1 /* CCL (Code Conversion Language) interpreter.
2 Copyright (C) 1995, 1997 Electrotechnical Laboratory, JAPAN.
3 Licensed to the Free Software Foundation.
5 This file is part of GNU Emacs.
7 GNU Emacs is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 2, or (at your option)
12 GNU Emacs is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
17 You should have received a copy of the GNU General Public License
18 along with GNU Emacs; see the file COPYING. If not, write to
19 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
20 Boston, MA 02111-1307, USA. */
22 /* Synched up with : FSF Emacs 21.0.90 except TranslateCharacter */
34 #include "mule-charset.h"
36 #include "file-coding.h"
42 #endif /* not emacs */
44 /* This contains all code conversion map available to CCL. */
45 Lisp_Object Vcode_conversion_map_vector;
47 /* Alist of fontname patterns vs corresponding CCL program. */
48 Lisp_Object Vfont_ccl_encoder_alist;
50 /* This symbol is a property which associates with ccl program vector.
51 Ex: (get 'ccl-big5-encoder 'ccl-program) returns ccl program vector. */
52 Lisp_Object Qccl_program;
54 /* These symbols are properties which associate with code conversion
55 map and their ID respectively. */
56 Lisp_Object Qcode_conversion_map;
57 Lisp_Object Qcode_conversion_map_id;
59 /* Symbols of ccl program have this property, a value of the property
60 is an index for Vccl_program_table. */
61 Lisp_Object Qccl_program_idx;
63 /* Table of registered CCL programs. Each element is a vector of
64 NAME, CCL_PROG, and RESOLVEDP where NAME (symbol) is the name of
65 the program, CCL_PROG (vector) is the compiled code of the program,
66 RESOLVEDP (t or nil) is the flag to tell if symbols in CCL_PROG is
67 already resolved to index numbers or not. */
68 Lisp_Object Vccl_program_table;
70 /* CCL (Code Conversion Language) is a simple language which has
71 operations on one input buffer, one output buffer, and 7 registers.
72 The syntax of CCL is described in `ccl.el'. Emacs Lisp function
73 `ccl-compile' compiles a CCL program and produces a CCL code which
74 is a vector of integers. The structure of this vector is as
75 follows: The 1st element: buffer-magnification, a factor for the
76 size of output buffer compared with the size of input buffer. The
77 2nd element: address of CCL code to be executed when encountered
78 with end of input stream. The 3rd and the remaining elements: CCL
81 /* Header of CCL compiled code */
82 #define CCL_HEADER_BUF_MAG 0
83 #define CCL_HEADER_EOF 1
84 #define CCL_HEADER_MAIN 2
86 /* CCL code is a sequence of 28-bit non-negative integers (i.e. the
87 MSB is always 0), each contains CCL command and/or arguments in the
90 |----------------- integer (28-bit) ------------------|
91 |------- 17-bit ------|- 3-bit --|- 3-bit --|- 5-bit -|
92 |--constant argument--|-register-|-register-|-command-|
93 ccccccccccccccccc RRR rrr XXXXX
95 |------- relative address -------|-register-|-command-|
96 cccccccccccccccccccc rrr XXXXX
98 |------------- constant or other args ----------------|
99 cccccccccccccccccccccccccccc
101 where, `cc...c' is a non-negative integer indicating constant value
102 (the left most `c' is always 0) or an absolute jump address, `RRR'
103 and `rrr' are CCL register number, `XXXXX' is one of the following
108 Each comment fields shows one or more lines for command syntax and
109 the following lines for semantics of the command. In semantics, IC
110 stands for Instruction Counter. */
112 #define CCL_SetRegister 0x00 /* Set register a register value:
113 1:00000000000000000RRRrrrXXXXX
114 ------------------------------
118 #define CCL_SetShortConst 0x01 /* Set register a short constant value:
119 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
120 ------------------------------
121 reg[rrr] = CCCCCCCCCCCCCCCCCCC;
124 #define CCL_SetConst 0x02 /* Set register a constant value:
125 1:00000000000000000000rrrXXXXX
127 ------------------------------
132 #define CCL_SetArray 0x03 /* Set register an element of array:
133 1:CCCCCCCCCCCCCCCCCRRRrrrXXXXX
137 ------------------------------
138 if (0 <= reg[RRR] < CC..C)
139 reg[rrr] = ELEMENT[reg[RRR]];
143 #define CCL_Jump 0x04 /* Jump:
144 1:A--D--D--R--E--S--S-000XXXXX
145 ------------------------------
149 /* Note: If CC..C is greater than 0, the second code is omitted. */
151 #define CCL_JumpCond 0x05 /* Jump conditional:
152 1:A--D--D--R--E--S--S-rrrXXXXX
153 ------------------------------
159 #define CCL_WriteRegisterJump 0x06 /* Write register and jump:
160 1:A--D--D--R--E--S--S-rrrXXXXX
161 ------------------------------
166 #define CCL_WriteRegisterReadJump 0x07 /* Write register, read, and jump:
167 1:A--D--D--R--E--S--S-rrrXXXXX
168 2:A--D--D--R--E--S--S-rrrYYYYY
169 -----------------------------
175 /* Note: If read is suspended, the resumed execution starts from the
176 second code (YYYYY == CCL_ReadJump). */
178 #define CCL_WriteConstJump 0x08 /* Write constant and jump:
179 1:A--D--D--R--E--S--S-000XXXXX
181 ------------------------------
186 #define CCL_WriteConstReadJump 0x09 /* Write constant, read, and jump:
187 1:A--D--D--R--E--S--S-rrrXXXXX
189 3:A--D--D--R--E--S--S-rrrYYYYY
190 -----------------------------
196 /* Note: If read is suspended, the resumed execution starts from the
197 second code (YYYYY == CCL_ReadJump). */
199 #define CCL_WriteStringJump 0x0A /* Write string and jump:
200 1:A--D--D--R--E--S--S-000XXXXX
202 3:0000STRIN[0]STRIN[1]STRIN[2]
204 ------------------------------
205 write_string (STRING, LENGTH);
209 #define CCL_WriteArrayReadJump 0x0B /* Write an array element, read, and jump:
210 1:A--D--D--R--E--S--S-rrrXXXXX
215 N:A--D--D--R--E--S--S-rrrYYYYY
216 ------------------------------
217 if (0 <= reg[rrr] < LENGTH)
218 write (ELEMENT[reg[rrr]]);
219 IC += LENGTH + 2; (... pointing at N+1)
223 /* Note: If read is suspended, the resumed execution starts from the
224 Nth code (YYYYY == CCL_ReadJump). */
226 #define CCL_ReadJump 0x0C /* Read and jump:
227 1:A--D--D--R--E--S--S-rrrYYYYY
228 -----------------------------
233 #define CCL_Branch 0x0D /* Jump by branch table:
234 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
235 2:A--D--D--R--E-S-S[0]000XXXXX
236 3:A--D--D--R--E-S-S[1]000XXXXX
238 ------------------------------
239 if (0 <= reg[rrr] < CC..C)
240 IC += ADDRESS[reg[rrr]];
242 IC += ADDRESS[CC..C];
245 #define CCL_ReadRegister 0x0E /* Read bytes into registers:
246 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
247 2:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
249 ------------------------------
254 #define CCL_WriteExprConst 0x0F /* write result of expression:
255 1:00000OPERATION000RRR000XXXXX
257 ------------------------------
258 write (reg[RRR] OPERATION CONSTANT);
262 /* Note: If the Nth read is suspended, the resumed execution starts
263 from the Nth code. */
265 #define CCL_ReadBranch 0x10 /* Read one byte into a register,
266 and jump by branch table:
267 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
268 2:A--D--D--R--E-S-S[0]000XXXXX
269 3:A--D--D--R--E-S-S[1]000XXXXX
271 ------------------------------
273 if (0 <= reg[rrr] < CC..C)
274 IC += ADDRESS[reg[rrr]];
