Annotation of OpenXM/src/ox_math/parse.c, Revision 1.2
1.1 ohara 1: /* -*- mode: C; coding: euc-japan -*- */
2: /* $OpenXM$ */
1.2 ! ohara 3: /* $Id: parse.c,v 1.6 1999/11/01 17:58:38 ohara Exp $ */
1.1 ohara 4: /* OX expression, CMO expression パーサ */
5:
6: #include <stdio.h>
7: #include <stdlib.h>
8: #include <string.h>
9: #include <sys/param.h>
1.2 ! ohara 10: #include <setjmp.h>
1.1 ohara 11: #include "oxtag.h"
12: #include "ox.h"
13: #include "parse.h"
14:
15: /* --- 構文解析部 --- */
1.2 ! ohara 16: /* (重要)セマンティックスについての注意.
! 17: CMO_LIST, CMO_STRING は、あらかじめ与えられた要素の個数を無視する.
! 18: CMO_MONOMIAL32 は無視しない. (つまりおかしいときは構文エラーになる)
! 19: */
! 20:
! 21:
1.1 ohara 22: /* parse.c, lex.c では, Lisp 表現された CMO 文字列を読み込み,
23: バイト列を出力する. 中間表現として、cmo *を利用する.
24: parse() はトークンの列から cmo *(の指す構造体)を生成する.
25: */
26:
27: /* 重要なことはパーサ(の各サブルーチン)は
28: 常にトークンをひとつ先読みしていると言うことである.
29: */
30:
31: /* 現在読み込み中のトークンを表す. */
32: static int token = 0;
33:
1.2 ! ohara 34: /* トークンの属性値. yylval は lex() によってセットされる. */
! 35: static union{
! 36: int d;
! 37: char* sym;
! 38: } yylval;
! 39:
! 40: /* pflag_cmo_addrev がセットされていれば、厳密には CMO expression では
! 41: ないもの, 例えば (CMO_STRING, "hello") も CMO に変換される. */
1.1 ohara 42:
1.2 ! ohara 43: static int pflag_cmo_addrev = 1; /* CMO の省略記法を許すか否かのフラグ */
1.1 ohara 44:
45: /* 関数の宣言 */
46: static int parse_error(char *s);
47: static int parse_lf();
48: static int parse_right_parenthesis();
49: static int parse_left_parenthesis();
50: static int parse_comma();
51: static int parse_integer();
52: static cmo *parse_cmo_null();
1.2 ! ohara 53: static cmo *parse_cmo_int32();
! 54: static cmo *parse_cmo_string();
! 55: static cmo *parse_cmo_mathcap();
! 56: static cmo *parse_cmo_dms_generic();
! 57: static cmo *parse_cmo_ring_by_name();
! 58: static cmo *parse_cmo_error2();
! 59: static cmo *parse_cmo_zero();
1.1 ohara 60: static cmo *parse_cmo_zz();
61: static cmo *parse_cmo_list();
62: static cmo *parse_cmo();
63: static char *parse_string();
64: static int parse_sm();
65: static ox* parse_ox();
66: static ox* parse_ox_command();
67: static ox* parse_ox_data();
68:
69: static int is_t_cmo(int token)
70: {
1.2 ! ohara 71: return (token >= MIN_T_CMO && token < MAX_T_CMO) || token == TOKEN(CMO_ERROR2);
1.1 ohara 72: }
73:
74: static int is_t_sm(int token)
75: {
76: return token >= MIN_T_SM && token < MAX_T_SM;
77: }
78:
79: static int is_t_ox(int token)
80: {
81: return token >= MIN_T_OX && token < MAX_T_OX;
82: }
83:
1.2 ! ohara 84: static jmp_buf env_parse;
! 85:
! 86: /* 構文解析に失敗したことを意味する. */
! 87: static int parse_error(char *s)
! 88: {
! 89: fprintf(stderr, "%s\n", s);
! 90: longjmp(env_parse, 1);
! 91: }
! 92:
1.1 ohara 93: /* この部分は書き換え予定. */
94: cmo *parse()
95: {
96: cmo *m;
97:
1.2 ! ohara 98: if (setjmp(env_parse) != 0) {
! 99: return NULL; /* 構文解析に失敗したら NULL を返す. */
! 100: }
! 101:
1.1 ohara 102: do{
103: token = lex();
104: }while (token == '\n');
105:
106: if (token == '(') {
107: token = lex();
108: if (is_t_cmo(token)) {
109: m = parse_cmo();
110: }else if(is_t_ox(token)) {
111: m = parse_ox();
112: }else {
