1 /*- 2 * Copyright (c) 1991, 1993 3 * The Regents of the University of California. All rights reserved. 4 * 5 * This code is derived from software contributed to Berkeley by 6 * Kenneth Almquist. 7 * 8 * Redistribution and use in source and binary forms, with or without 9 * modification, are permitted provided that the following conditions 10 * are met: 11 * 1. Redistributions of source code must retain the above copyright 12 * notice, this list of conditions and the following disclaimer. 13 * 2. Redistributions in binary form must reproduce the above copyright 14 * notice, this list of conditions and the following disclaimer in the 15 * documentation and/or other materials provided with the distribution. 16 * 3. All advertising materials mentioning features or use of this software 17 * must display the following acknowledgement: 18 * This product includes software developed by the University of 19 * California, Berkeley and its contributors. 20 * 4. Neither the name of the University nor the names of its contributors 21 * may be used to endorse or promote products derived from this software 22 * without specific prior written permission. 23 * 24 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 25 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 26 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 27 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 28 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 29 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 30 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 31 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 32 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 33 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 34 * SUCH DAMAGE. 35 */ 36 37 #ifndef lint 38 /*static char sccsid[] = "from: @(#)parser.c 8.1 (Berkeley) 5/31/93";*/ 39 static char *rcsid = "$Id: parser.c,v 1.22 1994/12/05 19:07:50 cgd Exp $"; 40 #endif /* not lint */ 41 42 #include "shell.h" 43 #include "parser.h" 44 #include "nodes.h" 45 #include "expand.h" /* defines rmescapes() */ 46 #include "redir.h" /* defines copyfd() */ 47 #include "syntax.h" 48 #include "options.h" 49 #include "input.h" 50 #include "output.h" 51 #include "var.h" 52 #include "error.h" 53 #include "memalloc.h" 54 #include "mystring.h" 55 #include "alias.h" 56 #ifndef NO_HISTORY 57 #include "myhistedit.h" 58 #endif 59 60 /* 61 * Shell command parser. 62 */ 63 64 #define EOFMARKLEN 79 65 66 /* values returned by readtoken */ 67 #include "token.def" 68 69 70 71 struct heredoc { 72 struct heredoc *next; /* next here document in list */ 73 union node *here; /* redirection node */ 74 char *eofmark; /* string indicating end of input */ 75 int striptabs; /* if set, strip leading tabs */ 76 }; 77 78 79 80 struct heredoc *heredoclist; /* list of here documents to read */ 81 int parsebackquote; /* nonzero if we are inside backquotes */ 82 int doprompt; /* if set, prompt the user */ 83 int needprompt; /* true if interactive and at start of line */ 84 int lasttoken; /* last token read */ 85 MKINIT int tokpushback; /* last token pushed back */ 86 char *wordtext; /* text of last word returned by readtoken */ 87 MKINIT int checkkwd; /* 1 == check for kwds, 2 == also eat newlines */ 88 struct nodelist *backquotelist; 89 union node *redirnode; 90 struct heredoc *heredoc; 91 int quoteflag; /* set if (part of) last token was quoted */ 92 int startlinno; /* line # where last token started */ 93 94 95 #define GDB_HACK 1 /* avoid local declarations which gdb can't handle */ 96 #ifdef GDB_HACK 97 static const char argvars[5] = {CTLVAR, VSNORMAL|VSQUOTE, '@', '=', '\0'}; 98 static const char types[] = "}-+?