Annotation of src/usr.bin/awk/run.c, Revision 1.63
1.63 ! millert 1: /* $OpenBSD: run.c,v 1.62 2020/07/01 13:32:27 millert Exp $ */
1.1 tholo 2: /****************************************************************
1.13 kstailey 3: Copyright (C) Lucent Technologies 1997
1.1 tholo 4: All Rights Reserved
5:
6: Permission to use, copy, modify, and distribute this software and
7: its documentation for any purpose and without fee is hereby
8: granted, provided that the above copyright notice appear in all
9: copies and that both that the copyright notice and this
10: permission notice and warranty disclaimer appear in supporting
1.13 kstailey 11: documentation, and that the name Lucent Technologies or any of
12: its entities not be used in advertising or publicity pertaining
13: to distribution of the software without specific, written prior
14: permission.
15:
16: LUCENT DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
17: INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS.
18: IN NO EVENT SHALL LUCENT OR ANY OF ITS ENTITIES BE LIABLE FOR ANY
19: SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
20: WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER
21: IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION,
22: ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF
23: THIS SOFTWARE.
1.1 tholo 24: ****************************************************************/
25:
26: #define DEBUG
27: #include <stdio.h>
28: #include <ctype.h>
1.63 ! millert 29: #include <errno.h>
1.57 millert 30: #include <wchar.h>
31: #include <wctype.h>
1.56 millert 32: #include <fcntl.h>
1.1 tholo 33: #include <setjmp.h>
1.25 millert 34: #include <limits.h>
1.1 tholo 35: #include <math.h>
36: #include <string.h>
37: #include <stdlib.h>
38: #include <time.h>
1.47 millert 39: #include <sys/types.h>
40: #include <sys/wait.h>
1.1 tholo 41: #include "awk.h"
1.13 kstailey 42: #include "ytab.h"
1.1 tholo 43:
1.57 millert 44: static void stdinit(void);
45: static void flush_all(void);
1.1 tholo 46:
1.57 millert 47: #if 1
48: #define tempfree(x) do { if (istemp(x)) tfree(x); } while (/*CONSTCOND*/0)
49: #else
1.1 tholo 50: void tempfree(Cell *p) {
51: if (p->ctype == OCELL && (p->csub < CUNK || p->csub > CFREE)) {
1.16 millert 52: WARNING("bad csub %d in Cell %d %s",
53: p->csub, p->ctype, p->sval);
1.1 tholo 54: }
55: if (istemp(p))
56: tfree(p);
57: }
1.57 millert 58: #endif
1.1 tholo 59:
1.30 millert 60: /* do we really need these? */
61: /* #ifdef _NFILE */
62: /* #ifndef FOPEN_MAX */
63: /* #define FOPEN_MAX _NFILE */
64: /* #endif */
65: /* #endif */
66: /* */
67: /* #ifndef FOPEN_MAX */
68: /* #define FOPEN_MAX 40 */ /* max number of open files */
69: /* #endif */
70: /* */
71: /* #ifndef RAND_MAX */
72: /* #define RAND_MAX 32767 */ /* all that ansi guarantees */
73: /* #endif */
1.1 tholo 74:
75: jmp_buf env;
1.13 kstailey 76: extern int pairstack[];
1.33 millert 77: extern Awkfloat srand_seed;
1.1 tholo 78:
79: Node *winner = NULL; /* root of parse tree */
80: Cell *tmps; /* free temporary cells for execution */
81:
1.57 millert 82: static Cell truecell ={ OBOOL, BTRUE, 0, 0, 1.0, NUM, NULL, NULL };
1.15 millert 83: Cell *True = &truecell;
1.57 millert 84: static Cell falsecell ={ OBOOL, BFALSE, 0, 0, 0.0, NUM, NULL, NULL };
1.15 millert 85: Cell *False = &falsecell;
1.57 millert 86: static Cell breakcell ={ OJUMP, JBREAK, 0, 0, 0.0, NUM, NULL, NULL };
1.1 tholo 87: Cell *jbreak = &breakcell;
1.57 millert 88: static Cell contcell ={ OJUMP, JCONT, 0, 0, 0.0, NUM, NULL, NULL };
1.1 tholo 89: Cell *jcont = &contcell;
1.57 millert 90: static Cell nextcell ={ OJUMP, JNEXT, 0, 0, 0.0, NUM, NULL, NULL };
1.1 tholo 91: Cell *jnext = &nextcell;
1.57 millert 92: static Cell nextfilecell ={ OJUMP, JNEXTFILE, 0, 0, 0.0, NUM, NULL, NULL };
1.1 tholo 93: Cell *jnextfile = &nextfilecell;
1.57 millert 94: static Cell exitcell ={ OJUMP, JEXIT, 0, 0, 0.0, NUM, NULL, NULL };
1.1 tholo 95: Cell *jexit = &exitcell;
1.57 millert 96: static Cell retcell ={ OJUMP, JRET, 0, 0, 0.0, NUM, NULL, NULL };
1.1 tholo 97: Cell *jret = &retcell;
1.57 millert 98: static Cell tempcell ={ OCELL, CTEMP, 0, EMPTY, 0.0, NUM|STR|DONTFREE, NULL, NULL };
1.1 tholo 99:
100: Node *curnode = NULL; /* the node being executed, for debugging */
1.22 deraadt 101:
1.13 kstailey 102: /* buffer memory management */
103: int adjbuf(char **pbuf, int *psiz, int minlen, int quantum, char **pbptr,
1.18 millert 104: const char *whatrtn)
1.13 kstailey 105: /* pbuf: address of pointer to buffer being managed
106: * psiz: address of buffer size variable
107: * minlen: minimum length of buffer needed
108: * quantum: buffer size quantum
109: * pbptr: address of movable pointer into buffer, or 0 if none
110: * whatrtn: name of the calling routine if failure should cause fatal error
111: *
112: * return 0 for realloc failure, !=0 for success
113: */
114: {
115: if (minlen > *psiz) {
116: char *tbuf;
117: int rminlen = quantum ? minlen % quantum : 0;
118: int boff = pbptr ? *pbptr - *pbuf : 0;
119: /* round up to next multiple of quantum */
120: if (rminlen)
121: minlen += quantum - rminlen;
1.53 millert 122: tbuf = realloc(*pbuf, minlen);
1.60 millert 123: DPRINTF("adjbuf %s: %d %d (pbuf=%p, tbuf=%p)\n", whatrtn, *psiz, minlen, *pbuf, tbuf);
1.13 kstailey 124: if (tbuf == NULL) {
125: if (whatrtn)
1.16 millert 126: FATAL("out of memory in %s", whatrtn);
1.13 kstailey 127: return 0;
128: }
129: *pbuf = tbuf;
130: *psiz = minlen;
131: if (pbptr)
132: *pbptr = tbuf + boff;
133: }
134: return 1;
135: }
136:
1.1 tholo 137: void run(Node *a) /* execution of parse tree starts here */
138: {
1.63 ! millert 139:
1.16 millert 140: stdinit();
1.1 tholo 141: execute(a);
142: closeall();
143: }
144:
145: Cell *execute(Node *u) /* execute a node of the parse tree */
146: {
147: Cell *(*proc)(Node **, int);
148: Cell *x;
149: Node *a;
150:
151: if (u == NULL)
1.15 millert 152: return(True);
1.1 tholo 153: for (a = u; ; a = a->nnext) {
154: curnode = a;
155: if (isvalue(a)) {
1.2 millert 156: x = (Cell *) (a->narg[0]);
1.13 kstailey 157: if (isfld(x) && !donefld)
1.1 tholo 158: fldbld();
1.13 kstailey 159: else if (isrec(x) && !donerec)
1.1 tholo 160: recbld();
161: return(x);
162: }
163: if (notlegal(a->nobj)) /* probably a Cell* but too risky to print */
1.16 millert 164: FATAL("illegal statement");
1.1 tholo 165: proc = proctab[a->nobj-FIRSTTOKEN];
166: x = (*proc)(a->narg, a->nobj);
1.13 kstailey 167: if (isfld(x) && !donefld)
1.1 tholo 168: fldbld();
1.13 kstailey 169: else if (isrec(x) && !donerec)
1.1 tholo 170: recbld();
171: if (isexpr(a))
172: return(x);
173: if (isjump(x))
174: return(x);
175: if (a->nnext == NULL)
176: return(x);
177: tempfree(x);
178: }
179: }
180:
181:
182: Cell *program(Node **a, int n) /* execute an awk program */
183: { /* a[0] = BEGIN, a[1] = body, a[2] = END */
184: Cell *x;
185:
186: if (setjmp(env) != 0)
187: goto ex;
188: if (a[0]) { /* BEGIN */
189: x = execute(a[0]);
190: if (isexit(x))
1.15 millert 191: return(True);
1.1 tholo 192: if (isjump(x))
1.16 millert 193: FATAL("illegal break, continue, next or nextfile from BEGIN");
1.1 tholo 194: tempfree(x);
195: }
196: if (a[1] || a[2])
1.54 millert 197: while (getrec(&record, &recsize, true) > 0) {
1.1 tholo 198: x = execute(a[1]);
199: if (isexit(x))
200: break;
201: tempfree(x);
202: }
203: ex:
204: if (setjmp(env) != 0) /* handles exit within END */
205: goto ex1;
206: if (a[2]) { /* END */
207: x = execute(a[2]);
208: if (isbreak(x) || isnext(x) || iscont(x))
1.16 millert 209: FATAL("illegal break, continue, next or nextfile from END");
1.1 tholo 210: tempfree(x);
211: }
212: ex1:
1.15 millert 213: return(True);
1.1 tholo 214: }
215:
216: struct Frame { /* stack frame for awk function calls */
217: int nargs; /* number of arguments in this call */
218: Cell *fcncell; /* pointer to Cell for function */
219: Cell **args; /* pointer to array of arguments after execute */
220: Cell *retval; /* return value */
221: };
222:
223: #define NARGS 50 /* max args in a call */
224:
225: struct Frame *frame = NULL; /* base of stack frames; dynamically allocated */
226: int nframe = 0; /* number of frames allocated */
1.57 millert 227: struct Frame *frp = NULL; /* frame pointer. bottom level unused */
1.1 tholo 228:
229: Cell *call(Node **a, int n) /* function call. very kludgy and fragile */
230: {
1.57 millert 231: static const Cell newcopycell = { OCELL, CCOPY, 0, EMPTY, 0.0, NUM|STR|DONTFREE, NULL, NULL };
1.1 tholo 232: int i, ncall, ndef;
1.25 millert 233: int freed = 0; /* handles potential double freeing when fcn & param share a tempcell */
1.1 tholo 234: Node *x;
1.13 kstailey 235: Cell *args[NARGS], *oargs[NARGS]; /* BUG: fixed size arrays */
236: Cell *y, *z, *fcn;
1.1 tholo 237: char *s;
238:
239: fcn = execute(a[0]); /* the function itself */
240: s = fcn->nval;
1.13 kstailey 241: if (!isfcn(fcn))
1.16 millert 242: FATAL("calling undefined function %s", s);
1.1 tholo 243: if (frame == NULL) {
