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 "character.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)
1275 case CCL_ReadMultibyteChar2:
1283 goto ccl_read_multibyte_character_suspend;
1291 reg[RRR] = LEADING_BYTE_ASCII;
1293 else if (i <= MAX_LEADING_BYTE_OFFICIAL_1)
1296 goto ccl_read_multibyte_character_suspend;
1298 reg[rrr] = (*src++ & 0x7F);
1300 else if (i <= MAX_LEADING_BYTE_OFFICIAL_2)
1302 if ((src + 1) >= src_end)
1303 goto ccl_read_multibyte_character_suspend;
1305 i = (*src++ & 0x7F);
1306 reg[rrr] = ((i << 7) | (*src & 0x7F));
1309 else if (i == PRE_LEADING_BYTE_PRIVATE_1)
1311 if ((src + 1) >= src_end)
1312 goto ccl_read_multibyte_character_suspend;
1314 reg[rrr] = (*src++ & 0x7F);
1316 else if (i == PRE_LEADING_BYTE_PRIVATE_2)
1318 if ((src + 2) >= src_end)
1319 goto ccl_read_multibyte_character_suspend;
1321 i = (*src++ & 0x7F);
1322 reg[rrr] = ((i << 7) | (*src & 0x7F));
1327 /* INVALID CODE. Return a single byte character. */
1328 reg[RRR] = LEADING_BYTE_ASCII;
1335 ccl_read_multibyte_character_suspend:
1337 if (ccl->last_block)
1343 CCL_SUSPEND (CCL_STAT_SUSPEND_BY_SRC);
1349 case CCL_WriteMultibyteChar2:
1350 i = reg[RRR]; /* charset */
1351 if (i == LEADING_BYTE_ASCII)
1352 i = reg[rrr] & 0xFF;
1353 else if (XCHARSET_DIMENSION (CHARSET_BY_LEADING_BYTE (i)) == 1)
1354 i = (((i - FIELD2_TO_OFFICIAL_LEADING_BYTE) << 7)
1355 | (reg[rrr] & 0x7F));
1356 else if (i < MAX_LEADING_BYTE_OFFICIAL_2)
1357 i = ((i - FIELD1_TO_OFFICIAL_LEADING_BYTE) << 14) | reg[rrr];
1359 i = ((i - FIELD1_TO_PRIVATE_LEADING_BYTE) << 14) | reg[rrr];
1366 case CCL_TranslateCharacter:
1368 /* XEmacs does not have translate_char, and its
1369 equivalent nor. We do nothing on this operation. */
1370 CCL_MAKE_CHAR (reg[RRR], reg[rrr], i);
1371 op = translate_char (GET_TRANSLATION_TABLE (reg[Rrr]),
1373 SPLIT_CHAR (op, reg[RRR], i, j);
1381 case CCL_TranslateCharacterConstTbl:
1383 /* XEmacs does not have translate_char, and its
1384 equivalent nor. We do nothing on this operation. */
1385 op = XINT (ccl_prog[ic]); /* table */
1387 CCL_MAKE_CHAR (reg[RRR], reg[rrr], i);
1388 op = translate_char (GET_TRANSLATION_TABLE (op), i, -1, 0, 0);
1389 SPLIT_CHAR (op, reg[RRR], i, j);
1397 case CCL_IterateMultipleMap:
1399 Lisp_Object map, content, attrib, value;
1400 int point, size, fin_ic;
1402 j = XINT (ccl_prog[ic++]); /* number of maps. */
1405 if ((j > reg[RRR]) && (j >= 0))
1420 size = XVECTOR (Vcode_conversion_map_vector)->size;
1421 point = XINT (ccl_prog[ic++]);
1422 if (point >= size) continue;
1424 XVECTOR (Vcode_conversion_map_vector)->contents[point];
1426 /* Check map validity. */
1427 if (!CONSP (map)) continue;
1429 if (!VECTORP (map)) continue;
1430 size = XVECTOR (map)->size;
1431 if (size <= 1) continue;
1433 content = XVECTOR (map)->contents[0];
1436 [STARTPOINT VAL1 VAL2 ...] or
1437 [t ELEMENT STARTPOINT ENDPOINT] */
1440 point = XUINT (content);
1441 point = op - point + 1;
1442 if (!((point >= 1) && (point < size))) continue;
1443 content = XVECTOR (map)->contents[point];
1445 else if (EQ (content, Qt))
1447 if (size != 4) continue;
