| 1 | /* Copyright (c) 1979 Regents of the University of California */ |
| 2 | #include "ex.h" |
| 3 | #include "ex_re.h" |
| 4 | |
| 5 | /* |
| 6 | * Global, substitute and regular expressions. |
| 7 | * Very similar to ed, with some re extensions and |
| 8 | * confirmed substitute. |
| 9 | */ |
| 10 | global(k) |
| 11 | bool k; |
| 12 | { |
| 13 | register char *gp; |
| 14 | register int c; |
| 15 | register line *a1; |
| 16 | char globuf[GBSIZE], *Cwas; |
| 17 | int lines = lineDOL(); |
| 18 | int oinglobal = inglobal; |
| 19 | char *oglobp = globp; |
| 20 | |
| 21 | Cwas = Command; |
| 22 | /* |
| 23 | * States of inglobal: |
| 24 | * 0: ordinary - not in a global command. |
| 25 | * 1: text coming from some buffer, not tty. |
| 26 | * 2: like 1, but the source of the buffer is a global command. |
| 27 | * Hence you're only in a global command if inglobal==2. This |
| 28 | * strange sounding convention is historically derived from |
| 29 | * everybody simulating a global command. |
| 30 | */ |
| 31 | if (inglobal==2) |
| 32 | error("Global within global@not allowed"); |
| 33 | markDOT(); |
| 34 | setall(); |
| 35 | nonzero(); |
| 36 | if (skipend()) |
| 37 | error("Global needs re|Missing regular expression for global"); |
| 38 | c = getchar(); |
| 39 | ignore(compile(c, 1)); |
| 40 | savere(scanre); |
| 41 | gp = globuf; |
| 42 | while ((c = getchar()) != '\n') { |
| 43 | switch (c) { |
| 44 | |
| 45 | case EOF: |
| 46 | c = '\n'; |
| 47 | goto brkwh; |
| 48 | |
| 49 | case '\\': |
| 50 | c = getchar(); |
| 51 | switch (c) { |
| 52 | |
| 53 | case '\\': |
| 54 | ungetchar(c); |
| 55 | break; |
| 56 | |
| 57 | case '\n': |
| 58 | break; |
| 59 | |
| 60 | default: |
| 61 | *gp++ = '\\'; |
| 62 | break; |
| 63 | } |
| 64 | break; |
| 65 | } |
| 66 | *gp++ = c; |
| 67 | if (gp >= &globuf[GBSIZE - 2]) |
| 68 | error("Global command too long"); |
| 69 | } |
| 70 | brkwh: |
| 71 | ungetchar(c); |
| 72 | out: |
| 73 | newline(); |
| 74 | *gp++ = c; |
| 75 | *gp++ = 0; |
| 76 | saveall(); |
| 77 | inglobal = 2; |
| 78 | for (a1 = one; a1 <= dol; a1++) { |
| 79 | *a1 &= ~01; |
| 80 | if (a1 >= addr1 && a1 <= addr2 && execute(0, a1) == k) |
| 81 | *a1 |= 01; |
| 82 | } |
| 83 | /* should use gdelete from ed to avoid n**2 here on g/.../d */ |
| 84 | if (inopen) |
| 85 | inopen = -1; |
| 86 | for (a1 = one; a1 <= dol; a1++) { |
| 87 | if (*a1 & 01) { |
| 88 | *a1 &= ~01; |
| 89 | dot = a1; |
| 90 | globp = globuf; |
| 91 | commands(1, 1); |
| 92 | a1 = zero; |
| 93 | } |
| 94 | } |
| 95 | globp = oglobp; |
| 96 | inglobal = oinglobal; |
| 97 | endline = 1; |
| 98 | Command = Cwas; |
| 99 | netchHAD(lines); |
| 100 | setlastchar(EOF); |
| 101 | if (inopen) { |
