| 1 | /* $Header: /p/tcsh/cvsroot/tcsh/sh.lex.c,v 3.81 2009/06/25 21:15:37 christos Exp $ */ |
| 2 | /* |
| 3 | * sh.lex.c: Lexical analysis into tokens |
| 4 | */ |
| 5 | /*- |
| 6 | * Copyright (c) 1980, 1991 The Regents of the University of California. |
| 7 | * All rights reserved. |
| 8 | * |
| 9 | * Redistribution and use in source and binary forms, with or without |
| 10 | * modification, are permitted provided that the following conditions |
| 11 | * are met: |
| 12 | * 1. Redistributions of source code must retain the above copyright |
| 13 | * notice, this list of conditions and the following disclaimer. |
| 14 | * 2. Redistributions in binary form must reproduce the above copyright |
| 15 | * notice, this list of conditions and the following disclaimer in the |
| 16 | * documentation and/or other materials provided with the distribution. |
| 17 | * 3. Neither the name of the University nor the names of its contributors |
| 18 | * may be used to endorse or promote products derived from this software |
| 19 | * without specific prior written permission. |
| 20 | * |
| 21 | * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND |
| 22 | * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
| 23 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE |
| 24 | * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE |
| 25 | * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
| 26 | * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS |
| 27 | * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) |
| 28 | * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT |
| 29 | * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY |
| 30 | * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF |
| 31 | * SUCH DAMAGE. |
| 32 | */ |
| 33 | #include "sh.h" |
| 34 | |
| 35 | RCSID("$tcsh: sh.lex.c,v 3.81 2009/06/25 21:15:37 christos Exp $") |
| 36 | |
| 37 | #include "ed.h" |
| 38 | |
| 39 | #include <assert.h> |
| 40 | /* #define DEBUG_INP */ |
| 41 | /* #define DEBUG_SEEK */ |
| 42 | |
| 43 | /* |
| 44 | * C shell |
| 45 | */ |
| 46 | |
| 47 | #define FLAG_G 1 |
| 48 | #define FLAG_A 2 |
| 49 | /* |
| 50 | * These lexical routines read input and form lists of words. |
| 51 | * There is some involved processing here, because of the complications |
| 52 | * of input buffering, and especially because of history substitution. |
| 53 | */ |
| 54 | static Char *word (int); |
| 55 | static eChar getC1 (int); |
| 56 | static void getdol (void); |
| 57 | static void getexcl (Char); |
| 58 | static struct Hist *findev (Char *, int); |
| 59 | static void setexclp (Char *); |
| 60 | static eChar bgetc (void); |
| 61 | static void balloc (int); |
| 62 | static void bfree (void); |
| 63 | static struct wordent *gethent (Char); |
| 64 | static int matchs (const Char *, const Char *); |
| 65 | static int getsel (int *, int *, int); |
| 66 | static struct wordent *getsub (struct wordent *); |
| 67 | static Char *subword (Char *, Char, int *, size_t *); |
| 68 | static struct wordent *dosub (Char, struct wordent *, int); |
| 69 | static ssize_t wide_read (int, Char *, size_t, int); |
| 70 | |
| 71 | /* |
| 72 | * Peekc is a peek character for getC, peekread for readc. |
| 73 | * There is a subtlety here in many places... history routines |
| 74 | * will read ahead and then insert stuff into the input stream. |
| 75 | * If they push back a character then they must push it behind |
| 76 | * the text substituted by the history substitution. On the other |
| 77 | * hand in several places we need 2 peek characters. To make this |
| 78 | * all work, the history routines read with getC, and make use both |
| 79 | * of ungetC and unreadc. The key observation is that the state |
| 80 | * of getC at the call of a history reference is such that calls |
| 81 | * to getC from the history routines will always yield calls of |
| 82 | * readc, unless this peeking is involved. That is to say that during |
| 83 | * getexcl the variables lap, exclp, and exclnxt are all zero. |
| 84 | * |
| 85 | * Getdol invokes history substitution, hence the extra peek, peekd, |
| 86 | * which it can ungetD to be before history substitutions. |
| 87 | */ |
| 88 | static Char peekc = 0, peekd = 0; |
| 89 | static Char peekread = 0; |
| 90 | |
| 91 | /* (Tail of) current word from ! subst */ |
| 92 | static Char *exclp = NULL; |
| 93 | |
| 94 | /* The rest of the ! subst words */ |
| 95 | static struct wordent *exclnxt = NULL; |
| 96 | |
| 97 | /* Count of remaining words in ! subst */ |
| 98 | static int exclc = 0; |
| 99 | |
| 100 | /* "Globp" for alias resubstitution */ |
| 101 | int aret = TCSH_F_SEEK; |
| 102 | |
| 103 | /* |
| 104 | * Labuf implements a general buffer for lookahead during lexical operations. |
| 105 | * Text which is to be placed in the input stream can be stuck here. |
| 106 | * We stick parsed ahead $ constructs during initial input, |
| 107 | * process id's from `$$', and modified variable values (from qualifiers |
| 108 | * during expansion in sh.dol.c) here. |
| 109 | */ |
| 110 | struct Strbuf labuf; /* = Strbuf_INIT; */ |
| 111 | |
| 112 | /* |
| 113 | * Lex returns to its caller not only a wordlist (as a "var" parameter) |
| 114 | * but also whether a history substitution occurred. This is used in |
| 115 | * the main (process) routine to determine whether to echo, and also |
| 116 | * when called by the alias routine to determine whether to keep the |
| 117 | * argument list. |
| 118 | */ |
| 119 | static int hadhist = 0; |
| 120 | |
| 121 | /* |
| 122 | * Avoid alias expansion recursion via \!# |
| 123 | */ |
| 124 | int hleft; |
| 125 | |
| 126 | struct Strbuf histline; /* = Strbuf_INIT; last line input */ |
| 127 | |
| 128 | int histvalid = 0; /* is histline valid */ |
| 129 | |
| 130 | static Char getCtmp; |
| 131 | |
| 132 | #define getC(f) (((getCtmp = peekc) != '\0') ? (peekc = 0, (eChar)getCtmp) : getC1(f)) |
| 133 | #define ungetC(c) peekc = (Char) c |
| 134 | #define ungetD(c) peekd = (Char) c |
| 135 | |
| 136 | /* Use Htime to store timestamps picked up from history file for enthist() |
| 137 | * if reading saved history (sg) |
| 138 | */ |
| 139 | time_t Htime = (time_t)0; |
| 140 | static time_t a2time_t (Char *); |
| 141 | |
| 142 | /* |
| 143 | * special parsing rules apply for source -h |
| 144 | */ |
| 145 | extern int enterhist; |
| 146 | |
| 147 | int |
| 148 | lex(struct wordent *hp) |
| 149 | { |
| 150 | struct wordent *wdp; |
| 151 | eChar c; |
| 152 | int parsehtime = enterhist; |