276 IC += ADDRESS[CC..C];
279 #define CCL_WriteRegister 0x11 /* Write registers:
280 1:CCCCCCCCCCCCCCCCCCCrrrXXXXX
281 2:CCCCCCCCCCCCCCCCCCCrrrXXXXX
283 ------------------------------
289 /* Note: If the Nth write is suspended, the resumed execution
290 starts from the Nth code. */
292 #define CCL_WriteExprRegister 0x12 /* Write result of expression
293 1:00000OPERATIONRrrRRR000XXXXX
294 ------------------------------
295 write (reg[RRR] OPERATION reg[Rrr]);
298 #define CCL_Call 0x13 /* Call the CCL program whose ID is
300 1:CCCCCCCCCCCCCCCCCCCCFFFXXXXX
301 [2:00000000cccccccccccccccccccc]
302 ------------------------------
310 #define CCL_WriteConstString 0x14 /* Write a constant or a string:
311 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
312 [2:0000STRIN[0]STRIN[1]STRIN[2]]
314 -----------------------------
318 write_string (STRING, CC..C);
319 IC += (CC..C + 2) / 3;
322 #define CCL_WriteArray 0x15 /* Write an element of array:
323 1:CCCCCCCCCCCCCCCCCCCCrrrXXXXX
327 ------------------------------
328 if (0 <= reg[rrr] < CC..C)
329 write (ELEMENT[reg[rrr]]);
333 #define CCL_End 0x16 /* Terminate:
334 1:00000000000000000000000XXXXX
335 ------------------------------
339 /* The following two codes execute an assignment arithmetic/logical
340 operation. The form of the operation is like REG OP= OPERAND. */
342 #define CCL_ExprSelfConst 0x17 /* REG OP= constant:
343 1:00000OPERATION000000rrrXXXXX
345 ------------------------------
346 reg[rrr] OPERATION= CONSTANT;
349 #define CCL_ExprSelfReg 0x18 /* REG1 OP= REG2:
350 1:00000OPERATION000RRRrrrXXXXX
351 ------------------------------
352 reg[rrr] OPERATION= reg[RRR];
355 /* The following codes execute an arithmetic/logical operation. The
356 form of the operation is like REG_X = REG_Y OP OPERAND2. */
358 #define CCL_SetExprConst 0x19 /* REG_X = REG_Y OP constant:
359 1:00000OPERATION000RRRrrrXXXXX
361 ------------------------------
362 reg[rrr] = reg[RRR] OPERATION CONSTANT;
366 #define CCL_SetExprReg 0x1A /* REG1 = REG2 OP REG3:
367 1:00000OPERATIONRrrRRRrrrXXXXX
368 ------------------------------
369 reg[rrr] = reg[RRR] OPERATION reg[Rrr];
372 #define CCL_JumpCondExprConst 0x1B /* Jump conditional according to
373 an operation on constant:
374 1:A--D--D--R--E--S--S-rrrXXXXX
377 -----------------------------
378 reg[7] = reg[rrr] OPERATION CONSTANT;
385 #define CCL_JumpCondExprReg 0x1C /* Jump conditional according to
386 an operation on register:
387 1:A--D--D--R--E--S--S-rrrXXXXX
390 -----------------------------
391 reg[7] = reg[rrr] OPERATION reg[RRR];
398 #define CCL_ReadJumpCondExprConst 0x1D /* Read and jump conditional according
399 to an operation on constant:
400 1:A--D--D--R--E--S--S-rrrXXXXX
403 -----------------------------
405 reg[7] = reg[rrr] OPERATION CONSTANT;
412 #define CCL_ReadJumpCondExprReg 0x1E /* Read and jump conditional according
413 to an operation on register:
414 1:A--D--D--R--E--S--S-rrrXXXXX
417 -----------------------------
419 reg[7] = reg[rrr] OPERATION reg[RRR];
426 #define CCL_Extention 0x1F /* Extended CCL code
427 1:ExtendedCOMMNDRrrRRRrrrXXXXX
430 ------------------------------
431 extended_command (rrr,RRR,Rrr,ARGS)
435 Here after, Extended CCL Instructions.
436 Bit length of extended command is 14.
437 Therefore, the instruction code range is 0..16384(0x3fff).
440 /* Read a multibyte characeter.
441 A code point is stored into reg[rrr]. A charset ID is stored into
444 #define CCL_ReadMultibyteChar2 0x00 /* Read Multibyte Character
445 1:ExtendedCOMMNDRrrRRRrrrXXXXX */
447 /* Write a multibyte character.
448 Write a character whose code point is reg[rrr] and the charset ID
451 #define CCL_WriteMultibyteChar2 0x01 /* Write Multibyte Character
452 1:ExtendedCOMMNDRrrRRRrrrXXXXX */
454 /* Translate a character whose code point is reg[rrr] and the charset
455 ID is reg[RRR] by a translation table whose ID is reg[Rrr].
457 A translated character is set in reg[rrr] (code point) and reg[RRR]
460 #define CCL_TranslateCharacter 0x02 /* Translate a multibyte character
461 1:ExtendedCOMMNDRrrRRRrrrXXXXX */
463 /* Translate a character whose code point is reg[rrr] and the charset
464 ID is reg[RRR] by a translation table whose ID is ARGUMENT.
466 A translated character is set in reg[rrr] (code point) and reg[RRR]
469 #define CCL_TranslateCharacterConstTbl 0x03 /* Translate a multibyte character
470 1:ExtendedCOMMNDRrrRRRrrrXXXXX
471 2:ARGUMENT(Translation Table ID)
474 /* Iterate looking up MAPs for reg[rrr] starting from the Nth (N =
475 reg[RRR]) MAP until some value is found.
477 Each MAP is a Lisp vector whose element is number, nil, t, or
479 If the element is nil, ignore the map and proceed to the next map.
480 If the element is t or lambda, finish without changing reg[rrr].
481 If the element is a number, set reg[rrr] to the number and finish.
483 Detail of the map structure is described in the comment for
484 CCL_MapMultiple below. */
486 #define CCL_IterateMultipleMap 0x10 /* Iterate multiple maps
487 1:ExtendedCOMMNDXXXRRRrrrXXXXX
494 /* Map the code in reg[rrr] by MAPs starting from the Nth (N =
497 MAPs are supplied in the succeeding CCL codes as follows:
499 When CCL program gives this nested structure of map to this command:
502 (MAP-ID121 MAP-ID122 MAP-ID123)
505 (MAP-ID211 (MAP-ID2111) MAP-ID212)
507 the compiled CCL code has this sequence:
508 CCL_MapMultiple (CCL code of this command)
509 16 (total number of MAPs and SEPARATORs)
527 A value of each SEPARATOR follows this rule:
528 MAP-SET := SEPARATOR [(MAP-ID | MAP-SET)]+
529 SEPARATOR := -(number of MAP-IDs and SEPARATORs in the MAP-SET)
531 (*)....Nest level of MAP-SET must not be over than MAX_MAP_SET_LEVEL.
533 When some map fails to map (i.e. it doesn't have a value for
534 reg[rrr]), the mapping is treated as identity.
536 The mapping is iterated for all maps in each map set (set of maps
537 separated by SEPARATOR) except in the case that lambda is
538 encountered. More precisely, the mapping proceeds as below:
540 At first, VAL0 is set to reg[rrr], and it is translated by the
541 first map to VAL1. Then, VAL1 is translated by the next map to
542 VAL2. This mapping is iterated until the last map is used. The
543 result of the mapping is the last value of VAL?. When the mapping
544 process reached to the end of the map set, it moves to the next
545 map set. If the next does not exit, the mapping process terminates,
546 and regard the last value as a result.
548 But, when VALm is mapped to VALn and VALn is not a number, the
549 mapping proceeds as follows:
551 If VALn is nil, the lastest map is ignored and the mapping of VALm
552 proceeds to the next map.
554 In VALn is t, VALm is reverted to reg[rrr] and the mapping of VALm
555 proceeds to the next map.
557 If VALn is lambda, move to the next map set like reaching to the
558 end of the current map set.
560 If VALn is a symbol, call the CCL program refered by it.
561 Then, use reg[rrr] as a mapped value except for -1, -2 and -3.
562 Such special values are regarded as nil, t, and lambda respectively.
564 Each map is a Lisp vector of the following format (a) or (b):
565 (a)......[STARTPOINT VAL1 VAL2 ...]