1448 if ((op >= XUINT (XVECTOR (map)->contents[2]))
1449 && (op < XUINT (XVECTOR (map)->contents[3])))
1450 content = XVECTOR (map)->contents[1];
1459 else if (INTP (content))
1462 reg[rrr] = XINT(content);
1465 else if (EQ (content, Qt) || EQ (content, Qlambda))
1470 else if (CONSP (content))
1472 attrib = XCAR (content);
1473 value = XCDR (content);
1474 if (!INTP (attrib) || !INTP (value))
1477 reg[rrr] = XUINT (value);
1480 else if (SYMBOLP (content))
1481 CCL_CALL_FOR_MAP_INSTRUCTION (content, fin_ic);
1491 case CCL_MapMultiple:
1493 Lisp_Object map, content, attrib, value;
1494 int point, size, map_vector_size;
1495 int map_set_rest_length, fin_ic;
1496 int current_ic = this_ic;
1498 /* inhibit recursive call on MapMultiple. */
1499 if (stack_idx_of_map_multiple > 0)
1501 if (stack_idx_of_map_multiple <= stack_idx)
1503 stack_idx_of_map_multiple = 0;
1504 mapping_stack_pointer = mapping_stack;
1509 mapping_stack_pointer = mapping_stack;
1510 stack_idx_of_map_multiple = 0;
1512 map_set_rest_length =
1513 XINT (ccl_prog[ic++]); /* number of maps and separators. */
1514 fin_ic = ic + map_set_rest_length;
1517 if ((map_set_rest_length > reg[RRR]) && (reg[RRR] >= 0))
1521 map_set_rest_length -= i;
1527 mapping_stack_pointer = mapping_stack;
1531 if (mapping_stack_pointer <= (mapping_stack + 1))
1533 /* Set up initial state. */
1534 mapping_stack_pointer = mapping_stack;
1535 PUSH_MAPPING_STACK (0, op);
1540 /* Recover after calling other ccl program. */
1543 POP_MAPPING_STACK (map_set_rest_length, orig_op);
1544 POP_MAPPING_STACK (map_set_rest_length, reg[rrr]);
1548 /* Regard it as Qnil. */
1552 map_set_rest_length--;
1555 /* Regard it as Qt. */
1559 map_set_rest_length--;
1562 /* Regard it as Qlambda. */
1564 i += map_set_rest_length;
1565 ic += map_set_rest_length;
1566 map_set_rest_length = 0;
1569 /* Regard it as normal mapping. */
1570 i += map_set_rest_length;
1571 ic += map_set_rest_length;
1572 POP_MAPPING_STACK (map_set_rest_length, reg[rrr]);
1576 map_vector_size = XVECTOR (Vcode_conversion_map_vector)->size;
1579 for (;map_set_rest_length > 0;i++, ic++, map_set_rest_length--)
1581 point = XINT(ccl_prog[ic]);
1584 /* +1 is for including separator. */
1586 if (mapping_stack_pointer
1587 >= &mapping_stack[MAX_MAP_SET_LEVEL])
1589 PUSH_MAPPING_STACK (map_set_rest_length - point,
1591 map_set_rest_length = point;
1596 if (point >= map_vector_size) continue;
1597 map = (XVECTOR (Vcode_conversion_map_vector)
1600 /* Check map validity. */
1601 if (!CONSP (map)) continue;
1603 if (!VECTORP (map)) continue;
1604 size = XVECTOR (map)->size;
1605 if (size <= 1) continue;
1607 content = XVECTOR (map)->contents[0];
1610 [STARTPOINT VAL1 VAL2 ...] or
1611 [t ELEMENT STARTPOINT ENDPOINT] */
1614 point = XUINT (content);
1615 point = op - point + 1;
1616 if (!((point >= 1) && (point < size))) continue;
1617 content = XVECTOR (map)->contents[point];
1619 else if (EQ (content, Qt))
1621 if (size != 4) continue;
1622 if ((op >= XUINT (XVECTOR (map)->contents[2])) &&
1623 (op < XUINT (XVECTOR (map)->contents[3])))
1624 content = XVECTOR (map)->contents[1];
