1 /**************************************************************** 2 Copyright (C) AT&T 1993 3 All Rights Reserved 4 5 Permission to use, copy, modify, and distribute this software and 6 its documentation for any purpose and without fee is hereby 7 granted, provided that the above copyright notice appear in all 8 copies and that both that the copyright notice and this 9 permission notice and warranty disclaimer appear in supporting 10 documentation, and that the name of AT&T or any of its entities 11 not be used in advertising or publicity pertaining to 12 distribution of the software without specific, written prior 13 permission. 14 15 AT&T DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, 16 INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS. 17 IN NO EVENT SHALL AT&T OR ANY OF ITS ENTITIES BE LIABLE FOR ANY 18 SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES 19 WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER 20 IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, 21 ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF 22 THIS SOFTWARE. 23 ****************************************************************/ 24 25 #define DEBUG 26 #include <stdio.h> 27 #include <string.h> 28 #include <ctype.h> 29 #include <errno.h> 30 #include <stdlib.h> 31 #include "awk.h" 32 #include "y.tab.h" 33 34 #define getfval(p) (((p)->tval & (ARR|FLD|REC|NUM)) == NUM ? (p)->fval : r_getfval(p)) 35 #define getsval(p) (((p)->tval & (ARR|FLD|REC|STR)) == STR ? (p)->sval : r_getsval(p)) 36 37 FILE *infile = NULL; 38 uchar *file = (uchar*) ""; 39 int recsize = RECSIZE; 40 uchar *recdata; 41 uchar *record; 42 uchar *fields; 43 Cell *fldtab; 44 45 #define MAXFLD 200 46 int nfields = MAXFLD; /* can be set from commandline in main */ 47 48 int donefld; /* 1 = implies rec broken into fields */ 49 int donerec; /* 1 = record is valid (no flds have changed) */ 50 51 int maxfld = 0; /* last used field */ 52 int argno = 1; /* current input argument number */ 53 extern Awkfloat *ARGC; 54 55 void recinit(unsigned int n) 56 { 57 static Cell dollar0 = { 58 OCELL, CFLD, (uchar*) "$0", /*recdata*/0, 0.0, REC|STR|DONTFREE }; 59 static Cell dollar1 = { 60 OCELL, CFLD, NULL, (uchar*) "", 0.0, FLD|STR|DONTFREE }; 61 int i; 62 63 record = recdata = (uchar *) malloc(n); 64 fields = (uchar *) malloc(n); 65 fldtab = (Cell *) malloc(nfields * sizeof(Cell)); 66 if (recdata == NULL || fields == NULL || fldtab == NULL) 67 ERROR "out of space for $0 and fields" FATAL; 68 fldtab[0] = dollar0; 69 fldtab[0].sval = recdata; 70 for (i = 1; i < nfields; i++) 71 fldtab[i] = dollar1; 72 } 73 74 void initgetrec(void) 75 { 76 int i; 77 uchar *p; 78 79 for (i = 1; i < *ARGC; i++) { 80 if (!isclvar(p = getargv(i))) { /* find 1st real filename */ 81 setsval(lookup("FILENAME", symtab), getargv(i)); 82 return; 83 } 84 setclvar(p); /* a commandline assignment before filename */ 85 argno++; 86 } 87 infile = stdin; /* no filenames, so use stdin */ 88 } 89 90 getrec(uchar *buf) /* get next input record from whatever source */ 91 { /* note: tests