566 (b)......[t VAL STARTPOINT ENDPOINT],
568 STARTPOINT is an offset to be used for indexing a map,
569 ENDPOINT is a maximum index number of a map,
570 VAL and VALn is a number, nil, t, or lambda.
572 Valid index range of a map of type (a) is:
573 STARTPOINT <= index < STARTPOINT + map_size - 1
574 Valid index range of a map of type (b) is:
575 STARTPOINT <= index < ENDPOINT */
577 #define CCL_MapMultiple 0x11 /* Mapping by multiple code conversion maps
578 1:ExtendedCOMMNDXXXRRRrrrXXXXX
590 #define MAX_MAP_SET_LEVEL 30
598 static tr_stack mapping_stack[MAX_MAP_SET_LEVEL];
599 static tr_stack *mapping_stack_pointer;
601 /* If this variable is non-zero, it indicates the stack_idx
602 of immediately called by CCL_MapMultiple. */
603 static int stack_idx_of_map_multiple = 0;
605 #define PUSH_MAPPING_STACK(restlen, orig) \
607 mapping_stack_pointer->rest_length = (restlen); \
608 mapping_stack_pointer->orig_val = (orig); \
609 mapping_stack_pointer++; \
612 #define POP_MAPPING_STACK(restlen, orig) \
614 mapping_stack_pointer--; \
615 (restlen) = mapping_stack_pointer->rest_length; \
616 (orig) = mapping_stack_pointer->orig_val; \
619 #define CCL_CALL_FOR_MAP_INSTRUCTION(symbol, ret_ic) \
621 struct ccl_program called_ccl; \
622 if (stack_idx >= 256 \
623 || (setup_ccl_program (&called_ccl, (symbol)) != 0)) \
627 ccl_prog = ccl_prog_stack_struct[0].ccl_prog; \
628 ic = ccl_prog_stack_struct[0].ic; \
632 ccl_prog_stack_struct[stack_idx].ccl_prog = ccl_prog; \
633 ccl_prog_stack_struct[stack_idx].ic = (ret_ic); \
635 ccl_prog = called_ccl.prog; \
636 ic = CCL_HEADER_MAIN; \
640 #define CCL_MapSingle 0x12 /* Map by single code conversion map
641 1:ExtendedCOMMNDXXXRRRrrrXXXXX
643 ------------------------------
644 Map reg[rrr] by MAP-ID.
645 If some valid mapping is found,
646 set reg[rrr] to the result,
651 /* CCL arithmetic/logical operators. */
652 #define CCL_PLUS 0x00 /* X = Y + Z */
653 #define CCL_MINUS 0x01 /* X = Y - Z */
654 #define CCL_MUL 0x02 /* X = Y * Z */
655 #define CCL_DIV 0x03 /* X = Y / Z */
656 #define CCL_MOD 0x04 /* X = Y % Z */
657 #define CCL_AND 0x05 /* X = Y & Z */
658 #define CCL_OR 0x06 /* X = Y | Z */
659 #define CCL_XOR 0x07 /* X = Y ^ Z */
660 #define CCL_LSH 0x08 /* X = Y << Z */
661 #define CCL_RSH 0x09 /* X = Y >> Z */
662 #define CCL_LSH8 0x0A /* X = (Y << 8) | Z */
663 #define CCL_RSH8 0x0B /* X = Y >> 8, r[7] = Y & 0xFF */
664 #define CCL_DIVMOD 0x0C /* X = Y / Z, r[7] = Y % Z */
665 #define CCL_LS 0x10 /* X = (X < Y) */
666 #define CCL_GT 0x11 /* X = (X > Y) */
667 #define CCL_EQ 0x12 /* X = (X == Y) */
668 #define CCL_LE 0x13 /* X = (X <= Y) */
669 #define CCL_GE 0x14 /* X = (X >= Y) */
670 #define CCL_NE 0x15 /* X = (X != Y) */
672 #define CCL_DECODE_SJIS 0x16 /* X = HIGHER_BYTE (DE-SJIS (Y, Z))
673 r[7] = LOWER_BYTE (DE-SJIS (Y, Z)) */
674 #define CCL_ENCODE_SJIS 0x17 /* X = HIGHER_BYTE (SJIS (Y, Z))
675 r[7] = LOWER_BYTE (SJIS (Y, Z) */
677 /* Terminate CCL program successfully. */
678 #define CCL_SUCCESS \
680 ccl->status = CCL_STAT_SUCCESS; \
684 /* Suspend CCL program because of reading from empty input buffer or
685 writing to full output buffer. When this program is resumed, the
686 same I/O command is executed. */
687 #define CCL_SUSPEND(stat) \
690 ccl->status = stat; \
694 /* Terminate CCL program because of invalid command. Should not occur
695 in the normal case. */
696 #define CCL_INVALID_CMD \
698 ccl->status = CCL_STAT_INVALID_CMD; \
699 goto ccl_error_handler; \
702 /* Encode one character CH to multibyte form and write to the current
703 output buffer. At encoding time, if CH is less than 256, CH is
704 written as is. At decoding time, if CH cannot be regarded as an
705 ASCII character, write it in multibyte form. */
706 #define CCL_WRITE_CHAR(ch) \
710 if (conversion_mode == CCL_MODE_ENCODING) \
714 if (ccl->eol_type == CCL_CODING_EOL_CRLF) \
716 Dynarr_add (destination, '\r'); \
717 Dynarr_add (destination, '\n'); \
719 else if (ccl->eol_type == CCL_CODING_EOL_CR) \
720 Dynarr_add (destination, '\r'); \
722 Dynarr_add (destination, '\n'); \
724 else if (ch < 0x100) \
726 Dynarr_add (destination, ch); \
730 Bufbyte work[MAX_EMCHAR_LEN]; \
732 len = non_ascii_set_charptr_emchar (work, ch); \
733 Dynarr_add_many (destination, work, len); \
738 if (!CHAR_MULTIBYTE_P(ch)) \
740 Dynarr_add (destination, ch); \
744 Bufbyte work[MAX_EMCHAR_LEN]; \
746 len = non_ascii_set_charptr_emchar (work, ch); \
747 Dynarr_add_many (destination, work, len); \
752 /* Write a string at ccl_prog[IC] of length LEN to the current output
753 buffer. But this macro treat this string as a binary. Therefore,
754 cannot handle a multibyte string except for Control-1 characters. */
755 #define CCL_WRITE_STRING(len) \
757 Bufbyte work[MAX_EMCHAR_LEN]; \
761 else if (conversion_mode == CCL_MODE_ENCODING) \
763 for (i = 0; i < len; i++) \
765 ch = ((XINT (ccl_prog[ic + (i / 3)])) \
766 >> ((2 - (i % 3)) * 8)) & 0xFF; \
769 if (ccl->eol_type == CCL_CODING_EOL_CRLF) \
771 Dynarr_add (destination, '\r'); \
772 Dynarr_add (destination, '\n'); \
774 else if (ccl->eol_type == CCL_CODING_EOL_CR) \
775 Dynarr_add (destination, '\r'); \
777 Dynarr_add (destination, '\n'); \
781 Dynarr_add (destination, ch); \
785 bytes = non_ascii_set_charptr_emchar (work, ch); \
786 Dynarr_add_many (destination, work, len); \
792 for (i = 0; i < len; i++) \
794 ch = ((XINT (ccl_prog[ic + (i / 3)])) \
795 >> ((2 - (i % 3)) * 8)) & 0xFF; \
796 if (!CHAR_MULTIBYTE_P(ch)) \
798 Dynarr_add (destination, ch); \
802 bytes = non_ascii_set_charptr_emchar (work, ch); \
803 Dynarr_add_many (destination, work, len); \
809 /* Read one byte from the current input buffer into Rth register. */
810 #define CCL_READ_CHAR(r) \
818 if (ccl->last_block) \
824 CCL_SUSPEND (CCL_STAT_SUSPEND_BY_SRC); \
829 /* Set C to the character code made from CHARSET and CODE. This is
830 like MAKE_CHAR but check the validity of CHARSET and CODE. If they
831 are not valid, set C to (CODE & 0xFF) because that is usually the
832 case that CCL_ReadMultibyteChar2 read an invalid code and it set