| 102 | ungetchar(EOF); |
| 103 | inopen = 1; |
| 104 | } |
| 105 | } |
| 106 | |
| 107 | bool xflag; |
| 108 | int scount, slines, stotal; |
| 109 | |
| 110 | substitute(c) |
| 111 | int c; |
| 112 | { |
| 113 | register line *addr; |
| 114 | register int n; |
| 115 | int gsubf; |
| 116 | |
| 117 | gsubf = compsub(c); |
| 118 | if(FIXUNDO) |
| 119 | save12(), undkind = UNDCHANGE; |
| 120 | stotal = 0; |
| 121 | slines = 0; |
| 122 | for (addr = addr1; addr <= addr2; addr++) { |
| 123 | scount = 0; |
| 124 | if (dosubcon(0, addr) == 0) |
| 125 | continue; |
| 126 | if (gsubf) { |
| 127 | #ifdef notdef |
| 128 | /* |
| 129 | * should check but loc2 is already munged. |
| 130 | * This needs a fancier check later. |
| 131 | */ |
| 132 | if (loc1 == loc2) |
| 133 | error("substitution loop"); |
| 134 | #endif |
| 135 | while (*loc2) |
| 136 | if (dosubcon(1, addr) == 0) |
| 137 | break; |
| 138 | } |
| 139 | if (scount) { |
| 140 | stotal += scount; |
| 141 | slines++; |
| 142 | putmark(addr); |
| 143 | n = append(getsub, addr); |
| 144 | addr += n; |
| 145 | addr2 += n; |
| 146 | } |
| 147 | } |
| 148 | if (stotal == 0 && !inglobal && !xflag) |
| 149 | error("Fail|Substitute pattern match failed"); |
| 150 | snote(stotal, slines); |
| 151 | return (stotal); |
| 152 | } |
| 153 | |
| 154 | compsub(ch) |
| 155 | { |
| 156 | register int seof, c, uselastre; |
| 157 | static int gsubf; |
| 158 | |
| 159 | if (!value(EDCOMPATIBLE)) |
| 160 | gsubf = xflag = 0; |
| 161 | uselastre = 0; |
| 162 | switch (ch) { |
| 163 | |
| 164 | case 's': |
| 165 | ignore(skipwh()); |
| 166 | seof = getchar(); |
| 167 | if (endcmd(seof) || any(seof, "gcr")) { |
| 168 | ungetchar(seof); |
| 169 | goto redo; |
| 170 | } |
| 171 | if (isalpha(seof) || isdigit(seof)) |
| 172 | error("Substitute needs re|Missing regular expression for substitute"); |
| 173 | seof = compile(seof, 1); |
| 174 | uselastre = 1; |
| 175 | comprhs(seof); |
| 176 | gsubf = 0; |
| 177 | xflag = 0; |
| 178 | break; |
| 179 | |
| 180 | case '~': |
| 181 | uselastre = 1; |
| 182 | /* fall into ... */ |
| 183 | case '&': |
| 184 | redo: |
| 185 | if (re.Expbuf[0] == 0) |
| 186 | error("No previous re|No previous regular expression"); |
| 187 | break; |
| 188 | } |
| 189 | for (;;) { |
| 190 | c = getchar(); |
| 191 | switch (c) { |
| 192 | |
| 193 | case 'g': |
| 194 | gsubf = !gsubf; |
| 195 | continue; |
| 196 | |
| 197 | case 'c': |
| 198 | xflag = !xflag; |
| 199 | continue; |
| 200 | |
| 201 | case 'r': |
| 202 | uselastre = 1; |
| 203 | continue; |
| 204 | |
| 205 | default: |
| 206 | ungetchar(c); |
| 207 | setcount(); |
| 208 | newline(); |
| 209 | if (uselastre) |
| 210 | savere(subre); |
| 211 | else |
