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 * @(#)parser.c 8.7 (Berkeley) 5/16/95 37 * $FreeBSD: src/bin/sh/parser.c,v 1.29.2.9 2002/10/18 11:24:04 tjr Exp $ 38 * $DragonFly: src/bin/sh/parser.c,v 1.6 2005/11/06 11:44:02 swildner Exp $ 39 */ 40 41 #include <stdlib.h> 42 43 #include "shell.h" 44 #include "parser.h" 45 #include "nodes.h" 46 #include "expand.h" /* defines rmescapes() */ 47 #include "redir.h" /* defines copyfd() */ 48 #include "syntax.h" 49 #include "options.h" 50 #include "input.h" 51 #include "output.h" 52 #include "var.h" 53 #include "error.h" 54 #include "memalloc.h" 55 #include "mystring.h" 56 #include "alias.h" 57 #include "show.h" 58 #include "eval.h" 59 #ifndef NO_HISTORY 60 #include "myhistedit.h" 61 #endif 62 63 /* 64 * Shell command parser. 65 */ 66 67 #define EOFMARKLEN 79 68 69 /* values returned by readtoken */ 70 #include "token.h" 71 72 73 74 struct heredoc { 75 struct heredoc *next; /* next here document in list */ 76 union node *here; /* redirection node */ 77 char *eofmark; /* string indicating end of input */ 78 int striptabs; /* if set, strip leading tabs */ 79 }; 80 81 82 83 STATIC struct heredoc *heredoclist; /* list of here documents to read */ 84 STATIC int parsebackquote; /* nonzero if we are inside backquotes */ 85 STATIC int doprompt; /* if set, prompt the user */ 86 STATIC int needprompt; /* true if interactive and at start of line */ 87 STATIC int lasttoken; /* last token read */ 88 MKINIT int tokpushback; /* last token pushed back */ 89 STATIC char *wordtext; /* text of last word returned by readtoken */ 90 MKINIT int checkkwd; /* 1 == check for kwds, 2 == also eat newlines */ 91 STATIC struct nodelist *backquotelist; 92 STATIC union node *redirnode; 93 STATIC struct heredoc *heredoc; 94 STATIC int quoteflag; /* set if (part of) last token was quoted */ 95 STATIC int startlinno; /* line # where last token started */ 96 97 /* XXX When 'noaliases' is set to one, no alias expansion takes place. */ 98 static int noaliases = 0; 99 100 #define GDB_HACK 1 /* avoid local declarations which gdb can't handle */ 101 #ifdef GDB_HACK 102 static const char argvars[5] = {CTLVAR, VSNORMAL|VSQUOTE, '@', '=', '\0'}; 103 static const char types[] = "}-+?="; 104 #endif 105 106 107 STATIC union node *list(int); 108 STATIC union node *andor(void); 109 STATIC union node *pipeline(void); 110 STATIC union node *command(void); 111 STATIC union node *simplecmd(union node **, union node *); 112 STATIC union node *makename(void); 113 STATIC void parsefname(void); 114 STATIC void parseheredoc(void); 115 STATIC int peektoken(void); 116 STATIC int readtoken(void); 117 STATIC int xxreadtoken(void); 118 STATIC int readtoken1(int, char const *, char *, int); 119 STATIC int noexpand(char *); 120 STATIC void synexpect(int); 121 STATIC void synerror(const char *); 122 STATIC void setprompt(int); 123 124 125 /* 126 * Read and parse a command. Returns NEOF on end of file. (NULL is a 127 * valid parse tree indicating a blank line.) 128 */ 129 130 union node * 131 parsecmd(int interact) 132 { 133 int t; 134 135 tokpushback = 0; 136 doprompt = interact; 137 if (doprompt) 138 setprompt(1); 139 else 140 setprompt(0); 141 needprompt = 0; 142 t = readtoken(); 143 if (t == TEOF) 144 return NEOF; 145 if (t == TNL) 146 return NULL; 147 tokpushback++; 148 return list(1); 149 } 150 151 152 STATIC union node * 153 list(int nlflag) 154 { 155 union node *n1, *n2, *n3; 156 int tok; 157 158 checkkwd = 2; 159 if (nlflag == 0 && tokendlist[peektoken()]) 160 return NULL; 161 n1 = NULL; 162 for (;;) { 163 n2 = andor(); 164 tok = readtoken(); 165 if (tok == TBACKGND) { 166 if (n2->type == NCMD || n2->type == NPIPE) { 167 n2->ncmd.backgnd = 1; 168 } else if (n2->type == NREDIR) { 169 n2->type = NBACKGND; 170 } else { 171 n3 = (union node *)stalloc(sizeof (struct nredir)); 172 n3->type = NBACKGND; 173 n3->nredir.n = n2; 174 n3->nredir.redirect = NULL; 175 n2 = n3; 176 } 177 } 178 if (n1 == NULL) { 179 n1 = n2; 180 } 181 else { 182 n3 = (union node *)stalloc(sizeof (struct nbinary)); 183 n3->type = NSEMI; 184 n3->nbinary.ch1 = n1; 185 n3->nbinary.ch2 = n2; 186 n1 = n3; 187 } 188 switch (tok) { 189 case TBACKGND: 190 case TSEMI: 191 tok = readtoken(); 192 /* fall through */ 193 case TNL: 194 if (tok == TNL) { 195 parseheredoc(); 196 if (nlflag) 197 return n1; 198 } else { 199 tokpushback++; 200 } 201 checkkwd = 2; 202 if (tokendlist[peektoken()]) 203 return n1; 204 break; 205 case TEOF: 206 if (heredoclist) 207 