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; \
640 ccl_prog_stack_struct[stack_idx].ccl_prog = ccl_prog; \
641 ccl_prog_stack_struct[stack_idx].ic = (ret_ic); \
643 ccl_prog = called_ccl.prog; \
644 ic = CCL_HEADER_MAIN; \
649 #define CCL_MapSingle 0x12 /* Map by single code conversion map
650 1:ExtendedCOMMNDXXXRRRrrrXXXXX
652 ------------------------------
653 Map reg[rrr] by MAP-ID.
654 If some valid mapping is found,
655 set reg[rrr] to the result,
660 #define CCL_LookupIntConstTbl 0x13 /* Lookup multibyte character by
661 integer key. Afterwards R7 set
662 to 1 iff lookup succeeded.
663 1:ExtendedCOMMNDRrrRRRXXXXXXXX
664 2:ARGUMENT(Hash table ID) */
666 #define CCL_LookupCharConstTbl 0x14 /* Lookup integer by multibyte
667 character key. Afterwards R7 set
668 to 1 iff lookup succeeded.
669 1:ExtendedCOMMNDRrrRRRrrrXXXXX
670 2:ARGUMENT(Hash table ID) */
672 /* CCL arithmetic/logical operators. */
673 #define CCL_PLUS 0x00 /* X = Y + Z */
674 #define CCL_MINUS 0x01 /* X = Y - Z */
675 #define CCL_MUL 0x02 /* X = Y * Z */
676 #define CCL_DIV 0x03 /* X = Y / Z */
677 #define CCL_MOD 0x04 /* X = Y % Z */
678 #define CCL_AND 0x05 /* X = Y & Z */
679 #define CCL_OR 0x06 /* X = Y | Z */
680 #define CCL_XOR 0x07 /* X = Y ^ Z */
681 #define CCL_LSH 0x08 /* X = Y << Z */
682 #define CCL_RSH 0x09 /* X = Y >> Z */
683 #define CCL_LSH8 0x0A /* X = (Y << 8) | Z */
684 #define CCL_RSH8 0x0B /* X = Y >> 8, r[7] = Y & 0xFF */
685 #define CCL_DIVMOD 0x0C /* X = Y / Z, r[7] = Y % Z */
686 #define CCL_LS 0x10 /* X = (X < Y) */
687 #define CCL_GT 0x11 /* X = (X > Y) */
688 #define CCL_EQ 0x12 /* X = (X == Y) */
689 #define CCL_LE 0x13 /* X = (X <= Y) */
690 #define CCL_GE 0x14 /* X = (X >= Y) */
691 #define CCL_NE 0x15 /* X = (X != Y) */
693 #define CCL_DECODE_SJIS 0x16 /* X = HIGHER_BYTE (DE-SJIS (Y, Z))
694 r[7] = LOWER_BYTE (DE-SJIS (Y, Z)) */
695 #define CCL_ENCODE_SJIS 0x17 /* X = HIGHER_BYTE (SJIS (Y, Z))
696 r[7] = LOWER_BYTE (SJIS (Y, Z) */
698 /* Terminate CCL program successfully. */
699 #define CCL_SUCCESS \
702 ccl->status = CCL_STAT_SUCCESS; \
707 /* Suspend CCL program because of reading from empty input buffer or
708 writing to full output buffer. When this program is resumed, the
709 same I/O command is executed. */
710 #define CCL_SUSPEND(stat) \
714 ccl->status = stat; \
719 /* Terminate CCL program because of invalid command. Should not occur
720 in the normal case. */
721 #define CCL_INVALID_CMD \
724 ccl->status = CCL_STAT_INVALID_CMD; \
725 goto ccl_error_handler; \
729 /* Encode one character CH to multibyte form and write to the current
730 output buffer. If CH is less than 256, CH is written as is. */
731 #define CCL_WRITE_CHAR(ch) \
735 else if (dst < dst_end) \
738 CCL_SUSPEND (CCL_STAT_SUSPEND_BY_DST); \
741 /* Write a string at ccl_prog[IC] of length LEN to the current output
743 #define CCL_WRITE_STRING(len) \
748 else if (dst + len <= dst_end) \
749 for (i = 0; i < len; i++) \
750 *dst++ = ((XFASTINT (ccl_prog[ic + (i / 3)])) \
751 >> ((2 - (i % 3)) * 8)) & 0xFF; \
753 CCL_SUSPEND (CCL_STAT_SUSPEND_BY_DST); \
756 /* Read one byte from the current input buffer into Rth register. */
757 #define CCL_READ_CHAR(r) \
761 else if (src < src_end) \
763 else if (ccl->last_block) \
769 CCL_SUSPEND (CCL_STAT_SUSPEND_BY_SRC); \
772 /* Decode CODE by a charset whose id is ID. If ID is 0, return CODE
773 as is for backward compatibility. Assume that we can use the
774 variable `charset'. */
776 #define CCL_DECODE_CHAR(id, code) \
777 ((id) == 0 ? (code) \
778 : (charset = CHARSET_FROM_ID ((id)), DECODE_CHAR (charset, (code))))
780 /* Encode character C by some of charsets in CHARSET_LIST. Set ID to
781 the id of the used charset, ENCODED to the resulf of encoding.
782 Assume that we can use the variable `charset'. */
784 #define CCL_ENCODE_CHAR(c, charset_list, id, encoded) \
788 charset = char_charset ((c), (charset_list), &code); \
789 if (! charset && ! NILP (charset_list)) \
790 charset = char_charset ((c), Qnil, &code); \
793 (id) = CHARSET_ID (charset); \
798 /* Execute CCL code on characters at SOURCE (length SRC_SIZE). The
799 resulting text goes to a place pointed by DESTINATION, the length
800 of which should not exceed DST_SIZE. As a side effect, how many
801 characters are consumed and produced are recorded in CCL->consumed
802 and CCL->produced, and the contents of CCL registers are updated.