1.2 ! ohara 113: parse_error("syntax error: unknown symbol.");
1.1 ohara 114: }
115: parse_lf();
116: return m;
117: }
118: return NULL;
119: }
120:
121: /* トークンを先読みしない(重要). */
122: static int parse_lf()
123: {
124: if (token != '\n') {
125: parse_error("syntax error: not new line.");
126: }
127: return 0;
128: }
129:
130:
131: static ox* parse_ox()
132: {
133: ox *m = NULL;
134:
135: switch(token) {
1.2 ! ohara 136: case TOKEN(OX_COMMAND):
1.1 ohara 137: token = lex();
138: m = parse_ox_command();
139: break;
1.2 ! ohara 140: case TOKEN(OX_DATA):
1.1 ohara 141: token = lex();
142: m = parse_ox_data();
143: break;
144: default:
145: parse_error("syntax error: invalid ox_tag.");
146: }
147: return m;
148: }
149:
150: static ox* parse_ox_data()
151: {
152: ox* m;
153:
154: parse_comma();
155: parse_left_parenthesis();
156: m = (ox *)new_ox_data(parse_cmo());
157: parse_right_parenthesis();
158: return m;
159: }
160:
161: static int parse_sm()
162: {
163: int sm_code;
164: if (!is_t_sm(token)) {
165: parse_error("syntax error: invalid sm code.");
166: }
167: sm_code = token - T_MAGIC;
168: token = lex();
169: return sm_code;
170: }
171:
172:
173: static ox* parse_ox_command()
174: {
175: ox* m;
176:
177: parse_comma();
178: m = (ox *)new_ox_command(parse_sm());
179: parse_right_parenthesis();
180: return m;
181: }
182:
183: /* 正しい入力ならば, parse_cmo を呼ぶ時点で, token には
1.2 ! ohara 184: TOKEN(CMO_xxx), TOKEN(OX_xxx) のいずれかがセットされている. */
1.1 ohara 185: static cmo *parse_cmo()
186: {
187: cmo *m = NULL;
188:
189: switch(token) {
1.2 ! ohara 190: case TOKEN(CMO_NULL):
1.1 ohara 191: token = lex();
192: m = parse_cmo_null();
193: break;
1.2 ! ohara 194: case TOKEN(CMO_INT32):
1.1 ohara 195: token = lex();
196: m = parse_cmo_int32();
197: break;
1.2 ! ohara 198: case TOKEN(CMO_STRING):
1.1 ohara 199: token = lex();
200: m = parse_cmo_string();
201: break;
1.2 ! ohara 202: case TOKEN(CMO_MATHCAP):
! 203: token = lex();
! 204: m = parse_cmo_mathcap();
! 205: break;
! 206: case TOKEN(CMO_LIST):
! 207: token = lex();
! 208: m = parse_cmo_list();
! 209: break;
! 210: case TOKEN(CMO_MONOMIAL32):
! 211: token = lex();
! 212: m = parse_cmo_monomial32();
! 213: break;
! 214: case TOKEN(CMO_ZZ):
1.1 ohara 215: token = lex();
216: m = parse_cmo_zz();
217: break;
1.2 ! ohara 218: case TOKEN(CMO_ZERO):
! 219: token = lex();
! 220: m = parse_cmo_zero();
! 221: break;
! 222: case TOKEN(CMO_DMS_GENERIC):
1.1 ohara 223: token = lex();
1.2 ! ohara 224: m = parse_cmo_dms_generic();
! 225: break;
! 226: case TOKEN(CMO_RING_BY_NAME):
! 227: token = lex();
! 228: m = parse_cmo_ring_by_name();
! 229: break;
! 230: case TOKEN(CMO_ERROR2):
! 231: token = lex();
! 232: m = parse_cmo_error2();
1.1 ohara 233: break;
234: default:
235: parse_error("syntax error: invalid cmo_tag.");
236: }
237: return m;
238: }
239:
1.2 ! ohara 240: static int parse_left_parenthesis()
1.1 ohara 241: {
1.2 ! ohara 242: if (token != '(') {
! 243: parse_error("syntax error: no left parenthesis.");
1.1 ohara 244: }
245: token = lex();
246: }
247:
1.2 ! ohara 248: static int parse_right_parenthesis()