1637 op = XINT (content);
1638 i += map_set_rest_length - 1;
1639 ic += map_set_rest_length - 1;
1640 POP_MAPPING_STACK (map_set_rest_length, reg[rrr]);
1641 map_set_rest_length++;
1643 else if (CONSP (content))
1645 attrib = XCAR (content);
1646 value = XCDR (content);
1647 if (!INTP (attrib) || !INTP (value))
1650 i += map_set_rest_length - 1;
1651 ic += map_set_rest_length - 1;
1652 POP_MAPPING_STACK (map_set_rest_length, reg[rrr]);
1653 map_set_rest_length++;
1655 else if (EQ (content, Qt))
1659 else if (EQ (content, Qlambda))
1661 i += map_set_rest_length;
1662 ic += map_set_rest_length;
1665 else if (SYMBOLP (content))
1667 if (mapping_stack_pointer
1668 >= &mapping_stack[MAX_MAP_SET_LEVEL])
1670 PUSH_MAPPING_STACK (map_set_rest_length, reg[rrr]);
1671 PUSH_MAPPING_STACK (map_set_rest_length, op);
1672 stack_idx_of_map_multiple = stack_idx + 1;
1673 CCL_CALL_FOR_MAP_INSTRUCTION (content, current_ic);
1678 if (mapping_stack_pointer <= (mapping_stack + 1))
1680 POP_MAPPING_STACK (map_set_rest_length, reg[rrr]);
1681 i += map_set_rest_length;
1682 ic += map_set_rest_length;
1683 POP_MAPPING_STACK (map_set_rest_length, reg[rrr]);
1693 Lisp_Object map, attrib, value, content;
1695 j = XINT (ccl_prog[ic++]); /* map_id */
1697 if (j >= XVECTOR (Vcode_conversion_map_vector)->size)
1702 map = XVECTOR (Vcode_conversion_map_vector)->contents[j];
1714 size = XVECTOR (map)->size;
1715 point = XUINT (XVECTOR (map)->contents[0]);
1716 point = op - point + 1;
1719 (!((point >= 1) && (point < size))))
1724 content = XVECTOR (map)->contents[point];
1727 else if (INTP (content))
1728 reg[rrr] = XINT (content);
1729 else if (EQ (content, Qt));
1730 else if (CONSP (content))
1732 attrib = XCAR (content);
1733 value = XCDR (content);
1734 if (!INTP (attrib) || !INTP (value))
1736 reg[rrr] = XUINT(value);
1739 else if (SYMBOLP (content))
1740 CCL_CALL_FOR_MAP_INSTRUCTION (content, ic);
1760 /* We can insert an error message only if DESTINATION is
1761 specified and we still have a room to store the message
1765 switch (ccl->status)
1767 case CCL_STAT_INVALID_CMD:
1768 sprintf(msg, "\nCCL: Invalid command %x (ccl_code = %x) at %d.",
1769 code & 0x1F, code, this_ic);
1772 int i = ccl_backtrace_idx - 1;
1775 Dynarr_add_many (destination, (unsigned char *) msg, strlen (msg));
1777 for (j = 0; j < CCL_DEBUG_BACKTRACE_LEN; j++, i--)
1779 if (i < 0) i = CCL_DEBUG_BACKTRACE_LEN - 1;
1780 if (ccl_backtrace_table[i] == 0)
1782 sprintf(msg, " %d", ccl_backtrace_table[i]);
1783 Dynarr_add_many (destination, (unsigned char *) msg, strlen (msg));
1791 sprintf(msg, "\nCCL: Exited.");
1795 sprintf(msg, "\nCCL: Unknown error type (%d).", ccl->status);
1798 Dynarr_add_many (destination, (unsigned char *) msg, strlen (msg));
1803 ccl->stack_idx = stack_idx;
1804 ccl->prog = ccl_prog;
1805 if (consumed) *consumed = src - source;
1808 return Dynarr_length (destination);
1811 /* Resolve symbols in the specified CCL code (Lisp vector). This
1812 function converts symbols of code conversion maps and character
1813 translation tables embedded in the CCL code into their ID numbers.