whether buf == record */ 92 int c; 93 static int firsttime = 1; 94 95 if (firsttime) { 96 firsttime = 0; 97 initgetrec(); 98 } 99 dprintf( ("RS=<%s>, FS=<%s>, ARGC=%g, FILENAME=%s\n", 100 *RS, *FS, *ARGC, *FILENAME) ); 101 donefld = 0; 102 donerec = 1; 103 buf[0] = 0; 104 while (argno < *ARGC || infile == stdin) { 105 dprintf( ("argno=%d, file=|%s|\n", argno, file) ); 106 if (infile == NULL) { /* have to open a new file */ 107 file = getargv(argno); 108 if (*file == '\0') { /* it's been zapped */ 109 argno++; 110 continue; 111 } 112 if (isclvar(file)) { /* a var=value arg */ 113 setclvar(file); 114 argno++; 115 continue; 116 } 117 *FILENAME = file; 118 dprintf( ("opening file %s\n", file) ); 119 if (*file == '-' && *(file+1) == '\0') 120 infile = stdin; 121 else if ((infile = fopen((char *)file, "r")) == NULL) 122 ERROR "can't open file %s", file FATAL; 123 setfval(fnrloc, 0.0); 124 } 125 c = readrec(buf, recsize, infile); 126 if (c != 0 || buf[0] != '\0') { /* normal record */ 127 if (buf == record) { 128 if (!(recloc->tval & DONTFREE)) 129 xfree(recloc->sval); 130 recloc->sval = record; 131 recloc->tval = REC | STR | DONTFREE; 132 if (isnumber(recloc->sval)) { 133 recloc->fval = atof(recloc->sval); 134 recloc->tval |= NUM; 135 } 136 } 137 setfval(nrloc, nrloc->fval+1); 138 setfval(fnrloc, fnrloc->fval+1); 139 return 1; 140 } 141 /* EOF arrived on this file; set up next */ 142 if (infile != stdin) 143 fclose(infile); 144 infile = NULL; 145 argno++; 146 } 147 return 0; /* true end of file */ 148 } 149 150 readrec(uchar *buf, int bufsize, FILE *inf) /* read one record into buf */ 151 { 152 register int sep, c; 153 register uchar *rr; 154 register int nrr; 155 156 if ((sep = **RS) == 0) { 157 sep = '\n'; 158 while ((c=getc(inf)) == '\n' && c != EOF) /* skip leading \n's */ 159 ; 160 if (c != EOF) 161 ungetc(c, inf); 162 } 163 for (rr = buf, nrr = bufsize; ; ) { 164 for (; (c=getc(inf)) != sep && c != EOF; *rr++ = c) 165 if (--nrr < 0) 166 ERROR "input record `%.30s...' too long; try -mr n", buf FATAL; 167 if (**RS == sep || c == EOF) 168 break; 169 if ((c = getc(inf)) == '\n' || c == EOF) /* 2 in a row */ 170 break; 171 *rr++ = '\n'; 172 *rr++ = c; 173 } 174 if (rr > buf + bufsize) 175 ERROR "input record `%.30s...' too long; try -mr n", buf FATAL; 176 *rr = 0; 177 dprintf( ("readrec saw <%s>, returns %d\n", buf, c == EOF && rr == buf ? 0 : 1) ); 178 return c == EOF && rr == buf ? 0 : 1; 179 } 180 181 uchar *getargv(int n) /* get ARGV[n] */ 182 { 183 Cell *x; 184 uchar *s, temp[10]; 185 extern Array *ARGVtab; 186 187 sprintf((char *)temp, "%d", n); 188 x = setsymtab(temp, "", 0.0, STR, ARGVtab); 189 s = getsval(x); 190 dprintf( ("getargv(%d) returns |%s|\n", n, s) ); 191 return s; 192 } 193 194 void setclvar(uchar *s) /* set var=value from s */ 195 { 196 uchar *p; 197 Cell *q; 198 199 for (p=s; *p != '='; p++) 200 ; 201 *p++ = 0; 202 p = qstring(p, '\0'); 203 q = setsymtab(s, p, 0.0, STR, symtab); 204 setsval(q, p); 205 if (isnumber(q->sval)) { 206 q->fval = atof(q->sval); 207 q->tval |= NUM; 208 } 209 dprintf( ("command line set %s to |%s|\n", s, p) ); 210 } 211 212 213 void fldbld(void) /* create fields from current record */ 214 { 215 register uchar *r, *fr, sep; 216 Cell *p; 217 int i; 218 219 if (donefld) 220 return; 221 if (!