prog
)
1854 Lisp_Object ccl_prog
;
1856 Lisp_Object val
, slot
;
1858 if (VECTORP (ccl_prog
))
1860 val
= resolve_symbol_ccl_program (ccl_prog
);
1861 return (VECTORP (val
) ? val
: Qnil
);
1863 if (!SYMBOLP (ccl_prog
))
1866 val
= Fget (ccl_prog
, Qccl_program_idx
);
1868 || XINT (val
) >= ASIZE (Vccl_program_table
))
1870 slot
= AREF (Vccl_program_table
, XINT (val
));
1871 if (! VECTORP (slot
)
1872 || ASIZE (slot
) != 3
1873 || ! VECTORP (AREF (slot
, 1)))
1875 if (NILP (AREF (slot
, 2)))
1877 val
= resolve_symbol_ccl_program (AREF (slot
, 1));
1878 if (! VECTORP (val
))
1880 AREF (slot
, 1) = val
;
1881 AREF (slot
, 2) = Qt
;
1883 return AREF (slot
, 1);
1886 /* Setup fields of the structure pointed by CCL appropriately for the
1887 execution of CCL program CCL_PROG. CCL_PROG is the name (symbol)
1888 of the CCL program or the already compiled code (vector).
1889 Return 0 if we succeed this setup, else return -1.
1891 If CCL_PROG is nil, we just reset the structure pointed by CCL. */
1893 setup_ccl_program (ccl
, ccl_prog
)
1894 struct ccl_program
*ccl
;
1895 Lisp_Object ccl_prog
;
1899 if (! NILP (ccl_prog
))
1901 struct Lisp_Vector
*vp
;
1903 ccl_prog
= ccl_get_compiled_code (ccl_prog
);
1904 if (! VECTORP (ccl_prog
))
1906 vp
= XVECTOR (ccl_prog
);
1907 ccl
->size
= vp
->size
;
1908 ccl
->prog
= vp
->contents
;
1909 ccl
->eof_ic
= XINT (vp
->contents
[CCL_HEADER_EOF
]);
1910 ccl
->buf_magnification
= XINT (vp
->contents
[CCL_HEADER_BUF_MAG
]);
1912 ccl
->ic
= CCL_HEADER_MAIN
;
1913 for (i
= 0; i
< 8; i
++)
1915 ccl
->last_block
= 0;
1916 ccl
->private_state
= 0;
1919 ccl
->suppress_error
= 0;
1920 ccl
->eight_bit_control
= 0;
1924 DEFUN ("ccl-program-p", Fccl_program_p
, Sccl_program_p
, 1, 1, 0,
1925 doc
: /* Return t if OBJECT is a CCL program name or a compiled CCL program code.
1926 See the documentation of `define-ccl-program' for the detail of CCL program. */)
1932 if (VECTORP (object
))
1934 val
= resolve_symbol_ccl_program (object
);
1935 return (VECTORP (val
) ? Qt
: Qnil
);
1937 if (!SYMBOLP (object
))
1940 val
= Fget (object
, Qccl_program_idx
);
1941 return ((! NATNUMP (val
)
1942 || XINT (val
) >= ASIZE (Vccl_program_table
))
1946 DEFUN ("ccl-execute", Fccl_execute
, Sccl_execute
, 2, 2, 0,
1947 doc
: /* Execute CCL-PROGRAM with registers initialized by REGISTERS.
1949 CCL-PROGRAM is a CCL program name (symbol)
1950 or compiled code generated by `ccl-compile' (for backward compatibility.
1951 In the latter case, the execution overhead is bigger than in the former).
1952 No I/O commands should appear in CCL-PROGRAM.
1954 REGISTERS is a vector of [R0 R1 ... R7] where RN is an initial value
1955 for the Nth register.
1957 As side effect, each element of REGISTERS holds the value of
1958 the corresponding register after the execution.