1.57 millert 244: frp = frame = calloc(nframe += 100, sizeof(*frame));
1.1 tholo 245: if (frame == NULL)
1.16 millert 246: FATAL("out of space for stack frames calling %s", s);
1.1 tholo 247: }
248: for (ncall = 0, x = a[1]; x != NULL; x = x->nnext) /* args in call */
249: ncall++;
1.12 millert 250: ndef = (int) fcn->fval; /* args in defn */
1.60 millert 251: DPRINTF("calling %s, %d args (%d in defn), frp=%d\n", s, ncall, ndef, (int) (frp-frame));
1.1 tholo 252: if (ncall > ndef)
1.16 millert 253: WARNING("function %s called with %d args, uses only %d",
254: s, ncall, ndef);
1.1 tholo 255: if (ncall + ndef > NARGS)
1.16 millert 256: FATAL("function %s has %d arguments, limit %d", s, ncall+ndef, NARGS);
1.1 tholo 257: for (i = 0, x = a[1]; x != NULL; i++, x = x->nnext) { /* get call args */
1.60 millert 258: DPRINTF("evaluate args[%d], frp=%d:\n", i, (int) (frp-frame));
1.1 tholo 259: y = execute(x);
260: oargs[i] = y;
1.60 millert 261: DPRINTF("args[%d]: %s %f <%s>, t=%o\n",
262: i, NN(y->nval), y->fval, isarr(y) ? "(array)" : NN(y->sval), y->tval);
1.13 kstailey 263: if (isfcn(y))
1.16 millert 264: FATAL("can't use function %s as argument in %s", y->nval, s);
1.1 tholo 265: if (isarr(y))
266: args[i] = y; /* arrays by ref */
267: else
268: args[i] = copycell(y);
269: tempfree(y);
270: }
271: for ( ; i < ndef; i++) { /* add null args for ones not provided */
272: args[i] = gettemp();
273: *args[i] = newcopycell;
274: }
1.57 millert 275: frp++; /* now ok to up frame */
276: if (frp >= frame + nframe) {
277: int dfp = frp - frame; /* old index */
278: frame = reallocarray(frame, (nframe += 100), sizeof(*frame));
1.1 tholo 279: if (frame == NULL)
1.16 millert 280: FATAL("out of space for stack frames in %s", s);
1.57 millert 281: frp = frame + dfp;
1.1 tholo 282: }
1.57 millert 283: frp->fcncell = fcn;
284: frp->args = args;
285: frp->nargs = ndef; /* number defined with (excess are locals) */
286: frp->retval = gettemp();
1.1 tholo 287:
1.60 millert 288: DPRINTF("start exec of %s, frp=%d\n", s, (int) (frp-frame));
1.1 tholo 289: y = execute((Node *)(fcn->sval)); /* execute body */
1.60 millert 290: DPRINTF("finished exec of %s, frp=%d\n", s, (int) (frp-frame));
1.1 tholo 291:
292: for (i = 0; i < ndef; i++) {
1.57 millert 293: Cell *t = frp->args[i];
1.1 tholo 294: if (isarr(t)) {
295: if (t->csub == CCOPY) {
296: if (i >= ncall) {
297: freesymtab(t);
298: t->csub = CTEMP;
1.14 millert 299: tempfree(t);
1.1 tholo 300: } else {
301: oargs[i]->tval = t->tval;
302: oargs[i]->tval &= ~(STR|NUM|DONTFREE);
303: oargs[i]->sval = t->sval;
304: tempfree(t);
305: }
306: }
307: } else if (t != y) { /* kludge to prevent freeing twice */
308: t->csub = CTEMP;
309: tempfree(t);
1.25 millert 310: } else if (t == y && t->csub == CCOPY) {
311: t->csub = CTEMP;
312: tempfree(t);
313: freed = 1;
1.1 tholo 314: }
315: }
316: tempfree(fcn);
1.17 millert 317: if (isexit(y) || isnext(y))
1.1 tholo 318: return y;
1.25 millert 319: if (freed == 0) {
320: tempfree(y); /* don't free twice! */
321: }
1.57 millert 322: z = frp->retval; /* return value */
1.60 millert 323: DPRINTF("%s returns %g |%s| %o\n", s, getfval(z), getsval(z), z->tval);
1.57 millert 324: frp--;
1.1 tholo 325: return(z);
326: }
327:
328: Cell *copycell(Cell *x) /* make a copy of a cell in a temp */
329: {
330: Cell *y;
331:
1.47 millert 332: /* copy is not constant or field */
333:
1.1 tholo 334: y = gettemp();
1.47 millert 335: y->tval = x->tval & ~(CON|FLD|REC);
1.1 tholo 336: y->csub = CCOPY; /* prevents freeing until call is over */
1.13 kstailey 337: y->nval = x->nval; /* BUG? */
1.47 millert 338: if (isstr(x) /* || x->ctype == OCELL */) {
1.17 millert 339: y->sval = tostring(x->sval);
1.47 millert 340: y->tval &= ~DONTFREE;
341: } else
342: y->tval |= DONTFREE;
1.1 tholo 343: y->fval = x->fval;
344: return y;
345: }
346:
347: Cell *arg(Node **a, int n) /* nth argument of a function */
348: {
349:
1.15 millert 350: n = ptoi(a[0]); /* argument number, counting from 0 */
1.60 millert 351: DPRINTF("arg(%d), frp->nargs=%d\n", n, frp->nargs);
1.57 millert 352: if (n+1 > frp->nargs)
1.16 millert 353: FATAL("argument #%d of function %s was not supplied",
1.57 millert 354: n+1, frp->fcncell->nval);
355: return frp->args[n];
1.1 tholo 356: }
357:
358: Cell *jump(Node **a, int n) /* break, continue, next, nextfile, return */
359: {
360: Cell *y;
361:
362: switch (n) {
363: case EXIT:
364: if (a[0] != NULL) {
365: y = execute(a[0]);
1.14 millert 366: errorflag = (int) getfval(y);
1.1 tholo 367: tempfree(y);
368: }
369: longjmp(env, 1);
370: case RETURN:
371: if (a[0] != NULL) {
372: y = execute(a[0]);
373: if ((y->tval & (STR|NUM)) == (STR|NUM)) {
1.57 millert 374: setsval(frp->retval, getsval(y));
375: frp->retval->fval = getfval(y);
376: frp->retval->tval |= NUM;
1.1 tholo 377: }
378: else if (y->tval & STR)
1.57 millert 379: setsval(frp->retval, getsval(y));
1.1 tholo 380: else if (y->tval & NUM)
1.57 millert 381: setfval(frp->retval, getfval(y));
1.1 tholo 382: else /* can't happen */
1.16 millert 383: FATAL("bad type variable %d", y->tval);
1.1 tholo 384: tempfree(y);
385: }
386: return(jret);
387: case NEXT:
388: return(jnext);
389: case NEXTFILE:
390: nextfile();
391: return(jnextfile);
392: case BREAK:
393: return(jbreak);
394: case CONTINUE:
395: return(jcont);
396: default: /* can't happen */
1.16 millert 397: FATAL("illegal jump type %d", n);
1.1 tholo 398: }
399: return 0; /* not reached */
400: }
401:
1.31 millert 402: Cell *awkgetline(Node **a, int n) /* get next line from specific input */
1.1 tholo 403: { /* a[0] is variable, a[1] is operator, a[2] is filename */
404: Cell *r, *x;
1.13 kstailey 405: extern Cell **fldtab;
1.1 tholo 406: FILE *fp;
1.13 kstailey 407: char *buf;
408: int bufsize = recsize;
1.15 millert 409: int mode;
1.57 millert 410: bool newflag;
1.13 kstailey 411:
1.53 millert 412: if ((buf = malloc(bufsize)) == NULL)
1.16 millert 413: FATAL("out of memory in getline");
1.1 tholo 414:
415: fflush(stdout); /* in case someone is waiting for a prompt */
416: r = gettemp();
417: if (a[1] != NULL) { /* getline < file */
418: x = execute(a[2]); /* filename */
1.15 millert 419: mode = ptoi(a[1]);
420: if (mode == '|') /* input pipe */
421: mode = LE; /* arbitrary flag */
1.57 millert 422: fp = openfile(mode, getsval(x), &newflag);
1.1 tholo 423: tempfree(x);
424: if (fp == NULL)
425: n = -1;
426: else
1.57 millert 427: n = readrec(&buf, &bufsize, fp, newflag);
1.1 tholo 428: if (n <= 0) {
429: ;
430: } else if (a[0] != NULL) { /* getline var <file */
1.13 kstailey 431: x = execute(a[0]);
432: setsval(x, buf);
1.49 millert 433: if (is_number(x->sval)) {
434: x->fval = atof(x->sval);
435: x->tval |= NUM;
436: }
1.13 kstailey 437: tempfree(x);
1.1 tholo 438: } else { /* getline <file */
1.13 kstailey 439: setsval(fldtab[0], buf);
1.14 millert 440: if (is_number(fldtab[0]->sval)) {
1.13 kstailey 441: fldtab[0]->fval = atof(fldtab[0]->sval);
442: fldtab[0]->tval |= NUM;
1.1 tholo 443: }
444: }
445: } else { /* bare getline; use current input */
446: if (a[0] == NULL) /* getline */
1.54 millert 447: n = getrec(&record, &recsize, true);
1.1 tholo 448: else { /* getline var */
1.54 millert 449: n = getrec(&buf, &bufsize, false);
1.13 kstailey 450: x = execute(a[0]);
451: setsval(x, buf);
1.49 millert 452: if (is_number(x->sval)) {
453: x->fval = atof(x->sval);
454: x->tval |= NUM;
455: }
1.13 kstailey 456: tempfree(x);
1.1 tholo 457: }
458: }
459: setfval(r, (Awkfloat) n);
1.13 kstailey 460: free(buf);
1.1 tholo 461: return r;
462: }
463:
464: Cell *getnf(Node **a, int n) /* get NF */
465: {
1.54 millert 466: if (!donefld)
1.1 tholo 467: fldbld();
468: return (Cell *) a[0];
469: }
470:
1.53 millert 471: static char *
472: makearraystring(Node *p, const char *func)
1.1 tholo 473: {
1.13 kstailey 474: char *buf;
475: int bufsz = recsize;
1.62 millert 476: size_t blen;
1.53 millert 477:
478: if ((buf = malloc(bufsz)) == NULL) {
479: FATAL("%s: out of memory", func);
480: }
1.13 kstailey 481:
1.53 millert 482: blen = 0;
483: buf[blen] = '\0';
484:
485: for (; p; p = p->nnext) {
486: Cell *x = execute(p); /* expr */
487: char *s = getsval(x);
1.62 millert 488: size_t seplen = strlen(getsval(subseploc));
1.53 millert 489: size_t nsub = p->nnext ? seplen : 0;
490: size_t slen = strlen(s);
491: size_t tlen = blen + slen + nsub;
492:
493: if (!adjbuf(&buf, &bufsz, tlen + 1, recsize, 0, func)) {
494: FATAL("%s: out of memory %s[%s...]",
495: func, x->nval, buf);
496: }
497: memcpy(buf + blen, s, slen);
498: if (nsub) {
499: memcpy(buf + blen + slen, *SUBSEP, nsub);
500: }
501: buf[tlen] = '\0';
502: blen = tlen;
503: tempfree(x);
504: }
505: return buf;
506: }
507:
508: Cell *array(Node **a, int n) /* a[0] is symtab, a[1] is list of subscripts */
509: {
510: Cell *x, *z;
511: char *buf;
1.1 tholo 512:
513: x = execute(a[0]); /* Cell* for symbol table */
1.53 millert 514: buf = makearraystring(a[1], __func__);
1.1 tholo 515: if (!isarr(x)) {
1.60 millert 516: DPRINTF("making %s into an array\n", NN(x->nval));