| 153 | |
| 154 | histvalid = 0; |
| 155 | histline.len = 0; |
| 156 | |
| 157 | btell(&lineloc); |
| 158 | hp->next = hp->prev = hp; |
| 159 | hp->word = STRNULL; |
| 160 | hadhist = 0; |
| 161 | do |
| 162 | c = readc(0); |
| 163 | while (c == ' ' || c == '\t'); |
| 164 | if (c == (eChar)HISTSUB && intty) |
| 165 | /* ^lef^rit from tty is short !:s^lef^rit */ |
| 166 | getexcl(c); |
| 167 | else |
| 168 | unreadc(c); |
| 169 | cleanup_push(hp, lex_cleanup); |
| 170 | wdp = hp; |
| 171 | /* |
| 172 | * The following loop is written so that the links needed by freelex will |
| 173 | * be ready and rarin to go even if it is interrupted. |
| 174 | */ |
| 175 | do { |
| 176 | struct wordent *new; |
| 177 | |
| 178 | new = xmalloc(sizeof(*new)); |
| 179 | new->word = NULL; |
| 180 | new->prev = wdp; |
| 181 | new->next = hp; |
| 182 | wdp->next = new; |
| 183 | hp->prev = new; |
| 184 | wdp = new; |
| 185 | wdp->word = word(parsehtime); |
| 186 | parsehtime = 0; |
| 187 | } while (wdp->word[0] != '\n'); |
| 188 | cleanup_ignore(hp); |
| 189 | cleanup_until(hp); |
| 190 | Strbuf_terminate(&histline); |
| 191 | if (histline.len != 0 && histline.s[histline.len - 1] == '\n') |
| 192 | histline.s[histline.len - 1] = '\0'; |
| 193 | histvalid = 1; |
| 194 | |
| 195 | return (hadhist); |
| 196 | } |
| 197 | |
| 198 | static time_t |
| 199 | a2time_t(Char *wordx) |
| 200 | { |
| 201 | /* Attempt to distinguish timestamps from other possible entries. |
| 202 | * Format: "+NNNNNNNNNN" (10 digits, left padded with ascii '0') */ |
| 203 | |
| 204 | time_t ret; |
| 205 | Char *s; |
| 206 | int ct; |
| 207 | |
| 208 | if (!wordx || *(s = wordx) != '+') |
| 209 | return (time_t)0; |
| 210 | |
| 211 | for (++s, ret = 0, ct = 0; *s; ++s, ++ct) { |
| 212 | if (!isdigit((unsigned char)*s)) |
| 213 | return (time_t)0; |
| 214 | ret = ret * 10 + (time_t)((unsigned char)*s - '0'); |
| 215 | } |
| 216 | |
| 217 | if (ct != 10) |
| 218 | return (time_t)0; |
| 219 | |
| 220 | return ret; |
| 221 | } |
| 222 | |
| 223 | void |
| 224 | prlex(struct wordent *sp0) |
| 225 | { |
| 226 | struct wordent *sp = sp0->next; |
| 227 | |
| 228 | for (;;) { |
| 229 | xprintf("%S", sp->word); |
| 230 | sp = sp->next; |
| 231 | if (sp == sp0) |
| 232 | break; |
| 233 | if (sp->word[0] != '\n') |
| 234 | xputchar(' '); |
| 235 | } |
| 236 | } |
| 237 | |
| 238 | void |
| 239 | copylex(struct wordent *hp, struct wordent *fp) |
| 240 | { |
| 241 | struct wordent *wdp; |
| 242 | |
| 243 | wdp = hp; |
| 244 | fp = fp->next; |
| 245 | do { |
| 246 | struct wordent *new; |
| 247 | |
| 248 | new = xmalloc(sizeof(*new)); |
| 249 | new->word = NULL; |
| 250 | new->prev = wdp; |
| 251 | new->next = hp; |
| 252 | wdp->next = new; |
| 253 | hp->prev = new; |
| 254 | wdp = new; |
| 255 | wdp->word = Strsave(fp->word); |
| 256 | fp = fp->next; |
| 257 | } while (wdp->word[0] != '\n'); |
| 258 | } |
| 259 | |
| 260 | void |
| 261 | freelex(struct wordent *vp) |
| 262 | { |
| 263 | struct wordent *fp; |
| 264 | |
| 265 | while (vp->next != vp) { |
| 266 | fp = vp->next; |
| 267 | vp->next = fp->next; |
| 268 | xfree(fp->word); |
| 269 | xfree(fp); |
| 270 | } |
| 271 | vp->prev = vp; |
| 272 | } |
| 273 | |
| 274 | void |
| 275 | lex_cleanup(void *xvp) |
| 276 | { |
| 277 | struct wordent *vp; |
| 278 | |
| 279 | vp = xvp; |
| 280 | freelex(vp); |
| 281 | } |
| 282 | |
| 283 | static Char * |
| 284 | word(int parsehtime) |
| 285 | { |
| 286 | eChar c, c1; |
| 287 | struct Strbuf wbuf = Strbuf_INIT; |
| 288 | Char hbuf[12]; |
| 289 | int h; |
| 290 | int dolflg; |
| 291 | |
| 292 | cleanup_push(&wbuf, Strbuf_cleanup); |
| 293 | loop: |
| 294 | while ((c = getC(DOALL)) == ' ' || c == '\t') |
| 295 | continue; |
| 296 | if (cmap(c, _META | _ESC)) |
| 297 | switch (c) { |
| 298 | case '&': |
| 299 | case '|': |
| 300 | case '<': |
| 301 | case '>': |
| 302 | Strbuf_append1(&wbuf, c); |
| 303 | c1 = getC(DOALL); |
| 304 | if (c1 == c) |
| 305 | Strbuf_append1(&wbuf, c1); |
| 306 | else |
| 307 | ungetC(c1); |
| 308 | goto ret; |
| 309 | |
| 310 | case '#': |
| 311 | if (intty || (enterhist && !parsehtime)) |
| 312 | break; |
| 313 | c = 0; |
| 314 | h = 0; |
| 315 | do { |
| 316 | c1 = c; |
| 317 | c = getC(0); |
| 318 | if (h < 11 && parsehtime) |
| 319 | hbuf[h++] = c; |
| 320 | } while (c != '\n'); |
| 321 | if (parsehtime) { |
| 322 | hbuf[11] = '\0'; |
| 323 | Htime = a2time_t(hbuf); |
| 324 | } |
| 325 | if (c1 == '\\') |
| 326 | goto loop; |
| 327 | /*FALLTHROUGH*/ |
| 328 | |
| 329 | case ';': |
| 330 | case '(': |
| 331 | case ')': |
| 332 | case '\n': |
| 333 | Strbuf_append1(&wbuf, c); |
| 334 | goto ret; |
| 335 | |
| 336 | case '\\': |
| 337 | c = getC(0); |
| 338 | if (c == '\n') { |
| 339 | if (onelflg == 1) |
| 340 | onelflg = 2; |
| 341 | goto loop; |
| 342 | } |
| 343 | if (c != (eChar)HIST) |
| 344 | Strbuf_append1(&wbuf, '\\'); |
| 345 | c |= QUOTE; |
| 346 | default: |
| 347 | break; |
| 348 | } |
| 349 | c1 = 0; |
| 350 | dolflg = DOALL; |
| 351 | for (;;) { |
| 352 | if (c1) { |
| 353 | if (c == c1) { |
| 354 | c1 = 0; |
| 355 | dolflg = DOALL; |
| 356 | } |
| 357 | else if (c == '\\') { |
| 358 | c = getC(0); |
| 359 | /* |
| 360 | * PWP: this is dumb, but how all of the other shells work. If \ quotes |
| 361 | * a character OUTSIDE of a set of ''s, why shouldn't it quote EVERY |
| 362 | * following character INSIDE a set of ''s. |
| 363 | * |
| 364 | * Actually, all I really want to be able to say is 'foo\'bar' --> foo'bar |
| 365 | */ |
| 366 | if (c == (eChar)HIST) |
| 367 | c |= QUOTE; |
| 368 | else { |
| 369 | if (bslash_quote && |
| 370 | ((c == '\'') || (c == '"') || |
| 371 | (c == '\\') || (c == '$'))) { |
| 372 | c |= QUOTE; |
| 373 | } |
| 374 | else { |
| 375 | if (c == '\n') |
| 376 | /* |
| 377 | * if (c1 == '`') c = ' '; else |
| 378 | */ |
| 379 | c |= QUOTE; |
| 380 | ungetC(c); |
| 381 | c = '\\'; |
| 382 | } |
| 383 | } |
| 384 | } |
| 385 | else if (c == '\n') { |
| 386 | seterror(ERR_UNMATCHED, c1); |
| 387 | ungetC(c); |
| 388 | break; |
| 389 | } |
| 390 | } |
| 391 | else if (cmap(c, _META | _QF | _QB | _ESC)) { |
| 392 | if (c == '\\') { |
| 393 | c = getC(0); |
| 394 | if (c == '\n') { |
| 395 | if (onelflg == 1) |
| 396 | onelflg = 2; |
| 397 | break; |
| 398 | } |
| 399 | if (c != (eChar)HIST) |
| 400 | Strbuf_append1(&wbuf, '\\'); |
| 401 | c |= QUOTE; |
| 402 | } |
| 403 | else if (cmap(c, _QF | _QB)) { /* '"` */ |
| 404 | c1 = c; |
| 405 | dolflg = c == '"' ? DOALL : DOEXCL; |