| 212 | resre(subre); |
| 213 | return (gsubf); |
| 214 | } |
| 215 | } |
| 216 | } |
| 217 | |
| 218 | comprhs(seof) |
| 219 | int seof; |
| 220 | { |
| 221 | register char *rp, *orp; |
| 222 | register int c; |
| 223 | char orhsbuf[LBSIZE / 2]; |
| 224 | |
| 225 | rp = rhsbuf; |
| 226 | CP(orhsbuf, rp); |
| 227 | for (;;) { |
| 228 | c = getchar(); |
| 229 | if (c == seof) |
| 230 | break; |
| 231 | switch (c) { |
| 232 | |
| 233 | case '\\': |
| 234 | c = getchar(); |
| 235 | if (c == EOF) { |
| 236 | ungetchar(c); |
| 237 | break; |
| 238 | } |
| 239 | if (value(MAGIC)) { |
| 240 | /* |
| 241 | * When "magic", \& turns into a plain &, |
| 242 | * and all other chars work fine quoted. |
| 243 | */ |
| 244 | if (c != '&') |
| 245 | c |= QUOTE; |
| 246 | break; |
| 247 | } |
| 248 | magic: |
| 249 | if (c == '~') { |
| 250 | for (orp = orhsbuf; *orp; *rp++ = *orp++) |
| 251 | if (rp >= &rhsbuf[LBSIZE / 2 + 1]) |
| 252 | goto toobig; |
| 253 | continue; |
| 254 | } |
| 255 | c |= QUOTE; |
| 256 | break; |
| 257 | |
| 258 | case '\n': |
| 259 | case EOF: |
| 260 | if (!(globp && globp[0])) { |
| 261 | ungetchar(c); |
| 262 | goto endrhs; |
| 263 | } |
| 264 | |
| 265 | case '~': |
| 266 | case '&': |
| 267 | if (value(MAGIC)) |
| 268 | goto magic; |
| 269 | break; |
| 270 | } |
| 271 | if (rp >= &rhsbuf[LBSIZE / 2 - 1]) |
| 272 | toobig: |
| 273 | error("Replacement pattern too long@- limit 256 characters"); |
| 274 | *rp++ = c; |
| 275 | } |
| 276 | endrhs: |
| 277 | *rp++ = 0; |
| 278 | } |
| 279 | |
| 280 | getsub() |
| 281 | { |
| 282 | register char *p; |
| 283 | |
| 284 | if ((p = linebp) == 0) |
| 285 | return (EOF); |
| 286 | strcLIN(p); |
| 287 | linebp = 0; |
| 288 | return (0); |
| 289 | } |
| 290 | |
| 291 | dosubcon(f, a) |
| 292 | bool f; |
| 293 | line *a; |
| 294 | { |
| 295 | |
| 296 | if (execute(f, a) == 0) |
| 297 | return (0); |
| 298 | if (confirmed(a)) { |
| 299 | dosub(); |
| 300 | scount++; |
| 301 | } |
| 302 | return (1); |
| 303 | } |
| 304 | |
| 305 | confirmed(a) |
| 306 | line *a; |
| 307 | { |
| 308 | register int c, ch; |
| 309 | |
| 310 | if (xflag == 0) |
| 311 | return (1); |
| 312 | pofix(); |
| 313 | pline(lineno(a)); |
| 314 | if (inopen) |
| 315 | putchar('\n' | QUOTE); |
| 316 | c = column(loc1 - 1); |
| 317 | ugo(c - 1 + (inopen ? 1 : 0), ' '); |
| 318 | ugo(column(loc2 - 1) - c, '^'); |
| 319 | flush(); |
| 320 | ch = c = getkey(); |
| 321 | again: |
| 322 | if (c == '\r') |
| 323 | c = '\n'; |
| 324 | if (inopen) |
| 325 | putchar(c), flush(); |
| 326 | if (c != '\n' && c != EOF) { |
| 327 | c = getkey(); |
| 328 | goto again; |
| 329 | } |
| 330 | noteinp(); |
| 331 | return (ch == 'y'); |
| 332 | } |
| 333 | |