="; 99 #endif 100 101 102 STATIC union node *list __P((int)); 103 STATIC union node *andor __P((void)); 104 STATIC union node *pipeline __P((void)); 105 STATIC union node *command __P((void)); 106 STATIC union node *simplecmd __P((union node **, union node *)); 107 STATIC union node *makename __P((void)); 108 STATIC void parsefname __P((void)); 109 STATIC void parseheredoc __P((void)); 110 STATIC int readtoken __P((void)); 111 STATIC int readtoken1 __P((int, char const *, char *, int)); 112 STATIC void attyline __P((void)); 113 STATIC int noexpand __P((char *)); 114 STATIC int peektoken __P((void)); 115 STATIC void synexpect __P((int)); 116 STATIC void synerror __P((char *)); 117 STATIC void setprompt __P((int)); 118 119 120 /* 121 * Read and parse a command. Returns NEOF on end of file. (NULL is a 122 * valid parse tree indicating a blank line.) 123 */ 124 125 union node * 126 parsecmd(interact) 127 int interact; 128 { 129 int t; 130 131 doprompt = interact; 132 if (doprompt) 133 setprompt(1); 134 else 135 setprompt(0); 136 needprompt = 0; 137 t = readtoken(); 138 if (t == TEOF) 139 return NEOF; 140 if (t == TNL) 141 return NULL; 142 tokpushback++; 143 return list(1); 144 } 145 146 147 STATIC union node * 148 list(nlflag) 149 int nlflag; 150 { 151 union node *n1, *n2, *n3; 152 int tok; 153 154 checkkwd = 2; 155 if (nlflag == 0 && tokendlist[peektoken()]) 156 return NULL; 157 n1 = NULL; 158 for (;;) { 159 n2 = andor(); 160 tok = readtoken(); 161 if (tok == TBACKGND) { 162 if (n2->type == NCMD || n2->type == NPIPE) { 163 n2->ncmd.backgnd = 1; 164 } else if (n2->type == NREDIR) { 165 n2->type = NBACKGND; 166 } else { 167 n3 = (union node *)stalloc(sizeof (struct nredir)); 168 n3->type = NBACKGND; 169 n3->nredir.n = n2; 170 n3->nredir.redirect = NULL; 171 n2 = n3; 172 } 173 } 174 if (n1 == NULL) { 175 n1 = n2; 176 } 177 else { 178 n3 = (union node *)stalloc(sizeof (struct nbinary)); 179 n3->type = NSEMI; 180 n3->nbinary.ch1 = n1; 181 n3->nbinary.ch2 = n2; 182 n1 = n3; 183 } 184 switch (tok) { 185 case TBACKGND: 186 case TSEMI: 187 tok = readtoken(); 188 /* fall through */ 189 case TNL: 190 if (tok == TNL) { 191 parseheredoc(); 192 if (nlflag) 193 return n1; 194 } else { 195 tokpushback++; 196 } 197 checkkwd = 2; 198 if (tokendlist[peektoken()]) 199 return n1; 200 break; 201 case TEOF: 202 if (heredoclist) 203 parseheredoc(); 204 else 205 pungetc(); /* push back EOF on input */ 206 return n1; 207 default: 208 if (nlflag) 209 synexpect(-1); 210 tokpushback++; 211 return n1; 212 } 213 } 214 } 215 216 217 218 STATIC union node * 219 andor() { 220 union node *n1, *n2, *n3; 221 int t; 222 223 n1 = pipeline(); 224 for (;;) { 225 if ((t = readtoken()) == TAND) { 226 t = NAND; 227 } else if (t == TOR) { 228 t = NOR; 229 } else { 230 tokpushback++; 231 return n1; 232 } 233 n2 = pipeline(); 234 n3 = (union node *)stalloc(sizeof (struct nbinary)); 235 n3->type = t; 236 n3->nbinary.ch1 = n1; 237 n3->nbinary.ch2 = n2; 238 n1 = n3; 239 } 240 } 241 242 243 244 STATIC union node * 245 pipeline() { 246 union node *n1, *pipenode, *notnode; 247 struct nodelist *lp, *prev; 248 int negate = 0; 249 250 TRACE(("pipeline: entered\n")); 251 while (readtoken() == TNOT) { 252 TRACE(("pipeline: TNOT recognized\n")); 253 negate = !negate; 254 } 255 tokpushback++; 256 n1 = command(); 257 if (readtoken() == TPIPE) { 258 pipenode = (union node *)stalloc(sizeof (struct npipe)); 259 pipenode->type = NPIPE; 260 pipenode->npipe.backgnd = 0; 261 lp = (struct nodelist *)stalloc(sizeof (struct nodelist)); 262 pipenode->npipe.cmdlist = lp; 263 lp->n = n1; 264 do { 265 prev = lp; 266 lp = (struct nodelist *)stalloc(sizeof (struct nodelist)); 267 lp->n = command(); 268 prev->next = lp; 269 } while (readtoken() == TPIPE); 270 lp->next = NULL; 271 n1 = pipenode; 272 } 273 tokpushback++; 274 if (negate) { 275 notnode = (union node *)stalloc(sizeof (struct nnot)); 276 notnode->type = NNOT; 277 notnode->nnot.com = n1; 278 n1 = notnode; 279 } 280 return n1; 281 } 282 283 284 285 STATIC union node * 286 command() { 287 union node *n1, *n2; 288 union node *ap, **app; 289 union node *cp, **cpp; 290 union node *redir, **rpp; 291 int t; 292 293 checkkwd = 2; 294 redir = 0; 295 rpp = &redir; 296 /* Check for redirection which may precede command */ 297 while (readtoken() == TREDIR) { 298 *rpp = n2 = redirnode; 299 rpp = &n2->nfile.next; 300 parsefname(); 301 } 302 tokpushback++; 303 304 switch (readtoken()) { 305 case TIF: 306 n1 = (union node *)stalloc(sizeof (struct nif)); 307 n1->type = NIF; 308 n1->nif.test = list(0); 309 if (readtoken() != TTHEN) 310 synexpect(TTHEN); 311 n1->nif.ifpart = list(0); 312 n2 = n1; 313 while (readtoken() == TELIF) { 314 n2->nif.elsepart = (union node *)stalloc(sizeof (struct nif)); 315 n2 = n2->nif.elsepart; 316 n2->type = NIF; 317 n2->nif.test = list(0); 318 if (readtoken() != TTHEN) 319 synexpect(TTHEN); 320 n2->nif.ifpart = list(0); 321 } 322 if (lasttoken == TELSE) 323 n2->nif.elsepart = list(0); 324 else { 325 n2->nif.elsepart = NULL; 326 tokpushback++; 327 } 328 if (readtoken() != TFI) 329 synexpect(TFI); 330 checkkwd = 1; 331 break; 332 case TWHILE: 333 case TUNTIL: { 334 int got; 335 n1 = (union node *)stalloc(sizeof (struct nbinary)); 336 n1->type = (lasttoken == TWHILE)? NWHILE : NUNTIL; 337 n1->nbinary.ch1 = list(0); 338 if ((got=readtoken()) != TDO) { 339 TRACE(("expecting DO got %s %s\n", tokname[got], got == TWORD ? wordtext : "")); 340 synexpect(TDO); 341 } 342 n1->nbinary.ch2 = list(0); 343 if (readtoken() != TDONE) 344 synexpect(TDONE); 345 checkkwd = 1; 346 break; 347 } 348 case TFOR: 349 if (readtoken() != TWORD || quoteflag || ! goodname(wordtext)) 350 synerror("Bad for loop variable"); 351 n1 = (union node *)stalloc(sizeof (struct nfor)); 352 n1->type = NFOR; 353 n1->nfor.var = wordtext; 354 if (readtoken() == TWORD && ! quoteflag && equal(wordtext, "in")) { 355 app = ≈ 356 while (readtoken() == TWORD) { 357 n2 = (union node *)stalloc(sizeof (struct narg)); 358 n2->type = NARG; 359 n2->narg.text = wordtext; 360 n2->narg.backquote = backquotelist; 361 *app = n2; 362 app = &n2->narg.next; 363 } 364 *app = NULL; 365 n1->nfor.args = ap; 366 if (lasttoken != TNL && lasttoken != TSEMI) 367 synexpect(-1); 368 } else { 369 #ifndef GDB_HACK 370 static const char argvars[5] = {CTLVAR, VSNORMAL|VSQUOTE, 371 '@', '=', '\0'}; 372 #endif 373 n2 = (union node *)stalloc(sizeof (struct narg)); 374 n2->type = NARG; 375 n2->narg.text = (char *)argvars; 376 n2->narg.backquote = NULL; 377 n2->narg.next = NULL; 378 n1->nfor.args = n2; 379 /* 380 * Newline or semicolon here is optional (but note 381 * that the original Bourne shell only allowed NL). 382 */ 383 if (lasttoken != TNL && lasttoken != TSEMI) 384 tokpushback++; 385 } 386 checkkwd = 2; 387 if ((t = readtoken()) == TDO) 388 t = TDONE; 389 else if (t == TBEGIN) 390 t = TEND; 391 else 392 synexpect(-1); 393 n1->nfor.body = list(0); 394 if (readtoken() != t) 395 synexpect(t); 396 checkkwd = 1; 397 break; 398 case TCASE: 399 n1 = (union node *)stalloc(sizeof (struct ncase)); 400 n1->type = NCASE; 401 if (readtoken() != TWORD) 402 synexpect(TWORD); 403 n1->ncase.expr = n2 = (union node *)stalloc(sizeof (struct narg)); 404 n2->type = NARG; 405 n2->narg.text = wordtext; 406 n2->narg.backquote = backquotelist; 407 n2->narg.next = NULL; 408 while (readtoken() == TNL); 409 if (lasttoken != TWORD || ! equal(wordtext, "in")) 410 synerror("expecting \"in\""); 411 cpp = &n1->ncase.cases; 412 checkkwd = 2, readtoken(); 413 do { 414 *cpp = cp = (union node *)stalloc(sizeof (struct nclist)); 415 cp->type = NCLIST; 416 app = &cp->nclist.pattern; 417 for (;;) { 418 *app = ap = (union node *)stalloc(sizeof (struct narg)); 419 ap->type = NARG; 420 ap->narg.text = wordtext; 421 ap->narg.backquote = backquotelist; 422 if (checkkwd = 2, readtoken() != TPIPE) 423 break; 424 app = &ap->narg.next; 425 readtoken(); 426 } 427 ap->narg.next = NULL; 428 if (lasttoken != TRP) 429 synexpect(TRP); 430 cp->nclist.body = list(0); 431 432 checkkwd = 2; 433 if ((t = readtoken()) != TESAC) { 434 if (t != TENDCASE) 435 synexpect(TENDCASE); 436 else 437 checkkwd = 2, readtoken(); 438 } 439 cpp = &cp->nclist.next; 440 } while(lasttoken != TESAC); 441 *cpp = NULL; 442 checkkwd = 1; 443 break; 444 case TLP: 445 n1 = (union node *)stalloc(sizeof (struct nredir)); 446 n1->type = NSUBSHELL; 447 n1->nredir.n = list(0); 448 n1->nredir.redirect = NULL; 449 if (readtoken() != TRP) 450 synexpect(TRP); 451 checkkwd = 1; 452 break; 453 case TBEGIN: 454 n1 = list(0); 455 if (readtoken() != TEND) 456 synexpect(TEND); 457 checkkwd = 1; 458 break; 459 /* Handle an empty command like other simple commands. */ 460 case TSEMI: 461 /* 462 * An empty command before a ; doesn't make much sense, and 463 * should certainly be disallowed in the case of `if ;'. 