(recloc->tval & STR)) 222 getsval(recloc); 223 r = recloc->sval; 224 fr = fields; 225 i = 0; /* number of fields accumulated here */ 226 if (strlen(*FS) > 1) { /* it's a regular expression */ 227 i = refldbld(r, *FS); 228 } else if ((sep = **FS) == ' ') { /* default whitespace */ 229 for (i = 0; ; ) { 230 while (*r == ' ' || *r == '\t' || *r == '\n') 231 r++; 232 if (*r == 0) 233 break; 234 i++; 235 if (i >= nfields) 236 break; 237 if (!(fldtab[i].tval & DONTFREE)) 238 xfree(fldtab[i].sval); 239 fldtab[i].sval = fr; 240 fldtab[i].tval = FLD | STR | DONTFREE; 241 do 242 *fr++ = *r++; 243 while (*r != ' ' && *r != '\t' && *r != '\n' && *r != '\0'); 244 *fr++ = 0; 245 } 246 *fr = 0; 247 } else if (*r != 0) { /* if 0, it's a null field */ 248 for (;;) { 249 i++; 250 if (i >= nfields) 251 break; 252 if (!(fldtab[i].tval & DONTFREE)) 253 xfree(fldtab[i].sval); 254 fldtab[i].sval = fr; 255 fldtab[i].tval = FLD | STR | DONTFREE; 256 while (*r != sep && *r != '\n' && *r != '\0') /* \n is always a separator */ 257 *fr++ = *r++; 258 *fr++ = 0; 259 if (*r++ == 0) 260 break; 261 } 262 *fr = 0; 263 } 264 if (i >= nfields) 265 ERROR "record `%.30s...' has too many fields; try -mf n", record FATAL; 266 /* clean out junk from previous record */ 267 cleanfld(i, maxfld); 268 maxfld = i; 269 donefld = 1; 270 for (p = fldtab+1; p <= fldtab+maxfld; p++) { 271 if(isnumber(p->sval)) { 272 p->fval = atof(p->sval); 273 p->tval |= NUM; 274 } 275 } 276 setfval(nfloc, (Awkfloat) maxfld); 277 if (dbg) 278 for (p = fldtab; p <= fldtab+maxfld; p++) 279 printf("field %d: |%s|\n", p-fldtab, p->sval); 280 } 281 282 void cleanfld(int n1, int n2) /* clean out fields n1..n2 inclusive */ 283 { 284 static uchar *nullstat = (uchar *) ""; 285 register Cell *p, *q; 286 287 for (p = &fldtab[n2], q = &fldtab[n1]; p > q; p--) { 288 if (!(p->tval & DONTFREE)) 289 xfree(p->sval); 290 p->tval = FLD | STR | DONTFREE; 291 p->sval = nullstat; 292 } 293 } 294 295 void newfld(int n) /* add field n (after end) */ 296 { 297 if (n >= nfields) 298 ERROR "creating too many fields (%d); try -mf n", n, record FATAL; 299 cleanfld(maxfld, n); 300 maxfld = n; 301 setfval(nfloc, (Awkfloat) n); 302 } 303 304 refldbld(uchar *rec, uchar *fs) /* build fields from reg expr in FS */ 305 { 306 uchar *fr; 307 int i, tempstat; 308 fa *pfa; 309 310 fr = fields; 311 *fr = '\0'; 312 if (*rec == '\0') 313 return 0; 314 pfa = makedfa(fs, 1); 315 dprintf( ("into refldbld, rec = <%s>, pat = <%s>\n", rec, fs) ); 316 tempstat = pfa->initstat; 317 for (i = 1; i < nfields; i++) { 318 if (!