1815 The return value is a vector (CCL itself or a new vector in which
1816 all symbols are resolved), Qt if resolving of some symbol failed,
1817 or nil if CCL contains invalid data. */
1820 resolve_symbol_ccl_program (Lisp_Object ccl)
1822 int i, veclen, unresolved = 0;
1823 Lisp_Object result, contents, val;
1826 veclen = XVECTOR (result)->size;
1828 for (i = 0; i < veclen; i++)
1830 contents = XVECTOR (result)->contents[i];
1831 if (INTP (contents))
1833 else if (CONSP (contents)
1834 && SYMBOLP (XCAR (contents))
1835 && SYMBOLP (XCDR (contents)))
1837 /* This is the new style for embedding symbols. The form is
1838 (SYMBOL . PROPERTY). (get SYMBOL PROPERTY) should give
1841 if (EQ (result, ccl))
1842 result = Fcopy_sequence (ccl);
1844 val = Fget (XCAR (contents), XCDR (contents), Qnil);
1846 XVECTOR (result)->contents[i] = val;
1851 else if (SYMBOLP (contents))
1853 /* This is the old style for embedding symbols. This style
1854 may lead to a bug if, for instance, a translation table
1855 and a code conversion map have the same name. */
1856 if (EQ (result, ccl))
1857 result = Fcopy_sequence (ccl);
1859 val = Fget (contents, Qcode_conversion_map_id, Qnil);
1861 XVECTOR (result)->contents[i] = val;
1864 val = Fget (contents, Qccl_program_idx, Qnil);
1866 XVECTOR (result)->contents[i] = val;
1875 return (unresolved ? Qt : result);
1878 /* Return the compiled code (vector) of CCL program CCL_PROG.
1879 CCL_PROG is a name (symbol) of the program or already compiled
1880 code. If necessary, resolve symbols in the compiled code to index
1881 numbers. If we failed to get the compiled code or to resolve
1882 symbols, return Qnil. */
1885 ccl_get_compiled_code (Lisp_Object ccl_prog)
1887 Lisp_Object val, slot;
1889 if (VECTORP (ccl_prog))
1891 val = resolve_symbol_ccl_program (ccl_prog);
1892 return (VECTORP (val) ? val : Qnil);
1894 if (!SYMBOLP (ccl_prog))
1897 val = Fget (ccl_prog, Qccl_program_idx, Qnil);
1899 || XINT (val) >= XVECTOR_LENGTH (Vccl_program_table))
1901 slot = XVECTOR_DATA (Vccl_program_table)[XINT (val)];
1902 if (! VECTORP (slot)
1903 || XVECTOR (slot)->size != 3
1904 || ! VECTORP (XVECTOR_DATA (slot)[1]))
1906 if (NILP (XVECTOR_DATA (slot)[2]))
1908 val = resolve_symbol_ccl_program (XVECTOR_DATA (slot)[1]);
1909 if (! VECTORP (val))
1911 XVECTOR_DATA (slot)[1] = val;
1912 XVECTOR_DATA (slot)[2] = Qt;
1914 return XVECTOR_DATA (slot)[1];
1917 /* Setup fields of the structure pointed by CCL appropriately for the
1918 execution of CCL program CCL_PROG. CCL_PROG is the name (symbol)
1919 of the CCL program or the already compiled code (vector).