parseheredoc(); 208 else 209 pungetc(); /* push back EOF on input */ 210 return n1; 211 default: 212 if (nlflag) 213 synexpect(-1); 214 tokpushback++; 215 return n1; 216 } 217 } 218 } 219 220 221 222 STATIC union node * 223 andor(void) 224 { 225 union node *n1, *n2, *n3; 226 int t; 227 228 n1 = pipeline(); 229 for (;;) { 230 if ((t = readtoken()) == TAND) { 231 t = NAND; 232 } else if (t == TOR) { 233 t = NOR; 234 } else { 235 tokpushback++; 236 return n1; 237 } 238 n2 = pipeline(); 239 n3 = (union node *)stalloc(sizeof (struct nbinary)); 240 n3->type = t; 241 n3->nbinary.ch1 = n1; 242 n3->nbinary.ch2 = n2; 243 n1 = n3; 244 } 245 } 246 247 248 249 STATIC union node * 250 pipeline(void) 251 { 252 union node *n1, *n2, *pipenode; 253 struct nodelist *lp, *prev; 254 int negate; 255 256 negate = 0; 257 TRACE(("pipeline: entered\n")); 258 while (readtoken() == TNOT) 259 negate = !negate; 260 tokpushback++; 261 n1 = command(); 262 if (readtoken() == TPIPE) { 263 pipenode = (union node *)stalloc(sizeof (struct npipe)); 264 pipenode->type = NPIPE; 265 pipenode->npipe.backgnd = 0; 266 lp = (struct nodelist *)stalloc(sizeof (struct nodelist)); 267 pipenode->npipe.cmdlist = lp; 268 lp->n = n1; 269 do { 270 prev = lp; 271 lp = (struct nodelist *)stalloc(sizeof (struct nodelist)); 272 lp->n = command(); 273 prev->next = lp; 274 } while (readtoken() == TPIPE); 275 lp->next = NULL; 276 n1 = pipenode; 277 } 278 tokpushback++; 279 if (negate) { 280 n2 = (union node *)stalloc(sizeof (struct nnot)); 281 n2->type = NNOT; 282 n2->nnot.com = n1; 283 return n2; 284 } else 285 return n1; 286 } 287 288 289 290 STATIC union node * 291 command(void) 292 { 293 union node *n1, *n2; 294 union node *ap, **app; 295 union node *cp, **cpp; 296 union node *redir, **rpp; 297 int t, negate = 0; 298 299 checkkwd = 2; 300 redir = NULL; 301 n1 = NULL; 302 rpp = &redir; 303 304 /* Check for redirection which may precede command */ 305 while (readtoken() == TREDIR) { 306 *rpp = n2 = redirnode; 307 rpp = &n2->nfile.next; 308 parsefname(); 309 } 310 tokpushback++; 311 312 while (readtoken() == TNOT) { 313 TRACE(("command: TNOT recognized\n")); 314 negate = !negate; 315 } 316 tokpushback++; 317 318 switch (readtoken()) { 319 case TIF: 320 n1 = (union node *)stalloc(sizeof (struct nif)); 321 n1->type = NIF; 322 n1->nif.test = list(0); 323 if (readtoken() != TTHEN) 324 synexpect(TTHEN); 325 n1->nif.ifpart = list(0); 326 n2 = n1; 327 while (readtoken() == TELIF) { 328 n2->nif.elsepart = (union node *)stalloc(sizeof (struct nif)); 329 n2 = n2->nif.elsepart; 330 n2->type = NIF; 331 n2->nif.test = list(0); 332 if (readtoken() != TTHEN) 333 synexpect(TTHEN); 334 n2->nif.ifpart = list(0); 335 } 336 if (lasttoken == TELSE) 337 n2->nif.elsepart = list(0); 338 else { 339 n2->nif.elsepart = NULL; 340 tokpushback++; 341 } 342 if (readtoken() != TFI) 343 synexpect(TFI); 344 checkkwd = 1; 345 break; 346 case TWHILE: 347 case TUNTIL: { 348 int got; 349 n1 = (union node *)stalloc(sizeof (struct nbinary)); 350 n1->type = (lasttoken == TWHILE)? NWHILE : NUNTIL; 351 n1->nbinary.ch1 = list(0); 352 if ((got=readtoken()) != TDO) { 353 TRACE(("expecting DO got %s %s\n", tokname[got], got == TWORD ? wordtext : "")); 354 synexpect(TDO); 355 } 356 n1->nbinary.ch2 = list(0); 357 if (readtoken() != TDONE) 358 synexpect(TDONE); 359 checkkwd = 1; 360 break; 361 } 362 case TFOR: 363 if (readtoken() != TWORD || quoteflag || ! goodname(wordtext)) 364 synerror("Bad for loop variable"); 365 n1 = (union node *)stalloc(sizeof (struct nfor)); 366 n1->type = NFOR; 367 n1->nfor.var = wordtext; 368 if (readtoken() == TWORD && ! quoteflag && equal(wordtext, "in")) { 369 app = ≈ 370 while (readtoken() == TWORD) { 371 n2 = (union node *)stalloc(sizeof (struct narg)); 372 n2->type = NARG; 373 n2->narg.text = wordtext; 374 n2->narg.backquote = backquotelist; 375 *app = n2; 376 app = &n2->narg.next; 377 } 378 *app = NULL; 379 n1->nfor.args = ap; 380 if (lasttoken != TNL && lasttoken != TSEMI) 381 synexpect(-1); 382 } else { 383 #ifndef GDB_HACK 384 static const char argvars[5] = {CTLVAR, VSNORMAL|VSQUOTE, 385 '@', '=', '\0'}; 386 #endif 387 n2 = (union node *)stalloc(sizeof (struct narg)); 388 n2->type = NARG; 389 n2->narg.text = (char *)argvars; 390 n2->narg.backquote = NULL; 391 n2->narg.next = NULL; 392 n1->nfor.args = n2; 393 /* 394 * Newline or semicolon here is optional (but note 395 * that the original Bourne shell only allowed NL). 