833 CODE to that invalid byte. */
835 /* On XEmacs, TranslateCharacter is not supported. Thus, this
836 macro is not used. */
838 #define CCL_MAKE_CHAR(charset, code, c) \
840 if (charset == CHARSET_ASCII) \
842 else if (CHARSET_DEFINED_P (charset) \
843 && (code & 0x7F) >= 32 \
844 && (code < 256 || ((code >> 7) & 0x7F) >= 32)) \
846 int c1 = code & 0x7F, c2 = 0; \
849 c2 = c1, c1 = (code >> 7) & 0x7F; \
850 c = MAKE_CHAR (charset, c1, c2); \
858 /* Execute CCL code on SRC_BYTES length text at SOURCE. The resulting
859 text goes to a place pointed by DESTINATION, the length of which
860 should not exceed DST_BYTES. The bytes actually processed is
861 returned as *CONSUMED. The return value is the length of the
862 resulting text. As a side effect, the contents of CCL registers
863 are updated. If SOURCE or DESTINATION is NULL, only operations on
864 registers are permitted. */
867 #define CCL_DEBUG_BACKTRACE_LEN 256
868 int ccl_backtrace_table[CCL_BACKTRACE_TABLE];
869 int ccl_backtrace_idx;
872 struct ccl_prog_stack
874 Lisp_Object *ccl_prog; /* Pointer to an array of CCL code. */
875 int ic; /* Instruction Counter. */
878 /* For the moment, we only support depth 256 of stack. */
879 static struct ccl_prog_stack ccl_prog_stack_struct[256];
882 ccl_driver (struct ccl_program *ccl,
883 const unsigned char *source,
884 unsigned_char_dynarr *destination,
889 register int *reg = ccl->reg;
890 register int ic = ccl->ic;
891 register int code = -1;
892 register int field1, field2;
893 register Lisp_Object *ccl_prog = ccl->prog;
894 const unsigned char *src = source, *src_end = src + src_bytes;
897 int stack_idx = ccl->stack_idx;
898 /* Instruction counter of the current CCL code. */
901 if (ic >= ccl->eof_ic)
902 ic = CCL_HEADER_MAIN;
904 if (ccl->buf_magnification ==0) /* We can't produce any bytes. */
907 /* Set mapping stack pointer. */
908 mapping_stack_pointer = mapping_stack;
911 ccl_backtrace_idx = 0;
918 ccl_backtrace_table[ccl_backtrace_idx++] = ic;
919 if (ccl_backtrace_idx >= CCL_DEBUG_BACKTRACE_LEN)
920 ccl_backtrace_idx = 0;
921 ccl_backtrace_table[ccl_backtrace_idx] = 0;
924 if (!NILP (Vquit_flag) && NILP (Vinhibit_quit))
926 /* We can't just signal Qquit, instead break the loop as if
927 the whole data is processed. Don't reset Vquit_flag, it
928 must be handled later at a safer place. */
930 src = source + src_bytes;
931 ccl->status = CCL_STAT_QUIT;
936 code = XINT (ccl_prog[ic]); ic++;
938 field2 = (code & 0xFF) >> 5;
941 #define RRR (field1 & 7)
942 #define Rrr ((field1 >> 3) & 7)
944 #define EXCMD (field1 >> 6)
948 case CCL_SetRegister: /* 00000000000000000RRRrrrXXXXX */
952 case CCL_SetShortConst: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
956 case CCL_SetConst: /* 00000000000000000000rrrXXXXX */
957 reg[rrr] = XINT (ccl_prog[ic]);
961 case CCL_SetArray: /* CCCCCCCCCCCCCCCCCCCCRRRrrrXXXXX */
964 if ((unsigned int) i < j)
965 reg[rrr] = XINT (ccl_prog[ic + i]);
969 case CCL_Jump: /* A--D--D--R--E--S--S-000XXXXX */
973 case CCL_JumpCond: /* A--D--D--R--E--S--S-rrrXXXXX */
978 case CCL_WriteRegisterJump: /* A--D--D--R--E--S--S-rrrXXXXX */
984 case CCL_WriteRegisterReadJump: /* A--D--D--R--E--S--S-rrrXXXXX */
988 CCL_READ_CHAR (reg[rrr]);
992 case CCL_WriteConstJump: /* A--D--D--R--E--S--S-000XXXXX */
993 i = XINT (ccl_prog[ic]);
998 case CCL_WriteConstReadJump: /* A--D--D--R--E--S--S-rrrXXXXX */
999 i = XINT (ccl_prog[ic]);
1002 CCL_READ_CHAR (reg[rrr]);
1006 case CCL_WriteStringJump: /* A--D--D--R--E--S--S-000XXXXX */
1007 j = XINT (ccl_prog[ic]);
1009 CCL_WRITE_STRING (j);
1013 case CCL_WriteArrayReadJump: /* A--D--D--R--E--S--S-rrrXXXXX */
1015 j = XINT (ccl_prog[ic]);
1016 if ((unsigned int) i < j)
1018 i = XINT (ccl_prog[ic + 1 + i]);
1022 CCL_READ_CHAR (reg[rrr]);
1023 ic += ADDR - (j + 2);
1026 case CCL_ReadJump: /* A--D--D--R--E--S--S-rrrYYYYY */
1027 CCL_READ_CHAR (reg[rrr]);
1031 case CCL_ReadBranch: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
1032 CCL_READ_CHAR (reg[rrr]);
1033 /* fall through ... */
1034 case CCL_Branch: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
1035 if ((unsigned int) reg[rrr] < field1)
1036 ic += XINT (ccl_prog[ic + reg[rrr]]);
1038 ic += XINT (ccl_prog[ic + field1]);
1041 case CCL_ReadRegister: /* CCCCCCCCCCCCCCCCCCCCrrXXXXX */
1044 CCL_READ_CHAR (reg[rrr]);
1046 code = XINT (ccl_prog[ic]); ic++;
1048 field2 = (code & 0xFF) >> 5;
1052 case CCL_WriteExprConst: /* 1:00000OPERATION000RRR000XXXXX */
1055 j = XINT (ccl_prog[ic]);
1057 jump_address = ic + 1;
1060 case CCL_WriteRegister: /* CCCCCCCCCCCCCCCCCCCrrrXXXXX */
1066 code = XINT (ccl_prog[ic]); ic++;
1068 field2 = (code & 0xFF) >> 5;
1072 case CCL_WriteExprRegister: /* 1:00000OPERATIONRrrRRR000XXXXX */
1080 case CCL_Call: /* 1:CCCCCCCCCCCCCCCCCCCCFFFXXXXX */
1085 /* If FFF is nonzero, the CCL program ID is in the
1089 prog_id = XINT (ccl_prog[ic]);
1095 if (stack_idx >= 256
1097 || prog_id >= XVECTOR (Vccl_program_table)->size
1098 || (slot = XVECTOR (Vccl_program_table)->contents[prog_id],
1100 || !VECTORP (XVECTOR (slot)->contents[1]))
1104 ccl_prog = ccl_prog_stack_struct[0].ccl_prog;
1105 ic = ccl_prog_stack_struct[0].ic;
1110 ccl_prog_stack_struct[stack_idx].ccl_prog = ccl_prog;
1111 ccl_prog_stack_struct[stack_idx].ic = ic;
1113 ccl_prog = XVECTOR (XVECTOR (slot)->contents[1])->contents;
1114 ic = CCL_HEADER_MAIN;
1118 case CCL_WriteConstString: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
1120 CCL_WRITE_CHAR (field1);
1123 CCL_WRITE_STRING (field1);
1124 ic += (field1 + 2) / 3;
1128 case CCL_WriteArray: /* CCCCCCCCCCCCCCCCCCCCrrrXXXXX */
1130 if ((unsigned int) i < field1)
1132 j = XINT (ccl_prog[ic + i]);
1138 case CCL_End: /* 0000000000000000000000XXXXX */
1142 ccl_prog = ccl_prog_stack_struct[stack_idx].ccl_prog;
1143 ic = ccl_prog_stack_struct[stack_idx].ic;
1148 /* ccl->ic should points to this command code again to
1149 suppress further processing. */
1153 case CCL_ExprSelfConst: /* 00000OPERATION000000rrrXXXXX */
1154 i = XINT (ccl_prog[ic]);
1159 case CCL_ExprSelfReg: /* 00000OPERATION000RRRrrrXXXXX */