803 If SOURCE or DESTINATION is NULL, only operations on registers are
807 #define CCL_DEBUG_BACKTRACE_LEN 256
808 int ccl_backtrace_table
[CCL_DEBUG_BACKTRACE_LEN
];
809 int ccl_backtrace_idx
;
812 struct ccl_prog_stack
814 Lisp_Object
*ccl_prog
; /* Pointer to an array of CCL code. */
815 int ic
; /* Instruction Counter. */
818 /* For the moment, we only support depth 256 of stack. */
819 static struct ccl_prog_stack ccl_prog_stack_struct
[256];
822 ccl_driver (ccl
, source
, destination
, src_size
, dst_size
, charset_list
)
823 struct ccl_program
*ccl
;
824 int *source
, *destination
;
825 int src_size
, dst_size
;
826 Lisp_Object charset_list
;
828 register int *reg
= ccl
->reg
;
829 register int ic
= ccl
->ic
;
830 register int code
= 0, field1
, field2
;
831 register Lisp_Object
*ccl_prog
= ccl
->prog
;
832 int *src
= source
, *src_end
= src
+ src_size
;
833 int *dst
= destination
, *dst_end
= dst
+ dst_size
;
836 int stack_idx
= ccl
->stack_idx
;
837 /* Instruction counter of the current CCL code. */
839 struct charset
*charset
;
841 if (ic
>= ccl
->eof_ic
)
842 ic
= CCL_HEADER_MAIN
;
844 if (ccl
->buf_magnification
== 0) /* We can't read/produce any bytes. */
847 /* Set mapping stack pointer. */
848 mapping_stack_pointer
= mapping_stack
;
851 ccl_backtrace_idx
= 0;
858 ccl_backtrace_table
[ccl_backtrace_idx
++] = ic
;
859 if (ccl_backtrace_idx
>= CCL_DEBUG_BACKTRACE_LEN
)
860 ccl_backtrace_idx
= 0;
861 ccl_backtrace_table
[ccl_backtrace_idx
] = 0;
864 if (!NILP (Vquit_flag
) && NILP (Vinhibit_quit
))
866 /* We can't just signal Qquit, instead break the loop as if
867 the whole data is processed. Don't reset Vquit_flag, it
868 must be handled later at a safer place. */
870 src
= source
+ src_size
;
871 ccl
->status
= CCL_STAT_QUIT
;
876 code
= XINT (ccl_prog
[ic
]); ic
++;
878 field2
= (code
& 0xFF) >> 5;
881 #define RRR (field1 & 7)
882 #define Rrr ((field1 >> 3) & 7)
884 #define EXCMD (field1 >> 6)
888 case CCL_SetRegister
: /* 00000000000000000RRRrrrXXXXX */
892 case CCL_SetShortConst
: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
896 case CCL_SetConst
: /* 00000000000000000000rrrXXXXX */
897 reg
[rrr
] = XINT (ccl_prog
[ic
]);
901 case CCL_SetArray
: /* CCCCCCCCCCCCCCCCCCCCRRRrrrXXXXX */
904 if ((unsigned int) i
< j
)
905 reg
[rrr
] = XINT (ccl_prog
[ic
+ i
]);
909 case CCL_Jump
: /* A--D--D--R--E--S--S-000XXXXX */
913 case CCL_JumpCond
: /* A--D--D--R--E--S--S-rrrXXXXX */
918 case CCL_WriteRegisterJump
: /* A--D--D--R--E--S--S-rrrXXXXX */
924 case CCL_WriteRegisterReadJump
: /* A--D--D--R--E--S--S-rrrXXXXX */
928 CCL_READ_CHAR (reg
[rrr
]);
932 case CCL_WriteConstJump
: /* A--D--D--R--E--S--S-000XXXXX */
933 i
= XINT (ccl_prog
[ic
]);
938 case CCL_WriteConstReadJump
: /* A--D--D--R--E--S--S-rrrXXXXX */
939 i
= XINT (ccl_prog
[ic
]);
942 CCL_READ_CHAR (reg
[rrr
]);
946 case CCL_WriteStringJump
: /* A--D--D--R--E--S--S-000XXXXX */
947 j
= XINT (ccl_prog
[ic
]);
949 CCL_WRITE_STRING (j
);
953 case CCL_WriteArrayReadJump
: /* A--D--D--R--E--S--S-rrrXXXXX */
955 j
= XINT (ccl_prog
[ic
]);
956 if ((unsigned int) i
< j
)
958 i
= XINT (ccl_prog
[ic
+ 1 + i
]);
962 CCL_READ_CHAR (reg
[rrr
]);
963 ic
+= ADDR
- (j
+ 2);
966 case CCL_ReadJump
: /* A--D--D--R--E--S--S-rrrYYYYY */
967 CCL_READ_CHAR (reg
[rrr
]);
971 case CCL_ReadBranch
: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
972 CCL_READ_CHAR (reg
[rrr
]);
973 /* fall through ... */
974 case CCL_Branch
: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
975 if ((unsigned int) reg
[rrr
] < field1
)
976 ic
+= XINT (ccl_prog
[ic
+ reg
[rrr
]]);
978 ic
+= XINT (ccl_prog
[ic
+ field1
]);
981 case CCL_ReadRegister
: /* CCCCCCCCCCCCCCCCCCCCrrXXXXX */
984 CCL_READ_CHAR (reg
[rrr
]);
986 code
= XINT (ccl_prog
[ic
]); ic
++;
988 field2
= (code
& 0xFF) >> 5;
992 case CCL_WriteExprConst
: /* 1:00000OPERATION000RRR000XXXXX */
995 j
= XINT (ccl_prog
[ic
]);
997 jump_address
= ic
+ 1;
1000 case CCL_WriteRegister
: /* CCCCCCCCCCCCCCCCCCCrrrXXXXX */
1006 code
= XINT (ccl_prog
[ic
]); ic
++;
1008 field2
= (code
& 0xFF) >> 5;
1012 case CCL_WriteExprRegister
: /* 1:00000OPERATIONRrrRRR000XXXXX */
1020 case CCL_Call
: /* 1:CCCCCCCCCCCCCCCCCCCCFFFXXXXX */
1025 /* If FFF is nonzero, the CCL program ID is in the
1029 prog_id
= XINT (ccl_prog
[ic
]);
1035 if (stack_idx
>= 256
1037 || prog_id
>= ASIZE (Vccl_program_table
)
1038 || (slot