396 */ 397 if (lasttoken != TNL && lasttoken != TSEMI) 398 tokpushback++; 399 } 400 checkkwd = 2; 401 if ((t = readtoken()) == TDO) 402 t = TDONE; 403 else if (t == TBEGIN) 404 t = TEND; 405 else 406 synexpect(-1); 407 n1->nfor.body = list(0); 408 if (readtoken() != t) 409 synexpect(t); 410 checkkwd = 1; 411 break; 412 case TCASE: 413 n1 = (union node *)stalloc(sizeof (struct ncase)); 414 n1->type = NCASE; 415 if (readtoken() != TWORD) 416 synexpect(TWORD); 417 n1->ncase.expr = n2 = (union node *)stalloc(sizeof (struct narg)); 418 n2->type = NARG; 419 n2->narg.text = wordtext; 420 n2->narg.backquote = backquotelist; 421 n2->narg.next = NULL; 422 while (readtoken() == TNL); 423 if (lasttoken != TWORD || ! equal(wordtext, "in")) 424 synerror("expecting \"in\""); 425 cpp = &n1->ncase.cases; 426 noaliases = 1; /* turn off alias expansion */ 427 checkkwd = 2, readtoken(); 428 while (lasttoken != TESAC) { 429 *cpp = cp = (union node *)stalloc(sizeof (struct nclist)); 430 cp->type = NCLIST; 431 app = &cp->nclist.pattern; 432 if (lasttoken == TLP) 433 readtoken(); 434 for (;;) { 435 *app = ap = (union node *)stalloc(sizeof (struct narg)); 436 ap->type = NARG; 437 ap->narg.text = wordtext; 438 ap->narg.backquote = backquotelist; 439 if (checkkwd = 2, readtoken() != TPIPE) 440 break; 441 app = &ap->narg.next; 442 readtoken(); 443 } 444 ap->narg.next = NULL; 445 if (lasttoken != TRP) 446 noaliases = 0, synexpect(TRP); 447 cp->nclist.body = list(0); 448 449 checkkwd = 2; 450 if ((t = readtoken()) != TESAC) { 451 if (t != TENDCASE) 452 noaliases = 0, synexpect(TENDCASE); 453 else 454 checkkwd = 2, readtoken(); 455 } 456 cpp = &cp->nclist.next; 457 } 458 noaliases = 0; /* reset alias expansion */ 459 *cpp = NULL; 460 checkkwd = 1; 461 break; 462 case TLP: 463 n1 = (union node *)stalloc(sizeof (struct nredir)); 464 n1->type = NSUBSHELL; 465 n1->nredir.n = list(0); 466 n1->nredir.redirect = NULL; 467 if (readtoken() != TRP) 468 synexpect(TRP); 469 checkkwd = 1; 470 break; 471 case TBEGIN: 472 n1 = list(0); 473 if (readtoken() != TEND) 474 synexpect(TEND); 475 checkkwd = 1; 476 break; 477 /* Handle an empty command like other simple commands. */ 478 case TSEMI: 479 case TAND: 480 case TOR: 481 /* 482 * An empty command before a ; doesn't make much sense, and 483 * should certainly be disallowed in the case of `if ;'. 484 */ 485 if (!redir) 486 synexpect(-1); 487 case TNL: 488 case TEOF: 489 case TWORD: 490 case TRP: 491 tokpushback++; 492 n1 = simplecmd(rpp, redir); 493 goto checkneg; 494 default: 495 synexpect(-1); 496 } 497 498 /* Now check for redirection which may follow command */ 499 while (readtoken() == TREDIR) { 500 *rpp = n2 = redirnode; 501 rpp = &n2->nfile.next; 502 parsefname(); 503 } 504 tokpushback++; 505 *rpp = NULL; 506 if (redir) { 507 if (n1->type != NSUBSHELL) { 508 n2 = (union node *)stalloc(sizeof (struct nredir)); 509 n2->type = NREDIR; 510 n2->nredir.n = n1; 511 n1 = n2; 512 } 513 n1->nredir.redirect = redir; 514 } 515 516 checkneg: 517 if (negate) { 518 n2 = (union node *)stalloc(sizeof (struct nnot)); 519 n2->type = NNOT; 520 n2->nnot.com = n1; 521 return n2; 522 } 523 else 524 return n1; 525 } 526 527 528 STATIC union node * 529 simplecmd(union node **rpp, union node *redir) 530 { 531 union node *args, **app; 532 union node **orig_rpp = rpp; 533 union node *n = NULL, *n2; 534 int negate = 0; 535 536 /* If we don't have any redirections already, then we must reset */ 537 /* rpp to be the address of the local redir variable. */ 538 if (redir == 0) 539 rpp = &redir; 540 541 args = NULL; 542 app = &args; 543 /* 544 * We save the incoming value, because we need this for shell 545 * functions. There can not be a redirect or an argument between 546 * the function name and the open parenthesis. 547 */ 548 orig_rpp = rpp; 549 550 while (readtoken() == TNOT) { 551 TRACE(("command: TNOT recognized\n")); 552 negate = !negate; 553 } 554 tokpushback++; 555 556 for (;;) { 557 if (readtoken() == TWORD) { 558 n = (union node *)stalloc(sizeof (struct narg)); 559 n->type = NARG; 560 n->narg.text = wordtext; 561 n->narg.backquote = backquotelist; 562 *app = n; 563 app = &n->narg.next; 564 } else if (lasttoken == TREDIR) { 565 *rpp = n = redirnode; 566 rpp = &n->nfile.next; 567 parsefname(); /* read name of redirection file */ 568 } else if (lasttoken == TLP && app == &args->narg.next 569 && rpp == orig_rpp) { 570 /* We have a function */ 571 if (readtoken() != TRP) 572 synexpect(TRP); 573 #ifdef notdef 574 if (! goodname(n->narg.text)) 575 synerror("Bad function