| 406 | } |
| 407 | else if (c != '#' || (!intty && !enterhist)) { |
| 408 | ungetC(c); |
| 409 | break; |
| 410 | } |
| 411 | } |
| 412 | Strbuf_append1(&wbuf, c); |
| 413 | c = getC(dolflg); |
| 414 | } |
| 415 | ret: |
| 416 | cleanup_ignore(&wbuf); |
| 417 | cleanup_until(&wbuf); |
| 418 | return Strbuf_finish(&wbuf); |
| 419 | } |
| 420 | |
| 421 | static eChar |
| 422 | getC1(int flag) |
| 423 | { |
| 424 | eChar c; |
| 425 | |
| 426 | for (;;) { |
| 427 | if ((c = peekc) != 0) { |
| 428 | peekc = 0; |
| 429 | return (c); |
| 430 | } |
| 431 | if (lap < labuf.len) { |
| 432 | c = labuf.s[lap++]; |
| 433 | if (cmap(c, _META | _QF | _QB)) |
| 434 | c |= QUOTE; |
| 435 | return (c); |
| 436 | } |
| 437 | if ((c = peekd) != 0) { |
| 438 | peekd = 0; |
| 439 | return (c); |
| 440 | } |
| 441 | if (exclp) { |
| 442 | if ((c = *exclp++) != 0) |
| 443 | return (c); |
| 444 | if (exclnxt && --exclc >= 0) { |
| 445 | exclnxt = exclnxt->next; |
| 446 | setexclp(exclnxt->word); |
| 447 | return (' '); |
| 448 | } |
| 449 | exclp = 0; |
| 450 | exclnxt = 0; |
| 451 | /* this will throw away the dummy history entries */ |
| 452 | savehist(NULL, 0); |
| 453 | |
| 454 | } |
| 455 | if (exclnxt) { |
| 456 | exclnxt = exclnxt->next; |
| 457 | if (--exclc < 0) |
| 458 | exclnxt = 0; |
| 459 | else |
| 460 | setexclp(exclnxt->word); |
| 461 | continue; |
| 462 | } |
| 463 | c = readc(0); |
| 464 | if (c == '$' && (flag & DODOL)) { |
| 465 | getdol(); |
| 466 | continue; |
| 467 | } |
| 468 | if (c == (eChar)HIST && (flag & DOEXCL)) { |
| 469 | getexcl(0); |
| 470 | continue; |
| 471 | } |
| 472 | break; |
| 473 | } |
| 474 | return (c); |
| 475 | } |
| 476 | |
| 477 | static void |
| 478 | getdol(void) |
| 479 | { |
| 480 | struct Strbuf name = Strbuf_INIT; |
| 481 | eChar c; |
| 482 | eChar sc; |
| 483 | int special = 0; |
| 484 | |
| 485 | c = sc = getC(DOEXCL); |
| 486 | if (any("\t \n", c)) { |
| 487 | ungetD(c); |
| 488 | ungetC('$' | QUOTE); |
| 489 | return; |
| 490 | } |
| 491 | cleanup_push(&name, Strbuf_cleanup); |
| 492 | Strbuf_append1(&name, '$'); |
| 493 | if (c == '{') |
| 494 | Strbuf_append1(&name, c), c = getC(DOEXCL); |
| 495 | if (c == '#' || c == '?' || c == '%') |
| 496 | special++, Strbuf_append1(&name, c), c = getC(DOEXCL); |
| 497 | Strbuf_append1(&name, c); |
| 498 | switch (c) { |
| 499 | |
| 500 | case '<': |
| 501 | case '$': |
| 502 | case '!': |
| 503 | if (special) |
| 504 | seterror(ERR_SPDOLLT); |
| 505 | goto end; |
| 506 | |
| 507 | case '\n': |
| 508 | ungetD(c); |
| 509 | name.len--; |
| 510 | if (!special) |
| 511 | seterror(ERR_NEWLINE); |
| 512 | goto end; |
| 513 | |
| 514 | case '*': |
| 515 | if (special) |
| 516 | seterror(ERR_SPSTAR); |
| 517 | goto end; |
| 518 | |
| 519 | default: |
| 520 | if (Isdigit(c)) { |
| 521 | #ifdef notdef |
| 522 | /* let $?0 pass for now */ |
| 523 | if (special) { |
| 524 | seterror(ERR_DIGIT); |
| 525 | goto end; |
| 526 | } |
| 527 | #endif |
| 528 | while ((c = getC(DOEXCL)) != 0) { |
| 529 | if (!Isdigit(c)) |
| 530 | break; |
| 531 | Strbuf_append1(&name, c); |
| 532 | } |
| 533 | } |
| 534 | else if (letter(c)) { |
| 535 | while ((c = getC(DOEXCL)) != 0) { |
| 536 | /* Bugfix for ${v123x} from Chris Torek, DAS DEC-90. */ |
| 537 | if (!letter(c) && !Isdigit(c)) |
| 538 | break; |
| 539 | Strbuf_append1(&name, c); |
| 540 | } |
| 541 | } |
| 542 | else { |
| 543 | if (!special) |
| 544 | seterror(ERR_VARILL); |
| 545 | else { |
| 546 | ungetD(c); |
| 547 | name.len--; |
| 548 | } |
| 549 | goto end; |
| 550 | } |
| 551 | break; |
| 552 | } |
| 553 | if (c == '[') { |
| 554 | Strbuf_append1(&name, c); |
| 555 | do { |
| 556 | /* |
| 557 | * Michael Greim: Allow $ expansion to take place in selector |
| 558 | * expressions. (limits the number of characters returned) |
| 559 | */ |
| 560 | c = getC(DOEXCL | DODOL); |
| 561 | if (c == '\n') { |
| 562 | ungetD(c); |
| 563 | name.len--; |
| 564 | seterror(ERR_NLINDEX); |
| 565 | goto end; |
| 566 | } |
| 567 | Strbuf_append1(&name, c); |
| 568 | } while (c != ']'); |
| 569 | c = getC(DOEXCL); |
| 570 | } |
| 571 | if (c == ':') { |
| 572 | /* |
| 573 | * if the :g modifier is followed by a newline, then error right away! |
| 574 | * -strike |
| 575 | */ |
| 576 | |
| 577 | int gmodflag = 0, amodflag = 0; |
| 578 | |
| 579 | do { |
| 580 | Strbuf_append1(&name, c), c = getC(DOEXCL); |
| 581 | if (c == 'g' || c == 'a') { |
| 582 | if (c == 'g') |
| 583 | gmodflag++; |
| 584 | else |
| 585 | amodflag++; |
| 586 | Strbuf_append1(&name, c); c = getC(DOEXCL); |
| 587 | } |
| 588 | if ((c == 'g' && !gmodflag) || (c == 'a' && !amodflag)) { |
| 589 | if (c == 'g') |
| 590 | gmodflag++; |
| 591 | else |
| 592 | amodflag++; |
| 593 | Strbuf_append1(&name, c); c = getC(DOEXCL); |
| 594 | } |
| 595 | Strbuf_append1(&name, c); |
| 596 | /* scan s// [eichin:19910926.0512EST] */ |
| 597 | if (c == 's') { |
| 598 | int delimcnt = 2; |
| 599 | eChar delim = getC(0); |
| 600 | |
| 601 | Strbuf_append1(&name, delim); |
| 602 | if (!delim || letter(delim) |
| 603 | || Isdigit(delim) || any(" \t\n", delim)) { |
| 604 | seterror(ERR_BADSUBST); |
| 605 | break; |
| 606 | } |
| 607 | while ((c = getC(0)) != CHAR_ERR) { |
| 608 | Strbuf_append1(&name, c); |
| 609 | if(c == delim) delimcnt--; |
| 610 | if(!delimcnt) break; |
| 611 | } |
| 612 | if(delimcnt) { |
| 613 | seterror(ERR_BADSUBST); |
| 614 | break; |
| 615 | } |
| 616 | c = 's'; |
| 617 | } |
| 618 | if (!any("htrqxesul", c)) { |
| 619 | if ((amodflag || gmodflag) && c == '\n') |
| 620 | stderror(ERR_VARSYN); /* strike */ |
| 621 | seterror(ERR_BADMOD, c); |
| 622 | goto end; |
| 623 | } |
| 624 | } |
| 625 | while ((c = getC(DOEXCL)) == ':'); |
| 626 | ungetD(c); |
| 627 | } |
| 628 | else |
| 629 | ungetD(c); |
| 630 | if (sc == '{') { |
| 631 | c = getC(DOEXCL); |
| 632 | if (c != '}') { |
| 633 | ungetD(c); |
| 634 | seterror(ERR_MISSING, '}'); |
| 635 | goto end; |
| 636 | } |
| 637 | Strbuf_append1(&name, c); |
| 638 | } |
| 639 | end: |
| 640 | cleanup_ignore(&name); |
| 641 | cleanup_until(&name); |
| 642 | addla(Strbuf_finish(&name)); |
| 643 | } |
| 644 | |
| 645 | /* xfree()'s its argument */ |
| 646 | void |
| 647 | addla(Char *cp) |
| 648 | { |
| 649 | static struct Strbuf buf; /* = Strbuf_INIT; */ |
| 650 | |
| 651 | buf.len = 0; |
| 652 | Strbuf_appendn(&buf, labuf.s + lap, labuf.len - lap); |
| 653 | labuf.len = 0; |
| 654 | Strbuf_append(&labuf, cp); |
| 655 | Strbuf_terminate(&labuf); |