= AREF (Vccl_program_table
, prog_id
), !VECTORP (slot
))
1039 || !VECTORP (AREF (slot
, 1)))
1043 ccl_prog
= ccl_prog_stack_struct
[0].ccl_prog
;
1044 ic
= ccl_prog_stack_struct
[0].ic
;
1049 ccl_prog_stack_struct
[stack_idx
].ccl_prog
= ccl_prog
;
1050 ccl_prog_stack_struct
[stack_idx
].ic
= ic
;
1052 ccl_prog
= XVECTOR (AREF (slot
, 1))->contents
;
1053 ic
= CCL_HEADER_MAIN
;
1057 case CCL_WriteConstString
: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
1059 CCL_WRITE_CHAR (field1
);
1062 CCL_WRITE_STRING (field1
);
1063 ic
+= (field1
+ 2) / 3;
1067 case CCL_WriteArray
: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
1069 if ((unsigned int) i
< field1
)
1071 j
= XINT (ccl_prog
[ic
+ i
]);
1077 case CCL_End
: /* 0000000000000000000000XXXXX */
1081 ccl_prog
= ccl_prog_stack_struct
[stack_idx
].ccl_prog
;
1082 ic
= ccl_prog_stack_struct
[stack_idx
].ic
;
1087 /* ccl->ic should points to this command code again to
1088 suppress further processing. */
1092 case CCL_ExprSelfConst
: /* 00000OPERATION000000rrrXXXXX */
1093 i
= XINT (ccl_prog
[ic
]);
1098 case CCL_ExprSelfReg
: /* 00000OPERATION000RRRrrrXXXXX */
1105 case CCL_PLUS
: reg
[rrr
] += i
; break;
1106 case CCL_MINUS
: reg
[rrr
] -= i
; break;
1107 case CCL_MUL
: reg
[rrr
] *= i
; break;
1108 case CCL_DIV
: reg
[rrr
] /= i
; break;
1109 case CCL_MOD
: reg
[rrr
] %= i
; break;
1110 case CCL_AND
: reg
[rrr
] &= i
; break;
1111 case CCL_OR
: reg
[rrr
] |= i
; break;
1112 case CCL_XOR
: reg
[rrr
] ^= i
; break;
1113 case CCL_LSH
: reg
[rrr
] <<= i
; break;
1114 case CCL_RSH
: reg
[rrr
] >>= i
; break;
1115 case CCL_LSH8
: reg
[rrr
] <<= 8; reg
[rrr
] |= i
; break;
1116 case CCL_RSH8
: reg
[7] = reg
[rrr
] & 0xFF; reg
[rrr
] >>= 8; break;
1117 case CCL_DIVMOD
: reg
[7] = reg
[rrr
] % i
; reg
[rrr
] /= i
; break;
1118 case CCL_LS
: reg
[rrr
] = reg
[rrr
] < i
; break;
1119 case CCL_GT
: reg
[rrr
] = reg
[rrr
] > i
; break;
1120 case CCL_EQ
: reg
[rrr
] = reg
[rrr
] == i
; break;
1121 case CCL_LE
: reg
[rrr
] = reg
[rrr
] <= i
; break;
1122 case CCL_GE
: reg
[rrr
] = reg
[rrr
] >= i
; break;
1123 case CCL_NE
: reg
[rrr
] = reg
[rrr
] != i
; break;
1124 default: CCL_INVALID_CMD
;
1128 case CCL_SetExprConst
: /* 00000OPERATION000RRRrrrXXXXX */
1130 j
= XINT (ccl_prog
[ic
]);
1132 jump_address
= ++ic
;
1135 case CCL_SetExprReg
: /* 00000OPERATIONRrrRRRrrrXXXXX */
1142 case CCL_ReadJumpCondExprConst
: /* A--D--D--R--E--S--S-rrrXXXXX */
1143 CCL_READ_CHAR (reg
[rrr
]);
1144 case CCL_JumpCondExprConst
: /* A--D--D--R--E--S--S-rrrXXXXX */
1146 op
= XINT (ccl_prog
[ic
]);
1147 jump_address
= ic
++ + ADDR
;
1148 j
= XINT (ccl_prog
[ic
]);
1153 case CCL_ReadJumpCondExprReg
: /* A--D--D--R--E--S--S-rrrXXXXX */
1154 CCL_READ_CHAR (reg
[rrr
]);
1155 case CCL_JumpCondExprReg
:
1157 op
= XINT (ccl_prog
[ic
]);
1158 jump_address
= ic
++ + ADDR
;
1159 j
= reg
[XINT (ccl_prog
[ic
])];
1166 case CCL_PLUS
: reg
[rrr
] = i
+ j
; break;
1167 case CCL_MINUS
: reg
[rrr
] = i
- j
; break;
1168 case CCL_MUL
: reg
[rrr
] = i
* j
; break;
1169 case CCL_DIV
: reg
[rrr
] = i
/ j
; break;
1170 case CCL_MOD
: reg
[rrr
] = i
% j
; break;
1171 case CCL_AND
: reg
[rrr
] = i
& j
; break;
1172 case CCL_OR
: reg
[rrr
] = i
| j
; break;
1173 case CCL_XOR
: reg
[rrr
] = i
^ j
;; break;
1174 case CCL_LSH
: reg
[rrr
] = i
<< j
; break;
1175 case CCL_RSH
: reg
[rrr
] = i
>> j
; break;
1176 case CCL_LSH8
: reg
[rrr
] = (i
<< 8) | j
; break;
1177 case CCL_RSH8
: reg
[rrr
] = i
>> 8; reg
[7] = i
& 0xFF; break;
1178 case CCL_DIVMOD
: reg
[rrr
] = i
/ j
; reg
[7] = i
% j
; break;
1179 case CCL_LS
: reg
[rrr
] = i
< j
; break;
1180 case CCL_GT
: reg
[rrr
] = i
> j
; break;
1181 case CCL_EQ
: reg
[rrr
] = i
== j
; break;
1182 case CCL_LE
: reg
[rrr
] = i
<= j
; break;
1183 case CCL_GE
: reg
[rrr
] = i
>= j
; break;
1184 case CCL_NE
: reg
[rrr
] = i
!= j
; break;
1185 case CCL_DECODE_SJIS
:
1193 case CCL_ENCODE_SJIS
:
1201 default: CCL_INVALID_CMD
;
1204 if (code
== CCL_WriteExprConst
|| code
== CCL_WriteExprRegister
)
1217 case CCL_ReadMultibyteChar2
:
1221 CCL_ENCODE_CHAR (i
, charset_list
, reg
[RRR
], reg
[rrr
]);
1224 case CCL_WriteMultibyteChar2
:
1227 i
= CCL_DECODE_CHAR (reg
[RRR
], reg
[rrr
]);
1231 case CCL_TranslateCharacter
:
1232 i
= CCL_DECODE_CHAR (reg
[RRR
], reg
[rrr
]);
1233 op
= translate_char (GET_TRANSLATION_TABLE (reg
[Rrr
]), i
);
1234 CCL_ENCODE_CHAR (op
, charset_list
, reg