464 */ 465 if (!redir) 466 synexpect(-1); 467 case TNL: 468 case TEOF: 469 case TWORD: 470 case TRP: 471 tokpushback++; 472 return simplecmd(rpp, redir); 473 default: 474 synexpect(-1); 475 } 476 477 /* Now check for redirection which may follow command */ 478 while (readtoken() == TREDIR) { 479 *rpp = n2 = redirnode; 480 rpp = &n2->nfile.next; 481 parsefname(); 482 } 483 tokpushback++; 484 *rpp = NULL; 485 if (redir) { 486 if (n1->type != NSUBSHELL) { 487 n2 = (union node *)stalloc(sizeof (struct nredir)); 488 n2->type = NREDIR; 489 n2->nredir.n = n1; 490 n1 = n2; 491 } 492 n1->nredir.redirect = redir; 493 } 494 return n1; 495 } 496 497 498 STATIC union node * 499 simplecmd(rpp, redir) 500 union node **rpp, *redir; 501 { 502 union node *args, **app; 503 union node **orig_rpp = rpp; 504 union node *n; 505 506 /* If we don't have any redirections already, then we must reset */ 507 /* rpp to be the address of the local redir variable. */ 508 if (redir == 0) 509 rpp = &redir; 510 511 args = NULL; 512 app = &args; 513 /* 514 * We save the incoming value, because we need this for shell 515 * functions. There can not be a redirect or an argument between 516 * the function name and the open parenthesis. 517 */ 518 orig_rpp = rpp; 519 520 for (;;) { 521 if (readtoken() == TWORD) { 522 n = (union node *)stalloc(sizeof (struct narg)); 523 n->type = NARG; 524 n->narg.text = wordtext; 525 n->narg.backquote = backquotelist; 526 *app = n; 527 app = &n->narg.next; 528 } else if (lasttoken == TREDIR) { 529 *rpp = n = redirnode; 530 rpp = &n->nfile.next; 531 parsefname(); /* read name of redirection file */ 532 } else if (lasttoken == TLP && app == &args->narg.next 533 && rpp == orig_rpp) { 534 /* We have a function */ 535 if (readtoken() != TRP) 536 synexpect(TRP); 537 #ifdef notdef 538 if (! goodname(n->narg.text)) 539 synerror("Bad function name"); 540 #endif 541 n->type = NDEFUN; 542 n->narg.next = command(); 543 return n; 544 } else { 545 tokpushback++; 546 break; 547 } 548 } 549 *app = NULL; 550 *rpp = NULL; 551 n = (union node *)stalloc(sizeof (struct ncmd)); 552 n->type = NCMD; 553 n->ncmd.backgnd = 0; 554 n->ncmd.args = args; 555 n->ncmd.redirect = redir; 556 return n; 557 } 558 559 STATIC union node * 560 makename() { 561 union node *n; 562 563 n = (union node *)stalloc(sizeof (struct narg)); 564 n->type = NARG; 565 n->narg.next = NULL; 566 n->narg.text = wordtext; 567 n->narg.backquote = backquotelist; 568 return n; 569 } 570 571 void fixredir(n, text, err) 572 union node *n; 573 const char *text; 574 int err; 575 { 576 TRACE(("Fix redir %s %d\n", text, err)); 577 if (!err) 578 n->ndup.vname = NULL; 579 580 if (is_digit(text[0]) && text[1] == '\0') 581 n->ndup.dupfd = digit_val(text[0]); 582 else if (text[0] == '-' && text[1] == '\0') 583 n->ndup.dupfd = -1; 584 else { 585 586 if (err) 587 synerror("Bad fd number"); 588 else 589 n->ndup.vname = makename(); 590 } 591 } 592 593 594 STATIC void 595 parsefname() { 596 union node *n = redirnode; 597 598 if (readtoken() != TWORD) 599 synexpect(-1); 600 if (n->type == NHERE) { 601 struct heredoc *here = heredoc; 602 struct heredoc *p; 603 int i; 604 605 if (quoteflag == 0) 606 n->type = NXHERE; 607 TRACE(("Here document %d\n", n->type)); 608 if (here->striptabs) { 609 while (*wordtext == '\t') 610 wordtext++; 611 } 612 if (! noexpand(wordtext) || (i = strlen(wordtext)) == 0 || i > EOFMARKLEN) 613 synerror("Illegal eof marker for << redirection"); 614 rmescapes(wordtext); 615 here->eofmark = wordtext; 616 here->next = NULL; 617 if (heredoclist == NULL) 618 heredoclist = here; 619 else { 620 for (p = heredoclist ; p->next ; p = p->next); 621 p->next = here; 622 } 623 } else if (n->type == NTOFD || n->type == NFROMFD) { 624 fixredir(n, wordtext, 0); 625 } else { 626 n->nfile.fname = makename(); 627 } 628 } 629 630 631 /* 632 * Input any here documents. 633 */ 634 635 STATIC void 636 parseheredoc() { 637 struct heredoc *here; 638 union node *n; 639 640 while (heredoclist) { 641 here = heredoclist; 642 heredoclist = here->next; 643 if (needprompt) { 644 setprompt(2); 645 needprompt = 0; 646 } 647 readtoken1(pgetc(), here->here->type == NHERE? SQSYNTAX : DQSYNTAX, 648 here->eofmark, here->striptabs); 649 n = (union node *)stalloc(sizeof (struct narg)); 650 n->narg.type = NARG; 651 n->narg.next = NULL; 652 n->narg.text = wordtext; 653 n->narg.backquote = backquotelist; 654 here->here->nhere.doc = n; 655 } 656 } 657 658 STATIC int 659 peektoken() { 660 int t; 661 662 t = readtoken(); 663 tokpushback++; 664 return (t); 665 } 666 667 STATIC int xxreadtoken(); 668 669 STATIC int 670 readtoken() { 671 int t; 672 int savecheckkwd = checkkwd; 673 struct