Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1 | /**************************************************************** |
| 2 | Copyright (C) Lucent Technologies 1997 |
| 3 | All Rights Reserved |
| 4 | |
| 5 | Permission to use, copy, modify, and distribute this software and |
| 6 | its documentation for any purpose and without fee is hereby |
| 7 | granted, provided that the above copyright notice appear in all |
| 8 | copies and that both that the copyright notice and this |
| 9 | permission notice and warranty disclaimer appear in supporting |
| 10 | documentation, and that the name Lucent Technologies or any of |
| 11 | its entities not be used in advertising or publicity pertaining |
| 12 | to distribution of the software without specific, written prior |
| 13 | permission. |
| 14 | |
| 15 | LUCENT DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, |
| 16 | INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS. |
| 17 | IN NO EVENT SHALL LUCENT OR ANY OF ITS ENTITIES BE LIABLE FOR ANY |
| 18 | SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES |
| 19 | WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER |
| 20 | IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, |
| 21 | ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF |
| 22 | THIS SOFTWARE. |
| 23 | ****************************************************************/ |
| 24 | |
| 25 | #define DEBUG |
| 26 | #include <stdio.h> |
| 27 | #include <ctype.h> |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 28 | #include <wchar.h> |
| 29 | #include <wctype.h> |
Arnold D. Robbins | 5a18f63 | 2020-01-22 02:10:59 -0700 | [diff] [blame] | 30 | #include <fcntl.h> |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 31 | #include <setjmp.h> |
| 32 | #include <limits.h> |
| 33 | #include <math.h> |
| 34 | #include <string.h> |
| 35 | #include <stdlib.h> |
| 36 | #include <time.h> |
Arnold D. Robbins | 32093f5 | 2018-08-22 20:40:26 +0300 | [diff] [blame] | 37 | #include <sys/types.h> |
| 38 | #include <sys/wait.h> |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 39 | #include "awk.h" |
| 40 | #include "ytab.h" |
| 41 | |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 42 | static void stdinit(void); |
| 43 | static void flush_all(void); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 44 | |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 45 | #if 1 |
| 46 | #define tempfree(x) do { if (istemp(x)) tfree(x); } while (/*CONSTCOND*/0) |
| 47 | #else |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 48 | void tempfree(Cell *p) { |
| 49 | if (p->ctype == OCELL && (p->csub < CUNK || p->csub > CFREE)) { |
| 50 | WARNING("bad csub %d in Cell %d %s", |
| 51 | p->csub, p->ctype, p->sval); |
| 52 | } |
| 53 | if (istemp(p)) |
| 54 | tfree(p); |
| 55 | } |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 56 | #endif |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 57 | |
| 58 | /* do we really need these? */ |
| 59 | /* #ifdef _NFILE */ |
| 60 | /* #ifndef FOPEN_MAX */ |
| 61 | /* #define FOPEN_MAX _NFILE */ |
| 62 | /* #endif */ |
| 63 | /* #endif */ |
| 64 | /* */ |
| 65 | /* #ifndef FOPEN_MAX */ |
| 66 | /* #define FOPEN_MAX 40 */ /* max number of open files */ |
| 67 | /* #endif */ |
| 68 | /* */ |
| 69 | /* #ifndef RAND_MAX */ |
| 70 | /* #define RAND_MAX 32767 */ /* all that ansi guarantees */ |
| 71 | /* #endif */ |
| 72 | |
| 73 | jmp_buf env; |
| 74 | extern int pairstack[]; |
| 75 | extern Awkfloat srand_seed; |
| 76 | |
| 77 | Node *winner = NULL; /* root of parse tree */ |
| 78 | Cell *tmps; /* free temporary cells for execution */ |
| 79 | |
Brian Kernighan | 3ed9e24 | 2018-08-15 10:45:03 -0400 | [diff] [blame] | 80 | static Cell truecell ={ OBOOL, BTRUE, 0, 0, 1.0, NUM, NULL }; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 81 | Cell *True = &truecell; |
Brian Kernighan | 3ed9e24 | 2018-08-15 10:45:03 -0400 | [diff] [blame] | 82 | static Cell falsecell ={ OBOOL, BFALSE, 0, 0, 0.0, NUM, NULL }; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 83 | Cell *False = &falsecell; |