[RRR
], reg
[rrr
]);
1237 case CCL_TranslateCharacterConstTbl
:
1238 op
= XINT (ccl_prog
[ic
]); /* table */
1240 i
= CCL_DECODE_CHAR (reg
[RRR
], reg
[rrr
]);
1241 op
= translate_char (GET_TRANSLATION_TABLE (op
), i
);
1242 CCL_ENCODE_CHAR (op
, charset_list
, reg
[RRR
], reg
[rrr
]);
1245 case CCL_LookupIntConstTbl
:
1246 op
= XINT (ccl_prog
[ic
]); /* table */
1249 struct Lisp_Hash_Table
*h
= GET_HASH_TABLE (op
);
1251 op
= hash_lookup (h
, make_number (reg
[RRR
]), NULL
);
1255 opl
= HASH_VALUE (h
, op
);
1256 if (! CHARACTERP (XINT (opl
)))
1258 reg
[RRR
] = charset_unicode
;
1260 reg
[7] = 1; /* r7 true for success */
1267 case CCL_LookupCharConstTbl
:
1268 op
= XINT (ccl_prog
[ic
]); /* table */
1270 i
= CCL_DECODE_CHAR (reg
[RRR
], reg
[rrr
]);
1272 struct Lisp_Hash_Table
*h
= GET_HASH_TABLE (op
);
1274 op
= hash_lookup (h
, make_number (i
), NULL
);
1278 opl
= HASH_VALUE (h
, op
);
1279 if (!INTEGERP (opl
))
1281 reg
[RRR
] = XINT (opl
);
1282 reg
[7] = 1; /* r7 true for success */
1289 case CCL_IterateMultipleMap
:
1291 Lisp_Object map
, content
, attrib
, value
;
1292 int point
, size
, fin_ic
;
1294 j
= XINT (ccl_prog
[ic
++]); /* number of maps. */
1297 if ((j
> reg
[RRR
]) && (j
>= 0))
1312 size
= ASIZE (Vcode_conversion_map_vector
);
1313 point
= XINT (ccl_prog
[ic
++]);
1314 if (point
>= size
) continue;
1315 map
= AREF (Vcode_conversion_map_vector
, point
);
1317 /* Check map varidity. */
1318 if (!CONSP (map
)) continue;
1320 if (!VECTORP (map
)) continue;
1322 if (size
<= 1) continue;
1324 content
= AREF (map
, 0);
1327 [STARTPOINT VAL1 VAL2 ...] or
1328 [t ELELMENT STARTPOINT ENDPOINT] */
1329 if (NUMBERP (content
))
1331 point
= XUINT (content
);
1332 point
= op
- point
+ 1;
1333 if (!((point
>= 1) && (point
< size
))) continue;
1334 content
= AREF (map
, point
);
1336 else if (EQ (content
, Qt
))
1338 if (size
!= 4) continue;
1339 if ((op
>= XUINT (AREF (map
, 2)))
1340 && (op
< XUINT (AREF (map
, 3))))
1341 content
= AREF (map
, 1);
1350 else if (NUMBERP (content
))
1353 reg
[rrr
] = XINT(content
);
1356 else if (EQ (content
, Qt
) || EQ (content
, Qlambda
))
1361 else if (CONSP (content
))
1363 attrib
= XCAR (content
);
1364 value
= XCDR (content
);
1365 if (!NUMBERP (attrib
) || !NUMBERP (value
))
1368 reg
[rrr
] = XUINT (value
);
1371 else if (SYMBOLP (content
))
1372 CCL_CALL_FOR_MAP_INSTRUCTION (content
, fin_ic
);
1382 case CCL_MapMultiple
:
1384 Lisp_Object map
, content
, attrib
, value
;
1385 int point
, size
, map_vector_size
;
1386 int map_set_rest_length
, fin_ic
;
1387 int current_ic
= this_ic
;
1389 /* inhibit recursive call on MapMultiple. */
1390 if (stack_idx_of_map_multiple
> 0)
1392 if (stack_idx_of_map_multiple
<= stack_idx
)
1394 stack_idx_of_map_multiple
= 0;
1395 mapping_stack_pointer
= mapping_stack
;
1400 mapping_stack_pointer
= mapping_stack
;
1401 stack_idx_of_map_multiple
= 0;
1403 map_set_rest_length
=
1404 XINT (ccl_prog
[ic
++]); /* number of maps and separators. */
1405 fin_ic
= ic
+ map_set_rest_length
;
1408 if ((map_set_rest_length
> reg
[RRR
]) && (reg
[RRR
] >= 0))
1412 map_set_rest_length
-= i
;
1418 mapping_stack_pointer
= mapping_stack
;
1422 if (mapping_stack_pointer
<= (mapping_stack
+ 1))
1424 /* Set up initial state. */
1425 mapping_stack_pointer
= mapping_stack
;
1426 PUSH_MAPPING_STACK (0, op
);
1431 /* Recover after calling other ccl program. */
1434 POP_MAPPING_STACK (map_set_rest_length
, orig_op
);
1435 POP_MAPPING_STACK (map_set_rest_length
, reg
[rrr
]);
1439 /* Regard it as Qnil. */
1443 map_set_rest_length
--;
1446 /* Regard it as Qt. */
1450 map_set_rest_length
--;
1453 /* Regard it as Qlambda. */
1455 i
+= map_set_rest_length
;
1456 ic
+= map_set_rest_length
;
1457 map_set_rest_length
= 0;
1460 /* Regard it as normal mapping. */
1461 i
+= map_set_rest_length
;
1462 ic
+= map_set_rest_length
;
1463 POP_MAPPING_STACK (map_set_rest_length
, reg
[rrr
]);
1467 map_vector_size
= ASIZE (Vcode_conversion_map_vector
);
1470 for (;map_set_rest_length
> 0;i
++, ic
++, map_set_rest_length
--)
1472 point
= XINT(ccl_prog
[ic
]);
1475 /* +1 is for including separator. */
1477 if (mapping_stack_pointer
1478 >= &mapping_stack
[MAX_MAP_SET_LEVEL
])
1480 PUSH_MAPPING_STACK (map_set_rest_length
- point
,
1482 map_set_rest_length
= point
;
1487 if (point
>= map_vector_size
) continue;
1488 map
= AREF (Vcode_conversion_map_vector
, point