1.1 ohara 249: {
1.2 ! ohara 250: if (token != ')') {
! 251: parse_error("syntax error: no right parenthesis.");
1.1 ohara 252: }
253: token = lex();
254: }
255:
256: static int parse_comma()
257: {
258: if (token != ',') {
1.2 ! ohara 259: parse_error("syntax error: no comma.");
1.1 ohara 260: }
261: token = lex();
262: }
263:
264: static int parse_integer()
265: {
266: int val;
267: if (token != T_INTEGER) {
1.2 ! ohara 268: parse_error("syntax error: no integer.");
1.1 ohara 269: }
270: val = yylval.d;
271: token = lex();
272: return val;
273: }
274:
275: static char *parse_string()
276: {
277: char *s;
278: if (token != T_STRING) {
1.2 ! ohara 279: parse_error("syntax error: no string.");
1.1 ohara 280: }
281: s = yylval.sym;
282: token = lex();
283: return s;
284: }
285:
286: static cmo *parse_cmo_null()
287: {
288: parse_right_parenthesis();
1.2 ! ohara 289: return (cmo *)new_cmo_null();
1.1 ohara 290: }
291:
292: static cmo *parse_cmo_int32()
293: {
294: int i;
295:
296: parse_comma();
297: i = parse_integer();
298: parse_right_parenthesis();
1.2 ! ohara 299: return (cmo *)new_cmo_int32(i);
1.1 ohara 300: }
301:
302: static cmo *parse_cmo_string()
303: {
304: cmo_string *m;
305: char *s;
306:
307: parse_comma();
308: if (token == T_INTEGER) {
1.2 ! ohara 309: parse_integer();
1.1 ohara 310: parse_comma();
1.2 ! ohara 311: }else if (!pflag_cmo_addrev) {
! 312: parse_error("syntax error: not a cmo string.");
! 313: }
! 314: s = parse_string();
! 315: m = new_cmo_string(s);
! 316: parse_right_parenthesis();
! 317: return (cmo *)m;
! 318: }
! 319:
! 320: static cmo *parse_cmo_mathcap()
! 321: {
! 322: cmo *ob;
! 323:
! 324: parse_comma();
! 325: parse_left_parenthesis();
! 326: ob = parse_cmo();
! 327: parse_right_parenthesis();
! 328: return (cmo *)new_cmo_mathcap(ob);
! 329: }
! 330:
! 331: static cmo *parse_cmo_list()
! 332: {
! 333: int length=0;
! 334: int i=0;
! 335: cmo_list *m = new_cmo_list();
! 336: cmo *newcmo;
! 337:
! 338: if (token == ',') {
! 339: parse_comma();
! 340:
! 341: if (token == T_INTEGER) {
! 342: parse_integer();
! 343: parse_comma();
! 344: }else if (!pflag_cmo_addrev) {
! 345: parse_error("syntax error: not a list.");
! 346: }
! 347:
! 348: while(token == '(') {
! 349: parse_left_parenthesis();
! 350: newcmo = parse_cmo();
! 351: append_cmo_list(m, newcmo);
! 352: if (token != ',') {
! 353: break;
! 354: }
! 355: parse_comma();
! 356: }
! 357: }else if (!pflag_cmo_addrev) {
! 358: parse_error("syntax error: not a list.");
! 359: }
! 360: parse_right_parenthesis();
! 361: return (cmo *)m;
! 362: }
! 363:
! 364: static cmo *parse_cmo_monomial32()
! 365: {
! 366: int size;
! 367: int *exps;
! 368: int i;
! 369: cmo_monomial32 *m;
! 370:
! 371: parse_comma();
! 372: size = parse_integer();
! 373: if (size <= 0) {
! 374: parse_error("syntax error: invalid value.");
! 375: }
! 376: m = new_cmo_monomial32_size(size);
! 377:
! 378: for(i=0; i<size; i++) {
! 379: parse_comma();
! 380: m->exps[i] = parse_integer();
! 381: }
! 382: parse_comma();