1.1 tholo 517: if (freeable(x))
518: xfree(x->sval);
519: x->tval &= ~(STR|NUM|DONTFREE);
520: x->tval |= ARR;
521: x->sval = (char *) makesymtab(NSYMTAB);
522: }
523: z = setsymtab(buf, "", 0.0, STR|NUM, (Array *) x->sval);
524: z->ctype = OCELL;
525: z->csub = CVAR;
526: tempfree(x);
1.13 kstailey 527: free(buf);
1.1 tholo 528: return(z);
529: }
530:
1.14 millert 531: Cell *awkdelete(Node **a, int n) /* a[0] is symtab, a[1] is list of subscripts */
1.1 tholo 532: {
1.53 millert 533: Cell *x;
1.1 tholo 534:
535: x = execute(a[0]); /* Cell* for symbol table */
1.50 millert 536: if (x == symtabloc) {
537: FATAL("cannot delete SYMTAB or its elements");
538: }
1.1 tholo 539: if (!isarr(x))
1.15 millert 540: return True;
1.51 millert 541: if (a[1] == NULL) { /* delete the elements, not the table */
1.1 tholo 542: freesymtab(x);
543: x->tval &= ~STR;
544: x->tval |= ARR;
545: x->sval = (char *) makesymtab(NSYMTAB);
546: } else {
1.53 millert 547: char *buf = makearraystring(a[1], __func__);
1.1 tholo 548: freeelem(x, buf);
1.13 kstailey 549: free(buf);
1.1 tholo 550: }
551: tempfree(x);
1.15 millert 552: return True;
1.1 tholo 553: }
554:
555: Cell *intest(Node **a, int n) /* a[0] is index (list), a[1] is symtab */
556: {
1.53 millert 557: Cell *ap, *k;
1.13 kstailey 558: char *buf;
1.1 tholo 559:
560: ap = execute(a[1]); /* array name */
561: if (!isarr(ap)) {
1.60 millert 562: DPRINTF("making %s into an array\n", ap->nval);
1.1 tholo 563: if (freeable(ap))
564: xfree(ap->sval);
565: ap->tval &= ~(STR|NUM|DONTFREE);
566: ap->tval |= ARR;
567: ap->sval = (char *) makesymtab(NSYMTAB);
568: }
1.53 millert 569: buf = makearraystring(a[0], __func__);
1.1 tholo 570: k = lookup(buf, (Array *) ap->sval);
571: tempfree(ap);
1.13 kstailey 572: free(buf);
1.1 tholo 573: if (k == NULL)
1.15 millert 574: return(False);
1.1 tholo 575: else
1.15 millert 576: return(True);
1.1 tholo 577: }
578:
579:
580: Cell *matchop(Node **a, int n) /* ~ and match() */
581: {
582: Cell *x, *y;
583: char *s, *t;
584: int i;
585: fa *pfa;
1.18 millert 586: int (*mf)(fa *, const char *) = match, mode = 0;
1.1 tholo 587:
588: if (n == MATCHFCN) {
589: mf = pmatch;
590: mode = 1;
591: }
592: x = execute(a[1]); /* a[1] = target text */
593: s = getsval(x);
1.51 millert 594: if (a[0] == NULL) /* a[1] == 0: already-compiled reg expr */
1.1 tholo 595: i = (*mf)((fa *) a[2], s);
596: else {
597: y = execute(a[2]); /* a[2] = regular expr */
598: t = getsval(y);
599: pfa = makedfa(t, mode);
600: i = (*mf)(pfa, s);
601: tempfree(y);
602: }
603: tempfree(x);
604: if (n == MATCHFCN) {
605: int start = patbeg - s + 1;
606: if (patlen < 0)
607: start = 0;
608: setfval(rstartloc, (Awkfloat) start);
609: setfval(rlengthloc, (Awkfloat) patlen);
610: x = gettemp();
611: x->tval = NUM;
612: x->fval = start;
613: return x;
614: } else if ((n == MATCH && i == 1) || (n == NOTMATCH && i == 0))
1.15 millert 615: return(True);
1.1 tholo 616: else
1.15 millert 617: return(False);
1.1 tholo 618: }
619:
620:
621: Cell *boolop(Node **a, int n) /* a[0] || a[1], a[0] && a[1], !a[0] */
622: {
623: Cell *x, *y;
624: int i;
625:
626: x = execute(a[0]);
627: i = istrue(x);
628: tempfree(x);
629: switch (n) {
630: case BOR:
1.15 millert 631: if (i) return(True);
1.1 tholo 632: y = execute(a[1]);
633: i = istrue(y);
634: tempfree(y);
1.15 millert 635: if (i) return(True);
636: else return(False);
1.1 tholo 637: case AND:
1.15 millert 638: if ( !i ) return(False);
1.1 tholo 639: y = execute(a[1]);
640: i = istrue(y);
641: tempfree(y);
1.15 millert 642: if (i) return(True);
643: else return(False);
1.1 tholo 644: case NOT:
1.15 millert 645: if (i) return(False);
646: else return(True);
1.1 tholo 647: default: /* can't happen */
1.16 millert 648: FATAL("unknown boolean operator %d", n);
1.1 tholo 649: }
650: return 0; /*NOTREACHED*/
651: }
652:
653: Cell *relop(Node **a, int n) /* a[0 < a[1], etc. */
654: {
655: int i;
656: Cell *x, *y;
657: Awkfloat j;
658:
659: x = execute(a[0]);
660: y = execute(a[1]);
661: if (x->tval&NUM && y->tval&NUM) {
662: j = x->fval - y->fval;
663: i = j<0? -1: (j>0? 1: 0);
664: } else {
665: i = strcmp(getsval(x), getsval(y));
666: }
667: tempfree(x);
668: tempfree(y);
669: switch (n) {
1.15 millert 670: case LT: if (i<0) return(True);
671: else return(False);
672: case LE: if (i<=0) return(True);
673: else return(False);
674: case NE: if (i!=0) return(True);
675: else return(False);
676: case EQ: if (i == 0) return(True);
677: else return(False);
678: case GE: if (i>=0) return(True);
679: else return(False);
680: case GT: if (i>0) return(True);
681: else return(False);
1.1 tholo 682: default: /* can't happen */
1.16 millert 683: FATAL("unknown relational operator %d", n);
1.1 tholo 684: }
685: return 0; /*NOTREACHED*/
686: }
687:
688: void tfree(Cell *a) /* free a tempcell */
689: {
1.13 kstailey 690: if (freeable(a)) {
1.60 millert 691: DPRINTF("freeing %s %s %o\n", NN(a->nval), NN(a->sval), a->tval);
1.1 tholo 692: xfree(a->sval);
1.13 kstailey 693: }
1.1 tholo 694: if (a == tmps)
1.16 millert 695: FATAL("tempcell list is curdled");
1.1 tholo 696: a->cnext = tmps;
697: tmps = a;
698: }
699:
700: Cell *gettemp(void) /* get a tempcell */
701: { int i;
702: Cell *x;
703:
704: if (!tmps) {
1.53 millert 705: tmps = calloc(100, sizeof(*tmps));
1.1 tholo 706: if (!tmps)
1.16 millert 707: FATAL("out of space for temporaries");
1.52 millert 708: for (i = 1; i < 100; i++)
1.1 tholo 709: tmps[i-1].cnext = &tmps[i];
1.51 millert 710: tmps[i-1].cnext = NULL;
1.1 tholo 711: }
712: x = tmps;
713: tmps = x->cnext;
714: *x = tempcell;
715: return(x);
716: }
717:
718: Cell *indirect(Node **a, int n) /* $( a[0] ) */
719: {
1.25 millert 720: Awkfloat val;
1.1 tholo 721: Cell *x;
722: int m;
723: char *s;
724:
725: x = execute(a[0]);
1.25 millert 726: val = getfval(x); /* freebsd: defend against super large field numbers */
727: if ((Awkfloat)INT_MAX < val)
728: FATAL("trying to access out of range field %s", x->nval);
729: m = (int) val;
1.14 millert 730: if (m == 0 && !is_number(s = getsval(x))) /* suspicion! */
1.16 millert 731: FATAL("illegal field $(%s), name \"%s\"", s, x->nval);
1.13 kstailey 732: /* BUG: can x->nval ever be null??? */
1.1 tholo 733: tempfree(x);
734: x = fieldadr(m);
1.13 kstailey 735: x->ctype = OCELL; /* BUG? why are these needed? */
1.1 tholo 736: x->csub = CFLD;
737: return(x);
738: }
739:
740: Cell *substr(Node **a, int nnn) /* substr(a[0], a[1], a[2]) */
741: {
742: int k, m, n;
743: char *s;
744: int temp;
1.51 millert 745: Cell *x, *y, *z = NULL;
1.1 tholo 746:
747: x = execute(a[0]);
748: y = execute(a[1]);
1.51 millert 749: if (a[2] != NULL)
1.1 tholo 750: z = execute(a[2]);
751: s = getsval(x);
752: k = strlen(s) + 1;
753: if (k <= 1) {
754: tempfree(x);
755: tempfree(y);
1.51 millert 756: if (a[2] != NULL) {
1.1 tholo 757: tempfree(z);
1.17 millert 758: }
1.1 tholo 759: x = gettemp();
760: setsval(x, "");
761: return(x);
762: }
1.14 millert 763: m = (int) getfval(y);
1.1 tholo 764: if (m <= 0)
765: m = 1;
766: else if (m > k)
767: m = k;
768: tempfree(y);
1.51 millert 769: if (a[2] != NULL) {
1.14 millert 770: n = (int) getfval(z);
1.1 tholo 771: tempfree(z);
772: } else
773: n = k - 1;
774: if (n < 0)
775: n = 0;
776: else if (n > k - m)
777: n = k - m;
1.60 millert 778: DPRINTF("substr: m=%d, n=%d, s=%s\n", m, n, s);
1.1 tholo 779: y = gettemp();
780: temp = s[n+m-1]; /* with thanks to John Linderman */
781: s[n+m-1] = '\0';
782: setsval(y, s + m - 1);
783: s[n+m-1] = temp;
784: tempfree(x);
785: return(y);
786: }
787:
788: Cell *sindex(Node **a, int nnn) /* index(a[0], a[1]) */
789: {
790: Cell *x, *y, *z;
791: char *s1, *s2, *p1, *p2, *q;
792: Awkfloat v = 0.0;
793:
794: x = execute(a[0]);
795: s1 = getsval(x);
796: y = execute(a[1]);
797: s2 = getsval(y);
798:
799: z = gettemp();
800: for (p1 = s1; *p1 != '\0'; p1++) {
1.57 millert 801: for (q = p1, p2 = s2; *p2 != '\0' && *q == *p2; q++, p2++)
802: continue;
1.1 tholo 803: if (*p2 == '\0') {
804: v = (Awkfloat) (p1 - s1 + 1); /* origin 1 */
805: break;
806: }
807: }
808: tempfree(x);
809: tempfree(y);
810: setfval(z, v);
811: return(z);
812: }
813:
1.13 kstailey 814: #define MAXNUMSIZE 50
815:
1.18 millert 816: int format(char **pbuf, int *pbufsize, const char *s, Node *a) /* printf-like conversions */
1.1 tholo 817: {
1.13 kstailey 818: char *fmt;
1.18 millert 819: char *p, *t;
820: const char *os;
1.1 tholo 821: Cell *x;
1.10 kstailey 822: int flag = 0, n;
1.13 kstailey 823: int fmtwd; /* format width */
824: int fmtsz = recsize;
825: char *buf = *pbuf;
826: int bufsize = *pbufsize;
1.53 millert 827: #define FMTSZ(a) (fmtsz - ((a) - fmt))
828: #define BUFSZ(a) (bufsize - ((a) - buf))
1.1 tholo 829:
1.54 millert 830: static bool first = true;
831: static bool have_a_format = false;
1.47 millert 832:
833: if (first) {
1.57 millert 834: char xbuf[100];
1.47 millert 835:
1.57 millert 836: snprintf(xbuf, sizeof(xbuf), "%a", 42.0);