);
1490 /* Check map varidity. */
1491 if (!CONSP (map
)) continue;
1493 if (!VECTORP (map
)) continue;
1495 if (size
<= 1) continue;
1497 content
= AREF (map
, 0);
1500 [STARTPOINT VAL1 VAL2 ...] or
1501 [t ELEMENT STARTPOINT ENDPOINT] */
1502 if (NUMBERP (content
))
1504 point
= XUINT (content
);
1505 point
= op
- point
+ 1;
1506 if (!((point
>= 1) && (point
< size
))) continue;
1507 content
= AREF (map
, point
);
1509 else if (EQ (content
, Qt
))
1511 if (size
!= 4) continue;
1512 if ((op
>= XUINT (AREF (map
, 2))) &&
1513 (op
< XUINT (AREF (map
, 3))))
1514 content
= AREF (map
, 1);
1525 if (NUMBERP (content
))
1527 op
= XINT (content
);
1528 i
+= map_set_rest_length
- 1;
1529 ic
+= map_set_rest_length
- 1;
1530 POP_MAPPING_STACK (map_set_rest_length
, reg
[rrr
]);
1531 map_set_rest_length
++;
1533 else if (CONSP (content
))
1535 attrib
= XCAR (content
);
1536 value
= XCDR (content
);
1537 if (!NUMBERP (attrib
) || !NUMBERP (value
))
1540 i
+= map_set_rest_length
- 1;
1541 ic
+= map_set_rest_length
- 1;
1542 POP_MAPPING_STACK (map_set_rest_length
, reg
[rrr
]);
1543 map_set_rest_length
++;
1545 else if (EQ (content
, Qt
))
1549 else if (EQ (content
, Qlambda
))
1551 i
+= map_set_rest_length
;
1552 ic
+= map_set_rest_length
;
1555 else if (SYMBOLP (content
))
1557 if (mapping_stack_pointer
1558 >= &mapping_stack
[MAX_MAP_SET_LEVEL
])
1560 PUSH_MAPPING_STACK (map_set_rest_length
, reg
[rrr
]);
1561 PUSH_MAPPING_STACK (map_set_rest_length
, op
);
1562 stack_idx_of_map_multiple
= stack_idx
+ 1;
1563 CCL_CALL_FOR_MAP_INSTRUCTION (content
, current_ic
);
1568 if (mapping_stack_pointer
<= (mapping_stack
+ 1))
1570 POP_MAPPING_STACK (map_set_rest_length
, reg
[rrr
]);
1571 i
+= map_set_rest_length
;
1572 ic
+= map_set_rest_length
;
1573 POP_MAPPING_STACK (map_set_rest_length
, reg
[rrr
]);
1583 Lisp_Object map
, attrib
, value
, content
;
1585 j
= XINT (ccl_prog
[ic
++]); /* map_id */
1587 if (j
>= ASIZE (Vcode_conversion_map_vector
))
1592 map
= AREF (Vcode_conversion_map_vector
, j
);
1605 point
= XUINT (AREF (map
, 0));
1606 point
= op
- point
+ 1;
1609 (!((point
>= 1) && (point
< size
))))
1614 content
= AREF (map
, point
);
1617 else if (NUMBERP (content
))
1618 reg
[rrr
] = XINT (content
);
1619 else if (EQ (content
, Qt
));
1620 else if (CONSP (content
))
1622 attrib
= XCAR (content
);
1623 value
= XCDR (content
);
1624 if (!NUMBERP (attrib
) || !NUMBERP (value
))
1626 reg
[rrr
] = XUINT(value
);
1629 else if (SYMBOLP (content
))
1630 CCL_CALL_FOR_MAP_INSTRUCTION (content
, ic
);
1648 /* The suppress_error member is set when e.g. a CCL-based coding
1649 system is used for terminal output. */
1650 if (!ccl
->suppress_error
&& destination
)
1652 /* We can insert an error message only if DESTINATION is
1653 specified and we still have a room to store the message
1661 switch (ccl
->status
)
1663 case CCL_STAT_INVALID_CMD
:
1664 sprintf(msg
, "\nCCL: Invalid command %x (ccl_code = %x) at %d.",
1665 code
& 0x1F, code
, this_ic
);
1668 int i
= ccl_backtrace_idx
- 1;
1671 msglen
= strlen (msg
);
1672 if (dst
+ msglen
<= (dst_bytes
? dst_end
: src
))
1674 bcopy (msg
, dst
, msglen
);
1678 for (j
= 0; j
< CCL_DEBUG_BACKTRACE_LEN
; j
++, i
--)
1680 if (i
< 0) i
= CCL_DEBUG_BACKTRACE_LEN
- 1;
1681 if (ccl_backtrace_table
[i
] == 0)
1683 sprintf(msg
, " %d", ccl_backtrace_table
[i
]);
1684 msglen
= strlen (msg
);
1685 if (dst
+ msglen
> (dst_bytes
? dst_end
: src
))
1687 bcopy (msg
, dst
, msglen
);
1696 sprintf(msg
, "\nCCL: Quited.");
1700 sprintf(msg
, "\nCCL: Unknown error type (%d).", ccl
->status
);
1703 msglen
= strlen (msg
);
1704 if (dst
+ msglen
<= dst_end
)
1706 for (i
= 0; i
< msglen
; i
++)
1710 if (ccl
->status
== CCL_STAT_INVALID_CMD
)
1712 #if 0 /* If the remaining bytes contain 0x80..0x9F, copying them
1713 results in an invalid multibyte sequence. */
1715 /* Copy the remaining source data. */
1716 int i
= src_end
- src
;
1717 if (dst_bytes
&& (dst_end
- dst
) < i
)
1719 bcopy (src
, dst
, i
);
1723 /* Signal that we've consumed everything. */
1731 ccl
->stack_idx
= stack_idx
;
1732 ccl
->prog
= ccl_prog
;
1733 ccl
->consumed
= src
- source
;
1734 ccl
->produced
= dst
- destination
;
1737 /* Resolve symbols in the specified CCL code (Lisp vector). This
1738 function converts symbols of code conversion maps and character
1739 translation tables embeded in the CCL code into their ID numbers.