| 656 | Strbuf_appendn(&labuf, buf.s, buf.len); |
| 657 | xfree(cp); |
| 658 | lap = 0; |
| 659 | } |
| 660 | |
| 661 | /* left-hand side of last :s or search string of last ?event? */ |
| 662 | static struct Strbuf lhsb; /* = Strbuf_INIT; */ |
| 663 | static struct Strbuf slhs; /* = Strbuf_INIT; left-hand side of last :s */ |
| 664 | static struct Strbuf rhsb; /* = Strbuf_INIT; right-hand side of last :s */ |
| 665 | static int quesarg; |
| 666 | |
| 667 | static void |
| 668 | getexcl(Char sc) |
| 669 | { |
| 670 | struct wordent *hp, *ip; |
| 671 | int left, right, dol; |
| 672 | eChar c; |
| 673 | |
| 674 | if (sc == 0) { |
| 675 | sc = getC(0); |
| 676 | if (sc != '{') { |
| 677 | ungetC(sc); |
| 678 | sc = 0; |
| 679 | } |
| 680 | } |
| 681 | quesarg = -1; |
| 682 | |
| 683 | lastev = eventno; |
| 684 | hp = gethent(sc); |
| 685 | if (hp == 0) |
| 686 | return; |
| 687 | hadhist = 1; |
| 688 | dol = 0; |
| 689 | if (hp == alhistp) |
| 690 | for (ip = hp->next->next; ip != alhistt; ip = ip->next) |
| 691 | dol++; |
| 692 | else |
| 693 | for (ip = hp->next->next; ip != hp->prev; ip = ip->next) |
| 694 | dol++; |
| 695 | left = 0, right = dol; |
| 696 | if (sc == HISTSUB) { |
| 697 | ungetC('s'), unreadc(HISTSUB), c = ':'; |
| 698 | goto subst; |
| 699 | } |
| 700 | c = getC(0); |
| 701 | if (!any(":^$*-%", c)) |
| 702 | goto subst; |
| 703 | left = right = -1; |
| 704 | if (c == ':') { |
| 705 | c = getC(0); |
| 706 | unreadc(c); |
| 707 | if (letter(c) || c == '&') { |
| 708 | c = ':'; |
| 709 | left = 0, right = dol; |
| 710 | goto subst; |
| 711 | } |
| 712 | } |
| 713 | else |
| 714 | ungetC(c); |
| 715 | if (!getsel(&left, &right, dol)) |
| 716 | return; |
| 717 | c = getC(0); |
| 718 | if (c == '*') |
| 719 | ungetC(c), c = '-'; |
| 720 | if (c == '-') { |
| 721 | if (!getsel(&left, &right, dol)) |
| 722 | return; |
| 723 | c = getC(0); |
| 724 | } |
| 725 | subst: |
| 726 | exclc = right - left + 1; |
| 727 | while (--left >= 0) |
| 728 | hp = hp->next; |
| 729 | if (sc == HISTSUB || c == ':') { |
| 730 | do { |
| 731 | hp = getsub(hp); |
| 732 | c = getC(0); |
| 733 | } while (c == ':'); |
| 734 | } |
| 735 | unreadc(c); |
| 736 | if (sc == '{') { |
| 737 | c = getC(0); |
| 738 | if (c != '}') |
| 739 | seterror(ERR_BADBANG); |
| 740 | } |
| 741 | exclnxt = hp; |
| 742 | } |
| 743 | |
| 744 | static struct wordent * |
| 745 | getsub(struct wordent *en) |
| 746 | { |
| 747 | eChar delim; |
| 748 | eChar c; |
| 749 | eChar sc; |
| 750 | int global; |
| 751 | |
| 752 | do { |
| 753 | exclnxt = 0; |
| 754 | global = 0; |
| 755 | sc = c = getC(0); |
| 756 | while (c == 'g' || c == 'a') { |
| 757 | global |= (c == 'g') ? FLAG_G : FLAG_A; |
| 758 | sc = c = getC(0); |
| 759 | } |
| 760 | |
| 761 | switch (c) { |
| 762 | case 'p': |
| 763 | justpr++; |
| 764 | return (en); |
| 765 | |
| 766 | case 'x': |
| 767 | case 'q': |
| 768 | global |= FLAG_G; |
| 769 | /*FALLTHROUGH*/ |
| 770 | |
| 771 | case 'h': |
| 772 | case 'r': |
| 773 | case 't': |
| 774 | case 'e': |
| 775 | case 'u': |
| 776 | case 'l': |
| 777 | break; |
| 778 | |
| 779 | case '&': |
| 780 | if (slhs.len == 0) { |
| 781 | seterror(ERR_NOSUBST); |
| 782 | return (en); |
| 783 | } |
| 784 | lhsb.len = 0; |
| 785 | Strbuf_append(&lhsb, slhs.s); |
| 786 | Strbuf_terminate(&lhsb); |
| 787 | break; |
| 788 | |
| 789 | #ifdef notdef |
| 790 | case '~': |
| 791 | if (lhsb.len == 0) |
| 792 | goto badlhs; |
| 793 | break; |
| 794 | #endif |
| 795 | |
| 796 | case 's': |
| 797 | delim = getC(0); |
| 798 | if (letter(delim) || Isdigit(delim) || any(" \t\n", delim)) { |
| 799 | unreadc(delim); |
| 800 | lhsb.len = 0; |
| 801 | seterror(ERR_BADSUBST); |
| 802 | return (en); |
| 803 | } |
| 804 | Strbuf_terminate(&lhsb); |
| 805 | lhsb.len = 0; |
| 806 | for (;;) { |
| 807 | c = getC(0); |
| 808 | if (c == '\n') { |
| 809 | unreadc(c); |
| 810 | break; |
| 811 | } |
| 812 | if (c == delim) |
| 813 | break; |
| 814 | if (c == '\\') { |
| 815 | c = getC(0); |
| 816 | if (c != delim && c != '\\') |
| 817 | Strbuf_append1(&lhsb, '\\'); |
| 818 | } |
| 819 | Strbuf_append1(&lhsb, c); |
| 820 | } |
| 821 | if (lhsb.len != 0) |
| 822 | Strbuf_terminate(&lhsb); |
| 823 | else if (lhsb.s[0] == 0) { |
| 824 | seterror(ERR_LHS); |
| 825 | return (en); |
| 826 | } else |
| 827 | lhsb.len = Strlen(lhsb.s); /* lhsb.s wasn't changed */ |
| 828 | rhsb.len = 0; |
| 829 | for (;;) { |
| 830 | c = getC(0); |
| 831 | if (c == '\n') { |
| 832 | unreadc(c); |
| 833 | break; |
| 834 | } |
| 835 | if (c == delim) |
| 836 | break; |
| 837 | if (c == '\\') { |
| 838 | c = getC(0); |
| 839 | if (c != delim /* && c != '~' */ ) |
| 840 | Strbuf_append1(&rhsb, '\\'); |
| 841 | } |
| 842 | Strbuf_append1(&rhsb, c); |
| 843 | } |
| 844 | Strbuf_terminate(&rhsb); |
| 845 | break; |
| 846 | |
| 847 | default: |
| 848 | if (c == '\n') |
| 849 | unreadc(c); |
| 850 | seterror(ERR_BADBANGMOD, (int)c); |
| 851 | return (en); |
| 852 | } |
| 853 | slhs.len = 0; |
| 854 | if (lhsb.s != NULL && lhsb.len != 0) |
| 855 | Strbuf_append(&slhs, lhsb.s); |
| 856 | Strbuf_terminate(&slhs); |
| 857 | if (exclc) |
| 858 | en = dosub(sc, en, global); |
| 859 | } |
| 860 | while ((c = getC(0)) == ':'); |
| 861 | unreadc(c); |
| 862 | return (en); |
| 863 | } |
| 864 | |
| 865 | /* |
| 866 | * |
| 867 | * From Beto Appleton (beto@aixwiz.austin.ibm.com) |
| 868 | * |
| 869 | * when using history substitution, and the variable |
| 870 | * 'history' is set to a value higher than 1000, |
| 871 | * the shell might either freeze (hang) or core-dump. |
| 872 | * We raise the limit to 50000000 |
| 873 | */ |
| 874 | |
| 875 | #define HIST_PURGE -50000000 |
| 876 | static struct wordent * |
| 877 | dosub(Char sc, struct wordent *en, int global) |
| 878 | { |
| 879 | struct wordent lexi; |
| 880 | int didsub = 0, didone = 0; |
| 881 | struct wordent *hp = &lexi; |
| 882 | struct wordent *wdp; |
| 883 | int i = exclc; |
| 884 | struct Hist *hst; |
| 885 | |
| 886 | wdp = hp; |
| 887 | while (--i >= 0) { |
| 888 | struct wordent *new = xcalloc(1, sizeof *wdp); |
| 889 | |
| 890 | new->word = 0; |
| 891 | new->prev = wdp; |
| 892 | new->next = hp; |
| 893 | wdp->next = new; |
| 894 | wdp = new; |
| 895 | en = en->next; |
| 896 | if (en->word) { |
| 897 | Char *tword, *otword; |
| 898 | |
| 899 | if ((global & FLAG_G) || didsub == 0) { |
| 900 | size_t pos; |
| 901 | |
| 902 | pos = 0; |
| 903 | tword = subword(en->word, sc, &didone, &pos); |
| 904 | if (didone) |
| 905 | didsub = 1; |