| 334 | getch() |
| 335 | { |
| 336 | char c; |
| 337 | |
| 338 | if (read(2, &c, 1) != 1) |
| 339 | return (EOF); |
| 340 | return (c & TRIM); |
| 341 | } |
| 342 | |
| 343 | ugo(cnt, with) |
| 344 | int with; |
| 345 | int cnt; |
| 346 | { |
| 347 | |
| 348 | if (cnt > 0) |
| 349 | do |
| 350 | putchar(with); |
| 351 | while (--cnt > 0); |
| 352 | } |
| 353 | |
| 354 | int casecnt; |
| 355 | bool destuc; |
| 356 | |
| 357 | dosub() |
| 358 | { |
| 359 | register char *lp, *sp, *rp; |
| 360 | int c; |
| 361 | |
| 362 | lp = linebuf; |
| 363 | sp = genbuf; |
| 364 | rp = rhsbuf; |
| 365 | while (lp < loc1) |
| 366 | *sp++ = *lp++; |
| 367 | casecnt = 0; |
| 368 | while (c = *rp++) { |
| 369 | if (c & QUOTE) |
| 370 | switch (c & TRIM) { |
| 371 | |
| 372 | case '&': |
| 373 | sp = place(sp, loc1, loc2); |
| 374 | if (sp == 0) |
| 375 | goto ovflo; |
| 376 | continue; |
| 377 | |
| 378 | case 'l': |
| 379 | casecnt = 1; |
| 380 | destuc = 0; |
| 381 | continue; |
| 382 | |
| 383 | case 'L': |
| 384 | casecnt = LBSIZE; |
| 385 | destuc = 0; |
| 386 | continue; |
| 387 | |
| 388 | case 'u': |
| 389 | casecnt = 1; |
| 390 | destuc = 1; |
| 391 | continue; |
| 392 | |
| 393 | case 'U': |
| 394 | casecnt = LBSIZE; |
| 395 | destuc = 1; |
| 396 | continue; |
| 397 | |
| 398 | case 'E': |
| 399 | case 'e': |
| 400 | casecnt = 0; |
| 401 | continue; |
| 402 | } |
| 403 | if (c < 0 && (c &= TRIM) >= '1' && c < nbra + '1') { |
| 404 | sp = place(sp, braslist[c - '1'], braelist[c - '1']); |
| 405 | if (sp == 0) |
| 406 | goto ovflo; |
| 407 | continue; |
| 408 | } |
| 409 | if (casecnt) |
| 410 | *sp++ = fixcase(c & TRIM); |
| 411 | else |
| 412 | *sp++ = c & TRIM; |
| 413 | if (sp >= &genbuf[LBSIZE]) |
| 414 | ovflo: |
| 415 | error("Line overflow@in substitute"); |
| 416 | } |
| 417 | lp = loc2; |
| 418 | loc2 = sp + (linebuf - genbuf); |
| 419 | while (*sp++ = *lp++) |
| 420 | if (sp >= &genbuf[LBSIZE]) |
| 421 | goto ovflo; |
| 422 | strcLIN(genbuf); |
| 423 | } |
| 424 | |
| 425 | fixcase(c) |
| 426 | register int c; |
| 427 | { |
| 428 | |
| 429 | if (casecnt == 0) |
| 430 | return (c); |
| 431 | casecnt--; |
| 432 | if (destuc) { |
| 433 | if (islower(c)) |
| 434 | c = toupper(c); |
| 435 | } else |
| 436 | if (isupper(c)) |
| 437 | c = tolower(c); |
| 438 | return (c); |
| 439 | } |
| 440 | |
| 441 | char * |
| 442 | place(sp, l1, l2) |
| 443 | register char *sp, *l1, *l2; |
| 444 | { |
| 445 | |
| 446 | while (l1 < l2) { |
| 447 | *sp++ = fixcase(*l1++); |
| 448 | if (sp >= &genbuf[LBSIZE]) |
| 449 | return (0); |
| 450 | } |
| 451 | return (sp); |
| 452 | } |
| 453 | |
| 454 | snote(total, lines) |
| 455 | register int total, lines; |
| 456 | { |
| 457 | |