1920 Return 0 if we succeed this setup, else return -1.
1922 If CCL_PROG is nil, we just reset the structure pointed by CCL. */
1924 setup_ccl_program (struct ccl_program *ccl, Lisp_Object ccl_prog)
1928 if (! NILP (ccl_prog))
1930 ccl_prog = ccl_get_compiled_code (ccl_prog);
1931 if (! VECTORP (ccl_prog))
1933 ccl->size = XVECTOR_LENGTH (ccl_prog);
1934 ccl->prog = XVECTOR_DATA (ccl_prog);
1935 ccl->eof_ic = XINT (XVECTOR_DATA (ccl_prog)[CCL_HEADER_EOF]);
1936 ccl->buf_magnification = XINT (XVECTOR_DATA (ccl_prog)[CCL_HEADER_BUF_MAG]);
1938 ccl->ic = CCL_HEADER_MAIN;
1939 for (i = 0; i < 8; i++)
1941 ccl->last_block = 0;
1942 ccl->private_state = 0;
1945 ccl->eol_type = CCL_CODING_EOL_LF;
1951 DEFUN ("ccl-program-p", Fccl_program_p, 1, 1, 0, /*
1952 Return t if OBJECT is a CCL program name or a compiled CCL program code.
1953 See the documentation of `define-ccl-program' for the detail of CCL program.
1959 if (VECTORP (object))
1961 val = resolve_symbol_ccl_program (object);
1962 return (VECTORP (val) ? Qt : Qnil);
1964 if (!SYMBOLP (object))
1967 val = Fget (object, Qccl_program_idx, Qnil);
1968 return ((! NATNUMP (val)
1969 || XINT (val) >= XVECTOR_LENGTH (Vccl_program_table))
1973 DEFUN ("ccl-execute", Fccl_execute, 2, 2, 0, /*
1974 Execute CCL-PROGRAM with registers initialized by REGISTERS.
1976 CCL-PROGRAM is a CCL program name (symbol)
1977 or a compiled code generated by `ccl-compile' (for backward compatibility,
1978 in this case, the overhead of the execution is bigger than the former case).
1979 No I/O commands should appear in CCL-PROGRAM.
1981 REGISTERS is a vector of [R0 R1 ... R7] where RN is an initial value
1984 As side effect, each element of REGISTERS holds the value of
1985 corresponding register after the execution.
1987 See the documentation of `define-ccl-program' for the detail of CCL program.
1991 struct ccl_program ccl;
1994 if (setup_ccl_program (&ccl, ccl_prog) < 0)
1995 error ("Invalid CCL program");
1998 if (XVECTOR_LENGTH (reg) != 8)
1999 error ("Length of vector REGISTERS is not 8");
2001 for (i = 0; i < 8; i++)
2002 ccl.reg[i] = (INTP (XVECTOR_DATA (reg)[i])
2003 ? XINT (XVECTOR_DATA (reg)[i])
2006 ccl_driver (&ccl, (const unsigned char *)0,
2007 (unsigned_char_dynarr *)0, 0, (int *)0,
2010 if (ccl.status != CCL_STAT_SUCCESS)
2011 error ("Error in CCL program at %dth code", ccl.ic);
2013 for (i = 0; i < 8; i++)
2014 XSETINT (XVECTOR (reg)->contents[i], ccl.reg[i]);
2018 DEFUN ("ccl-execute-on-string", Fccl_execute_on_string,
2020 Execute CCL-PROGRAM with initial STATUS on STRING.
2022 CCL-PROGRAM is a symbol registered by register-ccl-program,
2023 or a compiled code generated by `ccl-compile' (for backward compatibility,
2024 in this case, the execution is slower).