! 383: parse_left_parenthesis();
! 384: m->coef = parse_cmo();
! 385: /* 意味的チェックの必要あり */
! 386: parse_right_parenthesis();
! 387: return (cmo *)m;
! 388: }
! 389:
! 390: /* cmo_zz の内部を直接いじる. */
! 391: static cmo *parse_cmo_zz()
! 392: {
! 393: int length;
! 394: int i=0;
! 395: cmo_zz *m= NULL;
! 396:
! 397: parse_comma();
! 398: length = parse_integer();
! 399: if (token == ',') {
! 400: m = new_cmo_zz_size(length);
! 401:
! 402: length = abs(length);
! 403: for(i=0; i<length; i++) {
! 404: parse_comma();
! 405: m->mpz->_mp_d[i] = parse_integer();
1.1 ohara 406: }
1.2 ! ohara 407: }else if (pflag_cmo_addrev) {
! 408: m = new_cmo_zz_set_si(length);
1.1 ohara 409: }else {
1.2 ! ohara 410: parse_error("syntax error: invalid symbol.");
1.1 ohara 411: }
1.2 ! ohara 412:
1.1 ohara 413: parse_right_parenthesis();
414: return (cmo *)m;
415: }
416:
1.2 ! ohara 417: static cmo *parse_cmo_zero()
! 418: {
! 419: parse_right_parenthesis();
! 420: return (cmo *)new_cmo_zero();
! 421: }
! 422:
! 423: static cmo *parse_cmo_dms_generic()
! 424: {
! 425: parse_right_parenthesis();
! 426: return (cmo *)new_cmo_dms_generic();
! 427: }
! 428:
! 429: static cmo *parse_cmo_ring_by_name()
! 430: {
! 431: cmo *ob;
! 432:
! 433: parse_comma();
! 434: parse_left_parenthesis();
! 435: ob = parse_cmo();
! 436: /* 意味的チェックが必要(ob->tag == CMO_STRINGでなければいけない) */
! 437: parse_right_parenthesis();
! 438: return (cmo *)new_cmo_ring_by_name(ob);
! 439: }
! 440:
! 441: static cmo *parse_cmo_error2()
! 442: {
! 443: cmo *ob;
! 444:
! 445: parse_comma();
! 446: parse_left_parenthesis();
! 447: ob = parse_cmo();
! 448: parse_right_parenthesis();
! 449: return (cmo *)new_cmo_error2(ob);
! 450: }
! 451:
1.1 ohara 452: /* --- 字句解析部 --- */
453:
454: /* lexical analyzer で読み飛ばされる文字なら何を初期値にしてもよい */
455: static int c = ' ';
456:
457: /* 一文字読み込む関数 */
458: static int (*GETC)() = getchar;
459:
460: int setgetc(int (*foo)())
461: {
462: GETC = foo;
463: }
464:
465: int resetgetc()
466: {
467: GETC = getchar;
468: }
469:
1.2 ! ohara 470: #define SIZE_BUFFER 8192
1.1 ohara 471: static char buffer[SIZE_BUFFER];
472:
473: /* 桁溢れの場合の対策はない */
474: static int lex_digit()
475: {
476: int d = 0;
477: do {
478: d = 10*d + (c - '0');
479: c = GETC();
480: } while(isdigit(c));
481: return d;
482: }
483:
484: /* バッファあふれした場合の対策をちゃんと考えるべき */
485: static char *lex_quoted_string()
486: {
487: int i;
488: char c0 = ' ';
489: char *s = NULL;
490: for (i=0; i<SIZE_BUFFER; i++) {
491: c = GETC();
492: if(c == '"') {
493: s = malloc(i+1);
494: buffer[i]='\0';
495: strcpy(s, buffer);
496:
497: c = GETC();
498: return s;
499: }else if (c == '\\') {
500: c0 = c;
501: c = GETC();
502: if (c != '"') {
503: buffer[i++] = c0;
504: }
505: }
506: buffer[i]=c;
507: }
508: fprintf(stderr, "buffer overflow!\n");
509: exit(1);
510: /* return NULL; */
511: }
512:
1.2 ! ohara 513: typedef struct {
! 514: char *key;
! 515: int token;
! 516: } symbol;
! 517:
! 518: #define MK_KEY(x) { #x , TOKEN(x) }
! 519:
! 520: static symbol symbol_list[] = {
! 521: MK_KEY(CMO_NULL),
! 522: MK_KEY(CMO_INT32),
! 523: MK_KEY(CMO_DATUM),