1166 case CCL_PLUS: reg[rrr] += i; break;
1167 case CCL_MINUS: reg[rrr] -= i; break;
1168 case CCL_MUL: reg[rrr] *= i; break;
1169 case CCL_DIV: reg[rrr] /= i; break;
1170 case CCL_MOD: reg[rrr] %= i; break;
1171 case CCL_AND: reg[rrr] &= i; break;
1172 case CCL_OR: reg[rrr] |= i; break;
1173 case CCL_XOR: reg[rrr] ^= i; break;
1174 case CCL_LSH: reg[rrr] <<= i; break;
1175 case CCL_RSH: reg[rrr] >>= i; break;
1176 case CCL_LSH8: reg[rrr] <<= 8; reg[rrr] |= i; break;
1177 case CCL_RSH8: reg[7] = reg[rrr] & 0xFF; reg[rrr] >>= 8; break;
1178 case CCL_DIVMOD: reg[7] = reg[rrr] % i; reg[rrr] /= i; break;
1179 case CCL_LS: reg[rrr] = reg[rrr] < i; break;
1180 case CCL_GT: reg[rrr] = reg[rrr] > i; break;
1181 case CCL_EQ: reg[rrr] = reg[rrr] == i; break;
1182 case CCL_LE: reg[rrr] = reg[rrr] <= i; break;
1183 case CCL_GE: reg[rrr] = reg[rrr] >= i; break;
1184 case CCL_NE: reg[rrr] = reg[rrr] != i; break;
1185 default: CCL_INVALID_CMD;
1189 case CCL_SetExprConst: /* 00000OPERATION000RRRrrrXXXXX */
1191 j = XINT (ccl_prog[ic]);
1193 jump_address = ++ic;
1196 case CCL_SetExprReg: /* 00000OPERATIONRrrRRRrrrXXXXX */
1203 case CCL_ReadJumpCondExprConst: /* A--D--D--R--E--S--S-rrrXXXXX */
1204 CCL_READ_CHAR (reg[rrr]);
1205 case CCL_JumpCondExprConst: /* A--D--D--R--E--S--S-rrrXXXXX */
1207 op = XINT (ccl_prog[ic]);
1208 jump_address = ic++ + ADDR;
1209 j = XINT (ccl_prog[ic]);
1214 case CCL_ReadJumpCondExprReg: /* A--D--D--R--E--S--S-rrrXXXXX */
1215 CCL_READ_CHAR (reg[rrr]);
1216 case CCL_JumpCondExprReg:
1218 op = XINT (ccl_prog[ic]);
1219 jump_address = ic++ + ADDR;
1220 j = reg[XINT (ccl_prog[ic])];
1227 case CCL_PLUS: reg[rrr] = i + j; break;
1228 case CCL_MINUS: reg[rrr] = i - j; break;
1229 case CCL_MUL: reg[rrr] = i * j; break;
1230 case CCL_DIV: reg[rrr] = i / j; break;
1231 case CCL_MOD: reg[rrr] = i % j; break;
1232 case CCL_AND: reg[rrr] = i & j; break;
1233 case CCL_OR: reg[rrr] = i | j; break;
1234 case CCL_XOR: reg[rrr] = i ^ j;; break;
1235 case CCL_LSH: reg[rrr] = i << j; break;
1236 case CCL_RSH: reg[rrr] = i >> j; break;
1237 case CCL_LSH8: reg[rrr] = (i << 8) | j; break;
1238 case CCL_RSH8: reg[rrr] = i >> 8; reg[7] = i & 0xFF; break;
1239 case CCL_DIVMOD: reg[rrr] = i / j; reg[7] = i % j; break;
1240 case CCL_LS: reg[rrr] = i < j; break;
1241 case CCL_GT: reg[rrr] = i > j; break;
1242 case CCL_EQ: reg[rrr] = i == j; break;
1243 case CCL_LE: reg[rrr] = i <= j; break;
1244 case CCL_GE: reg[rrr] = i >= j; break;
1245 case CCL_NE: reg[rrr] = i != j; break;
1246 case CCL_DECODE_SJIS:
1247 /* DECODE_SJIS set MSB for internal format
1248 as opposed to Emacs. */
1249 DECODE_SJIS (i, j, reg[rrr], reg[7]);
1253 case CCL_ENCODE_SJIS:
1254 /* ENCODE_SJIS assumes MSB of SJIS-char is set
1255 as opposed to Emacs. */
1256 ENCODE_SJIS (i | 0x80, j | 0x80, reg[rrr], reg[7]);
1258 default: CCL_INVALID_CMD;
1261 if (code == CCL_WriteExprConst || code == CCL_WriteExprRegister)
1274 case CCL_ReadMultibyteChar2:
1282 goto ccl_read_multibyte_character_suspend;
1290 reg[RRR] = LEADING_BYTE_ASCII;
1292 else if (i <= MAX_LEADING_BYTE_OFFICIAL_1)
1295 goto ccl_read_multibyte_character_suspend;
1297 reg[rrr] = (*src++ & 0x7F);
1299 else if (i <= MAX_LEADING_BYTE_OFFICIAL_2)
1301 if ((src + 1) >= src_end)
1302 goto ccl_read_multibyte_character_suspend;
1304 i = (*src++ & 0x7F);
1305 reg[rrr] = ((i << 7) | (*src & 0x7F));
1308 else if (i == PRE_LEADING_BYTE_PRIVATE_1)
1310 if ((src + 1) >= src_end)
1311 goto ccl_read_multibyte_character_suspend;
1313 reg[rrr] = (*src++ & 0x7F);
1315 else if (i == PRE_LEADING_BYTE_PRIVATE_2)
1317 if ((src + 2) >= src_end)
1318 goto ccl_read_multibyte_character_suspend;
1320 i = (*src++ & 0x7F);
1321 reg[rrr] = ((i << 7) | (*src & 0x7F));
1326 /* INVALID CODE. Return a single byte character. */
1327 reg[RRR] = LEADING_BYTE_ASCII;
1334 ccl_read_multibyte_character_suspend:
1336 if (ccl->last_block)
1342 CCL_SUSPEND (CCL_STAT_SUSPEND_BY_SRC);
1346 case CCL_WriteMultibyteChar2:
1347 i = reg[RRR]; /* charset */
1348 if (i == LEADING_BYTE_ASCII)
1349 i = reg[rrr] & 0xFF;
1350 else if (XCHARSET_DIMENSION (CHARSET_BY_LEADING_BYTE (i)) == 1)
1351 i = (((i - FIELD2_TO_OFFICIAL_LEADING_BYTE) << 7)
1352 | (reg[rrr] & 0x7F));
1353 else if (i < MAX_LEADING_BYTE_OFFICIAL_2)
1354 i = ((i - FIELD1_TO_OFFICIAL_LEADING_BYTE) << 14) | reg[rrr];
1356 i = ((i - FIELD1_TO_PRIVATE_LEADING_BYTE) << 14) | reg[rrr];
1362 case CCL_TranslateCharacter:
1364 /* XEmacs does not have translate_char, and its
1365 equivalent nor. We do nothing on this operation. */
1366 CCL_MAKE_CHAR (reg[RRR], reg[rrr], i);
1367 op = translate_char (GET_TRANSLATION_TABLE (reg[Rrr]),
1369 SPLIT_CHAR (op, reg[RRR], i, j);
1377 case CCL_TranslateCharacterConstTbl:
1379 /* XEmacs does not have translate_char, and its
1380 equivalent nor. We do nothing on this operation. */
1381 op = XINT (ccl_prog[ic]); /* table */
1383 CCL_MAKE_CHAR (reg[RRR], reg[rrr], i);
1384 op = translate_char (GET_TRANSLATION_TABLE (op), i, -1, 0, 0);
1385 SPLIT_CHAR (op, reg[RRR], i, j);
1393 case CCL_IterateMultipleMap:
1395 Lisp_Object map, content, attrib, value;
1396 int point, size, fin_ic;
1398 j = XINT (ccl_prog[ic++]); /* number of maps. */
1401 if ((j > reg[RRR]) && (j >= 0))
1416 size = XVECTOR (Vcode_conversion_map_vector)->size;
1417 point = XINT (ccl_prog[ic++]);
1418 if (point >= size) continue;
1420 XVECTOR (Vcode_conversion_map_vector)->contents[point];
1422 /* Check map validity. */
1423 if (!CONSP (map)) continue;
1425 if (!VECTORP (map)) continue;
1426 size = XVECTOR (map)->size;
1427 if (size <= 1) continue;
1429 content = XVECTOR (map)->contents[0];
1432 [STARTPOINT VAL1 VAL2 ...] or
1433 [t ELEMENT STARTPOINT ENDPOINT] */
1436 point = XUINT (content);
1437 point = op - point + 1;
1438 if (!((point >= 1) && (point < size))) continue;
1439 content = XVECTOR (map)->contents[point];
1441 else if (EQ (content, Qt))
1443 if (size != 4) continue;
1444 if ((op >= XUINT (XVECTOR (map)->contents[2]))
1445 && (op < XUINT (XVECTOR (map)->contents[3])))
1446 content = XVECTOR (map)->contents[1];