| 458 | if (!notable(total)) |
| 459 | return; |
| 460 | printf(mesg("%d subs|%d substitutions"), total); |
| 461 | if (lines != 1 && lines != total) |
| 462 | printf(" on %d lines", lines); |
| 463 | noonl(); |
| 464 | flush(); |
| 465 | } |
| 466 | |
| 467 | compile(eof, oknl) |
| 468 | int eof; |
| 469 | int oknl; |
| 470 | { |
| 471 | register int c; |
| 472 | register char *ep; |
| 473 | char *lastep; |
| 474 | char bracket[NBRA], *bracketp, *rhsp; |
| 475 | int cclcnt; |
| 476 | |
| 477 | if (isalpha(eof) || isdigit(eof)) |
| 478 | error("Regular expressions cannot be delimited by letters or digits"); |
| 479 | ep = expbuf; |
| 480 | c = getchar(); |
| 481 | if (eof == '\\') |
| 482 | switch (c) { |
| 483 | |
| 484 | case '/': |
| 485 | case '?': |
| 486 | if (scanre.Expbuf[0] == 0) |
| 487 | error("No previous scan re|No previous scanning regular expression"); |
| 488 | resre(scanre); |
| 489 | return (c); |
| 490 | |
| 491 | case '&': |
| 492 | if (subre.Expbuf[0] == 0) |
| 493 | error("No previous substitute re|No previous substitute regular expression"); |
| 494 | resre(subre); |
| 495 | return (c); |
| 496 | |
| 497 | default: |
| 498 | error("Badly formed re|Regular expression \\ must be followed by / or ?"); |
| 499 | } |
| 500 | if (c == eof || c == '\n' || c == EOF) { |
| 501 | if (*ep == 0) |
| 502 | error("No previous re|No previous regular expression"); |
| 503 | if (c == '\n' && oknl == 0) |
| 504 | error("Missing closing delimiter@for regular expression"); |
| 505 | if (c != eof) |
| 506 | ungetchar(c); |
| 507 | return (eof); |
| 508 | } |
| 509 | bracketp = bracket; |
| 510 | nbra = 0; |
| 511 | circfl = 0; |
| 512 | if (c == '^') { |
| 513 | c = getchar(); |
| 514 | circfl++; |
| 515 | } |
| 516 | ungetchar(c); |
| 517 | for (;;) { |
| 518 | if (ep >= &expbuf[ESIZE - 2]) |
| 519 | complex: |
| 520 | cerror("Re too complex|Regular expression too complicated"); |
| 521 | c = getchar(); |
| 522 | if (c == eof || c == EOF) { |
| 523 | if (bracketp != bracket) |
| 524 | cerror("Unmatched \\(|More \\('s than \\)'s in regular expression"); |
| 525 | *ep++ = CEOF; |
| 526 | if (c == EOF) |
| 527 | ungetchar(c); |
| 528 | return (eof); |
| 529 | } |
| 530 | if (value(MAGIC)) { |
| 531 | if (c != '*' || ep == expbuf) |
| 532 | lastep = ep; |
| 533 | } else |
| 534 | if (c != '\\' || peekchar() != '*' || ep == expbuf) |
| 535 | lastep = ep; |
| 536 | switch (c) { |
| 537 | |
| 538 | case '\\': |
| 539 | c = getchar(); |
| 540 | switch (c) { |
| 541 | |
| 542 | case '(': |
| 543 | if (nbra >= NBRA) |
| 544 | cerror("Awash in \\('s!|Too many \\('d subexressions in a regular expression"); |
| 545 | *bracketp++ = nbra; |
| 546 | *ep++ = CBRA; |
| 547 | *ep++ = nbra++; |