837: have_a_format = (strcmp(xbuf, "0x1.5p+5") == 0);
1.54 millert 838: first = false;
1.47 millert 839: }
840:
1.1 tholo 841: os = s;
842: p = buf;
1.53 millert 843: if ((fmt = malloc(fmtsz)) == NULL)
1.16 millert 844: FATAL("out of memory in format()");
1.1 tholo 845: while (*s) {
1.30 millert 846: adjbuf(&buf, &bufsize, MAXNUMSIZE+1+p-buf, recsize, &p, "format1");
1.1 tholo 847: if (*s != '%') {
848: *p++ = *s++;
849: continue;
850: }
851: if (*(s+1) == '%') {
852: *p++ = '%';
853: s += 2;
854: continue;
855: }
1.13 kstailey 856: /* have to be real careful in case this is a huge number, eg, %100000d */
857: fmtwd = atoi(s+1);
858: if (fmtwd < 0)
859: fmtwd = -fmtwd;
1.30 millert 860: adjbuf(&buf, &bufsize, fmtwd+1+p-buf, recsize, &p, "format2");
1.13 kstailey 861: for (t = fmt; (*t++ = *s) != '\0'; s++) {
1.30 millert 862: if (!adjbuf(&fmt, &fmtsz, MAXNUMSIZE+1+t-fmt, recsize, &t, "format3"))
1.16 millert 863: FATAL("format item %.30s... ran format() out of memory", os);
1.55 millert 864: /* Ignore size specifiers */
865: if (strchr("hjLlqtz", *s) != NULL) { /* the ansi panoply */
866: t--;
867: continue;
868: }
869: if (isalpha((uschar)*s))
870: break;
1.48 millert 871: if (*s == '$') {
872: FATAL("'$' not permitted in awk formats");
873: }
1.1 tholo 874: if (*s == '*') {
1.49 millert 875: if (a == NULL) {
1.27 deraadt 876: FATAL("not enough args in printf(%s)", os);
1.49 millert 877: }
1.1 tholo 878: x = execute(a);
879: a = a->nnext;
1.53 millert 880: snprintf(t - 1, FMTSZ(t - 1),
881: "%d", fmtwd=(int) getfval(x));
1.13 kstailey 882: if (fmtwd < 0)
883: fmtwd = -fmtwd;
884: adjbuf(&buf, &bufsize, fmtwd+1+p-buf, recsize, &p, "format");
1.1 tholo 885: t = fmt + strlen(fmt);
886: tempfree(x);
887: }
888: }
889: *t = '\0';
1.13 kstailey 890: if (fmtwd < 0)
891: fmtwd = -fmtwd;
1.30 millert 892: adjbuf(&buf, &bufsize, fmtwd+1+p-buf, recsize, &p, "format4");
1.1 tholo 893: switch (*s) {
1.47 millert 894: case 'a': case 'A':
895: if (have_a_format)
896: flag = *s;
897: else
898: flag = 'f';
899: break;
1.1 tholo 900: case 'f': case 'e': case 'g': case 'E': case 'G':
1.18 millert 901: flag = 'f';
1.1 tholo 902: break;
1.55 millert 903: case 'd': case 'i': case 'o': case 'x': case 'X': case 'u':
904: flag = (*s == 'd' || *s == 'i') ? 'd' : 'u';
1.53 millert 905: *(t-1) = 'j';
906: *t = *s;
907: *++t = '\0';
1.1 tholo 908: break;
909: case 's':
1.18 millert 910: flag = 's';
1.1 tholo 911: break;
912: case 'c':
1.18 millert 913: flag = 'c';
1.1 tholo 914: break;
915: default:
1.16 millert 916: WARNING("weird printf conversion %s", fmt);
1.18 millert 917: flag = '?';
1.1 tholo 918: break;
919: }
920: if (a == NULL)
1.16 millert 921: FATAL("not enough args in printf(%s)", os);
1.1 tholo 922: x = execute(a);
923: a = a->nnext;
1.13 kstailey 924: n = MAXNUMSIZE;
925: if (fmtwd > n)
926: n = fmtwd;
1.30 millert 927: adjbuf(&buf, &bufsize, 1+n+p-buf, recsize, &p, "format5");
1.1 tholo 928: switch (flag) {
1.63 ! millert 929: case '?': snprintf(p, BUFSZ(p), "%s", fmt); /* unknown, so dump it too */
1.13 kstailey 930: t = getsval(x);
931: n = strlen(t);
932: if (fmtwd > n)
933: n = fmtwd;
1.30 millert 934: adjbuf(&buf, &bufsize, 1+strlen(p)+n+p-buf, recsize, &p, "format6");
1.5 kstailey 935: p += strlen(p);
1.53 millert 936: snprintf(p, BUFSZ(p), "%s", t);
1.1 tholo 937: break;
1.47 millert 938: case 'a':
939: case 'A':
1.53 millert 940: case 'f': snprintf(p, BUFSZ(p), fmt, getfval(x)); break;
1.55 millert 941: case 'd': snprintf(p, BUFSZ(p), fmt, (intmax_t) getfval(x)); break;
942: case 'u': snprintf(p, BUFSZ(p), fmt, (uintmax_t) getfval(x)); break;
1.18 millert 943: case 's':
1.1 tholo 944: t = getsval(x);
945: n = strlen(t);
1.13 kstailey 946: if (fmtwd > n)
947: n = fmtwd;
1.30 millert 948: if (!adjbuf(&buf, &bufsize, 1+n+p-buf, recsize, &p, "format7"))
1.16 millert 949: FATAL("huge string/format (%d chars) in printf %.30s... ran format() out of memory", n, t);
1.53 millert 950: snprintf(p, BUFSZ(p), fmt, t);
1.1 tholo 951: break;
1.18 millert 952: case 'c':
1.13 kstailey 953: if (isnum(x)) {
1.36 millert 954: if ((int)getfval(x))
1.53 millert 955: snprintf(p, BUFSZ(p), fmt, (int) getfval(x));
1.18 millert 956: else {
957: *p++ = '\0'; /* explicit null byte */
958: *p = '\0'; /* next output will start here */
959: }
1.13 kstailey 960: } else
1.53 millert 961: snprintf(p, BUFSZ(p), fmt, getsval(x)[0]);
1.1 tholo 962: break;
1.18 millert 963: default:
964: FATAL("can't happen: bad conversion %c in format()", flag);
1.1 tholo 965: }
966: tempfree(x);
1.5 kstailey 967: p += strlen(p);
1.1 tholo 968: s++;
969: }
970: *p = '\0';
1.13 kstailey 971: free(fmt);
1.1 tholo 972: for ( ; a; a = a->nnext) /* evaluate any remaining args */
973: execute(a);
1.13 kstailey 974: *pbuf = buf;
975: *pbufsize = bufsize;
976: return p - buf;
1.1 tholo 977: }
978:
979: Cell *awksprintf(Node **a, int n) /* sprintf(a[0]) */
980: {
981: Cell *x;
982: Node *y;
1.13 kstailey 983: char *buf;
984: int bufsz=3*recsize;
1.1 tholo 985:
1.53 millert 986: if ((buf = malloc(bufsz)) == NULL)
1.16 millert 987: FATAL("out of memory in awksprintf");
1.1 tholo 988: y = a[0]->nnext;
989: x = execute(a[0]);
1.13 kstailey 990: if (format(&buf, &bufsz, getsval(x), y) == -1)
1.16 millert 991: FATAL("sprintf string %.30s... too long. can't happen.", buf);
1.1 tholo 992: tempfree(x);
993: x = gettemp();
1.13 kstailey 994: x->sval = buf;
1.1 tholo 995: x->tval = STR;
996: return(x);
997: }
998:
999: Cell *awkprintf(Node **a, int n) /* printf */
1000: { /* a[0] is list of args, starting with format string */
1001: /* a[1] is redirection operator, a[2] is redirection file */
1002: FILE *fp;
1003: Cell *x;
1004: Node *y;
1.13 kstailey 1005: char *buf;
1.9 kstailey 1006: int len;
1.13 kstailey 1007: int bufsz=3*recsize;
1.1 tholo 1008:
1.53 millert 1009: if ((buf = malloc(bufsz)) == NULL)
1.16 millert 1010: FATAL("out of memory in awkprintf");
1.1 tholo 1011: y = a[0]->nnext;
1012: x = execute(a[0]);
1.13 kstailey 1013: if ((len = format(&buf, &bufsz, getsval(x), y)) == -1)
1.16 millert 1014: FATAL("printf string %.30s... too long. can't happen.", buf);
1.1 tholo 1015: tempfree(x);
1016: if (a[1] == NULL) {
1.13 kstailey 1017: /* fputs(buf, stdout); */
1.9 kstailey 1018: fwrite(buf, len, 1, stdout);
1.8 kstailey 1019: if (ferror(stdout))
1.16 millert 1020: FATAL("write error on stdout");
1.1 tholo 1021: } else {
1.15 millert 1022: fp = redirect(ptoi(a[1]), a[2]);
1.13 kstailey 1023: /* fputs(buf, fp); */
1.9 kstailey 1024: fwrite(buf, len, 1, fp);
1.8 kstailey 1025: fflush(fp);
1026: if (ferror(fp))
1.16 millert 1027: FATAL("write error on %s", filename(fp));
1.1 tholo 1028: }
1.13 kstailey 1029: free(buf);
1.15 millert 1030: return(True);
1.1 tholo 1031: }
1032:
1033: Cell *arith(Node **a, int n) /* a[0] + a[1], etc. also -a[0] */
1034: {
1035: Awkfloat i, j = 0;
1036: double v;
1037: Cell *x, *y, *z;
1038:
1039: x = execute(a[0]);
1040: i = getfval(x);
1041: tempfree(x);
1.47 millert 1042: if (n != UMINUS && n != UPLUS) {
1.1 tholo 1043: y = execute(a[1]);
1044: j = getfval(y);
1045: tempfree(y);
1046: }
1047: z = gettemp();
1048: switch (n) {
1049: case ADD:
1050: i += j;
1051: break;
1052: case MINUS:
1053: i -= j;
1054: break;
1055: case MULT:
1056: i *= j;
1057: break;
1058: case DIVIDE:
1059: if (j == 0)
1.16 millert 1060: FATAL("division by zero");
1.1 tholo 1061: i /= j;
1062: break;
1063: case MOD:
1064: if (j == 0)
1.16 millert 1065: FATAL("division by zero in mod");
1.1 tholo 1066: modf(i/j, &v);
1067: i = i - j * v;
1068: break;
1069: case UMINUS:
1070: i = -i;
1071: break;
1.57 millert 1072: case UPLUS: /* handled by getfval(), above */
1.47 millert 1073: break;
1.1 tholo 1074: case POWER:
1075: if (j >= 0 && modf(j, &v) == 0.0) /* pos integer exponent */
1.12 millert 1076: i = ipow(i, (int) j);
1.63 ! millert 1077: else {
1.45 guenther 1078: errno = 0;
1.1 tholo 1079: i = errcheck(pow(i, j), "pow");
1.63 ! millert 1080: }
1.1 tholo 1081: break;
1082: default: /* can't happen */
1.16 millert 1083: FATAL("illegal arithmetic operator %d", n);
1.1 tholo 1084: }
1085: setfval(z, i);
1086: return(z);
1087: }
1088:
1089: double ipow(double x, int n) /* x**n. ought to be done by pow, but isn't always */
1090: {
1091: double v;
1092:
1093: if (n <= 0)
1094: return 1;
1095: v = ipow(x, n/2);
1096: if (n % 2 == 0)
1097: return v * v;
1098: else
1099: return x * v * v;
1100: }
1101:
1102: Cell *incrdecr(Node **a, int n) /* a[0]++, etc. */
1103: {
1104: Cell *x, *z;
1105: int k;
1106: Awkfloat xf;
1107:
1108: x = execute(a[0]);
1109: xf = getfval(x);
1110: k = (n == PREINCR || n == POSTINCR) ? 1 : -1;
1111: if (n == PREINCR || n == PREDECR) {
1112: setfval(x, xf + k);
1113: return(x);
1114: }
1115: z = gettemp();
1116: setfval(z, xf);
1117: setfval(x, xf + k);
1118: tempfree(x);
1119: return(z);
1120: }
1121:
1122: Cell *assign(Node **a, int n) /* a[0] = a[1], a[0] += a[1], etc. */
1123: { /* this is subtle; don't muck with it. */
1124: Cell *x, *y;
1125: Awkfloat xf, yf;
1126: double v;
1127:
1128: y = execute(a[1]);
1129: x = execute(a[0]);
1130: if (n == ASSIGN) { /* ordinary assignment */
1.49 millert 1131: if (x == y && !(x->tval & (FLD|REC)) && x != nfloc)
1132: ; /* self-assignment: leave alone unless it's a field or NF */
1.1 tholo 1133: else if ((y->tval & (STR|NUM)) == (STR|NUM)) {
1134: setsval(x, getsval(y));
1135: x->fval = getfval(y);
1136: x->tval |= NUM;
1137: }
1.13 kstailey 1138: else if (isstr(y))
1.1 tholo 1139: setsval(x, getsval(y));
1.13 kstailey 1140: else if (isnum(y))
1.1 tholo 1141: setfval(x, getfval(y));
1142: else
1143: funnyvar(y, "read value of");
1144: tempfree(y);
1145: return(x);
1146: }
1147: xf = getfval(x);
1148: yf = getfval(y);
1149: switch (n) {
1150: case ADDEQ:
1151: xf += yf;
1152: break;
1153: case SUBEQ:
1154: xf -= yf;
1155: break;
1156: case MULTEQ:
1157: xf *= yf;
1158: break;
1159: case DIVEQ:
1160: if (yf == 0)
1.16 millert 1161: FATAL("division by zero in /=");
1.1 tholo 1162: xf /= yf;
1163: break;
1164: case MODEQ:
1165: if (yf == 0)
1.16 millert 1166: FATAL("division by zero in %%=");
1.1 tholo 1167: modf(xf/yf, &v);
1168: xf = xf - yf * v;
1169: break;
1170: case POWEQ:
1171: if (yf >= 0 && modf(yf, &v) == 0.0) /* pos integer exponent */
1.12 millert 1172: xf = ipow(xf, (int) yf);
1.63 ! millert 1173: else {
1.45 guenther 1174: errno = 0;
1.1 tholo 1175: xf = errcheck(pow(xf, yf), "pow");
1.63 ! millert 1176: }
1.1 tholo 1177: break;
1178: default:
1.16 millert 1179: FATAL("illegal assignment operator %d", n);
1.1 tholo 1180: break;
1181: }
1182: tempfree(y);
1183: setfval(x, xf);
1184: return(x);
1185: }
1186:
1187: Cell *cat(Node **a, int q) /* a[0] cat a[1] */
1188: {
1189: Cell *x, *y, *z;
1190: int n1, n2;
1.49 millert 1191: char *s = NULL;
1192: int ssz = 0;
1.1 tholo 1193:
1194: x = execute(a[0]);
1.49 millert 1195: n1 = strlen(getsval(x));
1.61 millert 1196: adjbuf(&s, &ssz, n1, recsize, 0, "cat1");
1197: memcpy(s, x->sval, n1);
1.49 millert 1198:
1.1 tholo 1199: y = execute(a[1]);
1.49 millert 1200: n2 = strlen(getsval(y));
1.61 millert 1201: adjbuf(&s, &ssz, n1 + n2 + 1, recsize, 0, "cat2");
1.53 millert 1202: memcpy(s + n1, y->sval, n2);
1203: s[n1 + n2] = '\0';
1.49 millert 1204:
1.31 millert 1205: tempfree(x);
1.1 tholo 1206: tempfree(y);
1.49 millert 1207:
1.1 tholo 1208: z = gettemp();
1209: z->sval = s;
1210: z->tval = STR;
1.49 millert 1211:
1.1 tholo 1212: return(z);
1213: }
1214:
1215: Cell *pastat(Node **a, int n) /* a[0] { a[1] } */
1216: {
1217: Cell *x;
1218:
1.51 millert 1219: if (a[0] == NULL)
1.1 tholo 1220: x = execute(a[1]);
1221: else {
1222: x = execute(a[0]);
1223: if (istrue(x)) {
1224: tempfree(x);
1225: x = execute(a[1]);
1226: }
1227: }
1228: return x;
1229: }
1230:
1231: Cell *dopa2(Node **a, int n) /* a[0], a[1] { a[2] } */
1232: {
1233: Cell *x;
1234: int pair;
1235:
1.15 millert 1236: pair = ptoi(a[3]);
1.1 tholo 1237: if (pairstack[pair] == 0) {
1238: x = execute(a[0]);
1239: if (istrue(x))
1240: pairstack[pair] = 1;
1241: tempfree(x);
1242: }
1243: if (pairstack[pair] == 1) {
1244: x = execute(a[1]);
1245: if (istrue(x))
1246: pairstack[pair] = 0;
1247: tempfree(x);
1248: x = execute(a[2]);
1249: return(x);
1250: }
1.15 millert 1251: return(False);
1.1 tholo 1252: }
1253:
1254: Cell *split(Node **a, int nnn) /* split(a[0], a[1], a[2]); a[3] is type */
1255: {
1.51 millert 1256: Cell *x = NULL, *y, *ap;
1.53 millert 1257: const char *s, *origs, *t;
1.56 millert 1258: const char *fs = NULL;
1259: char *origfs = NULL;
1.1 tholo 1260: int sep;
1.53 millert 1261: char temp, num[50];
1.15 millert 1262: int n, tempstat, arg3type;
1.1 tholo 1263:
1264: y = execute(a[0]); /* source string */
1.43 fcambus 1265: origs = s = strdup(getsval(y));
1.44 fcambus 1266: if (s == NULL)
1267: FATAL("out of space in split");
1.15 millert 1268: arg3type = ptoi(a[3]);
1.51 millert 1269: if (a[2] == NULL) /* fs string */
1.49 millert 1270: fs = getsval(fsloc);
1.15 millert 1271: else if (arg3type == STRING) { /* split(str,arr,"string") */
1.1 tholo 1272: x = execute(a[2]);
1.56 millert 1273: fs = origfs = strdup(getsval(x));
1.49 millert 1274: if (fs == NULL)
1275: FATAL("out of space in split");
1276: tempfree(x);
1.15 millert 1277: } else if (arg3type == REGEXPR)
1.13 kstailey 1278: fs = "(regexpr)"; /* split(str,arr,/regexpr/) */
1.1 tholo 1279: else
1.16 millert 1280: FATAL("illegal type of split");
1.1 tholo 1281: sep = *fs;
1282: ap = execute(a[1]); /* array name */
1283: freesymtab(ap);
1.60 millert 1284: DPRINTF("split: s=|%s|, a=%s, sep=|%s|\n", s, NN(ap->nval), fs);
1.1 tholo 1285: ap->tval &= ~STR;
1286: ap->tval |= ARR;
1287: ap->sval = (char *) makesymtab(NSYMTAB);
1288:
1289: n = 0;
1.33 millert 1290: if (arg3type == REGEXPR && strlen((char*)((fa*)a[2])->restr) == 0) {
1291: /* split(s, a, //); have to arrange that it looks like empty sep */
1292: arg3type = 0;
1293: fs = "";
1294: sep = 0;
1295: }
1.25 millert 1296: if (*s != '\0' && (strlen(fs) > 1 || arg3type == REGEXPR)) { /* reg expr */
1.1 tholo 1297: fa *pfa;
1.15 millert 1298: if (arg3type == REGEXPR) { /* it's ready already */
1.1 tholo 1299: pfa = (fa *) a[2];
1300: } else {
1301: pfa = makedfa(fs, 1);
1302: }
1303: if (nematch(pfa,s)) {
1304: tempstat = pfa->initstat;
1305: pfa->initstat = 2;
1306: do {
1307: n++;
1.53 millert 1308: snprintf(num, sizeof(num), "%d", n);
1.1 tholo 1309: temp = *patbeg;
1.53 millert 1310: setptr(patbeg, '\0');
1.14 millert 1311: if (is_number(s))
1.13 kstailey 1312: setsymtab(num, s, atof(s), STR|NUM, (Array *) ap->sval);
1.1 tholo 1313: else
1314: setsymtab(num, s, 0.0, STR, (Array *) ap->sval);
1.53 millert 1315: setptr(patbeg, temp);
1.1 tholo 1316: s = patbeg + patlen;
1.57 millert 1317: if (*(patbeg+patlen-1) == '\0' || *s == '\0') {
1.1 tholo 1318: n++;
1.53 millert 1319: snprintf(num, sizeof(num), "%d", n);
1.1 tholo 1320: setsymtab(num, "", 0.0, STR, (Array *) ap->sval);
1321: pfa->initstat = tempstat;
1322: goto spdone;
1323: }
1324: } while (nematch(pfa,s));
1.25 millert 1325: pfa->initstat = tempstat; /* bwk: has to be here to reset */
1326: /* cf gsub and refldbld */
1.1 tholo 1327: }
1328: n++;
1.53 millert 1329: snprintf(num, sizeof(num), "%d", n);
1.14 millert 1330: if (is_number(s))
1.13 kstailey 1331: setsymtab(num, s, atof(s), STR|NUM, (Array *) ap->sval);
1.1 tholo 1332: else
1333: setsymtab(num, s, 0.0, STR, (Array *) ap->sval);
1334: spdone:
1335: pfa = NULL;
1336: } else if (sep == ' ') {
1337: for (n = 0; ; ) {
1.57 millert 1338: #define ISWS(c) ((c) == ' ' || (c) == '\t' || (c) == '\n')
1339: while (ISWS(*s))
1.1 tholo 1340: s++;
1.57 millert 1341: if (*s == '\0')
1.1 tholo 1342: break;
1343: n++;
1344: t = s;
1345: do
1346: s++;
1.57 millert 1347: while (*s != '\0' && !ISWS(*s));
1.1 tholo 1348: temp = *s;
1.53 millert 1349: setptr(s, '\0');
1350: snprintf(num, sizeof(num), "%d", n);
1.14 millert 1351: if (is_number(t))
1.13 kstailey 1352: setsymtab(num, t, atof(t), STR|NUM, (Array *) ap->sval);
1.1 tholo 1353: else
1354: setsymtab(num, t, 0.0, STR, (Array *) ap->sval);
1.53 millert 1355: setptr(s, temp);
1.57 millert 1356: if (*s != '\0')
1.1 tholo 1357: s++;
1358: }
1359: } else if (sep == 0) { /* new: split(s, a, "") => 1 char/elem */
1.57 millert 1360: for (n = 0; *s != '\0'; s++) {
1.1 tholo 1361: char buf[2];
1362: n++;
1.53 millert 1363: snprintf(num, sizeof(num), "%d", n);
1.1 tholo 1364: buf[0] = *s;
1.57 millert 1365: buf[1] = '\0';
1.17 millert 1366: if (isdigit((uschar)buf[0]))
1.1 tholo 1367: setsymtab(num, buf, atof(buf), STR|NUM, (Array *) ap->sval);
1368: else
1369: setsymtab(num, buf, 0.0, STR, (Array *) ap->sval);
1370: }
1.57 millert 1371: } else if (*s != '\0') {
1.1 tholo 1372: for (;;) {
1373: n++;
1374: t = s;
1375: while (*s != sep && *s != '\n' && *s != '\0')
1376: s++;
1377: temp = *s;
1.53 millert 1378: setptr(s, '\0');
1379: snprintf(num, sizeof(num), "%d", n);
1.14 millert 1380: if (is_number(t))
1.13 kstailey 1381: setsymtab(num, t, atof(t), STR|NUM, (Array *) ap->sval);
1.1 tholo 1382: else
1383: setsymtab(num, t, 0.0, STR, (Array *) ap->sval);
1.53 millert 1384: setptr(s, temp);
1.57 millert 1385: if (*s++ == '\0')
1.1 tholo 1386: break;
1387: }
1388: }
1389: tempfree(ap);
1390: tempfree(y);
1.53 millert 1391: xfree(origs);
1392: xfree(origfs);
1.1 tholo 1393: x = gettemp();
1394: x->tval = NUM;
1395: x->fval = n;
1396: return(x);
1397: }
1398:
1399: Cell *condexpr(Node **a, int n) /* a[0] ? a[1] : a[2] */
1400: {
1401: Cell *x;
1402:
1403: x = execute(a[0]);
1404: if (istrue(x)) {
1405: tempfree(x);
1406: x = execute(a[1]);
1407: } else {
1408: tempfree(x);
1409: x = execute(a[2]);
1410: }
1411: return(x);
1412: }
1413:
1414: Cell *ifstat(Node **a, int n) /* if (a[0]) a[1]; else a[2] */
1415: {
1416: Cell *x;
1417:
1418: x = execute(a[0]);
1419: if (istrue(x)) {
1420: tempfree(x);
1421: x = execute(a[1]);
1.51 millert 1422: } else if (a[2] != NULL) {
1.1 tholo 1423: tempfree(x);
1424: x = execute(a[2]);
1425: }
1426: return(x);
1427: }
1428:
1429: Cell *whilestat(Node **a, int n) /* while (a[0]) a[1] */
1430: {
1431: Cell *x;
1432:
1433: for (;;) {
1434: x = execute(a[0]);
1435: if (!istrue(x))
1436: return(x);
1437: tempfree(x);
1438: x = execute(a[1]);
1439: if (isbreak(x)) {
1.15 millert 1440: x = True;
1.1 tholo 1441: return(x);
1442: }
1443: if (isnext(x) || isexit(x) || isret(x))
1444: return(x);
1445: tempfree(x);
1446: }
1447: }
1448:
1449: Cell *dostat(Node **a, int n) /* do a[0]; while(a[1]) */
1450: {
1451: Cell *x;
1452:
1453: for (;;) {
1454: x = execute(a[0]);
1455: if (isbreak(x))
1.15 millert 1456: return True;
1.17 millert 1457: if (isnext(x) || isexit(x) || isret(x))
1.1 tholo 1458: return(x);
1459: tempfree(x);
1460: x = execute(a[1]);
1461: if (!istrue(x))
1462: return(x);
1463: tempfree(x);
1464: }
1465: }
1466:
1467: Cell *forstat(Node **a, int n) /* for (a[0]; a[1]; a[2]) a[3] */
1468: {
1469: Cell *x;
1470:
1471: x = execute(a[0]);
1472: tempfree(x);
1473: for (;;) {
1.51 millert 1474: if (a[1]!=NULL) {
1.1 tholo 1475: x = execute(a[1]);
1476: if (!istrue(x)) return(x);
1477: else tempfree(x);
1478: }
1479: x = execute(a[3]);
1480: if (isbreak(x)) /* turn off break */
1.15 millert 1481: return True;
1.1 tholo 1482: if (isnext(x) || isexit(x) || isret(x))
1483: return(x);
1484: tempfree(x);
1485: x = execute(a[2]);
1486: tempfree(x);
1487: }
1488: }
1489:
1490: Cell *instat(Node **a, int n) /* for (a[0] in a[1]) a[2] */
1491: {
1492: Cell *x, *vp, *arrayp, *cp, *ncp;
1493: Array *tp;
1494: int i;
1495:
1496: vp = execute(a[0]);
1497: arrayp = execute(a[1]);
1498: if (!isarr(arrayp)) {
1.15 millert 1499: return True;
1.1 tholo 1500: }
1501: tp = (Array *) arrayp->sval;
1502: tempfree(arrayp);
1503: for (i = 0; i < tp->size; i++) { /* this routine knows too much */
1504: for (cp = tp->tab[i]; cp != NULL; cp = ncp) {
1505: setsval(vp, cp->nval);
1506: ncp = cp->cnext;
1507: x = execute(a[2]);
1508: if (isbreak(x)) {
1509: tempfree(vp);
1.15 millert 1510: return True;
1.1 tholo 1511: }
1512: if (isnext(x) || isexit(x) || isret(x)) {
1513: tempfree(vp);
1514: return(x);
1515: }
1516: tempfree(x);
1517: }
1518: }
1.15 millert 1519: return True;
1.1 tholo 1520: }
1521:
1.57 millert 1522: static char *nawk_convert(const char *s, int (*fun_c)(int),
1523: wint_t (*fun_wc)(wint_t))
1524: {
1525: char *buf = NULL;
1526: char *pbuf = NULL;
1527: const char *ps = NULL;
1528: size_t n = 0;
1529: mbstate_t mbs, mbs2;
1530: wchar_t wc;
1531: size_t sz = MB_CUR_MAX;
1532:
1533: if (sz == 1) {
1534: buf = tostring(s);
1535:
1536: for (pbuf = buf; *pbuf; pbuf++)
1537: *pbuf = fun_c((uschar)*pbuf);
1538:
1539: return buf;
1540: } else {
1541: /* upper/lower character may be shorter/longer */
1542: buf = tostringN(s, strlen(s) * sz + 1);
1543:
1544: memset(&mbs, 0, sizeof(mbs));
1545: memset(&mbs2, 0, sizeof(mbs2));
1546:
1547: ps = s;
1548: pbuf = buf;
1549: while (n = mbrtowc(&wc, ps, sz, &mbs),
1550: n > 0 && n != (size_t)-1 && n != (size_t)-2)
1551: {
1552: ps += n;
1553:
1554: n = wcrtomb(pbuf, fun_wc(wc), &mbs2);
1555: if (n == (size_t)-1)
1556: FATAL("illegal wide character %s", s);
1557:
1558: pbuf += n;
1559: }
1560:
1561: *pbuf = '\0';
1562:
1563: if (n)
1564: FATAL("illegal byte sequence %s", s);
1565:
1566: return buf;
1567: }
1568: }
1569:
1570: static char *nawk_toupper(const char *s)
1571: {
1572: return nawk_convert(s, toupper, towupper);
1573: }
1574:
1575: static char *nawk_tolower(const char *s)
1576: {
1577: return nawk_convert(s, tolower, towlower);
1578: }
1579:
1.1 tholo 1580: Cell *bltin(Node **a, int n) /* builtin functions. a[0] is type, a[1] is arg list */
1581: {
1582: Cell *x, *y;
1583: Awkfloat u;
1.59 millert 1584: int t, sz;
1.33 millert 1585: Awkfloat tmp;
1.59 millert 1586: char *buf, *fmt;
1.1 tholo 1587: Node *nextarg;
1588: FILE *fp;
1.47 millert 1589: int status = 0;
1.59 millert 1590: time_t tv;
1591: struct tm *tm;
1.1 tholo 1592:
1.15 millert 1593: t = ptoi(a[0]);
1.1 tholo 1594: x = execute(a[1]);
1595: nextarg = a[1]->nnext;
1596: switch (t) {
1597: case FLENGTH:
1.18 millert 1598: if (isarr(x))
1599: u = ((Array *) x->sval)->nelem; /* GROT. should be function*/
1600: else
1601: u = strlen(getsval(x));
1602: break;
1.1 tholo 1603: case FLOG:
1.45 guenther 1604: errno = 0;
1.63 ! millert 1605: u = errcheck(log(getfval(x)), "log");
! 1606: break;
1.1 tholo 1607: case FINT:
1608: modf(getfval(x), &u); break;
1609: case FEXP:
1.45 guenther 1610: errno = 0;
1.63 ! millert 1611: u = errcheck(exp(getfval(x)), "exp");
! 1612: break;
1.1 tholo 1613: case FSQRT:
1.45 guenther 1614: errno = 0;
1.63 ! millert 1615: u = errcheck(sqrt(getfval(x)), "sqrt");
! 1616: break;
1.1 tholo 1617: case FSIN:
1618: u = sin(getfval(x)); break;
1619: case FCOS:
1620: u = cos(getfval(x)); break;
1621: case FATAN:
1.51 millert 1622: if (nextarg == NULL) {
1.16 millert 1623: WARNING("atan2 requires two arguments; returning 1.0");
1.1 tholo 1624: u = 1.0;
1625: } else {
1626: y = execute(a[1]->nnext);
1627: u = atan2(getfval(x), getfval(y));
1628: tempfree(y);
1629: nextarg = nextarg->nnext;
1630: }
1.29 pyr 1631: break;
1632: case FCOMPL:
1633: u = ~((int)getfval(x));
1634: break;
1635: case FAND:
1636: if (nextarg == 0) {
1637: WARNING("and requires two arguments; returning 0");
1638: u = 0;
1639: break;
1640: }
1641: y = execute(a[1]->nnext);
1642: u = ((int)getfval(x)) & ((int)getfval(y));
1643: tempfree(y);
1644: nextarg = nextarg->nnext;
1645: break;
1646: case FFOR:
1647: if (nextarg == 0) {
1648: WARNING("or requires two arguments; returning 0");
1649: u = 0;
1650: break;
1651: }
1652: y = execute(a[1]->nnext);
1653: u = ((int)getfval(x)) | ((int)getfval(y));
1654: tempfree(y);
1655: nextarg = nextarg->nnext;
1656: break;
1657: case FXOR:
1658: if (nextarg == 0) {
1.41 ajacouto 1659: WARNING("xor requires two arguments; returning 0");
1.29 pyr 1660: u = 0;
1661: break;
1662: }
1663: y = execute(a[1]->nnext);
1664: u = ((int)getfval(x)) ^ ((int)getfval(y));
1665: tempfree(y);
1666: nextarg = nextarg->nnext;
1667: break;
1668: case FLSHIFT:
1669: if (nextarg == 0) {
1.41 ajacouto 1670: WARNING("lshift requires two arguments; returning 0");
1.29 pyr 1671: u = 0;
1672: break;
1673: }
1674: y = execute(a[1]->nnext);
1675: u = ((int)getfval(x)) << ((int)getfval(y));
1676: tempfree(y);
1677: nextarg = nextarg->nnext;
1678: break;
1679: case FRSHIFT:
1680: if (nextarg == 0) {
1.41 ajacouto 1681: WARNING("rshift requires two arguments; returning 0");
1.29 pyr 1682: u = 0;
1683: break;
1684: }
1685: y = execute(a[1]->nnext);
1686: u = ((int)getfval(x)) >> ((int)getfval(y));
1687: tempfree(y);
1688: nextarg = nextarg->nnext;
1.1 tholo 1689: break;
1690: case FSYSTEM:
1691: fflush(stdout); /* in case something is buffered already */
1.47 millert 1692: status = system(getsval(x));
1693: u = status;
1694: if (status != -1) {
1695: if (WIFEXITED(status)) {
1696: u = WEXITSTATUS(status);
1697: } else if (WIFSIGNALED(status)) {
1698: u = WTERMSIG(status) + 256;
1699: #ifdef WCOREDUMP
1700: if (WCOREDUMP(status))
1701: u += 256;
1702: #endif
1703: } else /* something else?!? */
1704: u = 0;
1705: }
1.1 tholo 1706: break;
1707: case FRAND:
1.51 millert 1708: /* random() returns numbers in [0..2^31-1]
1709: * in order to get a number in [0, 1), divide it by 2^31
1710: */
1711: u = (Awkfloat) random() / (0x7fffffffL + 0x1UL);
1.1 tholo 1712: break;
1713: case FSRAND:
1.39 deraadt 1714: if (isrec(x)) { /* no argument provided */
1715: u = time(NULL);
1716: tmp = u;
1717: srandom((unsigned int) u);
1718: } else {
1.33 millert 1719: u = getfval(x);
1720: tmp = u;
1.37 deraadt 1721: srandom_deterministic((unsigned int) u);
1.24 millert 1722: }
1.39 deraadt 1723: u = srand_seed;
1724: srand_seed = tmp;
1.1 tholo 1725: break;
1726: case FTOUPPER:
1727: case FTOLOWER:
1.57 millert 1728: if (t == FTOUPPER)
1729: buf = nawk_toupper(getsval(x));
1730: else
1731: buf = nawk_tolower(getsval(x));
1.1 tholo 1732: tempfree(x);
1733: x = gettemp();
1734: setsval(x, buf);
1.13 kstailey 1735: free(buf);
1.1 tholo 1736: return x;
1737: case FFLUSH:
1.18 millert 1738: if (isrec(x) || strlen(getsval(x)) == 0) {
1739: flush_all(); /* fflush() or fflush("") -> all */
1740: u = 0;
1.57 millert 1741: } else if ((fp = openfile(FFLUSH, getsval(x), NULL)) == NULL)
1.1 tholo 1742: u = EOF;
1743: else
1744: u = fflush(fp);
1745: break;
1.59 millert 1746: case FSYSTIME:
1747: u = time((time_t *) 0);
1748: break;
1749: case FSTRFTIME:
1750: /* strftime([format [,timestamp]]) */
1751: if (nextarg) {
1752: y = execute(nextarg);
1753: nextarg = nextarg->nnext;
1754: tv = (time_t) getfval(y);
1755: tempfree(y);
1756: } else
1757: tv = time((time_t *) 0);
1758: tm = localtime(&tv);
1759: if (tm == NULL)
1760: FATAL("bad time %ld", (long)tv);
1761:
1762: if (isrec(x)) {
1763: /* format argument not provided, use default */
1764: fmt = tostring("%a %b %d %H:%M:%S %Z %Y");
1765: } else
1766: fmt = tostring(getsval(x));
1767:
1768: sz = 32;
1769: buf = NULL;
1770: do {
1771: if ((buf = reallocarray(buf, 2, sz)) == NULL)