1455 else if (INTP (content))
1458 reg[rrr] = XINT(content);
1461 else if (EQ (content, Qt) || EQ (content, Qlambda))
1466 else if (CONSP (content))
1468 attrib = XCAR (content);
1469 value = XCDR (content);
1470 if (!INTP (attrib) || !INTP (value))
1473 reg[rrr] = XUINT (value);
1476 else if (SYMBOLP (content))
1477 CCL_CALL_FOR_MAP_INSTRUCTION (content, fin_ic);
1487 case CCL_MapMultiple:
1489 Lisp_Object map, content, attrib, value;
1490 int point, size, map_vector_size;
1491 int map_set_rest_length, fin_ic;
1492 int current_ic = this_ic;
1494 /* inhibit recursive call on MapMultiple. */
1495 if (stack_idx_of_map_multiple > 0)
1497 if (stack_idx_of_map_multiple <= stack_idx)
1499 stack_idx_of_map_multiple = 0;
1500 mapping_stack_pointer = mapping_stack;
1505 mapping_stack_pointer = mapping_stack;
1506 stack_idx_of_map_multiple = 0;
1508 map_set_rest_length =
1509 XINT (ccl_prog[ic++]); /* number of maps and separators. */
1510 fin_ic = ic + map_set_rest_length;
1513 if ((map_set_rest_length > reg[RRR]) && (reg[RRR] >= 0))
1517 map_set_rest_length -= i;
1523 mapping_stack_pointer = mapping_stack;
1527 if (mapping_stack_pointer <= (mapping_stack + 1))
1529 /* Set up initial state. */
1530 mapping_stack_pointer = mapping_stack;
1531 PUSH_MAPPING_STACK (0, op);
1536 /* Recover after calling other ccl program. */
1539 POP_MAPPING_STACK (map_set_rest_length, orig_op);
1540 POP_MAPPING_STACK (map_set_rest_length, reg[rrr]);
1544 /* Regard it as Qnil. */
1548 map_set_rest_length--;
1551 /* Regard it as Qt. */
1555 map_set_rest_length--;
1558 /* Regard it as Qlambda. */
1560 i += map_set_rest_length;
1561 ic += map_set_rest_length;
1562 map_set_rest_length = 0;
1565 /* Regard it as normal mapping. */
1566 i += map_set_rest_length;
1567 ic += map_set_rest_length;
1568 POP_MAPPING_STACK (map_set_rest_length, reg[rrr]);
1572 map_vector_size = XVECTOR (Vcode_conversion_map_vector)->size;
1575 for (;map_set_rest_length > 0;i++, ic++, map_set_rest_length--)
1577 point = XINT(ccl_prog[ic]);
1580 /* +1 is for including separator. */
1582 if (mapping_stack_pointer
1583 >= &mapping_stack[MAX_MAP_SET_LEVEL])
1585 PUSH_MAPPING_STACK (map_set_rest_length - point,
1587 map_set_rest_length = point;
1592 if (point >= map_vector_size) continue;
1593 map = (XVECTOR (Vcode_conversion_map_vector)
1596 /* Check map validity. */
1597 if (!CONSP (map)) continue;
1599 if (!VECTORP (map)) continue;
1600 size = XVECTOR (map)->size;
1601 if (size <= 1) continue;
1603 content = XVECTOR (map)->contents[0];
1606 [STARTPOINT VAL1 VAL2 ...] or
1607 [t ELEMENT STARTPOINT ENDPOINT] */
1610 point = XUINT (content);
1611 point = op - point + 1;
1612 if (!((point >= 1) && (point < size))) continue;
1613 content = XVECTOR (map)->contents[point];
1615 else if (EQ (content, Qt))
1617 if (size != 4) continue;
1618 if ((op >= XUINT (XVECTOR (map)->contents[2])) &&
1619 (op < XUINT (XVECTOR (map)->contents[3])))
1620 content = XVECTOR (map)->contents[1];
1633 op = XINT (content);
1634 i += map_set_rest_length - 1;
1635 ic += map_set_rest_length - 1;
1636 POP_MAPPING_STACK (map_set_rest_length, reg[rrr]);
1637 map_set_rest_length++;
1639 else if (CONSP (content))
1641 attrib = XCAR (content);
1642 value = XCDR (content);
1643 if (!INTP (attrib) || !INTP (value))
1646 i += map_set_rest_length - 1;
1647 ic += map_set_rest_length - 1;
1648 POP_MAPPING_STACK (map_set_rest_length, reg[rrr]);
1649 map_set_rest_length++;
1651 else if (EQ (content, Qt))
1655 else if (EQ (content, Qlambda))
1657 i += map_set_rest_length;
1658 ic += map_set_rest_length;
1661 else if (SYMBOLP (content))
1663 if (mapping_stack_pointer
1664 >= &mapping_stack[MAX_MAP_SET_LEVEL])
1666 PUSH_MAPPING_STACK (map_set_rest_length, reg[rrr]);
1667 PUSH_MAPPING_STACK (map_set_rest_length, op);
1668 stack_idx_of_map_multiple = stack_idx + 1;
1669 CCL_CALL_FOR_MAP_INSTRUCTION (content, current_ic);
1674 if (mapping_stack_pointer <= (mapping_stack + 1))
1676 POP_MAPPING_STACK (map_set_rest_length, reg[rrr]);
1677 i += map_set_rest_length;
1678 ic += map_set_rest_length;
1679 POP_MAPPING_STACK (map_set_rest_length, reg[rrr]);
1689 Lisp_Object map, attrib, value, content;
1691 j = XINT (ccl_prog[ic++]); /* map_id */
1693 if (j >= XVECTOR (Vcode_conversion_map_vector)->size)
1698 map = XVECTOR (Vcode_conversion_map_vector)->contents[j];
1710 size = XVECTOR (map)->size;
1711 point = XUINT (XVECTOR (map)->contents[0]);
1712 point = op - point + 1;
1715 (!((point >= 1) && (point < size))))
1720 content = XVECTOR (map)->contents[point];
1723 else if (INTP (content))
1724 reg[rrr] = XINT (content);
1725 else if (EQ (content, Qt));
1726 else if (CONSP (content))
1728 attrib = XCAR (content);
1729 value = XCDR (content);
1730 if (!INTP (attrib) || !INTP (value))
1732 reg[rrr] = XUINT(value);
1735 else if (SYMBOLP (content))
1736 CCL_CALL_FOR_MAP_INSTRUCTION (content, ic);
1756 /* We can insert an error message only if DESTINATION is
1757 specified and we still have a room to store the message
1761 switch (ccl->status)
1763 case CCL_STAT_INVALID_CMD:
1764 sprintf(msg, "\nCCL: Invalid command %x (ccl_code = %x) at %d.",
1765 code & 0x1F, code, this_ic);
1768 int i = ccl_backtrace_idx - 1;
1771 Dynarr_add_many (destination, (unsigned char *) msg, strlen (msg));
1773 for (j = 0; j < CCL_DEBUG_BACKTRACE_LEN; j++, i--)
1775 if (i < 0) i = CCL_DEBUG_BACKTRACE_LEN - 1;
1776 if (ccl_backtrace_table[i] == 0)
1778 sprintf(msg, " %d", ccl_backtrace_table[i]);
1779 Dynarr_add_many (destination, (unsigned char *) msg, strlen (msg));
1787 sprintf(msg, "\nCCL: Exited.");
1791 sprintf(msg, "\nCCL: Unknown error type (%d).", ccl->status);
1794 Dynarr_add_many (destination, (unsigned char *) msg, strlen (msg));
1799 ccl->stack_idx = stack_idx;
1800 ccl->prog = ccl_prog;
1801 if (consumed) *consumed = src - source;
1804 return Dynarr_length (destination);
1807 /* Resolve symbols in the specified CCL code (Lisp vector). This
1808 function converts symbols of code conversion maps and character
1809 translation tables embedded in the CCL code into their ID numbers.