1741 The return value is a vector (CCL itself or a new vector in which
1742 all symbols are resolved), Qt if resolving of some symbol failed,
1743 or nil if CCL contains invalid data. */
1746 resolve_symbol_ccl_program (ccl
)
1749 int i
, veclen
, unresolved
= 0;
1750 Lisp_Object result
, contents
, val
;
1753 veclen
= ASIZE (result
);
1755 for (i
= 0; i
< veclen
; i
++)
1757 contents
= AREF (result
, i
);
1758 if (INTEGERP (contents
))
1760 else if (CONSP (contents
)
1761 && SYMBOLP (XCAR (contents
))
1762 && SYMBOLP (XCDR (contents
)))
1764 /* This is the new style for embedding symbols. The form is
1765 (SYMBOL . PROPERTY). (get SYMBOL PROPERTY) should give
1768 if (EQ (result
, ccl
))
1769 result
= Fcopy_sequence (ccl
);
1771 val
= Fget (XCAR (contents
), XCDR (contents
));
1773 AREF (result
, i
) = val
;
1778 else if (SYMBOLP (contents
))
1780 /* This is the old style for embedding symbols. This style
1781 may lead to a bug if, for instance, a translation table
1782 and a code conversion map have the same name. */
1783 if (EQ (result
, ccl
))
1784 result
= Fcopy_sequence (ccl
);
1786 val
= Fget (contents
, Qtranslation_table_id
);
1788 AREF (result
, i
) = val
;
1791 val
= Fget (contents
, Qcode_conversion_map_id
);
1793 AREF (result
, i
) = val
;
1796 val
= Fget (contents
, Qccl_program_idx
);
1798 AREF (result
, i
) = val
;
1808 return (unresolved
? Qt
: result
);
1811 /* Return the compiled code (vector) of CCL program CCL_PROG.
1812 CCL_PROG is a name (symbol) of the program or already compiled
1813 code. If necessary, resolve symbols in the compiled code to index
1814 numbers. If we failed to get the compiled code or to resolve
1815 symbols, return Qnil. */
1818 ccl_get_compiled_code (ccl_prog
)
1819 Lisp_Object ccl_prog
;
1821 Lisp_Object val
, slot
;
1823 if (VECTORP (ccl_prog
))
1825 val
= resolve_symbol_ccl_program (ccl_prog
);
1826 return (VECTORP (val
) ? val
: Qnil
);
1828 if (!SYMBOLP (ccl_prog
))
1831 val
= Fget (ccl_prog
, Qccl_program_idx
);
1833 || XINT (val
) >= ASIZE (Vccl_program_table
))
1835 slot
= AREF (Vccl_program_table
, XINT (val
));
1836 if (! VECTORP (slot
)
1837 || ASIZE (slot
) != 3
1838 || ! VECTORP (AREF (slot
, 1)))
1840 if (NILP (AREF (slot
, 2)))
1842 val
= resolve_symbol_ccl_program (AREF (slot
, 1));
1843 if (! VECTORP (val
))
1845 AREF (slot
, 1) = val
;
1846 AREF (slot
, 2) = Qt
;
1848 return AREF (slot
, 1);
1851 /* Setup fields of the structure pointed by CCL appropriately for the
1852 execution of CCL program CCL_PROG. CCL_PROG is the name (symbol)
1853 of the CCL program or the already compiled code (vector).
1854 Return 0 if we succeed this setup, else return -1.
1856 If CCL_PROG is nil, we just reset the structure pointed by CCL. */
1858 setup_ccl_program (ccl
, ccl_prog
)
1859 struct ccl_program
*ccl
;
1860 Lisp_Object ccl_prog
;
1864 if (! NILP (ccl_prog
))
1866 struct Lisp_Vector
*vp
;
1868 ccl_prog
= ccl_get_compiled_code (ccl_prog
);
1869 if (! VECTORP (ccl_prog
))
1871 vp
= XVECTOR (ccl_prog
);
1872 ccl
->size
= vp
->size
;
1873 ccl
->prog
= vp
->contents
;
1874 ccl
->eof_ic
= XINT (vp
->contents
[CCL_HEADER_EOF
]);
1875 ccl
->buf_magnification
= XINT (vp
->contents
[CCL_HEADER_BUF_MAG
]);
1877 ccl
->ic
= CCL_HEADER_MAIN
;
1878 for (i
= 0; i
< 8; i
++)
1880 ccl
->last_block
= 0;
1881 ccl
->private_state
= 0;
1884 ccl
->suppress_error
= 0;
1885 ccl
->eight_bit_control
= 0;
1889 DEFUN ("ccl-program-p", Fccl_program_p
, Sccl_program_p
, 1, 1, 0,
1890 doc
: /* Return t if OBJECT is a CCL program name or a compiled CCL program code.
1891 See the documentation of `define-ccl-program' for the detail of CCL program. */)
1897 if (VECTORP (object
))
1899 val
= resolve_symbol_ccl_program (object
);
1900 return (VECTORP (val
) ? Qt
: Qnil
);
1902 if (!SYMBOLP (object
))
1905 val
= Fget (object
, Qccl_program_idx
);
1906 return ((! NATNUMP (val
)
1907 || XINT (val
) >= ASIZE (Vccl_program_table
))
1911 DEFUN ("ccl-execute", Fccl_execute
, Sccl_execute
, 2, 2, 0,
1912 doc
: /* Execute CCL-PROGRAM with registers initialized by REGISTERS.
1914 CCL-PROGRAM is a CCL program name (symbol)
1915 or compiled code generated by `ccl-compile' (for backward compatibility.
1916 In the latter case, the execution overhead is bigger than in the former).
1917 No I/O commands should appear in CCL-PROGRAM.
1919 REGISTERS is a vector of [R0 R1 ... R7] where RN is an initial value
1920 for the Nth register.
1922 As side effect, each element of REGISTERS holds the value of
1923 the corresponding register after the execution.