name"); 576 #endif 577 n->type = NDEFUN; 578 n->narg.next = command(); 579 goto checkneg; 580 } else { 581 tokpushback++; 582 break; 583 } 584 } 585 *app = NULL; 586 *rpp = NULL; 587 n = (union node *)stalloc(sizeof (struct ncmd)); 588 n->type = NCMD; 589 n->ncmd.backgnd = 0; 590 n->ncmd.args = args; 591 n->ncmd.redirect = redir; 592 593 checkneg: 594 if (negate) { 595 n2 = (union node *)stalloc(sizeof (struct nnot)); 596 n2->type = NNOT; 597 n2->nnot.com = n; 598 return n2; 599 } 600 else 601 return n; 602 } 603 604 STATIC union node * 605 makename(void) 606 { 607 union node *n; 608 609 n = (union node *)stalloc(sizeof (struct narg)); 610 n->type = NARG; 611 n->narg.next = NULL; 612 n->narg.text = wordtext; 613 n->narg.backquote = backquotelist; 614 return n; 615 } 616 617 void 618 fixredir(union node *n, const char *text, int err) 619 { 620 TRACE(("Fix redir %s %d\n", text, err)); 621 if (!err) 622 n->ndup.vname = NULL; 623 624 if (is_digit(text[0]) && text[1] == '\0') 625 n->ndup.dupfd = digit_val(text[0]); 626 else if (text[0] == '-' && text[1] == '\0') 627 n->ndup.dupfd = -1; 628 else { 629 630 if (err) 631 synerror("Bad fd number"); 632 else 633 n->ndup.vname = makename(); 634 } 635 } 636 637 638 STATIC void 639 parsefname(void) 640 { 641 union node *n = redirnode; 642 643 if (readtoken() != TWORD) 644 synexpect(-1); 645 if (n->type == NHERE) { 646 struct heredoc *here = heredoc; 647 struct heredoc *p; 648 int i; 649 650 if (quoteflag == 0) 651 n->type = NXHERE; 652 TRACE(("Here document %d\n", n->type)); 653 if (here->striptabs) { 654 while (*wordtext == '\t') 655 wordtext++; 656 } 657 if (! noexpand(wordtext) || (i = strlen(wordtext)) == 0 || i > EOFMARKLEN) 658 synerror("Illegal eof marker for << redirection"); 659 rmescapes(wordtext); 660 here->eofmark = wordtext; 661 here->next = NULL; 662 if (heredoclist == NULL) 663 heredoclist = here; 664 else { 665 for (p = heredoclist ; p->next ; p = p->next); 666 p->next = here; 667 } 668 } else if (n->type == NTOFD || n->type == NFROMFD) { 669 fixredir(n, wordtext, 0); 670 } else { 671 n->nfile.fname = makename(); 672 } 673 } 674 675 676 /* 677 * Input any here documents. 678 */ 679 680 STATIC void 681 parseheredoc(void) 682 { 683 struct heredoc *here; 684 union node *n; 685 686 while (heredoclist) { 687 here = heredoclist; 688 heredoclist = here->next; 689 if (needprompt) { 690 setprompt(2); 691 needprompt = 0; 692 } 693 readtoken1(pgetc(), here->here->type == NHERE? SQSYNTAX : DQSYNTAX, 694 here->eofmark, here->striptabs); 695 n = (union node *)stalloc(sizeof (struct narg)); 696 n->narg.type = NARG; 697 n->narg.next = NULL; 698 n->narg.text = wordtext; 699 n->narg.backquote = backquotelist; 700 here->here->nhere.doc = n; 701 } 702 } 703 704 STATIC int 705 peektoken(void) 706 { 707 int t; 708 709 t = readtoken(); 710 tokpushback++; 711 return (t); 712 } 713 714 STATIC int 715 readtoken(void) 716 { 717 int t; 718 int savecheckkwd = checkkwd; 719 struct alias *ap; 720 #ifdef DEBUG 721 int alreadyseen = tokpushback; 722 #endif 723 724 top: 725 t = xxreadtoken(); 726 727 if (checkkwd) { 728 /* 729 * eat newlines 730 */ 731 if (checkkwd == 2) { 732 checkkwd = 0; 733 while (t == TNL) { 734 parseheredoc(); 735 t = xxreadtoken(); 736 } 737 } else 738 checkkwd = 0; 739 /* 740 * check for keywords and aliases 741 */ 742 if (t == TWORD && !quoteflag) 743 { 744 const char * const *pp; 745 746 for (pp = parsekwd; *pp; pp++) { 747 if (**pp == *wordtext && equal(*pp, wordtext)) 748 { 749 lasttoken = t = pp - parsekwd + KWDOFFSET; 750 TRACE(("keyword %s recognized\n", tokname[t])); 751 goto out; 752 } 753 } 754 if (noaliases == 0 && 755 (ap = lookupalias(wordtext, 1)) != NULL) { 756 pushstring(ap->val, strlen(ap->val), ap); 757 checkkwd = savecheckkwd; 758 goto top; 759 } 760 } 761 out: 762 checkkwd = (t == TNOT) ? savecheckkwd : 0; 763 } 764 #ifdef DEBUG 765 if (!alreadyseen) 766 TRACE(("token %s %s\n", tokname[t], t == TWORD ? wordtext : "")); 767 else 768 TRACE(("reread token %s %s\n", tokname[t], t == TWORD ? wordtext : "")); 769 #endif 770 return (t); 771 } 772 773 774 /* 775 * Read the next input token. 776 * If the token is a word, we set backquotelist to the list of cmds in 777 * backquotes. We set quoteflag to true if any part of the word was 778 * quoted. 779 * If the token is TREDIR, then we set redirnode to a structure containing 780 * the redirection. 781 * In all cases, the variable startlinno is set to the number of the line 782 * on which the token starts. 