1960 See the documentation of `define-ccl-program' for a definition of CCL
1963 Lisp_Object ccl_prog
, reg
;
1965 struct ccl_program ccl
;
1968 if (setup_ccl_program (&ccl
, ccl_prog
) < 0)
1969 error ("Invalid CCL program");
1972 if (ASIZE (reg
) != 8)
1973 error ("Length of vector REGISTERS is not 8");
1975 for (i
= 0; i
< 8; i
++)
1976 ccl
.reg
[i
] = (INTEGERP (AREF (reg
, i
))
1977 ? XINT (AREF (reg
, i
))
1980 ccl_driver (&ccl
, NULL
, NULL
, 0, 0, Qnil
);
1982 if (ccl
.status
!= CCL_STAT_SUCCESS
)
1983 error ("Error in CCL program at %dth code", ccl
.ic
);
1985 for (i
= 0; i
< 8; i
++)
1986 XSETINT (AREF (reg
, i
), ccl
.reg
[i
]);
1990 DEFUN ("ccl-execute-on-string", Fccl_execute_on_string
, Sccl_execute_on_string
,
1992 doc
: /* Execute CCL-PROGRAM with initial STATUS on STRING.
1994 CCL-PROGRAM is a symbol registered by register-ccl-program,
1995 or a compiled code generated by `ccl-compile' (for backward compatibility,
1996 in this case, the execution is slower).
1998 Read buffer is set to STRING, and write buffer is allocated automatically.
2000 STATUS is a vector of [R0 R1 ... R7 IC], where
2001 R0..R7 are initial values of corresponding registers,
2002 IC is the instruction counter specifying from where to start the program.
2003 If R0..R7 are nil, they are initialized to 0.
2004 If IC is nil, it is initialized to head of the CCL program.
2006 If optional 4th arg CONTINUE is non-nil, keep IC on read operation
2007 when read buffer is exausted, else, IC is always set to the end of
2008 CCL-PROGRAM on exit.
2010 It returns the contents of write buffer as a string,
2011 and as side effect, STATUS is updated.
2012 If the optional 5th arg UNIBYTE-P is non-nil, the returned string
2013 is a unibyte string. By default it is a multibyte string.
2015 See the documentation of `define-ccl-program' for the detail of CCL program. */)
2016 (ccl_prog
, status
, str
, contin
, unibyte_p
)
2017 Lisp_Object ccl_prog
, status
, str
, contin
, unibyte_p
;
2020 struct ccl_program ccl
;
2023 unsigned char *outbuf
, *outp
;
2024 int str_chars
, str_bytes
;
2025 #define CCL_EXECUTE_BUF_SIZE 1024
2026 int source
[CCL_EXECUTE_BUF_SIZE
], destination
[CCL_EXECUTE_BUF_SIZE
];
2027 int consumed_chars
, consumed_bytes
, produced_chars
;
2029 if (setup_ccl_program (&ccl
, ccl_prog
) < 0)
2030 error ("Invalid CCL program");
2032 CHECK_VECTOR (status
);
2033 if (ASIZE (status
) != 9)
2034 error ("Length of vector STATUS is not 9");
2037 str_chars
= SCHARS (str
);
2038 str_bytes
= SBYTES (str
);
2040 for (i
= 0; i
< 8; i
++)
2042 if (NILP (AREF (status
, i
)))
2043 XSETINT (AREF (status
, i
), 0);
2044 if (INTEGERP (AREF (status
, i
)))
2045 ccl
.reg
[i
] = XINT (AREF (status
, i
));
2047 if (INTEGERP (AREF (status
, i
)))
2049 i
= XFASTINT (AREF (status
, 8));
2050 if (ccl
.ic
< i
&& i
< ccl
.size
)
2054 outbufsize
= (ccl
.buf_magnification
2055 ? str_bytes
* ccl
.buf_magnification
+ 256
2057 outp
= outbuf
= (unsigned char *) xmalloc (outbufsize
);
2059 consumed_chars
= consumed_bytes
= 0;
2063 const unsigned char *p
= SDATA (str
) + consumed_bytes
;
2064 const unsigned char *endp
= SDATA (str
) + str_bytes
;
2068 if (endp
- p
== str_chars
- consumed_chars
)
2069 while (i
< CCL_EXECUTE_BUF_SIZE
&& p
< endp
)
2072 while (i