| 548 | continue; |
| 549 | |
| 550 | case ')': |
| 551 | if (bracketp <= bracket) |
| 552 | cerror("Extra \\)|More \\)'s than \\('s in regular expression"); |
| 553 | *ep++ = CKET; |
| 554 | *ep++ = *--bracketp; |
| 555 | continue; |
| 556 | |
| 557 | case '<': |
| 558 | *ep++ = CBRC; |
| 559 | continue; |
| 560 | |
| 561 | case '>': |
| 562 | *ep++ = CLET; |
| 563 | continue; |
| 564 | } |
| 565 | if (value(MAGIC) == 0) |
| 566 | magic: |
| 567 | switch (c) { |
| 568 | |
| 569 | case '.': |
| 570 | *ep++ = CDOT; |
| 571 | continue; |
| 572 | |
| 573 | case '~': |
| 574 | rhsp = rhsbuf; |
| 575 | while (*rhsp) { |
| 576 | if (*rhsp & QUOTE) { |
| 577 | c = *rhsp & TRIM; |
| 578 | if (c == '&') |
| 579 | error("Replacement pattern contains &@- cannot use in re"); |
| 580 | if (c >= '1' && c <= '9') |
| 581 | error("Replacement pattern contains \\d@- cannot use in re"); |
| 582 | } |
| 583 | if (ep >= &expbuf[ESIZE-2]) |
| 584 | goto complex; |
| 585 | *ep++ = CCHR; |
| 586 | *ep++ = *rhsp++ & TRIM; |
| 587 | } |
| 588 | continue; |
| 589 | |
| 590 | case '*': |
| 591 | if (ep == expbuf) |
| 592 | break; |
| 593 | if (*lastep == CBRA || *lastep == CKET) |
| 594 | cerror("Illegal *|Can't * a \\( ... \\) in regular expression"); |
| 595 | if (*lastep == CCHR && (lastep[1] & QUOTE)) |
| 596 | cerror("Illegal *|Can't * a \\n in regular expression"); |
| 597 | *lastep |= STAR; |
| 598 | continue; |
| 599 | |
| 600 | case '[': |
| 601 | *ep++ = CCL; |
| 602 | *ep++ = 0; |
| 603 | cclcnt = 1; |
| 604 | c = getchar(); |
| 605 | if (c == '^') { |
| 606 | c = getchar(); |
| 607 | ep[-2] = NCCL; |
| 608 | } |
| 609 | if (c == ']') |
| 610 | cerror("Bad character class|Empty character class '[]' or '[^]' cannot match"); |
| 611 | while (c != ']') { |
| 612 | if (c == '\\' && any(peekchar(), "]-^\\")) |
| 613 | c = getchar() | QUOTE; |
| 614 | if (c == '\n' || c == EOF) |
| 615 | cerror("Missing ]"); |
| 616 | *ep++ = c; |
| 617 | cclcnt++; |
| 618 | if (ep >= &expbuf[ESIZE]) |
| 619 | goto complex; |
| 620 | c = getchar(); |
| 621 | } |
| 622 | lastep[1] = cclcnt; |
| 623 | continue; |
| 624 | } |
| 625 | if (c == EOF) { |
| 626 | ungetchar(EOF); |
| 627 | c = '\\'; |
| 628 | goto defchar; |
| 629 | } |
| 630 | *ep++ = CCHR; |
| 631 | if (c == '\n') |
| 632 | cerror("No newlines in re's|Can't escape newlines into regular expressions"); |
| 633 | /* |
| 634 | if (c < '1' || c > NBRA + '1') { |
| 635 | */ |
| 636 | *ep++ = c; |
| 637 | continue; |
| 638 | /* |
| 639 | } |
| 640 | c -= '1'; |
| 641 | if (c >= nbra) |
| 642 | cerror("Bad \\n|\\n in regular expression with n greater than the number of \\('s"); |
| 643 | *ep++ = c | QUOTE; |
| 644 | continue; |