783 * 784 * [Change comment: here documents and internal procedures] 785 * [Readtoken shouldn't have any arguments. Perhaps we should make the 786 * word parsing code into a separate routine. In this case, readtoken 787 * doesn't need to have any internal procedures, but parseword does. 788 * We could also make parseoperator in essence the main routine, and 789 * have parseword (readtoken1?) handle both words and redirection.] 790 */ 791 792 #define RETURN(token) return lasttoken = token 793 794 STATIC int 795 xxreadtoken(void) 796 { 797 int c; 798 799 if (tokpushback) { 800 tokpushback = 0; 801 return lasttoken; 802 } 803 if (needprompt) { 804 setprompt(2); 805 needprompt = 0; 806 } 807 startlinno = plinno; 808 for (;;) { /* until token or start of word found */ 809 c = pgetc_macro(); 810 if (c == ' ' || c == '\t') 811 continue; /* quick check for white space first */ 812 switch (c) { 813 case ' ': case '\t': 814 continue; 815 case '#': 816 while ((c = pgetc()) != '\n' && c != PEOF); 817 pungetc(); 818 continue; 819 case '\\': 820 if (pgetc() == '\n') { 821 startlinno = ++plinno; 822 if (doprompt) 823 setprompt(2); 824 else 825 setprompt(0); 826 continue; 827 } 828 pungetc(); 829 goto breakloop; 830 case '\n': 831 plinno++; 832 needprompt = doprompt; 833 RETURN(TNL); 834 case PEOF: 835 RETURN(TEOF); 836 case '&': 837 if (pgetc() == '&') 838 RETURN(TAND); 839 pungetc(); 840 RETURN(TBACKGND); 841 case '|': 842 if (pgetc() == '|') 843 RETURN(TOR); 844 pungetc(); 845 RETURN(TPIPE); 846 case ';': 847 if (pgetc() == ';') 848 RETURN(TENDCASE); 849 pungetc(); 850 RETURN(TSEMI); 851 case '(': 852 RETURN(TLP); 853 case ')': 854 RETURN(TRP); 855 default: 856 goto breakloop; 857 } 858 } 859 breakloop: 860 return readtoken1(c, BASESYNTAX, (char *)NULL, 0); 861 #undef RETURN 862 } 863 864 865 866 /* 867 * If eofmark is NULL, read a word or a redirection symbol. If eofmark 868 * is not NULL, read a here document. In the latter case, eofmark is the 869 * word which marks the end of the document and striptabs is true if 870 * leading tabs should be stripped from the document. The argument firstc 871 * is the first character of the input token or document. 872 * 873 * Because C does not have internal subroutines, I have simulated them 874 * using goto's to implement the subroutine linkage. The following macros 875 * will run code that appears at the end of readtoken1. 876 */ 877 878 #define CHECKEND() {goto checkend; checkend_return:;} 879 #define PARSEREDIR() {goto parseredir; parseredir_return:;} 880 #define PARSESUB() {goto parsesub; parsesub_return:;} 881 #define PARSEBACKQOLD() {oldstyle = 1; goto parsebackq; parsebackq_oldreturn:;} 882 #define PARSEBACKQNEW() {oldstyle = 0; goto parsebackq; parsebackq_newreturn:;} 883 #define PARSEARITH() {goto parsearith; parsearith_return:;} 884 885 STATIC int 886 readtoken1(int firstc, char const *syntax, char *eofmark, int striptabs) 887 { 888 int c = firstc; 889 char *out; 890 int len; 891 char line[EOFMARKLEN + 1]; 892 struct nodelist *bqlist; 893 int quotef; 894 int dblquote; 895 int varnest; /* levels of variables expansion */ 896 int arinest; /* levels of arithmetic expansion */ 897 int parenlevel; /* levels of parens in arithmetic */ 898 int oldstyle; 899 char const *prevsyntax; /* syntax before arithmetic */ 900 int synentry; 901 #if __GNUC__ 902 /* Avoid longjmp clobbering */ 903 (void) &out; 904 (void) "ef; 905 (void) &dblquote; 906 (void) &varnest; 907 (void) &arinest; 908 (void) &parenlevel; 909 (void) &oldstyle; 910 (void) &prevsyntax; 911 (void) &syntax; 912 (void) &synentry; 913 #endif 914 915 startlinno = plinno; 916 dblquote = 0; 917 if (syntax == DQSYNTAX) 918 dblquote = 1; 919 quotef = 0; 920 bqlist = NULL; 921 varnest = 0; 922 arinest = 0; 923 parenlevel = 0; 924 925 STARTSTACKSTR(out); 926 loop: { /* for each line, until end of word */ 927 CHECKEND(); /* set c to PEOF if at end of here document */ 928 for (;;) { /* until end of line or end of word */ 929 CHECKSTRSPACE(3, out); /* permit 3 calls to USTPUTC */ 930 931 synentry = syntax[c]; 932 933 switch(synentry) { 934 case CNL: /* '\n' */ 935 if (syntax == BASESYNTAX) 936 goto endword; /* exit outer loop */ 937 USTPUTC(c, out); 938 plinno++; 939 if (doprompt) 940 setprompt(2); 941 else 942 setprompt(0); 943 c = pgetc(); 944 goto loop; /* continue outer loop */ 945 case CWORD: 946 USTPUTC(c, out); 947 break; 948 case CCTL: 949 if (eofmark == NULL || dblquote) 950 USTPUTC(CTLESC, out); 951 USTPUTC(c, out); 