| 906 | if (global & FLAG_A) { |
| 907 | while (didone && tword != STRNULL) { |
| 908 | otword = tword; |
| 909 | tword = subword(otword, sc, &didone, &pos); |
| 910 | if (Strcmp(tword, otword) == 0) { |
| 911 | xfree(otword); |
| 912 | break; |
| 913 | } |
| 914 | else |
| 915 | xfree(otword); |
| 916 | } |
| 917 | } |
| 918 | } |
| 919 | else |
| 920 | tword = Strsave(en->word); |
| 921 | wdp->word = tword; |
| 922 | } |
| 923 | } |
| 924 | if (didsub == 0) |
| 925 | seterror(ERR_MODFAIL); |
| 926 | hp->prev = wdp; |
| 927 | /* |
| 928 | * ANSI mode HP/UX compiler chokes on |
| 929 | * return &enthist(HIST_PURGE, &lexi, 0)->Hlex; |
| 930 | */ |
| 931 | hst = enthist(HIST_PURGE, &lexi, 0, 0); |
| 932 | return &(hst->Hlex); |
| 933 | } |
| 934 | |
| 935 | /* Return a newly allocated result of one modification of CP using the |
| 936 | operation TYPE. Set ADID to 1 if a modification was performed. |
| 937 | If TYPE == 's', perform substitutions only from *START_POS on and set |
| 938 | *START_POS to the position of next substitution attempt. */ |
| 939 | static Char * |
| 940 | subword(Char *cp, Char type, int *adid, size_t *start_pos) |
| 941 | { |
| 942 | Char *wp; |
| 943 | const Char *mp, *np; |
| 944 | |
| 945 | switch (type) { |
| 946 | |
| 947 | case 'r': |
| 948 | case 'e': |
| 949 | case 'h': |
| 950 | case 't': |
| 951 | case 'q': |
| 952 | case 'x': |
| 953 | case 'u': |
| 954 | case 'l': |
| 955 | wp = domod(cp, type); |
| 956 | if (wp == 0) { |
| 957 | *adid = 0; |
| 958 | return (Strsave(cp)); |
| 959 | } |
| 960 | *adid = 1; |
| 961 | return (wp); |
| 962 | |
| 963 | default: |
| 964 | for (mp = cp + *start_pos; *mp; mp++) { |
| 965 | if (matchs(mp, lhsb.s)) { |
| 966 | struct Strbuf wbuf = Strbuf_INIT; |
| 967 | |
| 968 | Strbuf_appendn(&wbuf, cp, mp - cp); |
| 969 | for (np = rhsb.s; *np; np++) |
| 970 | switch (*np) { |
| 971 | |
| 972 | case '\\': |
| 973 | if (np[1] == '&') |
| 974 | np++; |
| 975 | /* fall into ... */ |
| 976 | |
| 977 | default: |
| 978 | Strbuf_append1(&wbuf, *np); |
| 979 | continue; |
| 980 | |
| 981 | case '&': |
| 982 | Strbuf_append(&wbuf, lhsb.s); |
| 983 | continue; |
| 984 | } |
| 985 | *start_pos = wbuf.len; |
| 986 | Strbuf_append(&wbuf, mp + lhsb.len); |
| 987 | *adid = 1; |
| 988 | return Strbuf_finish(&wbuf); |
| 989 | } |
| 990 | } |
| 991 | *adid = 0; |
| 992 | return (Strsave(cp)); |
| 993 | } |
| 994 | } |
| 995 | |
| 996 | Char * |
| 997 | domod(Char *cp, Char type) |
| 998 | { |
| 999 | Char *wp, *xp; |
| 1000 | int c; |
| 1001 | |
| 1002 | switch (type) { |
| 1003 | |
| 1004 | case 'x': |
| 1005 | case 'q': |
| 1006 | wp = Strsave(cp); |
| 1007 | for (xp = wp; (c = *xp) != 0; xp++) |
| 1008 | if ((c != ' ' && c != '\t') || type == 'q') |
| 1009 | *xp |= QUOTE; |
| 1010 | return (wp); |
| 1011 | |
| 1012 | case 'l': |
| 1013 | wp = NLSChangeCase(cp, 1); |
| 1014 | return wp ? wp : Strsave(cp); |
| 1015 | |
| 1016 | case 'u': |
| 1017 | wp = NLSChangeCase(cp, 0); |
| 1018 | return wp ? wp : Strsave(cp); |
| 1019 | |
| 1020 | case 'h': |
| 1021 | case 't': |
| 1022 | if (!any(short2str(cp), '/')) |
| 1023 | return (type == 't' ? Strsave(cp) : 0); |
| 1024 | wp = Strrchr(cp, '/'); |
| 1025 | if (type == 'h') |
| 1026 | xp = Strnsave(cp, wp - cp); |
| 1027 | else |
| 1028 | xp = Strsave(wp + 1); |
| 1029 | return (xp); |
| 1030 | |
| 1031 | case 'e': |
| 1032 | case 'r': |
| 1033 | wp = Strend(cp); |
| 1034 | for (wp--; wp >= cp && *wp != '/'; wp--) |
| 1035 | if (*wp == '.') { |
| 1036 | if (type == 'e') |
| 1037 | xp = Strsave(wp + 1); |
| 1038 | else |
| 1039 | xp = Strnsave(cp, wp - cp); |
| 1040 | return (xp); |
| 1041 | } |
| 1042 | return (Strsave(type == 'e' ? STRNULL : cp)); |
| 1043 | default: |
| 1044 | break; |
| 1045 | } |
| 1046 | return (0); |
| 1047 | } |
| 1048 | |
| 1049 | static int |
| 1050 | matchs(const Char *str, const Char *pat) |
| 1051 | { |
| 1052 | while (*str && *pat && *str == *pat) |
| 1053 | str++, pat++; |
| 1054 | return (*pat == 0); |
| 1055 | } |
| 1056 | |
| 1057 | static int |
| 1058 | getsel(int *al, int *ar, int dol) |
| 1059 | { |
| 1060 | eChar c = getC(0); |
| 1061 | int i; |
| 1062 | int first = *al < 0; |
| 1063 | |
| 1064 | switch (c) { |
| 1065 | |
| 1066 | case '%': |
| 1067 | if (quesarg == -1) { |
| 1068 | seterror(ERR_BADBANGARG); |
| 1069 | return (0); |
| 1070 | } |
| 1071 | if (*al < 0) |
| 1072 | *al = quesarg; |
| 1073 | *ar = quesarg; |
| 1074 | break; |
| 1075 | |
| 1076 | case '-': |
| 1077 | if (*al < 0) { |
| 1078 | *al = 0; |
| 1079 | *ar = dol - 1; |
| 1080 | unreadc(c); |
| 1081 | } |
| 1082 | return (1); |
| 1083 | |
| 1084 | case '^': |
| 1085 | if (*al < 0) |
| 1086 | *al = 1; |
| 1087 | *ar = 1; |
| 1088 | break; |
| 1089 | |
| 1090 | case '$': |
| 1091 | if (*al < 0) |
| 1092 | *al = dol; |
| 1093 | *ar = dol; |
| 1094 | break; |
| 1095 | |
| 1096 | case '*': |
| 1097 | if (*al < 0) |
| 1098 | *al = 1; |
| 1099 | *ar = dol; |
| 1100 | if (*ar < *al) { |
| 1101 | *ar = 0; |
| 1102 | *al = 1; |
| 1103 | return (1); |
| 1104 | } |
| 1105 | break; |
| 1106 | |
| 1107 | default: |
| 1108 | if (Isdigit(c)) { |
| 1109 | i = 0; |
| 1110 | while (Isdigit(c)) { |
| 1111 | i = i * 10 + c - '0'; |
| 1112 | c = getC(0); |
| 1113 | } |
| 1114 | if (i < 0) |
| 1115 | i = dol + 1; |
| 1116 | if (*al < 0) |
| 1117 | *al = i; |
| 1118 | *ar = i; |
| 1119 | } |
| 1120 | else if (*al < 0) |
| 1121 | *al = 0, *ar = dol; |
| 1122 | else |
| 1123 | *ar = dol - 1; |
| 1124 | unreadc(c); |
| 1125 | break; |
| 1126 | } |
| 1127 | if (first) { |
| 1128 | c = getC(0); |
| 1129 | unreadc(c); |
| 1130 | if (any("-$*", c)) |
| 1131 | return (1); |
| 1132 | } |
| 1133 | if (*al > *ar || *ar > dol) { |
| 1134 | seterror(ERR_BADBANGARG); |
| 1135 | return (0); |
| 1136 | } |
| 1137 | return (1); |
| 1138 | |
| 1139 | } |
| 1140 | |
| 1141 | static struct wordent * |
| 1142 | gethent(Char sc) |
| 1143 | { |
| 1144 | struct Hist *hp; |
| 1145 | Char *np; |
| 1146 | eChar c; |
| 1147 | int event; |
| 1148 | int back = 0; |
| 1149 | |
| 1150 | c = sc == HISTSUB ? (eChar)HIST : getC(0); |
| 1151 | if (c == (eChar)HIST) { |
| 1152 | if (alhistp) |
| 1153 | return (alhistp); |
| 1154 | event = eventno; |
| 1155 | } |
| 1156 | else |
| 1157 | switch (c) { |
| 1158 | |
| 1159 | case ':': |
| 1160 | case '^': |
| 1161 | case '$': |
| 1162 | case '*': |
| 1163 | case '%': |
| 1164 | ungetC(c); |
| 1165 | if (lastev == eventno && alhistp) |
| 1166 | return (alhistp); |
| 1167 | event = lastev; |