| 645 | */ |
| 646 | |
| 647 | case '\n': |
| 648 | if (oknl) { |
| 649 | ungetchar(c); |
| 650 | *ep++ = CEOF; |
| 651 | return (eof); |
| 652 | } |
| 653 | cerror("Badly formed re|Missing closing delimiter for regular expression"); |
| 654 | |
| 655 | case '$': |
| 656 | if (peekchar() == eof || peekchar() == EOF || oknl && peekchar() == '\n') { |
| 657 | *ep++ = CDOL; |
| 658 | continue; |
| 659 | } |
| 660 | goto defchar; |
| 661 | |
| 662 | case '.': |
| 663 | case '~': |
| 664 | case '*': |
| 665 | case '[': |
| 666 | if (value(MAGIC)) |
| 667 | goto magic; |
| 668 | defchar: |
| 669 | default: |
| 670 | *ep++ = CCHR; |
| 671 | *ep++ = c; |
| 672 | continue; |
| 673 | } |
| 674 | } |
| 675 | } |
| 676 | |
| 677 | cerror(s) |
| 678 | char *s; |
| 679 | { |
| 680 | |
| 681 | expbuf[0] = 0; |
| 682 | error(s); |
| 683 | } |
| 684 | |
| 685 | same(a, b) |
| 686 | register int a, b; |
| 687 | { |
| 688 | |
| 689 | return (a == b || value(IGNORECASE) && |
| 690 | ((islower(a) && toupper(a) == b) || (islower(b) && toupper(b) == a))); |
| 691 | } |
| 692 | |
| 693 | char *locs; |
| 694 | |
| 695 | execute(gf, addr) |
| 696 | line *addr; |
| 697 | { |
| 698 | register char *p1, *p2; |
| 699 | register int c; |
| 700 | |
| 701 | if (gf) { |
| 702 | if (circfl) |
| 703 | return (0); |
| 704 | #ifdef notdef |
| 705 | if (loc1 == loc2) |
| 706 | loc2++; |
| 707 | #endif |
| 708 | locs = p1 = loc2; |
| 709 | } else { |
| 710 | if (addr == zero) |
| 711 | return (0); |
| 712 | p1 = linebuf; |
| 713 | getline(*addr); |
| 714 | locs = 0; |
| 715 | } |
| 716 | p2 = expbuf; |
| 717 | if (circfl) { |
| 718 | loc1 = p1; |
| 719 | return (advance(p1, p2)); |
| 720 | } |
| 721 | /* fast check for first character */ |
| 722 | if (*p2 == CCHR) { |
| 723 | c = p2[1]; |
| 724 | do { |
| 725 | if (c != *p1 && (!value(IGNORECASE) || |
| 726 | !((islower(c) && toupper(c) == *p1) || |
| 727 | (islower(*p1) && toupper(*p1) == c)))) |
| 728 | continue; |
| 729 | if (advance(p1, p2)) { |
| 730 | loc1 = p1; |
| 731 | return (1); |
| 732 | } |
| 733 | } while (*p1++); |
| 734 | return (0); |
| 735 | } |
| 736 | /* regular algorithm */ |
| 737 | do { |
| 738 | if (advance(p1, p2)) { |
| 739 | loc1 = p1; |
| 740 | return (1); |
| 741 | } |
| 742 | } while (*p1++); |
| 743 | return (0); |
| 744 | } |
| 745 | |
| 746 | #define uletter(c) (isalpha(c) || c == '_') |
| 747 | |
| 748 | advance(lp, ep) |
| 749 | register char *lp, *ep; |
| 750 | { |
| 751 | register char *curlp; |
| 752 | char *sp, *sp1; |
| 753 | int c; |
| 754 | |
| 755 | for (;;) switch (*ep++) { |
| 756 | |
| 757 | case CCHR: |
| 758 | /* useless |
| 759 | if (*ep & QUOTE) { |
| 760 | c = *ep++ & TRIM; |
| 761 | sp = braslist[c]; |