952 break; 953 case CBACK: /* backslash */ 954 c = pgetc(); 955 if (c == PEOF) { 956 USTPUTC('\\', out); 957 pungetc(); 958 } else if (c == '\n') { 959 if (doprompt) 960 setprompt(2); 961 else 962 setprompt(0); 963 } else { 964 if (dblquote && c != '\\' && 965 c != '`' && c != '$' && 966 (c != '"' || eofmark != NULL)) 967 USTPUTC('\\', out); 968 if (SQSYNTAX[c] == CCTL) 969 USTPUTC(CTLESC, out); 970 else if (eofmark == NULL) 971 USTPUTC(CTLQUOTEMARK, out); 972 USTPUTC(c, out); 973 quotef++; 974 } 975 break; 976 case CSQUOTE: 977 if (eofmark == NULL) 978 USTPUTC(CTLQUOTEMARK, out); 979 syntax = SQSYNTAX; 980 break; 981 case CDQUOTE: 982 if (eofmark == NULL) 983 USTPUTC(CTLQUOTEMARK, out); 984 syntax = DQSYNTAX; 985 dblquote = 1; 986 break; 987 case CENDQUOTE: 988 if (eofmark != NULL && arinest == 0 && 989 varnest == 0) { 990 USTPUTC(c, out); 991 } else { 992 if (arinest) { 993 syntax = ARISYNTAX; 994 dblquote = 0; 995 } else if (eofmark == NULL) { 996 syntax = BASESYNTAX; 997 dblquote = 0; 998 } 999 quotef++; 1000 } 1001 break; 1002 case CVAR: /* '$' */ 1003 PARSESUB(); /* parse substitution */ 1004 break; 1005 case CENDVAR: /* '}' */ 1006 if (varnest > 0) { 1007 varnest--; 1008 USTPUTC(CTLENDVAR, out); 1009 } else { 1010 USTPUTC(c, out); 1011 } 1012 break; 1013 case CLP: /* '(' in arithmetic */ 1014 parenlevel++; 1015 USTPUTC(c, out); 1016 break; 1017 case CRP: /* ')' in arithmetic */ 1018 if (parenlevel > 0) { 1019 USTPUTC(c, out); 1020 --parenlevel; 1021 } else { 1022 if (pgetc() == ')') { 1023 if (--arinest == 0) { 1024 USTPUTC(CTLENDARI, out); 1025 syntax = prevsyntax; 1026 if (syntax == DQSYNTAX) 1027 dblquote = 1; 1028 else 1029 dblquote = 0; 1030 } else 1031 USTPUTC(')', out); 1032 } else { 1033 /* 1034 * unbalanced parens 1035 * (don't 2nd guess - no error) 1036 */ 1037 pungetc(); 1038 USTPUTC(')', out); 1039 } 1040 } 1041 break; 1042 case CBQUOTE: /* '`' */ 1043 PARSEBACKQOLD(); 1044 break; 1045 case CEOF: 1046 goto endword; /* exit outer loop */ 1047 default: 1048 if (varnest == 0) 1049 goto endword; /* exit outer loop */ 1050 USTPUTC(c, out); 1051 } 1052 c = pgetc_macro(); 1053 } 1054 } 1055 endword: 1056 if (syntax == ARISYNTAX) 1057 synerror("Missing '))'"); 1058 if (syntax != BASESYNTAX && ! parsebackquote && eofmark == NULL) 1059 synerror("Unterminated quoted string"); 1060 if (varnest != 0) { 1061 startlinno = plinno; 1062 synerror("Missing '}'"); 1063 } 1064 USTPUTC('\0', out); 1065 len = out - stackblock(); 1066 out = stackblock(); 1067 if (eofmark == NULL) { 1068 if ((c == '>' || c == '<') 1069 && quotef == 0 1070 && len <= 2 1071 && (*out == '\0' || is_digit(*out))) { 1072 PARSEREDIR(); 1073 return lasttoken = TREDIR; 1074 } else { 1075 pungetc(); 1076 } 1077 } 1078 quoteflag = quotef; 1079 backquotelist = bqlist; 1080 grabstackblock(len); 1081 wordtext = out; 1082 return lasttoken = TWORD; 1083 /* end of readtoken routine */ 1084 1085 1086 1087 /* 1088 * Check to see whether we are at the end of the here document. When this 1089 * is called, c is set to the first character of the next input line. If 1090 * we are at the end of the here document, this routine sets the c to PEOF. 1091 */ 1092 1093 checkend: { 1094 if (eofmark) { 1095 if (striptabs) { 1096 while (c == '\t') 1097 c = pgetc(); 1098 } 1099 if (c == *eofmark) { 1100 if (pfgets(line, sizeof line) != NULL) { 1101 char *p, *q; 1102 1103 p = line; 1104 for (q = eofmark + 1 ; *q && *p == *q ; p++, q++); 1105 if (*p == '\n' && *q == '\0') { 1106 c = PEOF; 1107 plinno++; 1108 needprompt = doprompt; 1109 } else { 1110 pushstring(line, strlen(line), NULL); 1111 } 1112 } 1113 } 1114 } 1115 goto checkend_return; 1116 } 1117 1118 1119 /* 1120 * Parse a redirection operator. The variable "out" points to a string 1121 * specifying the fd to be redirected. The variable "c" contains the 1122 * first character of the redirection operator. 