| 1168 | break; |
| 1169 | |
| 1170 | case '#': /* !# is command being typed in (mrh) */ |
| 1171 | if (--hleft == 0) { |
| 1172 | seterror(ERR_HISTLOOP); |
| 1173 | return (0); |
| 1174 | } |
| 1175 | else |
| 1176 | return (¶ml); |
| 1177 | /* NOTREACHED */ |
| 1178 | |
| 1179 | case '-': |
| 1180 | back = 1; |
| 1181 | c = getC(0); |
| 1182 | /* FALLSTHROUGH */ |
| 1183 | |
| 1184 | default: |
| 1185 | if (any("(=~", c)) { |
| 1186 | unreadc(c); |
| 1187 | ungetC(HIST); |
| 1188 | return (0); |
| 1189 | } |
| 1190 | Strbuf_terminate(&lhsb); |
| 1191 | lhsb.len = 0; |
| 1192 | event = 0; |
| 1193 | while (!cmap(c, _ESC | _META | _QF | _QB) && !any("^*-%${}:#", c)) { |
| 1194 | if (event != -1 && Isdigit(c)) |
| 1195 | event = event * 10 + c - '0'; |
| 1196 | else |
| 1197 | event = -1; |
| 1198 | Strbuf_append1(&lhsb, c); |
| 1199 | c = getC(0); |
| 1200 | } |
| 1201 | unreadc(c); |
| 1202 | if (lhsb.len == 0) { |
| 1203 | lhsb.len = Strlen(lhsb.s); /* lhsb.s wasn't changed */ |
| 1204 | ungetC(HIST); |
| 1205 | return (0); |
| 1206 | } |
| 1207 | Strbuf_terminate(&lhsb); |
| 1208 | if (event != -1) { |
| 1209 | /* |
| 1210 | * History had only digits |
| 1211 | */ |
| 1212 | if (back) |
| 1213 | event = eventno + (alhistp == 0) - event; |
| 1214 | break; |
| 1215 | } |
| 1216 | if (back) { |
| 1217 | Strbuf_append1(&lhsb, '\0'); /* Allocate space */ |
| 1218 | Strbuf_terminate(&lhsb); |
| 1219 | memmove(lhsb.s + 1, lhsb.s, (lhsb.len - 1) * sizeof (*lhsb.s)); |
| 1220 | lhsb.s[0] = '-'; |
| 1221 | } |
| 1222 | hp = findev(lhsb.s, 0); |
| 1223 | if (hp) |
| 1224 | lastev = hp->Hnum; |
| 1225 | return (&hp->Hlex); |
| 1226 | |
| 1227 | case '?': |
| 1228 | Strbuf_terminate(&lhsb); |
| 1229 | lhsb.len = 0; |
| 1230 | for (;;) { |
| 1231 | c = getC(0); |
| 1232 | if (c == '\n') { |
| 1233 | unreadc(c); |
| 1234 | break; |
| 1235 | } |
| 1236 | if (c == '?') |
| 1237 | break; |
| 1238 | Strbuf_append1(&lhsb, c); |
| 1239 | } |
| 1240 | if (lhsb.len == 0) { |
| 1241 | lhsb.len = Strlen(lhsb.s); /* lhsb.s wasn't changed */ |
| 1242 | if (lhsb.len == 0) { |
| 1243 | seterror(ERR_NOSEARCH); |
| 1244 | return (0); |
| 1245 | } |
| 1246 | } |
| 1247 | else |
| 1248 | Strbuf_terminate(&lhsb); |
| 1249 | hp = findev(lhsb.s, 1); |
| 1250 | if (hp) |
| 1251 | lastev = hp->Hnum; |
| 1252 | return (&hp->Hlex); |
| 1253 | } |
| 1254 | |
| 1255 | for (hp = Histlist.Hnext; hp; hp = hp->Hnext) |
| 1256 | if (hp->Hnum == event) { |
| 1257 | hp->Href = eventno; |
| 1258 | lastev = hp->Hnum; |
| 1259 | return (&hp->Hlex); |
| 1260 | } |
| 1261 | np = putn(event); |
| 1262 | seterror(ERR_NOEVENT, short2str(np)); |
| 1263 | xfree(np); |
| 1264 | return (0); |
| 1265 | } |
| 1266 | |
| 1267 | static struct Hist * |
| 1268 | findev(Char *cp, int anyarg) |
| 1269 | { |
| 1270 | struct Hist *hp; |
| 1271 | |
| 1272 | for (hp = Histlist.Hnext; hp; hp = hp->Hnext) { |
| 1273 | Char *dp; |
| 1274 | Char *p, *q; |
| 1275 | struct wordent *lp = hp->Hlex.next; |
| 1276 | int argno = 0; |
| 1277 | |
| 1278 | /* |
| 1279 | * The entries added by alias substitution don't have a newline but do |
| 1280 | * have a negative event number. Savehist() trims off these entries, |
| 1281 | * but it happens before alias expansion, too early to delete those |
| 1282 | * from the previous command. |
| 1283 | */ |
| 1284 | if (hp->Hnum < 0) |
| 1285 | continue; |
| 1286 | if (lp->word[0] == '\n') |
| 1287 | continue; |
| 1288 | if (!anyarg) { |
| 1289 | p = cp; |
| 1290 | q = lp->word; |
| 1291 | do |
| 1292 | if (!*p) |
| 1293 | return (hp); |
| 1294 | while (*p++ == *q++); |
| 1295 | continue; |
| 1296 | } |
| 1297 | do { |
| 1298 | for (dp = lp->word; *dp; dp++) { |
| 1299 | p = cp; |
| 1300 | q = dp; |
| 1301 | do |
| 1302 | if (!*p) { |
| 1303 | quesarg = argno; |
| 1304 | return (hp); |
| 1305 | } |
| 1306 | while (*p++ == *q++); |
| 1307 | } |
| 1308 | lp = lp->next; |
| 1309 | argno++; |
| 1310 | } while (lp->word[0] != '\n'); |
| 1311 | } |
| 1312 | seterror(ERR_NOEVENT, short2str(cp)); |
| 1313 | return (0); |
| 1314 | } |
| 1315 | |
| 1316 | |
| 1317 | static void |
| 1318 | setexclp(Char *cp) |
| 1319 | { |
| 1320 | if (cp && cp[0] == '\n') |
| 1321 | return; |
| 1322 | exclp = cp; |
| 1323 | } |
| 1324 | |
| 1325 | void |
| 1326 | unreadc(Char c) |
| 1327 | { |
| 1328 | peekread = (Char) c; |
| 1329 | } |
| 1330 | |
| 1331 | eChar |
| 1332 | readc(int wanteof) |
| 1333 | { |
| 1334 | eChar c; |
| 1335 | static int sincereal; /* Number of real EOFs we've seen */ |
| 1336 | |
| 1337 | #ifdef DEBUG_INP |
| 1338 | xprintf("readc\n"); |
| 1339 | #endif |
| 1340 | if ((c = peekread) != 0) { |
| 1341 | peekread = 0; |
| 1342 | return (c); |
| 1343 | } |
| 1344 | |
| 1345 | top: |
| 1346 | aret = TCSH_F_SEEK; |
| 1347 | if (alvecp) { |
| 1348 | arun = 1; |
| 1349 | #ifdef DEBUG_INP |
| 1350 | xprintf("alvecp %c\n", *alvecp & 0xff); |
| 1351 | #endif |
| 1352 | aret = TCSH_A_SEEK; |
| 1353 | if ((c = *alvecp++) != 0) |
| 1354 | return (c); |
| 1355 | if (alvec && *alvec) { |
| 1356 | alvecp = *alvec++; |
| 1357 | return (' '); |
| 1358 | } |
| 1359 | else { |
| 1360 | alvecp = NULL; |
| 1361 | aret = TCSH_F_SEEK; |
| 1362 | return('\n'); |
| 1363 | } |
| 1364 | } |
| 1365 | if (alvec) { |
| 1366 | arun = 1; |
| 1367 | if ((alvecp = *alvec) != 0) { |
| 1368 | alvec++; |
| 1369 | goto top; |
| 1370 | } |
| 1371 | /* Infinite source! */ |
| 1372 | return ('\n'); |
| 1373 | } |
| 1374 | arun = 0; |
| 1375 | if (evalp) { |
| 1376 | aret = TCSH_E_SEEK; |
| 1377 | if ((c = *evalp++) != 0) |
| 1378 | return (c); |
| 1379 | if (evalvec && *evalvec) { |
| 1380 | evalp = *evalvec++; |
| 1381 | return (' '); |
| 1382 | } |
| 1383 | aret = TCSH_F_SEEK; |
| 1384 | evalp = 0; |
| 1385 | } |
| 1386 | if (evalvec) { |
| 1387 | if (evalvec == INVPPTR) { |
| 1388 | doneinp = 1; |
| 1389 | reset(); |
| 1390 | } |
| 1391 | if ((evalp = *evalvec) != 0) { |
| 1392 | evalvec++; |
| 1393 | goto top; |
| 1394 | } |
| 1395 | evalvec = INVPPTR; |
| 1396 | return ('\n'); |
| 1397 | } |
| 1398 | do { |
| 1399 | if (arginp == INVPTR || onelflg == 1) { |
| 1400 | if (wanteof) |
| 1401 | return CHAR_ERR; |
| 1402 | exitstat(); |
| 1403 | } |
| 1404 | if (arginp) { |
| 1405 | if ((c = *arginp++) == 0) { |
| 1406 | arginp = INVPTR; |
| 1407 | return ('\n'); |
| 1408 | } |
| 1409 | return (c); |
| 1410 | } |
| 1411 | #ifdef BSDJOBS |
| 1412 | reread: |
| 1413 | #endif /* BSDJOBS */ |
| 1414 | c = bgetc(); |
| 1415 | if (c == CHAR_ERR) { |
| 1416 | #ifndef WINNT_NATIVE |