(fldtab[i].tval & DONTFREE)) 319 xfree(fldtab[i].sval); 320 fldtab[i].tval = FLD | STR | DONTFREE; 321 fldtab[i].sval = fr; 322 dprintf( ("refldbld: i=%d\n", i) ); 323 if (nematch(pfa, rec)) { 324 pfa->initstat = 2; /* horrible coupling */ 325 dprintf( ("match %s (%d chars)\n", patbeg, patlen) ); 326 strncpy(fr, rec, patbeg-rec); 327 fr += patbeg - rec + 1; 328 *(fr-1) = '\0'; 329 rec = patbeg + patlen; 330 } else { 331 dprintf( ("no match %s\n", rec) ); 332 strcpy(fr, rec); 333 pfa->initstat = tempstat; 334 break; 335 } 336 } 337 return i; 338 } 339 340 void recbld(void) /* create $0 from $1..$NF if necessary */ 341 { 342 register int i; 343 register uchar *r, *p; 344 static uchar *rec = 0; 345 346 if (donerec == 1) 347 return; 348 if (rec == 0) { 349 rec = (uchar *) malloc(recsize); 350 if (rec == 0) 351 ERROR "out of space building $0, record size %d", recsize FATAL; 352 } 353 r = rec; 354 for (i = 1; i <= *NF; i++) { 355 p = getsval(&fldtab[i]); 356 while (r < rec+recsize-1 && (*r = *p++)) 357 r++; 358 if (i < *NF) 359 for (p = *OFS; r < rec+recsize-1 && (*r = *p++); ) 360 r++; 361 } 362 if (r > rec + recsize - 1) 363 ERROR "built giant record `%.30s...'; try -mr n", record FATAL; 364 *r = '\0'; 365 dprintf( ("in recbld FS=%o, recloc=%o\n", **FS, recloc) ); 366 recloc->tval = REC | STR | DONTFREE; 367 recloc->sval = record = rec; 368 dprintf( ("in recbld FS=%o, recloc=%o\n", **FS, recloc) ); 369 dprintf( ("recbld = |%s|\n", record) ); 370 donerec = 1; 371 } 372 373 Cell *fieldadr(int n) 374 { 375 if (n < 0 || n >= nfields) 376 ERROR "trying to access field %d; try -mf n", n FATAL; 377 return(&fldtab[n]); 378 } 379 380 int errorflag = 0; 381 char errbuf[200]; 382 383 void yyerror(uchar *s) 384 { 385 extern uchar *cmdname, *curfname; 386 static int been_here = 0; 387 388 if (been_here++ > 2) 389 return; 390 fprintf(stderr, "%s: %s", cmdname, s); 391 fprintf(stderr, " at source line %d", lineno); 392 if (curfname != NULL) 393 fprintf(stderr, " in function %s", curfname); 394 fprintf(stderr, "\n"); 395 errorflag = 2; 396 eprint(); 397 } 398 399 void fpecatch(int n) 400 { 401 ERROR "floating point exception %d", n FATAL; 402 } 403 404 extern int bracecnt, brackcnt, parencnt; 405 406 void bracecheck(void) 407 { 408 int c; 409 static int beenhere = 0; 410 411 if (beenhere++) 412 return; 413 while ((c = input()) != EOF && c != '\0') 414 bclass(c); 415 bcheck2(bracecnt, '{', '}'); 416 bcheck2(brackcnt, '[', ']'); 417 bcheck2(parencnt, '(', ')'); 418 } 419 420 void bcheck2(int n, int c1, int c2) 421 { 422 if (n == 1) 423 fprintf(stderr, "\tmissing %c\n", c2); 424 else if (n > 1) 425 fprintf(stderr, "\t%d missing %c's\n", n, c2); 426 else if (n == -1) 427 fprintf(stderr, "\textra %c\n", c2); 428 else if (n < -1) 429 fprintf(stderr, "\t%d extra %c's\n", -n, c2); 430 } 431 432 void error(int f, char *s) 433 { 434 extern Node *curnode; 435 extern uchar *cmdname; 436 437 fflush(stdout); 438 fprintf(stderr, "%s: ", cmdname); 439 fprintf(stderr, "%s", s); 440 fprintf(stderr, "\n"); 441 if (compile_time != 2 && NR && *NR > 0) { 442 fprintf(stderr, " input record number %g", *FNR); 443 if (strcmp(*FILENAME, "-") != 0) 444 fprintf(stderr, ", file %s", *FILENAME); 445 fprintf(stderr, "\n"); 446 } 447 if (compile_time != 2 && curnode) 448 fprintf(stderr, " source line number %d\n", curnode->lineno); 