1123 */ 1124 1125 parseredir: { 1126 char fd = *out; 1127 union node *np; 1128 1129 np = (union node *)stalloc(sizeof (struct nfile)); 1130 if (c == '>') { 1131 np->nfile.fd = 1; 1132 c = pgetc(); 1133 if (c == '>') 1134 np->type = NAPPEND; 1135 else if (c == '&') 1136 np->type = NTOFD; 1137 else if (c == '|') 1138 np->type = NCLOBBER; 1139 else { 1140 np->type = NTO; 1141 pungetc(); 1142 } 1143 } else { /* c == '<' */ 1144 np->nfile.fd = 0; 1145 c = pgetc(); 1146 if (c == '<') { 1147 if (sizeof (struct nfile) != sizeof (struct nhere)) { 1148 np = (union node *)stalloc(sizeof (struct nhere)); 1149 np->nfile.fd = 0; 1150 } 1151 np->type = NHERE; 1152 heredoc = (struct heredoc *)stalloc(sizeof (struct heredoc)); 1153 heredoc->here = np; 1154 if ((c = pgetc()) == '-') { 1155 heredoc->striptabs = 1; 1156 } else { 1157 heredoc->striptabs = 0; 1158 pungetc(); 1159 } 1160 } else if (c == '&') 1161 np->type = NFROMFD; 1162 else if (c == '>') 1163 np->type = NFROMTO; 1164 else { 1165 np->type = NFROM; 1166 pungetc(); 1167 } 1168 } 1169 if (fd != '\0') 1170 np->nfile.fd = digit_val(fd); 1171 redirnode = np; 1172 goto parseredir_return; 1173 } 1174 1175 1176 /* 1177 * Parse a substitution. At this point, we have read the dollar sign 1178 * and nothing else. 1179 */ 1180 1181 parsesub: { 1182 int subtype; 1183 int typeloc; 1184 int flags; 1185 char *p; 1186 #ifndef GDB_HACK 1187 static const char types[] = "}-+?="; 1188 #endif 1189 int bracketed_name = 0; /* used to handle ${[0-9]*} variables */ 1190 1191 c = pgetc(); 1192 if (c != '(' && c != '{' && !is_name(c) && !is_special(c)) { 1193 USTPUTC('$', out); 1194 pungetc(); 1195 } else if (c == '(') { /* $(command) or $((arith)) */ 1196 if (pgetc() == '(') { 1197 PARSEARITH(); 1198 } else { 1199 pungetc(); 1200 PARSEBACKQNEW(); 1201 } 1202 } else { 1203 USTPUTC(CTLVAR, out); 1204 typeloc = out - stackblock(); 1205 USTPUTC(VSNORMAL, out); 1206 subtype = VSNORMAL; 1207 if (c == '{') { 1208 bracketed_name = 1; 1209 c = pgetc(); 1210 if (c == '#') { 1211 if ((c = pgetc()) == '}') 1212 c = '#'; 1213 else 1214 subtype = VSLENGTH; 1215 } 1216 else 1217 subtype = 0; 1218 } 1219 if (is_name(c)) { 1220 do { 1221 STPUTC(c, out); 1222 c = pgetc(); 1223 } while (is_in_name(c)); 1224 } else if (is_digit(c)) { 1225 if (bracketed_name) { 1226 do { 1227 STPUTC(c, out); 1228 c = pgetc(); 1229 } while (is_digit(c)); 1230 } else { 1231 STPUTC(c, out); 1232 c = pgetc(); 1233 } 1234 } else { 1235 if (! is_special(c)) 1236 badsub: synerror("Bad substitution"); 1237 USTPUTC(c, out); 1238 c = pgetc(); 1239 } 1240 STPUTC('=', out); 1241 flags = 0; 1242 if (subtype == 0) { 1243 switch (c) { 1244 case ':': 1245 flags = VSNUL; 1246 c = pgetc(); 1247 /*FALLTHROUGH*/ 1248 default: 1249 p = strchr(types, c); 1250 if (p == NULL) 1251 goto badsub; 1252 subtype = p - types + VSNORMAL; 1253 break; 1254 case '%': 1255 case '#': 1256 { 1257 int cc = c; 1258 subtype = c == '#' ? VSTRIMLEFT : 1259 VSTRIMRIGHT; 1260 c = pgetc(); 1261 if (c == cc) 1262 subtype++; 1263 else 1264 pungetc(); 1265 break; 1266 } 1267 } 1268 } else { 1269 pungetc(); 1270 } 1271 if (subtype != VSLENGTH && (dblquote || arinest)) 1272 flags |= VSQUOTE; 1273 *(stackblock() + typeloc) = subtype | flags; 1274 if (subtype != VSNORMAL) 1275 varnest++; 1276 } 1277 goto parsesub_return; 1278 } 1279 1280 1281 /* 1282 * Called to parse command substitutions. Newstyle is set if the command 1283 * is enclosed inside $(...); nlpp is a pointer to the head of the linked 1284 * list of commands (passed by reference), and savelen is the number of 1285 * characters on the top of the stack which must be preserved. 1286 */ 1287 1288 parsebackq: { 1289 struct nodelist **nlpp; 1290 int savepbq; 1291 union node *n; 1292 char *volatile str; 1293 struct jmploc jmploc; 1294 struct jmploc *volatile savehandler; 1295 int savelen; 1296 int saveprompt; 1297 #if __GNUC__ 1298 /* Avoid longjmp clobbering */ 1299 (void) &saveprompt; 1300 #endif 1301 1302 savepbq = parsebackquote; 1303 if (setjmp(jmploc.loc)) { 1304 if (str) 1305 ckfree(str); 1306 parsebackquote = 0; 1307 handler = savehandler; 1308 longjmp(handler->loc, 1); 1309 } 1310 INTOFF; 1311 str = NULL; 1312 savelen = out - stackblock(); 1313 if (savelen > 0) { 1314 str = ckmalloc(savelen); 1315 memcpy(str, stackblock(), savelen); 1316 } 1317 savehandler = handler; 1318 handler = &jmploc; 1319 INTON; 1320 if (oldstyle) { 1321 /* We must read until the closing backquote, giving special 1322 treatment to some slashes, and then push the string and 1323 reread it as input, interpreting it normally. */ 1324 char *out; 1325 int c; 1326 int savelen; 1327 char *str; 1328 1329 1330 STARTSTACKSTR(out); 1331 for (;;) { 1332 if (needprompt) { 1333 setprompt(2); 1334 needprompt = 0; 1335 } 1336 switch (c = pgetc()) { 1337 case '`': 1338 goto done; 1339 1340 case '\\': 1341 if ((c = pgetc()) == '\n') { 1342 plinno++; 1343 if (doprompt) 1344 setprompt(2); 1345 else 1346 setprompt(0); 1347 /* 1348 * If eating a newline, avoid putting 1349 * the newline into the new character 1350 * stream (via the STPUTC after the 1351 * switch). 