! 524: MK_KEY(CMO_STRING),
! 525: MK_KEY(CMO_MATHCAP),
! 526: MK_KEY(CMO_LIST),
! 527: MK_KEY(CMO_MONOMIAL32),
! 528: MK_KEY(CMO_ZZ),
! 529: MK_KEY(CMO_ZERO),
! 530: MK_KEY(CMO_DMS_GENERIC),
! 531: MK_KEY(CMO_RING_BY_NAME),
! 532: MK_KEY(CMO_INDETERMINATE),
! 533: MK_KEY(CMO_ERROR2),
! 534: MK_KEY(SM_popCMO),
! 535: MK_KEY(SM_popString),
! 536: MK_KEY(SM_mathcap),
! 537: MK_KEY(SM_pops),
! 538: MK_KEY(SM_executeStringByLocalParser),
! 539: MK_KEY(SM_executeFunction),
! 540: MK_KEY(SM_setMathCap),
! 541: MK_KEY(SM_control_kill),
! 542: MK_KEY(SM_control_reset_connection),
! 543: MK_KEY(OX_COMMAND), MK_KEY(OX_DATA),
! 544: {NULL, 0} /* a gate keeper */
! 545: };
! 546:
! 547: static int token_of_symbol(char *key)
! 548: {
! 549: symbol *kp;
! 550: for(kp = symbol_list; kp->key != NULL; kp++) {
! 551: if (strcmp(key, kp->key)==0) {
! 552: return kp->token;
! 553: }
! 554: }
! 555: #if DEBUG
1.1 ohara 556: fprintf(stderr, "lex error\n");
1.2 ! ohara 557: #endif
1.1 ohara 558: return 0;
559: }
560:
1.2 ! ohara 561: static int lex_symbol()
1.1 ohara 562: {
563: int i;
564: for (i=0; i<SIZE_BUFFER; i++) {
565: if (!isalnum(c) && c != '_') {
566: buffer[i]='\0';
1.2 ! ohara 567: return token_of_symbol(buffer);
1.1 ohara 568: }
569: buffer[i]=c;
570: c = GETC();
571: }
572: fprintf(stderr, "buffer overflow!\n");
1.2 ! ohara 573: return 0;
1.1 ohara 574: }
575:
576: /* return する前に一文字先読みしておく。 */
577: int lex()
578: {
579: int c_dash = 0;
580:
581: /* 空白をスキップする. */
582: while (isspace(c) && c != '\n') {
583: c = GETC();
584: }
585:
586: switch(c) {
587: case '(':
588: case ')':
589: case ',':
590: case '\n':
591: c_dash = c;
592: c = ' ';
593: return c_dash;
594: case EOF:
595: c = GETC();
596: return c_dash;
597: case '"': /* a quoted string! */
598: yylval.sym = lex_quoted_string();
599: return T_STRING;
600: default:
601: }
602:
603: if (isalpha(c)) { /* 識別子 */
1.2 ! ohara 604: return lex_symbol();
1.1 ohara 605: }
606:
607: /* 32bit 整数値 */
608: if (isdigit(c)){
609: yylval.d = lex_digit();
610: return T_INTEGER;
611: }
612: if (c == '-') {
613: c = GETC();
614: while (isspace(c) && c != '\n') {
615: c = GETC();
616: }
617: if (isdigit(c)){
618: yylval.d = - lex_digit();
619: return T_INTEGER;
620: }
621: return 0;
622: }
623:
624: c = GETC();
625: return 0;
626: }
627:
628: static char *mygetc_line;
629: static int mygetc_counter;
630: static int mygetc_counter_max;
631:
632: int mygetc()
633: {
634: if (mygetc_counter <= mygetc_counter_max) {
635: return mygetc_line[mygetc_counter++];
636: }
637: return 0;
638: }
639:
640: int setmode_mygetc(char *s, int len)
641: {
642: mygetc_counter=0;
643: mygetc_counter_max=len;
644: mygetc_line=s;
1.2 ! ohara 645: }
! 646:
! 647: int setflag_parse(int flag)
! 648: {
! 649: pflag_cmo_addrev = flag;
1.1 ohara 650: }
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>
![[BACK]](/icons/cvsweb/back.gif){kind=icon}
Return to parse.c CVS log ![[TXT]](/icons/cvsweb/text.gif){kind=icon}
![[DIR]](/icons/cvsweb/dir.gif){kind=icon}
Up to [local] / OpenXM / src / ox_math