alias *ap; 674 #ifdef DEBUG 675 int alreadyseen = tokpushback; 676 #endif 677 678 top: 679 t = xxreadtoken(); 680 681 if (checkkwd) { 682 /* 683 * eat newlines 684 */ 685 if (checkkwd == 2) { 686 checkkwd = 0; 687 while (t == TNL) { 688 parseheredoc(); 689 t = xxreadtoken(); 690 } 691 } else 692 checkkwd = 0; 693 /* 694 * check for keywords and aliases 695 */ 696 if (t == TWORD && !quoteflag) 697 { 698 register char * const *pp; 699 700 for (pp = (char **)parsekwd; *pp; pp++) { 701 if (**pp == *wordtext && equal(*pp, wordtext)) 702 { 703 lasttoken = t = pp - parsekwd + KWDOFFSET; 704 TRACE(("keyword %s recognized\n", tokname[t])); 705 goto out; 706 } 707 } 708 if (ap = lookupalias(wordtext, 1)) { 709 pushstring(ap->val, strlen(ap->val), ap); 710 checkkwd = savecheckkwd; 711 goto top; 712 } 713 } 714 out: 715 checkkwd = 0; 716 } 717 #ifdef DEBUG 718 if (!alreadyseen) 719 TRACE(("token %s %s\n", tokname[t], t == TWORD ? wordtext : "")); 720 else 721 TRACE(("reread token %s %s\n", tokname[t], t == TWORD ? wordtext : "")); 722 #endif 723 return (t); 724 } 725 726 727 /* 728 * Read the next input token. 729 * If the token is a word, we set backquotelist to the list of cmds in 730 * backquotes. We set quoteflag to true if any part of the word was 731 * quoted. 732 * If the token is TREDIR, then we set redirnode to a structure containing 733 * the redirection. 734 * In all cases, the variable startlinno is set to the number of the line 735 * on which the token starts. 736 * 737 * [Change comment: here documents and internal procedures] 738 * [Readtoken shouldn't have any arguments. Perhaps we should make the 739 * word parsing code into a separate routine. In this case, readtoken 740 * doesn't need to have any internal procedures, but parseword does. 741 * We could also make parseoperator in essence the main routine, and 742 * have parseword (readtoken1?) handle both words and redirection.] 743 */ 744 745 #define RETURN(token) return lasttoken = token 746 747 STATIC int 748 xxreadtoken() { 749 register c; 750 751 if (tokpushback) { 752 tokpushback = 0; 753 return lasttoken; 754 } 755 if (needprompt) { 756 setprompt(2); 757 needprompt = 0; 758 } 759 startlinno = plinno; 760 for (;;) { /* until token or start of word found */ 761 c = pgetc_macro(); 762 if (c == ' ' || c == '\t') 763 continue; /* quick check for white space first */ 764 switch (c) { 765 case ' ': case '\t': 766 continue; 767 case '#': 768 while ((c = pgetc()) != '\n' && c != PEOF); 769 pungetc(); 770 continue; 771 case '\\': 772 if (pgetc() == '\n') { 773 startlinno = ++plinno; 774 if (doprompt) 775 setprompt(2); 776 else 777 setprompt(0); 778 continue; 779 } 780 pungetc(); 781 goto breakloop; 782 case '\n': 783 plinno++; 784 needprompt = doprompt; 785 RETURN(TNL); 786 case PEOF: 787 RETURN(TEOF); 788 case '&': 789 if (pgetc() == '&') 790 RETURN(TAND); 791 pungetc(); 792 RETURN(TBACKGND); 793 case '|': 794 if (pgetc() == '|') 795 RETURN(TOR); 796 pungetc(); 797 RETURN(TPIPE); 798 case ';': 799 if (pgetc() == ';') 800 RETURN(TENDCASE); 801 pungetc(); 802 RETURN(TSEMI); 803 case '(': 804 RETURN(TLP); 805 case ')': 806 RETURN(TRP); 807 default: 808 goto breakloop; 809 } 810 } 811 breakloop: 812 return readtoken1(c, BASESYNTAX, (char *)NULL, 0); 813 #undef RETURN 814 } 815 816 817 818 /* 819 * If eofmark is NULL, read a word or a redirection symbol. If eofmark 820 * is not NULL, read a here document. In the latter case, eofmark is the 821 * word which marks the end of the document and striptabs is true if 822 * leading tabs should be stripped from the document. The argument firstc 823 * is the first character of the input token or document. 824 * 825 * Because C does not have internal subroutines, I have simulated them 826 * using goto's to implement the subroutine linkage. The following macros 827 * will run code that appears at the end of readtoken1. 