1352 */ 1353 continue; 1354 } 1355 if (c != '\\' && c != '`' && c != '$' 1356 && (!dblquote || c != '"')) 1357 STPUTC('\\', out); 1358 break; 1359 1360 case '\n': 1361 plinno++; 1362 needprompt = doprompt; 1363 break; 1364 1365 case PEOF: 1366 startlinno = plinno; 1367 synerror("EOF in backquote substitution"); 1368 break; 1369 1370 default: 1371 break; 1372 } 1373 STPUTC(c, out); 1374 } 1375 done: 1376 STPUTC('\0', out); 1377 savelen = out - stackblock(); 1378 if (savelen > 0) { 1379 str = ckmalloc(savelen); 1380 memcpy(str, stackblock(), savelen); 1381 setinputstring(str, 1); 1382 } 1383 } 1384 nlpp = &bqlist; 1385 while (*nlpp) 1386 nlpp = &(*nlpp)->next; 1387 *nlpp = (struct nodelist *)stalloc(sizeof (struct nodelist)); 1388 (*nlpp)->next = NULL; 1389 parsebackquote = oldstyle; 1390 1391 if (oldstyle) { 1392 saveprompt = doprompt; 1393 doprompt = 0; 1394 } 1395 1396 n = list(0); 1397 1398 if (oldstyle) 1399 doprompt = saveprompt; 1400 else { 1401 if (readtoken() != TRP) 1402 synexpect(TRP); 1403 } 1404 1405 (*nlpp)->n = n; 1406 if (oldstyle) { 1407 /* 1408 * Start reading from old file again, ignoring any pushed back 1409 * tokens left from the backquote parsing 1410 */ 1411 popfile(); 1412 tokpushback = 0; 1413 } 1414 while (stackblocksize() <= savelen) 1415 growstackblock(); 1416 STARTSTACKSTR(out); 1417 if (str) { 1418 memcpy(out, str, savelen); 1419 STADJUST(savelen, out); 1420 INTOFF; 1421 ckfree(str); 1422 str = NULL; 1423 INTON; 1424 } 1425 parsebackquote = savepbq; 1426 handler = savehandler; 1427 if (arinest || dblquote) 1428 USTPUTC(CTLBACKQ | CTLQUOTE, out); 1429 else 1430 USTPUTC(CTLBACKQ, out); 1431 if (oldstyle) 1432 goto parsebackq_oldreturn; 1433 else 1434 goto parsebackq_newreturn; 1435 } 1436 1437 /* 1438 * Parse an arithmetic expansion (indicate start of one and set state) 1439 */ 1440 parsearith: { 1441 1442 if (++arinest == 1) { 1443 prevsyntax = syntax; 1444 syntax = ARISYNTAX; 1445 USTPUTC(CTLARI, out); 1446 if (dblquote) 1447 USTPUTC('"',out); 1448 else 1449 USTPUTC(' ',out); 1450 } else { 1451 /* 1452 * we collapse embedded arithmetic expansion to 1453 * parenthesis, which should be equivalent 1454 */ 1455 USTPUTC('(', out); 1456 } 1457 goto parsearith_return; 1458 } 1459 1460 } /* end of readtoken */ 1461 1462 1463 1464 #ifdef mkinit 1465 RESET { 1466 tokpushback = 0; 1467 checkkwd = 0; 1468 } 1469 #endif 1470 1471 /* 1472 * Returns true if the text contains nothing to expand (no dollar signs 1473 * or backquotes). 1474 */ 1475 1476 STATIC int 1477 noexpand(char *text) 1478 { 1479 char *p; 1480 char c; 1481 1482 p = text; 1483 while ((c = *p++) != '\0') { 1484 if ( c == CTLQUOTEMARK) 1485 continue; 1486 if (c == CTLESC) 1487 p++; 1488 else if (BASESYNTAX[(int)c] == CCTL) 1489 return 0; 1490 } 1491 return 1; 1492 } 1493 1494 1495 /* 1496 * Return true if the argument is a legal variable name (a letter or 1497 * underscore followed by zero or more letters, underscores, and digits). 1498 */ 1499 1500 int 1501 goodname(char *name) 1502 { 1503 char *p; 1504 1505 p = name; 1506 if (! is_name(*p)) 1507 return 0; 1508 while (*++p) { 1509 if (! is_in_name(*p)) 1510 return 0; 1511 } 1512 return 1; 1513 } 1514 1515 1516 /* 1517 * Called when an unexpected token is read during the parse. The argument 1518 * is the token that is expected, or -1 if more than one type of token can 1519 * occur at this point. 1520 */ 1521 1522 STATIC void 1523 synexpect(int token) 1524 { 1525 char msg[64]; 1526 1527 if (token >= 0) { 1528 fmtstr(msg, 64, "%s unexpected (expecting %s)", 1529 tokname[lasttoken], tokname[token]); 1530 } else { 1531 fmtstr(msg, 64, "%s unexpected", tokname[lasttoken]); 1532 } 1533 synerror(msg); 1534 } 1535 1536 1537 STATIC void 1538 synerror(const char *msg) 1539 { 1540 if (commandname) 1541 outfmt(&errout, "%s: %d: ", commandname, startlinno); 1542 outfmt(&errout, "Syntax error: %s\n", msg); 1543 error((char *)NULL); 1544 } 1545 1546 STATIC void 1547 setprompt(int which) 1548 { 1549 whichprompt = which; 1550 1551 #ifndef NO_HISTORY 1552 if (!el) 1553 #endif 1554 out2str(getprompt(NULL)); 1555 } 1556 1557 /* 1558 * called by editline -- any expansions to the prompt 1559 * should be added here. 1560 */ 1561 const char * 1562 getprompt(void *unused __unused) 1563 { 1564 switch (whichprompt) { 1565 case 0: 1566 return ""; 1567 case 1: 1568 return ps1val(); 1569 case 2: 1570 return ps2val(); 1571 default: 1572 return "<internal prompt error>"; 1573 } 1574 } 1575