1811 The return value is a vector (CCL itself or a new vector in which
1812 all symbols are resolved), Qt if resolving of some symbol failed,
1813 or nil if CCL contains invalid data. */
1816 resolve_symbol_ccl_program (Lisp_Object ccl)
1818 int i, veclen, unresolved = 0;
1819 Lisp_Object result, contents, val;
1822 veclen = XVECTOR (result)->size;
1824 for (i = 0; i < veclen; i++)
1826 contents = XVECTOR (result)->contents[i];
1827 if (INTP (contents))
1829 else if (CONSP (contents)
1830 && SYMBOLP (XCAR (contents))
1831 && SYMBOLP (XCDR (contents)))
1833 /* This is the new style for embedding symbols. The form is
1834 (SYMBOL . PROPERTY). (get SYMBOL PROPERTY) should give
1837 if (EQ (result, ccl))
1838 result = Fcopy_sequence (ccl);
1840 val = Fget (XCAR (contents), XCDR (contents), Qnil);
1842 XVECTOR (result)->contents[i] = val;
1847 else if (SYMBOLP (contents))
1849 /* This is the old style for embedding symbols. This style
1850 may lead to a bug if, for instance, a translation table
1851 and a code conversion map have the same name. */
1852 if (EQ (result, ccl))
1853 result = Fcopy_sequence (ccl);
1855 val = Fget (contents, Qcode_conversion_map_id, Qnil);
1857 XVECTOR (result)->contents[i] = val;
1860 val = Fget (contents, Qccl_program_idx, Qnil);
1862 XVECTOR (result)->contents[i] = val;
1871 return (unresolved ? Qt : result);
1874 /* Return the compiled code (vector) of CCL program CCL_PROG.
1875 CCL_PROG is a name (symbol) of the program or already compiled
1876 code. If necessary, resolve symbols in the compiled code to index
1877 numbers. If we failed to get the compiled code or to resolve
1878 symbols, return Qnil. */
1881 ccl_get_compiled_code (Lisp_Object ccl_prog)
1883 Lisp_Object val, slot;
1885 if (VECTORP (ccl_prog))
1887 val = resolve_symbol_ccl_program (ccl_prog);
1888 return (VECTORP (val) ? val : Qnil);
1890 if (!SYMBOLP (ccl_prog))
1893 val = Fget (ccl_prog, Qccl_program_idx, Qnil);
1895 || XINT (val) >= XVECTOR_LENGTH (Vccl_program_table))
1897 slot = XVECTOR_DATA (Vccl_program_table)[XINT (val)];
1898 if (! VECTORP (slot)
1899 || XVECTOR (slot)->size != 3
1900 || ! VECTORP (XVECTOR_DATA (slot)[1]))
1902 if (NILP (XVECTOR_DATA (slot)[2]))
1904 val = resolve_symbol_ccl_program (XVECTOR_DATA (slot)[1]);
1905 if (! VECTORP (val))
1907 XVECTOR_DATA (slot)[1] = val;
1908 XVECTOR_DATA (slot)[2] = Qt;
1910 return XVECTOR_DATA (slot)[1];
1913 /* Setup fields of the structure pointed by CCL appropriately for the
1914 execution of CCL program CCL_PROG. CCL_PROG is the name (symbol)
1915 of the CCL program or the already compiled code (vector).
1916 Return 0 if we succeed this setup, else return -1.
1918 If CCL_PROG is nil, we just reset the structure pointed by CCL. */
1920 setup_ccl_program (struct ccl_program *ccl, Lisp_Object ccl_prog)
1924 if (! NILP (ccl_prog))
1926 ccl_prog = ccl_get_compiled_code (ccl_prog);
1927 if (! VECTORP (ccl_prog))
1929 ccl->size = XVECTOR_LENGTH (ccl_prog);
1930 ccl->prog = XVECTOR_DATA (ccl_prog);
1931 ccl->eof_ic = XINT (XVECTOR_DATA (ccl_prog)[CCL_HEADER_EOF]);
1932 ccl->buf_magnification = XINT (XVECTOR_DATA (ccl_prog)[CCL_HEADER_BUF_MAG]);
1934 ccl->ic = CCL_HEADER_MAIN;
1935 for (i = 0; i < 8; i++)
1937 ccl->last_block = 0;
1938 ccl->private_state = 0;
1941 ccl->eol_type = CCL_CODING_EOL_LF;
1947 DEFUN ("ccl-program-p", Fccl_program_p, 1, 1, 0, /*
1948 Return t if OBJECT is a CCL program name or a compiled CCL program code.
1949 See the documentation of `define-ccl-program' for the detail of CCL program.
1955 if (VECTORP (object))
1957 val = resolve_symbol_ccl_program (object);
1958 return (VECTORP (val) ? Qt : Qnil);
1960 if (!SYMBOLP (object))
1963 val = Fget (object, Qccl_program_idx, Qnil);
1964 return ((! NATNUMP (val)
1965 || XINT (val) >= XVECTOR_LENGTH (Vccl_program_table))
1969 DEFUN ("ccl-execute", Fccl_execute, 2, 2, 0, /*
1970 Execute CCL-PROGRAM with registers initialized by REGISTERS.
1972 CCL-PROGRAM is a CCL program name (symbol)
1973 or a compiled code generated by `ccl-compile' (for backward compatibility,
1974 in this case, the overhead of the execution is bigger than the former case).
1975 No I/O commands should appear in CCL-PROGRAM.
1977 REGISTERS is a vector of [R0 R1 ... R7] where RN is an initial value
1980 As side effect, each element of REGISTERS holds the value of
1981 corresponding register after the execution.
1983 See the documentation of `define-ccl-program' for the detail of CCL program.
1987 struct ccl_program ccl;
1990 if (setup_ccl_program (&ccl, ccl_prog) < 0)
1991 error ("Invalid CCL program");
1994 if (XVECTOR_LENGTH (reg) != 8)
1995 error ("Length of vector REGISTERS is not 8");
1997 for (i = 0; i < 8; i++)
1998 ccl.reg[i] = (INTP (XVECTOR_DATA (reg)[i])
1999 ? XINT (XVECTOR_DATA (reg)[i])
2002 ccl_driver (&ccl, (const unsigned char *)0,
2003 (unsigned_char_dynarr *)0, 0, (int *)0,
2006 if (ccl.status != CCL_STAT_SUCCESS)
2007 error ("Error in CCL program at %dth code", ccl.ic);
2009 for (i = 0; i < 8; i++)
2010 XSETINT (XVECTOR (reg)->contents[i], ccl.reg[i]);
2014 DEFUN ("ccl-execute-on-string", Fccl_execute_on_string,
2016 Execute CCL-PROGRAM with initial STATUS on STRING.
2018 CCL-PROGRAM is a symbol registered by register-ccl-program,
2019 or a compiled code generated by `ccl-compile' (for backward compatibility,
2020 in this case, the execution is slower).