828 */ 829 830 #define CHECKEND() {goto checkend; checkend_return:;} 831 #define PARSEREDIR() {goto parseredir; parseredir_return:;} 832 #define PARSESUB() {goto parsesub; parsesub_return:;} 833 #define PARSEBACKQOLD() {oldstyle = 1; goto parsebackq; parsebackq_oldreturn:;} 834 #define PARSEBACKQNEW() {oldstyle = 0; goto parsebackq; parsebackq_newreturn:;} 835 #define PARSEARITH() {goto parsearith; parsearith_return:;} 836 837 STATIC int 838 readtoken1(firstc, syntax, eofmark, striptabs) 839 int firstc; 840 char const *syntax; 841 char *eofmark; 842 int striptabs; 843 { 844 register c = firstc; 845 register char *out; 846 int len; 847 char line[EOFMARKLEN + 1]; 848 struct nodelist *bqlist; 849 int quotef; 850 int dblquote; 851 int varnest; /* levels of variables expansion */ 852 int arinest; /* levels of arithmetic expansion */ 853 int parenlevel; /* levels of parens in arithmetic */ 854 int oldstyle; 855 char const *prevsyntax; /* syntax before arithmetic */ 856 857 startlinno = plinno; 858 dblquote = 0; 859 if (syntax == DQSYNTAX) 860 dblquote = 1; 861 quotef = 0; 862 bqlist = NULL; 863 varnest = 0; 864 arinest = 0; 865 parenlevel = 0; 866 867 STARTSTACKSTR(out); 868 loop: { /* for each line, until end of word */ 869 #if ATTY 870 if (c == '\034' && doprompt 871 && attyset() && ! equal(termval(), "emacs")) { 872 attyline(); 873 if (syntax == BASESYNTAX) 874 return readtoken(); 875 c = pgetc(); 876 goto loop; 877 } 878 #endif 879 CHECKEND(); /* set c to PEOF if at end of here document */ 880 for (;;) { /* until end of line or end of word */ 881 CHECKSTRSPACE(3, out); /* permit 3 calls to USTPUTC */ 882 switch(syntax[c]) { 883 case CNL: /* '\n' */ 884 if (syntax == BASESYNTAX) 885 goto endword; /* exit outer loop */ 886 USTPUTC(c, out); 887 plinno++; 888 if (doprompt) 889 setprompt(2); 890 else 891 setprompt(0); 892 c = pgetc(); 893 goto loop; /* continue outer loop */ 894 case CWORD: 895 USTPUTC(c, out); 896 break; 897 case CCTL: 898 if (eofmark == NULL || dblquote) 899 USTPUTC(CTLESC, out); 900 USTPUTC(c, out); 901 break; 902 case CBACK: /* backslash */ 903 c = pgetc(); 904 if (c == PEOF) { 905 USTPUTC('\\', out); 906 pungetc(); 907 } else if (c == '\n') { 908 if (doprompt) 909 setprompt(2); 910 else 911 setprompt(0); 912 } else { 913 if (dblquote && c != '\\' && c != '`' && c != '$' 914 && (c != '"' || eofmark != NULL)) 915 USTPUTC('\\', out); 916 if (SQSYNTAX[c] == CCTL) 917 USTPUTC(CTLESC, out); 918 USTPUTC(c, out); 919 quotef++; 920 } 921 break; 922 case CSQUOTE: 923 syntax = SQSYNTAX; 924 break; 925 case CDQUOTE: 926 syntax = DQSYNTAX; 927 dblquote = 1; 928 break; 929 case CENDQUOTE: 930 if (eofmark) { 931 USTPUTC(c, out); 932 } else { 933 if (arinest) 934 syntax = ARISYNTAX; 935 else 936 syntax = BASESYNTAX; 937 quotef++; 938 dblquote = 0; 939 } 940 break; 941 case CVAR: /* '$' */ 942 PARSESUB(); /* parse substitution */ 943 break; 944 case CENDVAR: /* '}' */ 945 if (varnest > 0) { 946 varnest--; 947 USTPUTC(CTLENDVAR, out); 948 } else { 949 USTPUTC(c, out); 950 } 951 break; 952 case CLP: /* '(' in arithmetic */ 953 parenlevel++; 954 USTPUTC(c, out); 955 break; 956 case CRP: /* ')' in arithmetic */ 957 if (parenlevel > 0) { 958 USTPUTC(c, out); 959 --parenlevel; 960 } else { 961 if (pgetc() == ')') { 962 if (--arinest == 0) { 963 USTPUTC(CTLENDARI, out); 964 syntax = prevsyntax; 965 } else 966 USTPUTC(')', out); 967 } else { 968 /* 969 * unbalanced parens 970 * (don't 2nd guess - no error) 971 */ 972 pungetc(); 973 USTPUTC(')', out); 974 } 975 } 976 break; 977 case CBQUOTE: /* '`' */ 978 PARSEBACKQOLD(); 979 break; 980 case CEOF: 981 goto endword; /* exit outer loop */ 982 default: 983 if (varnest == 0) 984 goto endword; /* exit outer loop */ 985 USTPUTC(c, out); 986 } 987 c = pgetc_macro(); 988 } 989 } 990 endword: 991 if (syntax == ARISYNTAX) 992 synerror("Missing '))'"); 993 if (syntax != BASESYNTAX && ! parsebackquote && eofmark == NULL) 994 synerror("Unterminated quoted string"); 995 if (varnest != 0) { 996 startlinno = plinno; 997 synerror("Missing '}'"); 998 } 999 USTPUTC('\0', out); 1000 len = out - stackblock(); 1001 out = stackblock(); 1002 if (eofmark == NULL) { 1003 if ((c == '>' || c == '<') 1004 && quotef == 0 1005 && len <= 2 1006 && (*out == '\0' || is_digit(*out))) { 1007 PARSEREDIR(); 1008 return lasttoken = TREDIR; 1009 } else { 1010 pungetc(); 1011 } 1012 } 1013 quoteflag = quotef; 1014 backquotelist = bqlist; 1015 grabstackblock(len); 1016 wordtext = out; 1017 return lasttoken = TWORD; 1018 /* end of readtoken routine */ 1019 1020 1021 1022 /* 1023 * Check to see whether we are at the end of the here document. When this 1024 * is called, c is set to the first character of the