1925 See the documentation of `define-ccl-program' for a definition of CCL
1928 Lisp_Object ccl_prog
, reg
;
1930 struct ccl_program ccl
;
1933 if (setup_ccl_program (&ccl
, ccl_prog
) < 0)
1934 error ("Invalid CCL program");
1937 if (ASIZE (reg
) != 8)
1938 error ("Length of vector REGISTERS is not 8");
1940 for (i
= 0; i
< 8; i
++)
1941 ccl
.reg
[i
] = (INTEGERP (AREF (reg
, i
))
1942 ? XINT (AREF (reg
, i
))
1945 ccl_driver (&ccl
, NULL
, NULL
, 0, 0, Qnil
);
1947 if (ccl
.status
!= CCL_STAT_SUCCESS
)
1948 error ("Error in CCL program at %dth code", ccl
.ic
);
1950 for (i
= 0; i
< 8; i
++)
1951 XSETINT (AREF (reg
, i
), ccl
.reg
[i
]);
1955 DEFUN ("ccl-execute-on-string", Fccl_execute_on_string
, Sccl_execute_on_string
,
1957 doc
: /* Execute CCL-PROGRAM with initial STATUS on STRING.
1959 CCL-PROGRAM is a symbol registered by register-ccl-program,
1960 or a compiled code generated by `ccl-compile' (for backward compatibility,
1961 in this case, the execution is slower).
1963 Read buffer is set to STRING, and write buffer is allocated automatically.
1965 STATUS is a vector of [R0 R1 ... R7 IC], where
1966 R0..R7 are initial values of corresponding registers,
1967 IC is the instruction counter specifying from where to start the program.
1968 If R0..R7 are nil, they are initialized to 0.
1969 If IC is nil, it is initialized to head of the CCL program.
1971 If optional 4th arg CONTINUE is non-nil, keep IC on read operation
1972 when read buffer is exausted, else, IC is always set to the end of
1973 CCL-PROGRAM on exit.
1975 It returns the contents of write buffer as a string,
1976 and as side effect, STATUS is updated.
1977 If the optional 5th arg UNIBYTE-P is non-nil, the returned string
1978 is a unibyte string. By default it is a multibyte string.
1980 See the documentation of `define-ccl-program' for the detail of CCL program. */)
1981 (ccl_prog
, status
, str
, contin
, unibyte_p
)
1982 Lisp_Object ccl_prog
, status
, str
, contin
, unibyte_p
;
1985 struct ccl_program ccl
;
1988 unsigned char *outbuf
, *outp
;
1989 int str_chars
, str_bytes
;
1990 #define CCL_EXECUTE_BUF_SIZE 1024
1991 int source
[CCL_EXECUTE_BUF_SIZE
], destination
[CCL_EXECUTE_BUF_SIZE
];
1992 int consumed_chars
, consumed_bytes
, produced_chars
;
1994 if (setup_ccl_program (&ccl
, ccl_prog
) < 0)
1995 error ("Invalid CCL program");
1997 CHECK_VECTOR (status
);
1998 if (ASIZE (status
) != 9)
1999 error ("Length of vector STATUS is not 9");
2002 str_chars
= SCHARS (str
);
2003 str_bytes
= SBYTES (str
);
2005 for (i
= 0; i
< 8; i
++)
2007 if (NILP (AREF (status
, i
)))
2008 XSETINT (AREF (status
, i
), 0);
2009 if (INTEGERP (AREF (status
, i
)))
2010 ccl
.reg
[i
] = XINT (AREF (status
, i
));
2012 if (INTEGERP (AREF (status
, i
)))
2014 i
= XFASTINT (AREF (status
, 8));
2015 if (ccl
.ic
< i
&& i
< ccl
.size
)
2019 outbufsize
= (ccl
.buf_magnification
2020 ? str_bytes
* ccl
.buf_magnification
+ 256
2022 outp
= outbuf
= (unsigned char *) xmalloc (outbufsize
);
2024 consumed_chars
= consumed_bytes
= 0;
2026 while (consumed_bytes
< str_bytes
)
2028 const unsigned char *p
= SDATA (str
) + consumed_bytes
;
2029 const unsigned char *endp
= SDATA (str
) + str_bytes
;
2033 if (endp
- p
== str_chars
- consumed_chars
)
2034 while (i
< CCL_EXECUTE_BUF_SIZE
&& p
< endp
)
2037 while (i
< CCL_EXECUTE_BUF_SIZE
&& p
< endp
)
2038 source
[i
++] = STRING_CHAR_ADVANCE (p
);
2039 consumed_chars
+= i
;
2040 consumed_bytes
= p
- SDATA (str
);
2042 if (consumed_bytes
== str_bytes
)
2043 ccl
.last_block
= NILP (contin
);
2048 ccl_driver (&ccl
, src
, destination
, src_size
, CCL_EXECUTE_BUF_SIZE
,
2050 if (ccl
.status
!= CCL_STAT_SUSPEND_BY_DST
)
2052 produced_chars
+= ccl
.produced
;
2053 if (NILP (unibyte_p
))
2055 if (outp
- outbuf
+ MAX_MULTIBYTE_LENGTH
* ccl
.produced
2058 int offset
= outp
- outbuf
;
2059 outbufsize
+= MAX_MULTIBYTE_LENGTH
* ccl
.produced
;
2060 outbuf
= (unsigned char *) xrealloc (outbuf
, outbufsize
);
2061 outp
= outbuf
+ offset
;
2063 for (i
= 0; i
< ccl
.produced
; i
++)
2064 CHAR_STRING_ADVANCE (destination
[i
], outp
);
2068 if (outp
- outbuf
+ ccl
.produced
> outbufsize
)
2070 int offset
= outp
- outbuf
;
2071 outbufsize
+= ccl
.produced
;
2072 outbuf
= (unsigned char *) xrealloc (outbuf
, outbufsize
);
2073 outp
= outbuf
+ offset
;
2075 for (i
= 0; i
< ccl
.produced
; i
++)
2076 *outp
++ = destination
[i
];
2078 src
+= ccl
.consumed
;
2079 src_size
-= ccl
.consumed
;
2082 if (ccl
.status
!= CCL_STAT_SUSPEND_BY_SRC
)
2086 if (ccl
.status
!= CCL_STAT_SUCCESS
2087 && ccl
.status
!= CCL_STAT_SUSPEND_BY_SRC
)
2088 error ("Error in CCL program at %dth code", ccl
.ic
);
2090 for (i
= 0; i
< 8; i
++)
2091 XSET (XVECTOR (status
)->contents
[i
], Lisp_Int
, ccl
.reg
[i
]);
2092 XSETINT (XVECTOR (status
)->contents
[8], ccl
.ic
);
2094 if (NILP (unibyte_p
))
2095 val
= make_multibyte_string ((char *) outbuf
, produced_chars
,
2098 val
= make_unibyte_string ((char *) outbuf
, produced_chars
);
2104 DEFUN ("register-ccl-program", Fregister_ccl_program
, Sregister_ccl_program
,
2106 doc
: /* Register CCL program CCL_PROG as NAME in `ccl-program-table'.