< CCL_EXECUTE_BUF_SIZE
&& p
< endp
)
2073 source
[i
++] = STRING_CHAR_ADVANCE (p
);
2074 consumed_chars
+= i
;
2075 consumed_bytes
= p
- SDATA (str
);
2077 if (consumed_bytes
== str_bytes
)
2078 ccl
.last_block
= NILP (contin
);
2083 ccl_driver (&ccl
, src
, destination
, src_size
, CCL_EXECUTE_BUF_SIZE
,
2085 produced_chars
+= ccl
.produced
;
2086 if (NILP (unibyte_p
))
2088 if (outp
- outbuf
+ MAX_MULTIBYTE_LENGTH
* ccl
.produced
2091 int offset
= outp
- outbuf
;
2092 outbufsize
+= MAX_MULTIBYTE_LENGTH
* ccl
.produced
;
2093 outbuf
= (unsigned char *) xrealloc (outbuf
, outbufsize
);
2094 outp
= outbuf
+ offset
;
2096 for (i
= 0; i
< ccl
.produced
; i
++)
2097 CHAR_STRING_ADVANCE (destination
[i
], outp
);
2101 if (outp
- outbuf
+ ccl
.produced
> outbufsize
)
2103 int offset
= outp
- outbuf
;
2104 outbufsize
+= ccl
.produced
;
2105 outbuf
= (unsigned char *) xrealloc (outbuf
, outbufsize
);
2106 outp
= outbuf
+ offset
;
2108 for (i
= 0; i
< ccl
.produced
; i
++)
2109 *outp
++ = destination
[i
];
2111 src
+= ccl
.consumed
;
2112 src_size
-= ccl
.consumed
;
2113 if (ccl
.status
!= CCL_STAT_SUSPEND_BY_DST
)
2117 if (ccl
.status
!= CCL_STAT_SUSPEND_BY_SRC
)
2121 if (ccl
.status
!= CCL_STAT_SUCCESS
2122 && ccl
.status
!= CCL_STAT_SUSPEND_BY_SRC
)
2123 error ("Error in CCL program at %dth code", ccl
.ic
);
2125 for (i
= 0; i
< 8; i
++)
2126 XSET (XVECTOR (status
)->contents
[i
], Lisp_Int
, ccl
.reg
[i
]);
2127 XSETINT (XVECTOR (status
)->contents
[8], ccl
.ic
);
2129 if (NILP (unibyte_p
))
2130 val
= make_multibyte_string ((char *) outbuf
, produced_chars
,
2133 val
= make_unibyte_string ((char *) outbuf
, produced_chars
);
2139 DEFUN ("register-ccl-program", Fregister_ccl_program
, Sregister_ccl_program
,
2141 doc
: /* Register CCL program CCL_PROG as NAME in `ccl-program-table'.
2142 CCL_PROG should be a compiled CCL program (vector), or nil.
2143 If it is nil, just reserve NAME as a CCL program name.
2144 Return index number of the registered CCL program. */)
2146 Lisp_Object name
, ccl_prog
;
2148 int len
= ASIZE (Vccl_program_table
);
2150 Lisp_Object resolved
;
2152 CHECK_SYMBOL (name
);
2154 if (!NILP (ccl_prog
))
2156 CHECK_VECTOR (ccl_prog
);
2157 resolved
= resolve_symbol_ccl_program (ccl_prog
);
2158 if (NILP (resolved
))
2159 error ("Error in CCL program");
2160 if (VECTORP (resolved
))
2162 ccl_prog
= resolved
;
2169 for (idx
= 0; idx
< len
; idx
++)
2173 slot
= AREF (Vccl_program_table
, idx
);
2174 if (!VECTORP (slot
))
2175 /* This is the first unsed slot. Register NAME here. */
2178 if (EQ (name
, AREF (slot
, 0)))
2180 /* Update this slot. */
2181 AREF (slot
, 1) = ccl_prog
;
2182 AREF (slot
, 2) = resolved
;
2183 return make_number (idx
);
2189 /* Extend the table. */
2190 Lisp_Object new_table
;
2193 new_table
= Fmake_vector (make_number (len
* 2), Qnil
);
2194 for (j
= 0; j
< len
; j
++)
2196 = AREF (Vccl_program_table
, j
);
2197 Vccl_program_table
= new_table
;
2203 elt
= Fmake_vector (make_number (3), Qnil
);
2204 AREF (elt
, 0) = name
;
2205 AREF (elt
, 1) = ccl_prog
;
2206 AREF (elt
, 2) = resolved
;
2207 AREF (Vccl_program_table
, idx
) = elt
;
2210 Fput (name
, Qccl_program_idx
, make_number (idx
));
2211 return make_number (idx
);
2214 /* Register code conversion map.