next input line. If 1025 * we are at the end of the here document, this routine sets the c to PEOF. 1026 */ 1027 1028 checkend: { 1029 if (eofmark) { 1030 if (striptabs) { 1031 while (c == '\t') 1032 c = pgetc(); 1033 } 1034 if (c == *eofmark) { 1035 if (pfgets(line, sizeof line) != NULL) { 1036 register char *p, *q; 1037 1038 p = line; 1039 for (q = eofmark + 1 ; *q && *p == *q ; p++, q++); 1040 if (*p == '\n' && *q == '\0') { 1041 c = PEOF; 1042 plinno++; 1043 needprompt = doprompt; 1044 } else { 1045 pushstring(line, strlen(line), NULL); 1046 } 1047 } 1048 } 1049 } 1050 goto checkend_return; 1051 } 1052 1053 1054 /* 1055 * Parse a redirection operator. The variable "out" points to a string 1056 * specifying the fd to be redirected. The variable "c" contains the 1057 * first character of the redirection operator. 1058 */ 1059 1060 parseredir: { 1061 char fd = *out; 1062 union node *np; 1063 1064 np = (union node *)stalloc(sizeof (struct nfile)); 1065 if (c == '>') { 1066 np->nfile.fd = 1; 1067 c = pgetc(); 1068 if (c == '>') 1069 np->type = NAPPEND; 1070 else if (c == '&') 1071 np->type = NTOFD; 1072 else { 1073 np->type = NTO; 1074 pungetc(); 1075 } 1076 } else { /* c == '<' */ 1077 np->nfile.fd = 0; 1078 c = pgetc(); 1079 if (c == '<') { 1080 if (sizeof (struct nfile) != sizeof (struct nhere)) { 1081 np = (union node *)stalloc(sizeof (struct nhere)); 1082 np->nfile.fd = 0; 1083 } 1084 np->type = NHERE; 1085 heredoc = (struct heredoc *)stalloc(sizeof (struct heredoc)); 1086 heredoc->here = np; 1087 if ((c = pgetc()) == '-') { 1088 heredoc->striptabs = 1; 1089 } else { 1090 heredoc->striptabs = 0; 1091 pungetc(); 1092 } 1093 } else if (c == '&') 1094 np->type = NFROMFD; 1095 else { 1096 np->type = NFROM; 1097 pungetc(); 1098 } 1099 } 1100 if (fd != '\0') 1101 np->nfile.fd = digit_val(fd); 1102 redirnode = np; 1103 goto parseredir_return; 1104 } 1105 1106 1107 /* 1108 * Parse a substitution. At this point, we have read the dollar sign 1109 * and nothing else. 1110 */ 1111 1112 parsesub: { 1113 int subtype; 1114 int typeloc; 1115 int flags; 1116 char *p; 1117 #ifndef GDB_HACK 1118 static const char types[] = "}-+?="; 1119 #endif 1120 1121 c = pgetc(); 1122 if (c != '(' && c != '{' && !is_name(c) && !is_special(c)) { 1123 USTPUTC('$', out); 1124 pungetc(); 1125 } else if (c == '(') { /* $(command) or $((arith)) */ 1126 if (pgetc() == '(') { 1127 PARSEARITH(); 1128 } else { 1129 pungetc(); 1130 PARSEBACKQNEW(); 1131 } 1132 } else { 1133 USTPUTC(CTLVAR, out); 1134 typeloc = out - stackblock(); 1135 USTPUTC(VSNORMAL, out); 1136 subtype = VSNORMAL; 1137 if (c == '{') { 1138 c = pgetc(); 1139 subtype = 0; 1140 } 1141 if (is_name(c)) { 1142 do { 1143 STPUTC(c, out); 1144 c = pgetc(); 1145 } while (is_in_name(c)); 1146 } else { 1147 if (! is_special(c)) 1148 badsub: synerror("Bad substitution"); 1149 USTPUTC(c, out); 1150 c = pgetc(); 1151 } 1152 STPUTC('=', out); 1153 flags = 0; 1154 if (subtype == 0) { 1155 if (c == ':') { 1156 flags = VSNUL; 1157 c = pgetc(); 1158 } 1159 p = strchr(types, c); 1160 if (p == NULL) 1161 goto badsub; 1162 subtype = p - types + VSNORMAL; 1163 } else { 1164 pungetc(); 1165 } 1166 if (dblquote || arinest) 1167 flags |= VSQUOTE; 1168 *(stackblock() + typeloc) = subtype | flags; 1169 if (subtype != VSNORMAL) 1170 varnest++; 1171 } 1172 goto parsesub_return; 1173 } 1174 1175 1176 /* 1177 * Called to parse command substitutions. Newstyle is set if the command 1178 * is enclosed inside $(...); nlpp is a pointer to the head of the linked 1179 * list of commands (passed by reference), and savelen is the number of 1180 * characters on the top of the stack which must be preserved. 