2107 CCL_PROG should be a compiled CCL program (vector), or nil.
2108 If it is nil, just reserve NAME as a CCL program name.
2109 Return index number of the registered CCL program. */)
2111 Lisp_Object name
, ccl_prog
;
2113 int len
= ASIZE (Vccl_program_table
);
2115 Lisp_Object resolved
;
2117 CHECK_SYMBOL (name
);
2119 if (!NILP (ccl_prog
))
2121 CHECK_VECTOR (ccl_prog
);
2122 resolved
= resolve_symbol_ccl_program (ccl_prog
);
2123 if (NILP (resolved
))
2124 error ("Error in CCL program");
2125 if (VECTORP (resolved
))
2127 ccl_prog
= resolved
;
2134 for (idx
= 0; idx
< len
; idx
++)
2138 slot
= AREF (Vccl_program_table
, idx
);
2139 if (!VECTORP (slot
))
2140 /* This is the first unsed slot. Register NAME here. */
2143 if (EQ (name
, AREF (slot
, 0)))
2145 /* Update this slot. */
2146 AREF (slot
, 1) = ccl_prog
;
2147 AREF (slot
, 2) = resolved
;
2148 return make_number (idx
);
2154 /* Extend the table. */
2155 Lisp_Object new_table
;
2158 new_table
= Fmake_vector (make_number (len
* 2), Qnil
);
2159 for (j
= 0; j
< len
; j
++)
2161 = AREF (Vccl_program_table
, j
);
2162 Vccl_program_table
= new_table
;
2168 elt
= Fmake_vector (make_number (3), Qnil
);
2169 AREF (elt
, 0) = name
;
2170 AREF (elt
, 1) = ccl_prog
;
2171 AREF (elt
, 2) = resolved
;
2172 AREF (Vccl_program_table
, idx
) = elt
;
2175 Fput (name
, Qccl_program_idx
, make_number (idx
));
2176 return make_number (idx
);
2179 /* Register code conversion map.
2180 A code conversion map consists of numbers, Qt, Qnil, and Qlambda.
2181 The first element is the start code point.
2182 The other elements are mapped numbers.
2183 Symbol t means to map to an original number before mapping.
2184 Symbol nil means that the corresponding element is empty.
2185 Symbol lambda means to terminate mapping here.
2188 DEFUN ("register-code-conversion-map", Fregister_code_conversion_map
,
2189 Sregister_code_conversion_map
,
2191 doc
: /* Register SYMBOL as code conversion map MAP.
2192 Return index number of the registered map. */)
2194 Lisp_Object symbol
, map
;
2196 int len
= ASIZE (Vcode_conversion_map_vector
);
2200 CHECK_SYMBOL (symbol
);
2203 for (i
= 0; i
< len
; i
++)
2205 Lisp_Object slot
= AREF (Vcode_conversion_map_vector
, i
);
2210 if (EQ (symbol
, XCAR (slot
)))
2212 index
= make_number (i
);
2213 XSETCDR (slot
, map
);
2214 Fput (symbol
, Qcode_conversion_map
, map
);
2215 Fput (symbol
, Qcode_conversion_map_id
, index
);
2222 Lisp_Object new_vector
= Fmake_vector (make_number (len
* 2), Qnil
);
2225 for (j
= 0; j
< len
; j
++)
2226 AREF (new_vector
, j
)
2227 = AREF (Vcode_conversion_map_vector
, j
);
2228 Vcode_conversion_map_vector
= new_vector
;
2231 index
= make_number (i
);
2232 Fput (symbol
, Qcode_conversion_map
, map
);
2233 Fput (symbol
, Qcode_conversion_map_id
, index
);
2234 AREF (Vcode_conversion_map_vector
, i
) = Fcons (symbol
, map
);
2242 staticpro (&Vccl_program_table
);
2243 Vccl_program_table
= Fmake_vector (make_number (32), Qnil
);
2245 Qccl
= intern ("ccl");
2248 Qcclp
= intern ("cclp");
2251 Qccl_program
= intern ("ccl-program");
2252 staticpro (&Qccl_program
);
2254 Qccl_program_idx
= intern ("ccl-program-idx");
2255 staticpro (&Qccl_program_idx
);
2257 Qcode_conversion_map
= intern ("code-conversion-map");
2258 staticpro (&Qcode_conversion_map
);
2260 Qcode_conversion_map_id
= intern ("code-conversion-map-id");
2261 staticpro (&Qcode_conversion_map_id
);
2263 DEFVAR_LISP ("code-conversion-map-vector", &Vcode_conversion_map_vector
,
2264 doc
: /* Vector of code conversion maps. */);
2265 Vcode_conversion_map_vector
= Fmake_vector (make_number (16), Qnil
);
2267 DEFVAR_LISP ("font-ccl-encoder-alist", &Vfont_ccl_encoder_alist
,
2268 doc
: /* Alist of fontname patterns vs corresponding CCL program.
2269 Each element looks like (REGEXP . CCL-CODE),
2270 where CCL-CODE is a compiled CCL program.
2271 When a font whose name matches REGEXP is used for displaying a character,
2272 CCL-CODE is executed to calculate the code point in the font
2273 from the charset number and position code(s) of the character which are set
2274 in CCL registers R0, R1, and R2 before the execution.
2275 The code point in the font is set in CCL registers R1 and R2
2276 when the execution terminated.
2277 If the font is single-byte font, the register R2 is not used. */);
2278 Vfont_ccl_encoder_alist
= Qnil
;
2280 DEFVAR_LISP ("translation-hash-table-vector", &Vtranslation_hash_table_vector
,
2281 doc
: /* Vector containing all translation hash tables ever defined.
2282 Comprises pairs (SYMBOL . TABLE) where SYMBOL and TABLE were set up by calls
2283 to `define-translation-hash-table'. The vector is indexed by the table id
2285 Vtranslation_hash_table_vector
= Qnil
;
2287 defsubr (&Sccl_program_p
);
2288 defsubr (&Sccl_execute
);
2289 defsubr (&Sccl_execute_on_string
);
2290 defsubr (&Sregister_ccl_program
);
2291 defsubr (&Sregister_code_conversion_map
);