2022 Read buffer is set to STRING, and write buffer is allocated automatically.
2024 STATUS is a vector of [R0 R1 ... R7 IC], where
2025 R0..R7 are initial values of corresponding registers,
2026 IC is the instruction counter specifying from where to start the program.
2027 If R0..R7 are nil, they are initialized to 0.
2028 If IC is nil, it is initialized to head of the CCL program.
2030 If optional 4th arg CONTINUE is non-nil, keep IC on read operation
2031 when read buffer is exhausted, else, IC is always set to the end of
2032 CCL-PROGRAM on exit.
2034 It returns the contents of write buffer as a string,
2035 and as side effect, STATUS is updated.
2037 See the documentation of `define-ccl-program' for the detail of CCL program.
2039 (ccl_prog, status, string, continue_))
2042 struct ccl_program ccl;
2044 unsigned_char_dynarr *outbuf;
2045 struct gcpro gcpro1, gcpro2;
2047 if (setup_ccl_program (&ccl, ccl_prog) < 0)
2048 error ("Invalid CCL program");
2050 CHECK_VECTOR (status);
2051 if (XVECTOR (status)->size != 9)
2052 error ("Length of vector STATUS is not 9");
2053 CHECK_STRING (string);
2055 GCPRO2 (status, string);
2057 for (i = 0; i < 8; i++)
2059 if (NILP (XVECTOR_DATA (status)[i]))
2060 XSETINT (XVECTOR_DATA (status)[i], 0);
2061 if (INTP (XVECTOR_DATA (status)[i]))
2062 ccl.reg[i] = XINT (XVECTOR_DATA (status)[i]);
2064 if (INTP (XVECTOR (status)->contents[i]))
2066 i = XINT (XVECTOR_DATA (status)[8]);
2067 if (ccl.ic < i && i < ccl.size)
2070 outbuf = Dynarr_new (unsigned_char);
2071 ccl.last_block = NILP (continue_);
2072 produced = ccl_driver (&ccl, XSTRING_DATA (string), outbuf,
2073 XSTRING_LENGTH (string),
2076 for (i = 0; i < 8; i++)
2077 XSETINT (XVECTOR_DATA (status)[i], ccl.reg[i]);
2078 XSETINT (XVECTOR_DATA (status)[8], ccl.ic);
2081 val = make_string (Dynarr_atp (outbuf, 0), produced);
2082 Dynarr_free (outbuf);
2084 if (ccl.status == CCL_STAT_SUSPEND_BY_DST)
2085 error ("Output buffer for the CCL programs overflow");
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);
2093 DEFUN ("register-ccl-program", Fregister_ccl_program,
2095 Register CCL program CCL-PROG as NAME in `ccl-program-table'.
2096 CCL-PROG should be a compiled CCL program (vector), or nil.
2097 If it is nil, just reserve NAME as a CCL program name.
2098 Return index number of the registered CCL program.
2102 int len = XVECTOR_LENGTH (Vccl_program_table);
2104 Lisp_Object resolved;
2106 CHECK_SYMBOL (name);
2108 if (!NILP (ccl_prog))
2110 CHECK_VECTOR (ccl_prog);
2111 resolved = resolve_symbol_ccl_program (ccl_prog);
2112 if (! NILP (resolved))
2114 ccl_prog = resolved;
2119 for (idx = 0; idx < len; idx++)
2123 slot = XVECTOR_DATA (Vccl_program_table)[idx];
2124 if (!VECTORP (slot))
2125 /* This is the first unused slot. Register NAME here. */
2128 if (EQ (name, XVECTOR_DATA (slot)[0]))
2130 /* Update this slot. */
2131 XVECTOR_DATA (slot)[1] = ccl_prog;
2132 XVECTOR_DATA (slot)[2] = resolved;
2133 return make_int (idx);
2139 /* Extend the table. */
2140 Lisp_Object new_table;
2143 new_table = Fmake_vector (make_int (len * 2), Qnil);
2144 for (j = 0; j < len; j++)
2145 XVECTOR_DATA (new_table)[j]
2146 = XVECTOR_DATA (Vccl_program_table)[j];
2147 Vccl_program_table = new_table;
2153 elt = Fmake_vector (make_int (3), Qnil);
2154 XVECTOR_DATA (elt)[0] = name;
2155 XVECTOR_DATA (elt)[1] = ccl_prog;
2156 XVECTOR_DATA (elt)[2] = resolved;
2157 XVECTOR_DATA (Vccl_program_table)[idx] = elt;
2160 Fput (name, Qccl_program_idx, make_int (idx));
2161 return make_int (idx);
2164 /* Register code conversion map.
2165 A code conversion map consists of numbers, Qt, Qnil, and Qlambda.
2166 The first element is start code point.
2167 The rest elements are mapped numbers.
2168 Symbol t means to map to an original number before mapping.
2169 Symbol nil means that the corresponding element is empty.
2170 Symbol lambda means to terminate mapping here.
2173 DEFUN ("register-code-conversion-map", Fregister_code_conversion_map,
2175 Register SYMBOL as code conversion map MAP.
2176 Return index number of the registered map.
2180 int len = XVECTOR_LENGTH (Vcode_conversion_map_vector);
2184 CHECK_SYMBOL (symbol);
2187 for (i = 0; i < len; i++)
2189 Lisp_Object slot = XVECTOR_DATA (Vcode_conversion_map_vector)[i];
2194 if (EQ (symbol, XCAR (slot)))
2198 Fput (symbol, Qcode_conversion_map, map);
2199 Fput (symbol, Qcode_conversion_map_id, idx);
2206 Lisp_Object new_vector = Fmake_vector (make_int (len * 2), Qnil);
2209 for (j = 0; j < len; j++)
2210 XVECTOR_DATA (new_vector)[j]
2211 = XVECTOR_DATA (Vcode_conversion_map_vector)[j];
2212 Vcode_conversion_map_vector = new_vector;
2216 Fput (symbol, Qcode_conversion_map, map);
2217 Fput (symbol, Qcode_conversion_map_id, idx);
2218 XVECTOR_DATA (Vcode_conversion_map_vector)[i] = Fcons (symbol, map);
2224 syms_of_mule_ccl (void)
2226 DEFSUBR (Fccl_program_p);
2227 DEFSUBR (Fccl_execute);
2228 DEFSUBR (Fccl_execute_on_string);
2229 DEFSUBR (Fregister_ccl_program);
2230 DEFSUBR (Fregister_code_conversion_map);
2234 vars_of_mule_ccl (void)
2236 staticpro (&Vccl_program_table);
2237 Vccl_program_table = Fmake_vector (make_int (32), Qnil);
2239 defsymbol (&Qccl_program, "ccl-program");
2240 defsymbol (&Qccl_program_idx, "ccl-program-idx");
2241 defsymbol (&Qcode_conversion_map, "code-conversion-map");
2242 defsymbol (&Qcode_conversion_map_id, "code-conversion-map-id");
2244 DEFVAR_LISP ("code-conversion-map-vector", &Vcode_conversion_map_vector /*
2245 Vector of code conversion maps.
2247 Vcode_conversion_map_vector = Fmake_vector (make_int (16), Qnil);
2249 DEFVAR_LISP ("font-ccl-encoder-alist", &Vfont_ccl_encoder_alist /*
2250 Alist of fontname patterns vs corresponding CCL program.
2251 Each element looks like (REGEXP . CCL-CODE),
2252 where CCL-CODE is a compiled CCL program.
2253 When a font whose name matches REGEXP is used for displaying a character,
2254 CCL-CODE is executed to calculate the code point in the font
2255 from the charset number and position code(s) of the character which are set
2256 in CCL registers R0, R1, and R2 before the execution.
2257 The code point in the font is set in CCL registers R1 and R2
2258 when the execution terminated.
2259 If the font is single-byte font, the register R2 is not used.
2261 Vfont_ccl_encoder_alist = Qnil;