1772: FATAL("out of memory in strftime");
1773: sz *= 2;
1774: } while (strftime(buf, sz, fmt, tm) == 0 && fmt[0] != '\0');
1775:
1776: y = gettemp();
1777: setsval(y, buf);
1778: free(fmt);
1779: free(buf);
1780:
1781: return y;
1.1 tholo 1782: default: /* can't happen */
1.16 millert 1783: FATAL("illegal function type %d", t);
1.1 tholo 1784: break;
1785: }
1786: tempfree(x);
1787: x = gettemp();
1788: setfval(x, u);
1.51 millert 1789: if (nextarg != NULL) {
1.16 millert 1790: WARNING("warning: function has too many arguments");
1.1 tholo 1791: for ( ; nextarg; nextarg = nextarg->nnext)
1792: execute(nextarg);
1793: }
1794: return(x);
1795: }
1796:
1797: Cell *printstat(Node **a, int n) /* print a[0] */
1798: {
1799: Node *x;
1800: Cell *y;
1801: FILE *fp;
1802:
1.51 millert 1803: if (a[1] == NULL) /* a[1] is redirection operator, a[2] is file */
1.1 tholo 1804: fp = stdout;
1805: else
1.15 millert 1806: fp = redirect(ptoi(a[1]), a[2]);
1.1 tholo 1807: for (x = a[0]; x != NULL; x = x->nnext) {
1808: y = execute(x);
1.18 millert 1809: fputs(getpssval(y), fp);
1.1 tholo 1810: tempfree(y);
1811: if (x->nnext == NULL)
1.49 millert 1812: fputs(getsval(orsloc), fp);
1.1 tholo 1813: else
1.49 millert 1814: fputs(getsval(ofsloc), fp);
1.1 tholo 1815: }
1.51 millert 1816: if (a[1] != NULL)
1.1 tholo 1817: fflush(fp);
1818: if (ferror(fp))
1.16 millert 1819: FATAL("write error on %s", filename(fp));
1.15 millert 1820: return(True);
1.1 tholo 1821: }
1822:
1823: Cell *nullproc(Node **a, int n)
1824: {
1825: return 0;
1826: }
1827:
1828:
1829: FILE *redirect(int a, Node *b) /* set up all i/o redirections */
1830: {
1831: FILE *fp;
1832: Cell *x;
1833: char *fname;
1834:
1835: x = execute(b);
1836: fname = getsval(x);
1.57 millert 1837: fp = openfile(a, fname, NULL);
1.1 tholo 1838: if (fp == NULL)
1.16 millert 1839: FATAL("can't open file %s", fname);
1.1 tholo 1840: tempfree(x);
1841: return fp;
1842: }
1843:
1844: struct files {
1845: FILE *fp;
1.18 millert 1846: const char *fname;
1.1 tholo 1847: int mode; /* '|', 'a', 'w' => LE/LT, GT */
1.33 millert 1848: } *files;
1849:
1.57 millert 1850: size_t nfiles;
1.1 tholo 1851:
1.57 millert 1852: static void stdinit(void) /* in case stdin, etc., are not constants */
1.16 millert 1853: {
1.33 millert 1854: nfiles = FOPEN_MAX;
1855: files = calloc(nfiles, sizeof(*files));
1856: if (files == NULL)
1.57 millert 1857: FATAL("can't allocate file memory for %zu files", nfiles);
1.33 millert 1858: files[0].fp = stdin;
1859: files[0].fname = "/dev/stdin";
1860: files[0].mode = LT;
1861: files[1].fp = stdout;
1862: files[1].fname = "/dev/stdout";
1863: files[1].mode = GT;
1864: files[2].fp = stderr;
1865: files[2].fname = "/dev/stderr";
1866: files[2].mode = GT;
1.16 millert 1867: }
1868:
1.57 millert 1869: FILE *openfile(int a, const char *us, bool *pnewflag)
1.1 tholo 1870: {
1.18 millert 1871: const char *s = us;
1.57 millert 1872: size_t i;
1873: int m;
1.51 millert 1874: FILE *fp = NULL;
1.1 tholo 1875:
1876: if (*s == '\0')
1.16 millert 1877: FATAL("null file name in print or getline");
1.57 millert 1878: for (i = 0; i < nfiles; i++)
1879: if (files[i].fname && strcmp(s, files[i].fname) == 0 &&
1880: (a == files[i].mode || (a==APPEND && files[i].mode==GT) ||
1881: a == FFLUSH)) {
1882: if (pnewflag)
1883: *pnewflag = false;
1884: return files[i].fp;
1.13 kstailey 1885: }
1886: if (a == FFLUSH) /* didn't find it, so don't create it! */
1887: return NULL;
1888:
1.57 millert 1889: for (i = 0; i < nfiles; i++)
1.51 millert 1890: if (files[i].fp == NULL)
1.1 tholo 1891: break;
1.33 millert 1892: if (i >= nfiles) {
1893: struct files *nf;
1.57 millert 1894: size_t nnf = nfiles + FOPEN_MAX;
1.35 doug 1895: nf = reallocarray(files, nnf, sizeof(*nf));
1.33 millert 1896: if (nf == NULL)
1.57 millert 1897: FATAL("cannot grow files for %s and %zu files", s, nnf);
1.33 millert 1898: memset(&nf[nfiles], 0, FOPEN_MAX * sizeof(*nf));
1899: nfiles = nnf;
1900: files = nf;
1901: }
1.1 tholo 1902: fflush(stdout); /* force a semblance of order */
1903: m = a;
1904: if (a == GT) {
1905: fp = fopen(s, "w");
1906: } else if (a == APPEND) {
1907: fp = fopen(s, "a");
1908: m = GT; /* so can mix > and >> */
1909: } else if (a == '|') { /* output pipe */
1910: fp = popen(s, "w");
1911: } else if (a == LE) { /* input pipe */
1912: fp = popen(s, "r");
1913: } else if (a == LT) { /* getline <file */
1914: fp = strcmp(s, "-") == 0 ? stdin : fopen(s, "r"); /* "-" is stdin */
1915: } else /* can't happen */
1.16 millert 1916: FATAL("illegal redirection %d", a);
1.1 tholo 1917: if (fp != NULL) {
1918: files[i].fname = tostring(s);
1919: files[i].fp = fp;
1920: files[i].mode = m;
1.57 millert 1921: if (pnewflag)
1922: *pnewflag = true;
1.56 millert 1923: if (fp != stdin && fp != stdout && fp != stderr)
1924: (void) fcntl(fileno(fp), F_SETFD, FD_CLOEXEC);
1.1 tholo 1925: }
1926: return fp;
1927: }
1928:
1.18 millert 1929: const char *filename(FILE *fp)
1.1 tholo 1930: {
1.57 millert 1931: size_t i;
1.1 tholo 1932:
1.33 millert 1933: for (i = 0; i < nfiles; i++)
1.1 tholo 1934: if (fp == files[i].fp)
1935: return files[i].fname;
1936: return "???";
1937: }
1938:
1.57 millert 1939: Cell *closefile(Node **a, int n)
1940: {
1941: Cell *x;
1942: size_t i;
1943: bool stat;
1944:
1945: x = execute(a[0]);
1946: getsval(x);
1947: stat = true;
1948: for (i = 0; i < nfiles; i++) {
1949: if (!files[i].fname || strcmp(x->sval, files[i].fname) != 0)
1950: continue;
1951: if (ferror(files[i].fp))
1952: FATAL("i/o error occurred on %s", files[i].fname);
1953: if (files[i].mode == '|' || files[i].mode == LE)
1954: stat = pclose(files[i].fp) == -1;
1955: else
1956: stat = fclose(files[i].fp) == EOF;
1957: if (stat)
1958: FATAL("i/o error occurred closing %s", files[i].fname);
1959: if (i > 2) /* don't do /dev/std... */
1960: xfree(files[i].fname);
1961: files[i].fname = NULL; /* watch out for ref thru this */
1962: files[i].fp = NULL;
1963: }
1964: tempfree(x);
1965: x = gettemp();
1966: setfval(x, (Awkfloat) (stat ? -1 : 0));
1967: return(x);
1968: }
1.1 tholo 1969:
1970: void closeall(void)
1971: {
1.57 millert 1972: size_t i;
1973: bool stat = false;
1.1 tholo 1974:
1.57 millert 1975: for (i = 0; i < nfiles; i++) {
1976: if (! files[i].fp)
1977: continue;
1978: if (ferror(files[i].fp))
1979: FATAL( "i/o error occurred on %s", files[i].fname );
1980: if (files[i].mode == '|' || files[i].mode == LE)
1981: stat = pclose(files[i].fp) == -1;
1982: else
1983: stat = fclose(files[i].fp) == EOF;
1984: if (stat)
1985: FATAL( "i/o error occurred while closing %s", files[i].fname );
1.17 millert 1986: }
1.18 millert 1987: }
1988:
1.57 millert 1989: static void flush_all(void)
1.18 millert 1990: {
1.57 millert 1991: size_t i;
1.18 millert 1992:
1.33 millert 1993: for (i = 0; i < nfiles; i++)
1.18 millert 1994: if (files[i].fp)
1995: fflush(files[i].fp);
1.1 tholo 1996: }
1997:
1.53 millert 1998: void backsub(char **pb_ptr, const char **sptr_ptr);
1.1 tholo 1999:
2000: Cell *sub(Node **a, int nnn) /* substitute command */
2001: {
1.53 millert 2002: const char *sptr, *q;
1.1 tholo 2003: Cell *x, *y, *result;
1.53 millert 2004: char *t, *buf, *pb;
1.1 tholo 2005: fa *pfa;
1.13 kstailey 2006: int bufsz = recsize;
1.1 tholo 2007:
1.53 millert 2008: if ((buf = malloc(bufsz)) == NULL)
1.16 millert 2009: FATAL("out of memory in sub");
1.1 tholo 2010: x = execute(a[3]); /* target string */
2011: t = getsval(x);
1.51 millert 2012: if (a[0] == NULL) /* 0 => a[1] is already-compiled regexpr */
1.1 tholo 2013: pfa = (fa *) a[1]; /* regular expression */
2014: else {
2015: y = execute(a[1]);
2016: pfa = makedfa(getsval(y), 1);
2017: tempfree(y);
2018: }
2019: y = execute(a[2]); /* replacement string */
1.15 millert 2020: result = False;
1.1 tholo 2021: if (pmatch(pfa, t)) {
1.13 kstailey 2022: sptr = t;
2023: adjbuf(&buf, &bufsz, 1+patbeg-sptr, recsize, 0, "sub");
1.1 tholo 2024: pb = buf;
2025: while (sptr < patbeg)
2026: *pb++ = *sptr++;
2027: sptr = getsval(y);
1.57 millert 2028: while (*sptr != '\0') {
1.13 kstailey 2029: adjbuf(&buf, &bufsz, 5+pb-buf, recsize, &pb, "sub");
2030: if (*sptr == '\\') {
2031: backsub(&pb, &sptr);
1.1 tholo 2032: } else if (*sptr == '&') {
2033: sptr++;
1.13 kstailey 2034: adjbuf(&buf, &bufsz, 1+patlen+pb-buf, recsize, &pb, "sub");
1.1 tholo 2035: for (q = patbeg; q < patbeg+patlen; )
2036: *pb++ = *q++;
2037: } else
2038: *pb++ = *sptr++;
1.13 kstailey 2039: }
1.1 tholo 2040: *pb = '\0';
1.13 kstailey 2041: if (pb > buf + bufsz)
1.16 millert 2042: FATAL("sub result1 %.30s too big; can't happen", buf);
1.1 tholo 2043: sptr = patbeg + patlen;
1.13 kstailey 2044: if ((patlen == 0 && *patbeg) || (patlen && *(sptr-1))) {
2045: adjbuf(&buf, &bufsz, 1+strlen(sptr)+pb-buf, 0, &pb, "sub");
1.57 millert 2046: while ((*pb++ = *sptr++) != '\0')
2047: continue;
1.13 kstailey 2048: }
2049: if (pb > buf + bufsz)
1.16 millert 2050: FATAL("sub result2 %.30s too big; can't happen", buf);