1181 */ 1182 1183 parsebackq: { 1184 struct nodelist **nlpp; 1185 int savepbq; 1186 union node *n; 1187 char *volatile str; 1188 struct jmploc jmploc; 1189 struct jmploc *volatile savehandler; 1190 int savelen; 1191 1192 savepbq = parsebackquote; 1193 if (setjmp(jmploc.loc)) { 1194 if (str) 1195 ckfree(str); 1196 parsebackquote = 0; 1197 handler = savehandler; 1198 longjmp(handler->loc, 1); 1199 } 1200 INTOFF; 1201 str = NULL; 1202 savelen = out - stackblock(); 1203 if (savelen > 0) { 1204 str = ckmalloc(savelen); 1205 memcpy(str, stackblock(), savelen); 1206 } 1207 savehandler = handler; 1208 handler = &jmploc; 1209 INTON; 1210 if (oldstyle) { 1211 /* We must read until the closing backquote, giving special 1212 treatment to some slashes, and then push the string and 1213 reread it as input, interpreting it normally. */ 1214 register char *out; 1215 register c; 1216 int savelen; 1217 char *str; 1218 1219 STARTSTACKSTR(out); 1220 while ((c = pgetc ()) != '`') { 1221 if (c == '\\') { 1222 c = pgetc (); 1223 if (c != '\\' && c != '`' && c != '$' 1224 && (!dblquote || c != '"')) 1225 STPUTC('\\', out); 1226 } 1227 STPUTC(c, out); 1228 } 1229 STPUTC('\0', out); 1230 savelen = out - stackblock(); 1231 if (savelen > 0) { 1232 str = ckmalloc(savelen); 1233 memcpy(str, stackblock(), savelen); 1234 } 1235 setinputstring(str, 1); 1236 } 1237 nlpp = &bqlist; 1238 while (*nlpp) 1239 nlpp = &(*nlpp)->next; 1240 *nlpp = (struct nodelist *)stalloc(sizeof (struct nodelist)); 1241 (*nlpp)->next = NULL; 1242 parsebackquote = oldstyle; 1243 n = list(0); 1244 if (!oldstyle && (readtoken() != TRP)) 1245 synexpect(TRP); 1246 (*nlpp)->n = n; 1247 /* Start reading from old file again. */ 1248 if (oldstyle) 1249 popfile(); 1250 while (stackblocksize() <= savelen) 1251 growstackblock(); 1252 STARTSTACKSTR(out); 1253 if (str) { 1254 memcpy(out, str, savelen); 1255 STADJUST(savelen, out); 1256 INTOFF; 1257 ckfree(str); 1258 str = NULL; 1259 INTON; 1260 } 1261 parsebackquote = savepbq; 1262 handler = savehandler; 1263 if (arinest || dblquote) 1264 USTPUTC(CTLBACKQ | CTLQUOTE, out); 1265 else 1266 USTPUTC(CTLBACKQ, out); 1267 if (oldstyle) 1268 goto parsebackq_oldreturn; 1269 else 1270 goto parsebackq_newreturn; 1271 } 1272 1273 /* 1274 * Parse an arithmetic expansion (indicate start of one and set state) 1275 */ 1276 parsearith: { 1277 1278 if (++arinest == 1) { 1279 prevsyntax = syntax; 1280 syntax = ARISYNTAX; 1281 USTPUTC(CTLARI, out); 1282 } else { 1283 /* 1284 * we collapse embedded arithmetic expansion to 1285 * parenthesis, which should be equivalent 1286 */ 1287 USTPUTC('(', out); 1288 } 1289 goto parsearith_return; 1290 } 1291 1292 } /* end of readtoken */ 1293 1294 1295 1296 #ifdef mkinit 1297 RESET { 1298 tokpushback = 0; 1299 checkkwd = 0; 1300 } 1301 #endif 1302 1303 /* 1304 * Returns true if the text contains nothing to expand (no dollar signs 1305 * or backquotes). 1306 */ 1307 1308 STATIC int 1309 noexpand(text) 1310 char *text; 1311 { 1312 register char *p; 1313 register char c; 1314 1315 p = text; 1316 while ((c = *p++) != '\0') { 1317 if (c == CTLESC) 1318 p++; 1319 else if (BASESYNTAX[c] == CCTL) 1320 return 0; 1321 } 1322 return 1; 1323 } 1324 1325 1326 /* 1327 * Return true if the argument is a legal variable name (a letter or 1328 * underscore followed by zero or more letters, underscores, and digits). 1329 */ 1330 1331 int 1332 goodname(name) 1333 char *name; 1334 { 1335 register char *p; 1336 1337 p = name; 1338 if (! is_name(*p)) 1339 return 0; 1340 while (*++p) { 1341 if (! is_in_name(*p)) 1342 return 0; 1343 } 1344 return 1; 1345 } 1346 1347 1348 /* 1349 * Called when an unexpected token is read during the parse. The argument 1350 * is the token that is expected, or -1 if more than one type of token can 1351 * occur at this point. 1352 */ 1353 1354 STATIC void 1355 synexpect(token) 1356 int token; 1357 { 1358 char msg[64]; 1359 1360 if (token >= 0) { 1361 fmtstr(msg, 64, "%s unexpected (expecting %s)", 1362 tokname[lasttoken], tokname[token]); 1363 } else { 1364 fmtstr(msg, 64, "%s unexpected", tokname[lasttoken]); 1365 } 1366 synerror(msg); 1367 } 1368 1369 1370 STATIC void 1371 synerror(msg) 1372 char *msg; 1373 { 1374 if (commandname) 1375 outfmt(&errout, "%s: %d: ", commandname, startlinno); 1376 outfmt(&errout, "Syntax error: %s\n", msg); 1377 error((char *)NULL); 1378 } 1379 1380 STATIC void 1381 setprompt(which) 1382 int which; 1383 { 1384 whichprompt = which; 1385 1386 #ifndef NO_HISTORY 1387 if (!el) 1388 #endif 1389 out2str(getprompt(NULL)); 1390 } 1391 1392 /* 1393 * called by editline -- any expansions to the prompt 1394 * should be added here. 1395 */ 1396 char * 1397 getprompt(unused) 1398 void *unused; 1399 { 1400 switch (whichprompt) { 1401 case 0: 1402 return ""; 1403 case 1: 1404 return ps1val(); 1405 case 2: 1406 return ps2val(); 1407 default: 1408 return "<internal prompt error>"; 1409 } 1410 } 1411