2026 Read buffer is set to STRING, and write buffer is allocated automatically.
2028 STATUS is a vector of [R0 R1 ... R7 IC], where
2029 R0..R7 are initial values of corresponding registers,
2030 IC is the instruction counter specifying from where to start the program.
2031 If R0..R7 are nil, they are initialized to 0.
2032 If IC is nil, it is initialized to head of the CCL program.
2034 If optional 4th arg CONTINUE is non-nil, keep IC on read operation
2035 when read buffer is exhausted, else, IC is always set to the end of
2036 CCL-PROGRAM on exit.
2038 It returns the contents of write buffer as a string,
2039 and as side effect, STATUS is updated.
2041 See the documentation of `define-ccl-program' for the detail of CCL program.
2043 (ccl_prog, status, string, continue_))
2046 struct ccl_program ccl;
2048 unsigned_char_dynarr *outbuf;
2049 struct gcpro gcpro1, gcpro2;
2051 if (setup_ccl_program (&ccl, ccl_prog) < 0)
2052 error ("Invalid CCL program");
2054 CHECK_VECTOR (status);
2055 if (XVECTOR (status)->size != 9)
2056 error ("Length of vector STATUS is not 9");
2057 CHECK_STRING (string);
2059 GCPRO2 (status, string);
2061 for (i = 0; i < 8; i++)
2063 if (NILP (XVECTOR_DATA (status)[i]))
2064 XSETINT (XVECTOR_DATA (status)[i], 0);
2065 if (INTP (XVECTOR_DATA (status)[i]))
2066 ccl.reg[i] = XINT (XVECTOR_DATA (status)[i]);
2068 if (INTP (XVECTOR (status)->contents[i]))
2070 i = XINT (XVECTOR_DATA (status)[8]);
2071 if (ccl.ic < i && i < ccl.size)
2074 outbuf = Dynarr_new (unsigned_char);
2075 ccl.last_block = NILP (continue_);
2076 produced = ccl_driver (&ccl, XSTRING_DATA (string), outbuf,
2077 XSTRING_LENGTH (string),
2080 for (i = 0; i < 8; i++)
2081 XSETINT (XVECTOR_DATA (status)[i], ccl.reg[i]);
2082 XSETINT (XVECTOR_DATA (status)[8], ccl.ic);
2085 val = make_string (Dynarr_atp (outbuf, 0), produced);
2086 Dynarr_free (outbuf);
2088 if (ccl.status == CCL_STAT_SUSPEND_BY_DST)
2089 error ("Output buffer for the CCL programs overflow");
2090 if (ccl.status != CCL_STAT_SUCCESS
2091 && ccl.status != CCL_STAT_SUSPEND_BY_SRC)
2092 error ("Error in CCL program at %dth code", ccl.ic);
2097 DEFUN ("register-ccl-program", Fregister_ccl_program,
2099 Register CCL program CCL-PROG as NAME in `ccl-program-table'.
2100 CCL-PROG should be a compiled CCL program (vector), or nil.
2101 If it is nil, just reserve NAME as a CCL program name.
2102 Return index number of the registered CCL program.
2106 int len = XVECTOR_LENGTH (Vccl_program_table);
2108 Lisp_Object resolved;
2110 CHECK_SYMBOL (name);
2112 if (!NILP (ccl_prog))
2114 CHECK_VECTOR (ccl_prog);
2115 resolved = resolve_symbol_ccl_program (ccl_prog);
2116 if (! NILP (resolved))
2118 ccl_prog = resolved;
2123 for (idx = 0; idx < len; idx++)
2127 slot = XVECTOR_DATA (Vccl_program_table)[idx];
2128 if (!VECTORP (slot))
2129 /* This is the first unused slot. Register NAME here. */
2132 if (EQ (name, XVECTOR_DATA (slot)[0]))
2134 /* Update this slot. */
2135 XVECTOR_DATA (slot)[1] = ccl_prog;
2136 XVECTOR_DATA (slot)[2] = resolved;
2137 return make_int (idx);
2143 /* Extend the table. */
2144 Lisp_Object new_table;
2147 new_table = Fmake_vector (make_int (len * 2), Qnil);
2148 for (j = 0; j < len; j++)
2149 XVECTOR_DATA (new_table)[j]
2150 = XVECTOR_DATA (Vccl_program_table)[j];
2151 Vccl_program_table = new_table;
2157 elt = Fmake_vector (make_int (3), Qnil);
2158 XVECTOR_DATA (elt)[0] = name;
2159 XVECTOR_DATA (elt)[1] = ccl_prog;
2160 XVECTOR_DATA (elt)[2] = resolved;
2161 XVECTOR_DATA (Vccl_program_table)[idx] = elt;
2164 Fput (name, Qccl_program_idx, make_int (idx));
2165 return make_int (idx);
2168 /* Register code conversion map.