449 else if (compile_time != 2 && lineno) 450 fprintf(stderr, " source line number %d\n", lineno); 451 eprint(); 452 if (f) { 453 if (dbg > 1) /* core dump if serious debugging on */ 454 abort(); 455 exit(2); 456 } 457 } 458 459 void eprint(void) /* try to print context around error */ 460 { 461 uchar *p, *q; 462 int c; 463 static int been_here = 0; 464 extern uchar ebuf[], *ep; 465 466 if (compile_time == 2 || compile_time == 0 || been_here++ > 0) 467 return; 468 p = ep - 1; 469 if (p > ebuf && *p == '\n') 470 p--; 471 for ( ; p > ebuf && *p != '\n' && *p != '\0'; p--) 472 ; 473 while (*p == '\n') 474 p++; 475 fprintf(stderr, " context is\n\t"); 476 for (q=ep-1; q>=p && *q!=' ' && *q!='\t' && *q!='\n'; q--) 477 ; 478 for ( ; p < q; p++) 479 if (*p) 480 putc(*p, stderr); 481 fprintf(stderr, " >>> "); 482 for ( ; p < ep; p++) 483 if (*p) 484 putc(*p, stderr); 485 fprintf(stderr, " <<< "); 486 if (*ep) 487 while ((c = input()) != '\n' && c != '\0' && c != EOF) { 488 putc(c, stderr); 489 bclass(c); 490 } 491 putc('\n', stderr); 492 ep = ebuf; 493 } 494 495 void bclass(int c) 496 { 497 switch (c) { 498 case '{': bracecnt++; break; 499 case '}': bracecnt--; break; 500 case '[': brackcnt++; break; 501 case ']': brackcnt--; break; 502 case '(': parencnt++; break; 503 case ')': parencnt--; break; 504 } 505 } 506 507 double errcheck(double x, uchar *s) 508 { 509 extern int errno; 510 511 if (errno == EDOM) { 512 errno = 0; 513 ERROR "%s argument out of domain", s WARNING; 514 x = 1; 515 } else if (errno == ERANGE) { 516 errno = 0; 517 ERROR "%s result out of range", s WARNING; 518 x = 1; 519 } 520 return x; 521 } 522 523 isclvar(uchar *s) /* is s of form var=something ? */ 524 { 525 uchar *os = s; 526 527 if (!isalpha(*s) && *s != '_') 528 return 0; 529 for ( ; *s; s++) 530 if (!(isalnum(*s) || *s == '_')) 531 break; 532 return *s == '=' && s > os && *(s+1) != '='; 533 } 534 535 #define MAXEXPON 38 /* maximum exponent for fp number. should be IEEE */ 536 537 isnumber(uchar *s) /* probably should be done by a library function */ 538 { 539 register int d1, d2; 540 int point; 541 uchar *es; 542 543 d1 = d2 = point = 0; 544 while (*s == ' ' || *s == '\t' || *s == '\n') 545 s++; 546 if (*s == '\0') 547 return(0); /* empty stuff isn't number */ 548 if (*s == '+' || *s == '-') 549 s++; 550 if (!isdigit(*s) && *s != '.') 551 return(0); 552 if (isdigit(*s)) { 553 do { 554 d1++; 555 s++; 556 } while (isdigit(*s)); 557 } 558 if (*s == '.') { 559 point++; 560 s++; 561 } 562 if (isdigit(*s)) { 563 d2++; 564 do { 565 s++; 566 } while (isdigit(*s)); 567 } 568 if (!(d1 || point && d2)) 569 return(0); 570 if (*s == 'e' || *s == 'E') { 571 s++; 572 if (*s == '+' || *s == '-') 573 s++; 574 if (!isdigit(*s)) 575 return(0); 576 es = s; 577 do { 578 s++; 579 } while (isdigit(*s)); 580 if (s - es > 2) 581 return(0); 582 else if (s - es == 2 && (int)(10 * (*es-'0') + *(es+1)-'0') >= MAXEXPON) 583 return(0); 584 } 585 while (*s == ' ' || *s == '\t' || *s == '\n') 586 s++; 587 if (*s == '\0') 588 return(1); 589 else 590 return(0); 591 } 592