| 762 | sp1 = braelist[c]; |
| 763 | while (sp < sp1) { |
| 764 | if (!same(*sp, *lp)) |
| 765 | return (0); |
| 766 | sp++, lp++; |
| 767 | } |
| 768 | continue; |
| 769 | } |
| 770 | */ |
| 771 | if (!same(*ep, *lp)) |
| 772 | return (0); |
| 773 | ep++, lp++; |
| 774 | continue; |
| 775 | |
| 776 | case CDOT: |
| 777 | if (*lp++) |
| 778 | continue; |
| 779 | return (0); |
| 780 | |
| 781 | case CDOL: |
| 782 | if (*lp == 0) |
| 783 | continue; |
| 784 | return (0); |
| 785 | |
| 786 | case CEOF: |
| 787 | loc2 = lp; |
| 788 | return (1); |
| 789 | |
| 790 | case CCL: |
| 791 | if (cclass(ep, *lp++, 1)) { |
| 792 | ep += *ep; |
| 793 | continue; |
| 794 | } |
| 795 | return (0); |
| 796 | |
| 797 | case NCCL: |
| 798 | if (cclass(ep, *lp++, 0)) { |
| 799 | ep += *ep; |
| 800 | continue; |
| 801 | } |
| 802 | return (0); |
| 803 | |
| 804 | case CBRA: |
| 805 | braslist[*ep++] = lp; |
| 806 | continue; |
| 807 | |
| 808 | case CKET: |
| 809 | braelist[*ep++] = lp; |
| 810 | continue; |
| 811 | |
| 812 | case CDOT|STAR: |
| 813 | curlp = lp; |
| 814 | while (*lp++) |
| 815 | continue; |
| 816 | goto star; |
| 817 | |
| 818 | case CCHR|STAR: |
| 819 | curlp = lp; |
| 820 | while (same(*lp, *ep)) |
| 821 | lp++; |
| 822 | lp++; |
| 823 | ep++; |
| 824 | goto star; |
| 825 | |
| 826 | case CCL|STAR: |
| 827 | case NCCL|STAR: |
| 828 | curlp = lp; |
| 829 | while (cclass(ep, *lp++, ep[-1] == (CCL|STAR))) |
| 830 | continue; |
| 831 | ep += *ep; |
| 832 | goto star; |
| 833 | star: |
| 834 | do { |
| 835 | lp--; |
| 836 | if (lp == locs) |
| 837 | break; |
| 838 | if (advance(lp, ep)) |
| 839 | return (1); |
| 840 | } while (lp > curlp); |
| 841 | return (0); |
| 842 | |
| 843 | case CBRC: |
| 844 | if (lp == expbuf) |
| 845 | continue; |
| 846 | if ((isdigit(*lp) || uletter(*lp)) && !uletter(lp[-1]) && !isdigit(lp[-1])) |
| 847 | continue; |
| 848 | return (0); |
| 849 | |
| 850 | case CLET: |
| 851 | if (!uletter(*lp) && !isdigit(*lp)) |
| 852 | continue; |
| 853 | return (0); |
| 854 | |
| 855 | default: |
| 856 | error("Re internal error"); |
| 857 | } |
| 858 | } |
| 859 | |
| 860 | cclass(set, c, af) |
| 861 | register char *set; |
| 862 | register int c; |
| 863 | int af; |
| 864 | { |
| 865 | register int n; |
| 866 | |
| 867 | if (c == 0) |
| 868 | return (0); |
| 869 | if (value(IGNORECASE) && isupper(c)) |
| 870 | c = tolower(c); |
| 871 | n = *set++; |
| 872 | while (--n) |
| 873 | if (n > 2 && set[1] == '-') { |
| 874 | if (c >= (set[0] & TRIM) && c <= (set[2] & TRIM)) |
| 875 | return (af); |
| 876 | set += 3; |
| 877 | n -= 2; |
| 878 | } else |
| 879 | if ((*set++ & TRIM) == c) |
| 880 | return (af); |
| 881 | return (!af); |
| 882 | } |