| 1417 | # ifndef POSIX |
| 1418 | # ifdef TERMIO |
| 1419 | struct termio tty; |
| 1420 | # else /* SGTTYB */ |
| 1421 | struct sgttyb tty; |
| 1422 | # endif /* TERMIO */ |
| 1423 | # else /* POSIX */ |
| 1424 | struct termios tty; |
| 1425 | # endif /* POSIX */ |
| 1426 | #endif /* !WINNT_NATIVE */ |
| 1427 | if (wanteof) |
| 1428 | return CHAR_ERR; |
| 1429 | /* was isatty but raw with ignoreeof yields problems */ |
| 1430 | #ifndef WINNT_NATIVE |
| 1431 | # ifndef POSIX |
| 1432 | # ifdef TERMIO |
| 1433 | if (ioctl(SHIN, TCGETA, (ioctl_t) & tty) == 0 && |
| 1434 | (tty.c_lflag & ICANON)) |
| 1435 | # else /* GSTTYB */ |
| 1436 | if (ioctl(SHIN, TIOCGETP, (ioctl_t) & tty) == 0 && |
| 1437 | (tty.sg_flags & RAW) == 0) |
| 1438 | # endif /* TERMIO */ |
| 1439 | # else /* POSIX */ |
| 1440 | if (tcgetattr(SHIN, &tty) == 0 && |
| 1441 | (tty.c_lflag & ICANON)) |
| 1442 | # endif /* POSIX */ |
| 1443 | #else /* WINNT_NATIVE */ |
| 1444 | if (isatty(SHIN)) |
| 1445 | #endif /* !WINNT_NATIVE */ |
| 1446 | { |
| 1447 | #ifdef BSDJOBS |
| 1448 | pid_t ctpgrp; |
| 1449 | #endif /* BSDJOBS */ |
| 1450 | |
| 1451 | if (numeof != 0 && ++sincereal >= numeof) /* Too many EOFs? Bye! */ |
| 1452 | goto oops; |
| 1453 | #ifdef BSDJOBS |
| 1454 | if (tpgrp != -1 && |
| 1455 | (ctpgrp = tcgetpgrp(FSHTTY)) != -1 && |
| 1456 | tpgrp != ctpgrp) { |
| 1457 | (void) tcsetpgrp(FSHTTY, tpgrp); |
| 1458 | # ifdef _SEQUENT_ |
| 1459 | if (ctpgrp) |
| 1460 | # endif /* _SEQUENT */ |
| 1461 | (void) killpg(ctpgrp, SIGHUP); |
| 1462 | # ifdef notdef |
| 1463 | /* |
| 1464 | * With the walking process group fix, this message |
| 1465 | * is now obsolete. As the foreground process group |
| 1466 | * changes, the shell needs to adjust. Well too bad. |
| 1467 | */ |
| 1468 | xprintf(CGETS(16, 1, "Reset tty pgrp from %d to %d\n"), |
| 1469 | (int)ctpgrp, (int)tpgrp); |
| 1470 | # endif /* notdef */ |
| 1471 | goto reread; |
| 1472 | } |
| 1473 | #endif /* BSDJOBS */ |
| 1474 | /* What follows is complicated EOF handling -- sterling@netcom.com */ |
| 1475 | /* First, we check to see if we have ignoreeof set */ |
| 1476 | if (adrof(STRignoreeof)) { |
| 1477 | /* If so, we check for any stopped jobs only on the first EOF */ |
| 1478 | if ((sincereal == 1) && (chkstop == 0)) { |
| 1479 | panystop(1); |
| 1480 | } |
| 1481 | } else { |
| 1482 | /* If we don't have ignoreeof set, always check for stopped jobs */ |
| 1483 | if (chkstop == 0) { |
| 1484 | panystop(1); |
| 1485 | } |
| 1486 | } |
| 1487 | /* At this point, if there were stopped jobs, we would have already |
| 1488 | * called reset(). If we got this far, assume we can print an |
| 1489 | * exit/logout message if we ignoreeof, or just exit. |
| 1490 | */ |
| 1491 | if (adrof(STRignoreeof)) { |
| 1492 | /* If so, tell the user to use exit or logout */ |
| 1493 | if (loginsh) { |
| 1494 | xprintf("%s", CGETS(16, 2, |
| 1495 | "\nUse \"logout\" to logout.\n")); |
| 1496 | } else { |
| 1497 | xprintf(CGETS(16, 3, |
| 1498 | "\nUse \"exit\" to leave %s.\n"), |
| 1499 | progname); |
| 1500 | } |
| 1501 | reset(); |
| 1502 | } else { |
| 1503 | /* If we don't have ignoreeof set, just fall through */ |
| 1504 | ; /* EMPTY */ |
| 1505 | } |
| 1506 | } |
| 1507 | oops: |
| 1508 | doneinp = 1; |
| 1509 | reset(); |
| 1510 | } |
| 1511 | sincereal = 0; |
| 1512 | if (c == '\n' && onelflg) |
| 1513 | onelflg--; |
| 1514 | } while (c == 0); |
| 1515 | Strbuf_append1(&histline, c); |
| 1516 | return (c); |
| 1517 | } |
| 1518 | |
| 1519 | static void |
| 1520 | balloc(int buf) |
| 1521 | { |
| 1522 | Char **nfbuf; |
| 1523 | |
| 1524 | while (buf >= fblocks) { |
| 1525 | nfbuf = xcalloc(fblocks + 2, sizeof(Char **)); |
| 1526 | if (fbuf) { |
| 1527 | (void) blkcpy(nfbuf, fbuf); |
| 1528 | xfree(fbuf); |
| 1529 | } |
| 1530 | fbuf = nfbuf; |
| 1531 | fbuf[fblocks] = xcalloc(BUFSIZE, sizeof(Char)); |
| 1532 | fblocks++; |
| 1533 | } |
| 1534 | } |
| 1535 | |
| 1536 | static ssize_t |
| 1537 | wide_read(int fildes, Char *buf, size_t nchars, int use_fclens) |
| 1538 | { |
| 1539 | char cbuf[BUFSIZE + 1]; |
| 1540 | ssize_t res, r = 0; |
| 1541 | size_t partial; |
| 1542 | int err; |
| 1543 | |
| 1544 | if (nchars == 0) |
| 1545 | return 0; |
| 1546 | assert (nchars <= sizeof(cbuf) / sizeof(*cbuf)); |
| 1547 | USE(use_fclens); |
| 1548 | res = 0; |
| 1549 | partial = 0; |
| 1550 | do { |
| 1551 | size_t i; |
| 1552 | size_t len = nchars > partial ? nchars - partial : 1; |
| 1553 | |
| 1554 | if (partial + len >= sizeof(cbuf) / sizeof(*cbuf)) |
| 1555 | break; |
| 1556 | |
| 1557 | r = xread(fildes, cbuf + partial, len); |
| 1558 | |
| 1559 | if (partial == 0 && r <= 0) |
| 1560 | break; |
| 1561 | partial += r; |
| 1562 | i = 0; |
| 1563 | while (i < partial && nchars != 0) { |
| 1564 | int tlen; |
| 1565 | |
| 1566 | tlen = normal_mbtowc(buf + res, cbuf + i, partial - i); |
| 1567 | if (tlen == -1) { |
| 1568 | reset_mbtowc(); |
| 1569 | if ((partial - i) < MB_LEN_MAX && r > 0) |
| 1570 | /* Maybe a partial character and there is still a chance |
| 1571 | to read more */ |
| 1572 | break; |
| 1573 | buf[res] = (unsigned char)cbuf[i] | INVALID_BYTE; |
| 1574 | } |
| 1575 | if (tlen <= 0) |
| 1576 | tlen = 1; |
| 1577 | #ifdef WIDE_STRINGS |
| 1578 | if (use_fclens) |
| 1579 | fclens[res] = tlen; |
| 1580 | #endif |
| 1581 | i += tlen; |
| 1582 | res++; |
| 1583 | nchars--; |
| 1584 | } |
| 1585 | if (i != partial) |
| 1586 | memmove(cbuf, cbuf + i, partial - i); |
| 1587 | partial -= i; |
| 1588 | } while (partial != 0 && nchars > 0); |
| 1589 | /* Throwing away possible partial multibyte characters on error if the |
| 1590 | stream is not seekable */ |
| 1591 | err = errno; |
| 1592 | lseek(fildes, -(off_t)partial, L_INCR); |
| 1593 | errno = err; |
| 1594 | return res != 0 ? res : r; |
| 1595 | } |
| 1596 | |
| 1597 | static eChar |
| 1598 | bgetc(void) |
| 1599 | { |
| 1600 | Char ch; |
| 1601 | int c, off, buf; |
| 1602 | int numleft = 0, roomleft; |
| 1603 | |
| 1604 | if (cantell) { |
| 1605 | if (fseekp < fbobp || fseekp > feobp) { |
| 1606 | fbobp = feobp = fseekp; |
| 1607 | (void) lseek(SHIN, fseekp, L_SET); |
| 1608 | } |
| 1609 | if (fseekp == feobp) { |
| 1610 | #ifdef WIDE_STRINGS |
| 1611 | off_t bytes; |
| 1612 | size_t i; |
| 1613 | |
| 1614 | bytes = fbobp; |
| 1615 | for (i = 0; i < (size_t)(feobp - fbobp); i++) |
| 1616 | bytes += fclens[i]; |
| 1617 | fseekp = feobp = bytes; |
| 1618 | #endif |
| 1619 | fbobp = feobp; |