2169 A code conversion map consists of numbers, Qt, Qnil, and Qlambda.
2170 The first element is start code point.
2171 The rest elements are mapped numbers.
2172 Symbol t means to map to an original number before mapping.
2173 Symbol nil means that the corresponding element is empty.
2174 Symbol lambda means to terminate mapping here.
2177 DEFUN ("register-code-conversion-map", Fregister_code_conversion_map,
2179 Register SYMBOL as code conversion map MAP.
2180 Return index number of the registered map.
2184 int len = XVECTOR_LENGTH (Vcode_conversion_map_vector);
2188 CHECK_SYMBOL (symbol);
2191 for (i = 0; i < len; i++)
2193 Lisp_Object slot = XVECTOR_DATA (Vcode_conversion_map_vector)[i];
2198 if (EQ (symbol, XCAR (slot)))
2202 Fput (symbol, Qcode_conversion_map, map);
2203 Fput (symbol, Qcode_conversion_map_id, idx);
2210 Lisp_Object new_vector = Fmake_vector (make_int (len * 2), Qnil);
2213 for (j = 0; j < len; j++)
2214 XVECTOR_DATA (new_vector)[j]
2215 = XVECTOR_DATA (Vcode_conversion_map_vector)[j];
2216 Vcode_conversion_map_vector = new_vector;
2220 Fput (symbol, Qcode_conversion_map, map);
2221 Fput (symbol, Qcode_conversion_map_id, idx);
2222 XVECTOR_DATA (Vcode_conversion_map_vector)[i] = Fcons (symbol, map);
2228 syms_of_mule_ccl (void)
2230 DEFSUBR (Fccl_program_p);
2231 DEFSUBR (Fccl_execute);
2232 DEFSUBR (Fccl_execute_on_string);
2233 DEFSUBR (Fregister_ccl_program);
2234 DEFSUBR (Fregister_code_conversion_map);
2238 vars_of_mule_ccl (void)
2240 staticpro (&Vccl_program_table);
2241 Vccl_program_table = Fmake_vector (make_int (32), Qnil);
2243 defsymbol (&Qccl_program, "ccl-program");
2244 defsymbol (&Qccl_program_idx, "ccl-program-idx");
2245 defsymbol (&Qcode_conversion_map, "code-conversion-map");
2246 defsymbol (&Qcode_conversion_map_id, "code-conversion-map-id");
2248 DEFVAR_LISP ("code-conversion-map-vector", &Vcode_conversion_map_vector /*
2249 Vector of code conversion maps.
2251 Vcode_conversion_map_vector = Fmake_vector (make_int (16), Qnil);
2253 DEFVAR_LISP ("font-ccl-encoder-alist", &Vfont_ccl_encoder_alist /*
2254 Alist of fontname patterns vs corresponding CCL program.
2255 Each element looks like (REGEXP . CCL-CODE),
2256 where CCL-CODE is a compiled CCL program.
2257 When a font whose name matches REGEXP is used for displaying a character,
2258 CCL-CODE is executed to calculate the code point in the font
2259 from the charset number and position code(s) of the character which are set
2260 in CCL registers R0, R1, and R2 before the execution.
2261 The code point in the font is set in CCL registers R1 and R2
2262 when the execution terminated.
2263 If the font is single-byte font, the register R2 is not used.
2265 Vfont_ccl_encoder_alist = Qnil;