2215 A code conversion map consists of numbers, Qt, Qnil, and Qlambda.
2216 The first element is the start code point.
2217 The other elements are mapped numbers.
2218 Symbol t means to map to an original number before mapping.
2219 Symbol nil means that the corresponding element is empty.
2220 Symbol lambda means to terminate mapping here.
2223 DEFUN ("register-code-conversion-map", Fregister_code_conversion_map
,
2224 Sregister_code_conversion_map
,
2226 doc
: /* Register SYMBOL as code conversion map MAP.
2227 Return index number of the registered map. */)
2229 Lisp_Object symbol
, map
;
2231 int len
= ASIZE (Vcode_conversion_map_vector
);
2235 CHECK_SYMBOL (symbol
);
2238 for (i
= 0; i
< len
; i
++)
2240 Lisp_Object slot
= AREF (Vcode_conversion_map_vector
, i
);
2245 if (EQ (symbol
, XCAR (slot
)))
2247 index
= make_number (i
);
2248 XSETCDR (slot
, map
);
2249 Fput (symbol
, Qcode_conversion_map
, map
);
2250 Fput (symbol
, Qcode_conversion_map_id
, index
);
2257 Lisp_Object new_vector
= Fmake_vector (make_number (len
* 2), Qnil
);
2260 for (j
= 0; j
< len
; j
++)
2261 AREF (new_vector
, j
)
2262 = AREF (Vcode_conversion_map_vector
, j
);
2263 Vcode_conversion_map_vector
= new_vector
;
2266 index
= make_number (i
);
2267 Fput (symbol
, Qcode_conversion_map
, map
);
2268 Fput (symbol
, Qcode_conversion_map_id
, index
);
2269 AREF (Vcode_conversion_map_vector
, i
) = Fcons (symbol
, map
);
2277 staticpro (&Vccl_program_table
);
2278 Vccl_program_table
= Fmake_vector (make_number (32), Qnil
);
2280 Qccl
= intern ("ccl");
2283 Qcclp
= intern ("cclp");
2286 Qccl_program
= intern ("ccl-program");
2287 staticpro (&Qccl_program
);
2289 Qccl_program_idx
= intern ("ccl-program-idx");
2290 staticpro (&Qccl_program_idx
);
2292 Qcode_conversion_map
= intern ("code-conversion-map");
2293 staticpro (&Qcode_conversion_map
);
2295 Qcode_conversion_map_id
= intern ("code-conversion-map-id");
2296 staticpro (&Qcode_conversion_map_id
);
2298 DEFVAR_LISP ("code-conversion-map-vector", &Vcode_conversion_map_vector
,
2299 doc
: /* Vector of code conversion maps. */);
2300 Vcode_conversion_map_vector
= Fmake_vector (make_number (16), Qnil
);
2302 DEFVAR_LISP ("font-ccl-encoder-alist", &Vfont_ccl_encoder_alist
,
2303 doc
: /* Alist of fontname patterns vs corresponding CCL program.
2304 Each element looks like (REGEXP . CCL-CODE),
2305 where CCL-CODE is a compiled CCL program.
2306 When a font whose name matches REGEXP is used for displaying a character,
2307 CCL-CODE is executed to calculate the code point in the font
2308 from the charset number and position code(s) of the character which are set
2309 in CCL registers R0, R1, and R2 before the execution.
2310 The code point in the font is set in CCL registers R1 and R2
2311 when the execution terminated.
2312 If the font is single-byte font, the register R2 is not used. */);
2313 Vfont_ccl_encoder_alist
= Qnil
;
2315 DEFVAR_LISP ("translation-hash-table-vector", &Vtranslation_hash_table_vector
,
2316 doc
: /* Vector containing all translation hash tables ever defined.
2317 Comprises pairs (SYMBOL . TABLE) where SYMBOL and TABLE were set up by calls
2318 to `define-translation-hash-table'. The vector is indexed by the table id
2320 Vtranslation_hash_table_vector
= Qnil
;
2322 defsubr (&Sccl_program_p
);
2323 defsubr (&Sccl_execute
);
2324 defsubr (&Sccl_execute_on_string
);
2325 defsubr (&Sregister_ccl_program
);
2326 defsubr (&Sregister_code_conversion_map
);
2329 /* arch-tag: bb9a37be-68ce-4576-8d3d-15d750e4a860
2330 (do not change this comment) */