| 1620 | c = wide_read(SHIN, fbuf[0], BUFSIZE, 1); |
| 1621 | #ifdef convex |
| 1622 | if (c < 0) |
| 1623 | stderror(ERR_SYSTEM, progname, strerror(errno)); |
| 1624 | #endif /* convex */ |
| 1625 | if (c <= 0) |
| 1626 | return CHAR_ERR; |
| 1627 | feobp += c; |
| 1628 | } |
| 1629 | #if !defined(WINNT_NATIVE) && !defined(__CYGWIN__) |
| 1630 | ch = fbuf[0][fseekp - fbobp]; |
| 1631 | fseekp++; |
| 1632 | #else |
| 1633 | do { |
| 1634 | ch = fbuf[0][fseekp - fbobp]; |
| 1635 | fseekp++; |
| 1636 | } while(ch == '\r'); |
| 1637 | #endif /* !WINNT_NATIVE && !__CYGWIN__ */ |
| 1638 | return (ch); |
| 1639 | } |
| 1640 | |
| 1641 | while (fseekp >= feobp) { |
| 1642 | if ((editing |
| 1643 | #if defined(FILEC) && defined(TIOCSTI) |
| 1644 | || filec |
| 1645 | #endif /* FILEC && TIOCSTI */ |
| 1646 | ) && intty) { /* then use twenex routine */ |
| 1647 | fseekp = feobp; /* where else? */ |
| 1648 | #if defined(FILEC) && defined(TIOCSTI) |
| 1649 | if (!editing) |
| 1650 | c = numleft = tenex(InputBuf, BUFSIZE); |
| 1651 | else |
| 1652 | #endif /* FILEC && TIOCSTI */ |
| 1653 | c = numleft = Inputl(); /* PWP: get a line */ |
| 1654 | while (numleft > 0) { |
| 1655 | off = (int) feobp % BUFSIZE; |
| 1656 | buf = (int) feobp / BUFSIZE; |
| 1657 | balloc(buf); |
| 1658 | roomleft = BUFSIZE - off; |
| 1659 | if (roomleft > numleft) |
| 1660 | roomleft = numleft; |
| 1661 | (void) memcpy(fbuf[buf] + off, InputBuf + c - numleft, |
| 1662 | roomleft * sizeof(Char)); |
| 1663 | numleft -= roomleft; |
| 1664 | feobp += roomleft; |
| 1665 | } |
| 1666 | } else { |
| 1667 | off = (int) feobp % BUFSIZE; |
| 1668 | buf = (int) feobp / BUFSIZE; |
| 1669 | balloc(buf); |
| 1670 | roomleft = BUFSIZE - off; |
| 1671 | c = wide_read(SHIN, fbuf[buf] + off, roomleft, 0); |
| 1672 | if (c > 0) |
| 1673 | feobp += c; |
| 1674 | } |
| 1675 | if (c == 0 || (c < 0 && fixio(SHIN, errno) == -1)) |
| 1676 | return CHAR_ERR; |
| 1677 | } |
| 1678 | #ifdef SIG_WINDOW |
| 1679 | if (windowchg) |
| 1680 | (void) check_window_size(0); /* for window systems */ |
| 1681 | #endif /* SIG_WINDOW */ |
| 1682 | #if !defined(WINNT_NATIVE) && !defined(__CYGWIN__) |
| 1683 | ch = fbuf[(int) fseekp / BUFSIZE][(int) fseekp % BUFSIZE]; |
| 1684 | fseekp++; |
| 1685 | #else |
| 1686 | do { |
| 1687 | ch = fbuf[(int) fseekp / BUFSIZE][(int) fseekp % BUFSIZE]; |
| 1688 | fseekp++; |
| 1689 | } while(ch == '\r'); |
| 1690 | #endif /* !WINNT_NATIVE && !__CYGWIN__ */ |
| 1691 | return (ch); |
| 1692 | } |
| 1693 | |
| 1694 | static void |
| 1695 | bfree(void) |
| 1696 | { |
| 1697 | int sb, i; |
| 1698 | |
| 1699 | if (cantell) |
| 1700 | return; |
| 1701 | if (whyles) |
| 1702 | return; |
| 1703 | sb = (int) (fseekp - 1) / BUFSIZE; |
| 1704 | if (sb > 0) { |
| 1705 | for (i = 0; i < sb; i++) |
| 1706 | xfree(fbuf[i]); |
| 1707 | (void) blkcpy(fbuf, &fbuf[sb]); |
| 1708 | fseekp -= BUFSIZE * sb; |
| 1709 | feobp -= BUFSIZE * sb; |
| 1710 | fblocks -= sb; |
| 1711 | } |
| 1712 | } |
| 1713 | |
| 1714 | void |
| 1715 | bseek(struct Ain *l) |
| 1716 | { |
| 1717 | switch (aret = l->type) { |
| 1718 | case TCSH_E_SEEK: |
| 1719 | evalvec = l->a_seek; |
| 1720 | evalp = l->c_seek; |
| 1721 | #ifdef DEBUG_SEEK |
| 1722 | xprintf(CGETS(16, 4, "seek to eval %x %x\n"), evalvec, evalp); |
| 1723 | #endif |
| 1724 | return; |
| 1725 | case TCSH_A_SEEK: |
| 1726 | alvec = l->a_seek; |
| 1727 | alvecp = l->c_seek; |
| 1728 | #ifdef DEBUG_SEEK |
| 1729 | xprintf(CGETS(16, 5, "seek to alias %x %x\n"), alvec, alvecp); |
| 1730 | #endif |
| 1731 | return; |
| 1732 | case TCSH_F_SEEK: |
| 1733 | #ifdef DEBUG_SEEK |
| 1734 | xprintf(CGETS(16, 6, "seek to file %x\n"), fseekp); |
| 1735 | #endif |
| 1736 | fseekp = l->f_seek; |
| 1737 | #ifdef WIDE_STRINGS |
| 1738 | if (cantell) { |
| 1739 | if (fseekp >= fbobp && feobp >= fbobp) { |
| 1740 | size_t i; |
| 1741 | off_t o; |
| 1742 | |
| 1743 | o = fbobp; |
| 1744 | for (i = 0; i < (size_t)(feobp - fbobp); i++) { |
| 1745 | if (fseekp == o) { |
| 1746 | fseekp = fbobp + i; |
| 1747 | return; |
| 1748 | } |
| 1749 | o += fclens[i]; |
| 1750 | } |
| 1751 | if (fseekp == o) { |
| 1752 | fseekp = feobp; |
| 1753 | return; |
| 1754 | } |
| 1755 | } |
| 1756 | fbobp = feobp = fseekp + 1; /* To force lseek() */ |
| 1757 | } |
| 1758 | #endif |
| 1759 | return; |
| 1760 | default: |
| 1761 | xprintf(CGETS(16, 7, "Bad seek type %d\n"), aret); |
| 1762 | abort(); |
| 1763 | } |
| 1764 | } |
| 1765 | |
| 1766 | /* any similarity to bell telephone is purely accidental */ |
| 1767 | void |
| 1768 | btell(struct Ain *l) |
| 1769 | { |
| 1770 | switch (l->type = aret) { |
| 1771 | case TCSH_E_SEEK: |
| 1772 | l->a_seek = evalvec; |
| 1773 | l->c_seek = evalp; |
| 1774 | #ifdef DEBUG_SEEK |
| 1775 | xprintf(CGETS(16, 8, "tell eval %x %x\n"), evalvec, evalp); |
| 1776 | #endif |
| 1777 | return; |
| 1778 | case TCSH_A_SEEK: |
| 1779 | l->a_seek = alvec; |
| 1780 | l->c_seek = alvecp; |
| 1781 | #ifdef DEBUG_SEEK |
| 1782 | xprintf(CGETS(16, 9, "tell alias %x %x\n"), alvec, alvecp); |
| 1783 | #endif |
| 1784 | return; |
| 1785 | case TCSH_F_SEEK: |
| 1786 | #ifdef WIDE_STRINGS |
| 1787 | if (cantell && fseekp >= fbobp && fseekp <= feobp) { |
| 1788 | size_t i; |
| 1789 | |
| 1790 | l->f_seek = fbobp; |
| 1791 | for (i = 0; i < (size_t)(fseekp - fbobp); i++) |
| 1792 | l->f_seek += fclens[i]; |
| 1793 | } else |
| 1794 | #endif |
| 1795 | /*SUPPRESS 112*/ |
| 1796 | l->f_seek = fseekp; |
| 1797 | l->a_seek = NULL; |
| 1798 | #ifdef DEBUG_SEEK |
| 1799 | xprintf(CGETS(16, 10, "tell file %x\n"), fseekp); |
| 1800 | #endif |
| 1801 | return; |
| 1802 | default: |
| 1803 | xprintf(CGETS(16, 7, "Bad seek type %d\n"), aret); |
| 1804 | abort(); |
| 1805 | } |
| 1806 | } |
| 1807 | |
| 1808 | void |
| 1809 | btoeof(void) |
| 1810 | { |
| 1811 | (void) lseek(SHIN, (off_t) 0, L_XTND); |
| 1812 | aret = TCSH_F_SEEK; |
| 1813 | fseekp = feobp; |
| 1814 | alvec = NULL; |
| 1815 | alvecp = NULL; |
| 1816 | evalvec = NULL; |
| 1817 | evalp = NULL; |
| 1818 | wfree(); |
| 1819 | bfree(); |
| 1820 | } |
| 1821 | |
| 1822 | void |
| 1823 | settell(void) |
| 1824 | { |
| 1825 | off_t x; |
| 1826 | cantell = 0; |
| 1827 | if (arginp || onelflg || intty) |
| 1828 | return; |
| 1829 | if ((x = lseek(SHIN, (off_t) 0, L_INCR)) == -1) |
| 1830 | return; |
| 1831 | fbuf = xcalloc(2, sizeof(Char **)); |
| 1832 | fblocks = 1; |
| 1833 | fbuf[0] = xcalloc(BUFSIZE, sizeof(Char)); |
| 1834 | fseekp = fbobp = feobp = x; |
| 1835 | cantell = 1; |
| 1836 | } |