1.13 kstailey 2051: setsval(x, buf); /* BUG: should be able to avoid copy */
1.23 millert 2052: result = True;
1.1 tholo 2053: }
2054: tempfree(x);
2055: tempfree(y);
1.13 kstailey 2056: free(buf);
1.1 tholo 2057: return result;
2058: }
2059:
2060: Cell *gsub(Node **a, int nnn) /* global substitute */
2061: {
2062: Cell *x, *y;
1.53 millert 2063: char *rptr, *pb;
2064: const char *q, *t, *sptr;
1.13 kstailey 2065: char *buf;
1.1 tholo 2066: fa *pfa;
2067: int mflag, tempstat, num;
1.13 kstailey 2068: int bufsz = recsize;
1.1 tholo 2069:
1.53 millert 2070: if ((buf = malloc(bufsz)) == NULL)
1.16 millert 2071: FATAL("out of memory in gsub");
1.1 tholo 2072: mflag = 0; /* if mflag == 0, can replace empty string */
2073: num = 0;
2074: x = execute(a[3]); /* target string */
2075: t = getsval(x);
1.51 millert 2076: if (a[0] == NULL) /* 0 => a[1] is already-compiled regexpr */
1.1 tholo 2077: pfa = (fa *) a[1]; /* regular expression */
2078: else {
2079: y = execute(a[1]);
2080: pfa = makedfa(getsval(y), 1);
2081: tempfree(y);
2082: }
2083: y = execute(a[2]); /* replacement string */
2084: if (pmatch(pfa, t)) {
2085: tempstat = pfa->initstat;
2086: pfa->initstat = 2;
2087: pb = buf;
2088: rptr = getsval(y);
2089: do {
1.57 millert 2090: if (patlen == 0 && *patbeg != '\0') { /* matched empty string */
1.1 tholo 2091: if (mflag == 0) { /* can replace empty */
2092: num++;
2093: sptr = rptr;
1.57 millert 2094: while (*sptr != '\0') {
1.13 kstailey 2095: adjbuf(&buf, &bufsz, 5+pb-buf, recsize, &pb, "gsub");
2096: if (*sptr == '\\') {
2097: backsub(&pb, &sptr);
1.1 tholo 2098: } else if (*sptr == '&') {
2099: sptr++;
1.13 kstailey 2100: adjbuf(&buf, &bufsz, 1+patlen+pb-buf, recsize, &pb, "gsub");
1.1 tholo 2101: for (q = patbeg; q < patbeg+patlen; )
2102: *pb++ = *q++;
2103: } else
2104: *pb++ = *sptr++;
1.13 kstailey 2105: }
1.1 tholo 2106: }
1.57 millert 2107: if (*t == '\0') /* at end */
1.1 tholo 2108: goto done;
1.13 kstailey 2109: adjbuf(&buf, &bufsz, 2+pb-buf, recsize, &pb, "gsub");
1.1 tholo 2110: *pb++ = *t++;
1.13 kstailey 2111: if (pb > buf + bufsz) /* BUG: not sure of this test */
1.16 millert 2112: FATAL("gsub result0 %.30s too big; can't happen", buf);
1.1 tholo 2113: mflag = 0;
2114: }
2115: else { /* matched nonempty string */
2116: num++;
2117: sptr = t;
1.13 kstailey 2118: adjbuf(&buf, &bufsz, 1+(patbeg-sptr)+pb-buf, recsize, &pb, "gsub");
2119: while (sptr < patbeg)
1.1 tholo 2120: *pb++ = *sptr++;
2121: sptr = rptr;
1.57 millert 2122: while (*sptr != '\0') {
1.13 kstailey 2123: adjbuf(&buf, &bufsz, 5+pb-buf, recsize, &pb, "gsub");
2124: if (*sptr == '\\') {
2125: backsub(&pb, &sptr);
1.1 tholo 2126: } else if (*sptr == '&') {
2127: sptr++;
1.13 kstailey 2128: adjbuf(&buf, &bufsz, 1+patlen+pb-buf, recsize, &pb, "gsub");
1.1 tholo 2129: for (q = patbeg; q < patbeg+patlen; )
2130: *pb++ = *q++;
2131: } else
2132: *pb++ = *sptr++;
1.13 kstailey 2133: }
1.1 tholo 2134: t = patbeg + patlen;
1.57 millert 2135: if (patlen == 0 || *t == '\0' || *(t-1) == '\0')
1.1 tholo 2136: goto done;
1.13 kstailey 2137: if (pb > buf + bufsz)
1.16 millert 2138: FATAL("gsub result1 %.30s too big; can't happen", buf);
1.1 tholo 2139: mflag = 1;
2140: }
2141: } while (pmatch(pfa,t));
2142: sptr = t;
1.13 kstailey 2143: adjbuf(&buf, &bufsz, 1+strlen(sptr)+pb-buf, 0, &pb, "gsub");
1.57 millert 2144: while ((*pb++ = *sptr++) != '\0')
2145: continue;
1.31 millert 2146: done: if (pb < buf + bufsz)
2147: *pb = '\0';
2148: else if (*(pb-1) != '\0')
2149: FATAL("gsub result2 %.30s truncated; can't happen", buf);
1.13 kstailey 2150: setsval(x, buf); /* BUG: should be able to avoid copy + free */
1.1 tholo 2151: pfa->initstat = tempstat;
2152: }
2153: tempfree(x);
2154: tempfree(y);
2155: x = gettemp();
2156: x->tval = NUM;
2157: x->fval = num;
1.13 kstailey 2158: free(buf);
1.1 tholo 2159: return(x);
1.59 millert 2160: }
2161:
2162: Cell *gensub(Node **a, int nnn) /* global selective substitute */
2163: /* XXX incomplete - doesn't support backreferences \0 ... \9 */
2164: {
2165: Cell *x, *y, *res, *h;
2166: char *rptr;
2167: const char *sptr;
2168: char *buf, *pb;
2169: const char *t, *q;
2170: fa *pfa;
2171: int mflag, tempstat, num, whichm;
2172: int bufsz = recsize;
2173:
2174: if ((buf = malloc(bufsz)) == NULL)
2175: FATAL("out of memory in gensub");
2176: mflag = 0; /* if mflag == 0, can replace empty string */
2177: num = 0;
2178: x = execute(a[4]); /* source string */
2179: t = getsval(x);
2180: res = copycell(x); /* target string - initially copy of source */
2181: res->csub = CTEMP; /* result values are temporary */
2182: if (a[0] == 0) /* 0 => a[1] is already-compiled regexpr */
2183: pfa = (fa *) a[1]; /* regular expression */
2184: else {
2185: y = execute(a[1]);
2186: pfa = makedfa(getsval(y), 1);
2187: tempfree(y);
2188: }
2189: y = execute(a[2]); /* replacement string */
2190: h = execute(a[3]); /* which matches should be replaced */
2191: sptr = getsval(h);
2192: if (sptr[0] == 'g' || sptr[0] == 'G')
2193: whichm = -1;
2194: else {
2195: /*
2196: * The specified number is index of replacement, starting
2197: * from 1. GNU awk treats index lower than 0 same as
2198: * 1, we do same for compatibility.
2199: */
2200: whichm = (int) getfval(h) - 1;
2201: if (whichm < 0)
2202: whichm = 0;
2203: }
2204: tempfree(h);
2205:
2206: if (pmatch(pfa, t)) {
2207: char *sl;
2208:
2209: tempstat = pfa->initstat;
2210: pfa->initstat = 2;
2211: pb = buf;
2212: rptr = getsval(y);
2213: /*
2214: * XXX if there are any backreferences in subst string,
2215: * complain now.
2216: */
2217: for (sl = rptr; (sl = strchr(sl, '\\')) && sl[1]; sl++) {
2218: if (strchr("0123456789", sl[1])) {
2219: FATAL("gensub doesn't support backreferences (subst \"%s\")", rptr);
2220: }
2221: }
2222:
2223: do {
2224: if (whichm >= 0 && whichm != num) {
2225: num++;
2226: adjbuf(&buf, &bufsz, (pb - buf) + (patbeg - t) + patlen, recsize, &pb, "gensub");
2227:
2228: /* copy the part of string up to and including
2229: * match to output buffer */
2230: while (t < patbeg + patlen)
2231: *pb++ = *t++;
2232: continue;
2233: }
2234:
2235: if (patlen == 0 && *patbeg != 0) { /* matched empty string */
2236: if (mflag == 0) { /* can replace empty */
2237: num++;
2238: sptr = rptr;
2239: while (*sptr != 0) {
2240: adjbuf(&buf, &bufsz, 5+pb-buf, recsize, &pb, "gensub");
2241: if (*sptr == '\\') {
2242: backsub(&pb, &sptr);
2243: } else if (*sptr == '&') {
2244: sptr++;
2245: adjbuf(&buf, &bufsz, 1+patlen+pb-buf, recsize, &pb, "gensub");
2246: for (q = patbeg; q < patbeg+patlen; )
2247: *pb++ = *q++;
2248: } else
2249: *pb++ = *sptr++;
2250: }
2251: }
2252: if (*t == 0) /* at end */
2253: goto done;
2254: adjbuf(&buf, &bufsz, 2+pb-buf, recsize, &pb, "gensub");
2255: *pb++ = *t++;
2256: if (pb > buf + bufsz) /* BUG: not sure of this test */
2257: FATAL("gensub result0 %.30s too big; can't happen", buf);
2258: mflag = 0;
2259: }
2260: else { /* matched nonempty string */
2261: num++;
2262: sptr = t;
2263: adjbuf(&buf, &bufsz, 1+(patbeg-sptr)+pb-buf, recsize, &pb, "gensub");
2264: while (sptr < patbeg)
2265: *pb++ = *sptr++;
2266: sptr = rptr;
2267: while (*sptr != 0) {
2268: adjbuf(&buf, &bufsz, 5+pb-buf, recsize, &pb, "gensub");
2269: if (*sptr == '\\') {
2270: backsub(&pb, &sptr);
2271: } else if (*sptr == '&') {
2272: sptr++;
2273: adjbuf(&buf, &bufsz, 1+patlen+pb-buf, recsize, &pb, "gensub");
2274: for (q = patbeg; q < patbeg+patlen; )
2275: *pb++ = *q++;
2276: } else
2277: *pb++ = *sptr++;
2278: }
2279: t = patbeg + patlen;
2280: if (patlen == 0 || *t == 0 || *(t-1) == 0)
2281: goto done;
2282: if (pb > buf + bufsz)
2283: FATAL("gensub result1 %.30s too big; can't happen", buf);
2284: mflag = 1;
2285: }
2286: } while (pmatch(pfa,t));
2287: sptr = t;
2288: adjbuf(&buf, &bufsz, 1+strlen(sptr)+pb-buf, 0, &pb, "gensub");
2289: while ((*pb++ = *sptr++) != 0)
2290: ;
2291: done: if (pb > buf + bufsz)
2292: FATAL("gensub result2 %.30s too big; can't happen", buf);
2293: *pb = '\0';
2294: setsval(res, buf);
2295: pfa->initstat = tempstat;
2296: }
2297: tempfree(x);
2298: tempfree(y);
2299: free(buf);
2300: return(res);
1.13 kstailey 2301: }
2302:
1.53 millert 2303: void backsub(char **pb_ptr, const char **sptr_ptr) /* handle \\& variations */
1.13 kstailey 2304: { /* sptr[0] == '\\' */
1.53 millert 2305: char *pb = *pb_ptr;
2306: const char *sptr = *sptr_ptr;
1.13 kstailey 2307:
2308: if (sptr[1] == '\\') {
2309: if (sptr[2] == '\\' && sptr[3] == '&') { /* \\\& -> \& */
2310: *pb++ = '\\';
2311: *pb++ = '&';
2312: sptr += 4;
2313: } else if (sptr[2] == '&') { /* \\& -> \ + matched */
2314: *pb++ = '\\';
2315: sptr += 2;
1.56 millert 2316: } else if (do_posix) { /* \\x -> \x */
2317: sptr++;
2318: *pb++ = *sptr++;
1.13 kstailey 2319: } else { /* \\x -> \\x */
2320: *pb++ = *sptr++;
2321: *pb++ = *sptr++;
2322: }
2323: } else if (sptr[1] == '&') { /* literal & */
2324: sptr++;
2325: *pb++ = *sptr++;
2326: } else /* literal \ */
2327: *pb++ = *sptr++;
2328:
2329: *pb_ptr = pb;
2330: *sptr_ptr = sptr;
1.1 tholo 2331: }