Brian Kernighan | 3ed9e24 | 2018-08-15 10:45:03 -0400 | [diff] [blame] | 84 | static Cell breakcell ={ OJUMP, JBREAK, 0, 0, 0.0, NUM, NULL }; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 85 | Cell *jbreak = &breakcell; |
Brian Kernighan | 3ed9e24 | 2018-08-15 10:45:03 -0400 | [diff] [blame] | 86 | static Cell contcell ={ OJUMP, JCONT, 0, 0, 0.0, NUM, NULL }; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 87 | Cell *jcont = &contcell; |
Brian Kernighan | 3ed9e24 | 2018-08-15 10:45:03 -0400 | [diff] [blame] | 88 | static Cell nextcell ={ OJUMP, JNEXT, 0, 0, 0.0, NUM, NULL }; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 89 | Cell *jnext = &nextcell; |
Brian Kernighan | 3ed9e24 | 2018-08-15 10:45:03 -0400 | [diff] [blame] | 90 | static Cell nextfilecell ={ OJUMP, JNEXTFILE, 0, 0, 0.0, NUM, NULL }; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 91 | Cell *jnextfile = &nextfilecell; |
Brian Kernighan | 3ed9e24 | 2018-08-15 10:45:03 -0400 | [diff] [blame] | 92 | static Cell exitcell ={ OJUMP, JEXIT, 0, 0, 0.0, NUM, NULL }; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 93 | Cell *jexit = &exitcell; |
Brian Kernighan | 3ed9e24 | 2018-08-15 10:45:03 -0400 | [diff] [blame] | 94 | static Cell retcell ={ OJUMP, JRET, 0, 0, 0.0, NUM, NULL }; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 95 | Cell *jret = &retcell; |
zoulasc | 6a87709 | 2020-01-24 04:11:59 -0500 | [diff] [blame] | 96 | static Cell tempcell ={ OCELL, CTEMP, 0, EMPTY, 0.0, NUM|STR|DONTFREE, NULL }; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 97 | |
| 98 | Node *curnode = NULL; /* the node being executed, for debugging */ |
| 99 | |
| 100 | /* buffer memory management */ |
| 101 | int adjbuf(char **pbuf, int *psiz, int minlen, int quantum, char **pbptr, |
| 102 | const char *whatrtn) |
| 103 | /* pbuf: address of pointer to buffer being managed |
| 104 | * psiz: address of buffer size variable |
| 105 | * minlen: minimum length of buffer needed |
| 106 | * quantum: buffer size quantum |
| 107 | * pbptr: address of movable pointer into buffer, or 0 if none |
| 108 | * whatrtn: name of the calling routine if failure should cause fatal error |
| 109 | * |
| 110 | * return 0 for realloc failure, !=0 for success |
| 111 | */ |
| 112 | { |
| 113 | if (minlen > *psiz) { |
| 114 | char *tbuf; |
| 115 | int rminlen = quantum ? minlen % quantum : 0; |
| 116 | int boff = pbptr ? *pbptr - *pbuf : 0; |
| 117 | /* round up to next multiple of quantum */ |
| 118 | if (rminlen) |
| 119 | minlen += quantum - rminlen; |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 120 | tbuf = realloc(*pbuf, minlen); |
| 121 | dprintf( ("adjbuf %s: %d %d (pbuf=%p, tbuf=%p)\n", whatrtn, *psiz, minlen, *pbuf, tbuf) ); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 122 | if (tbuf == NULL) { |
| 123 | if (whatrtn) |
| 124 | FATAL("out of memory in %s", whatrtn); |
| 125 | return 0; |
| 126 | } |
| 127 | *pbuf = tbuf; |
| 128 | *psiz = minlen; |
| 129 | if (pbptr) |
| 130 | *pbptr = tbuf + boff; |
| 131 | } |
| 132 | return 1; |
| 133 | } |
| 134 | |
| 135 | void run(Node *a) /* execution of parse tree starts here */ |
| 136 | { |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 137 | |
| 138 | stdinit(); |
| 139 | execute(a); |
| 140 | closeall(); |
| 141 | } |
| 142 | |
| 143 | Cell *execute(Node *u) /* execute a node of the parse tree */ |
| 144 | { |
| 145 | Cell *(*proc)(Node **, int); |
| 146 | Cell *x; |
| 147 | Node *a; |
| 148 | |
| 149 | if (u == NULL) |
| 150 | return(True); |
| 151 | for (a = u; ; a = a->nnext) { |
| 152 | curnode = a; |
| 153 | if (isvalue(a)) { |
| 154 | x = (Cell *) (a->narg[0]); |
| 155 | if (isfld(x) && !donefld) |
| 156 | fldbld(); |
| 157 | else if (isrec(x) && !donerec) |
| 158 | recbld(); |
| 159 | return(x); |
| 160 | } |
| 161 | if (notlegal(a->nobj)) /* probably a Cell* but too risky to print */ |
| 162 | FATAL("illegal statement"); |
| 163 | proc = proctab[a->nobj-FIRSTTOKEN]; |
| 164 | x = (*proc)(a->narg, a->nobj); |
| 165 | if (isfld(x) && !donefld) |
| 166 | fldbld(); |
| 167 | else if (isrec(x) && !donerec) |
| 168 | recbld(); |
| 169 | if (isexpr(a)) |
| 170 | return(x); |
| 171 | if (isjump(x)) |
| 172 | return(x); |
| 173 | if (a->nnext == NULL) |
| 174 | return(x); |
| 175 | tempfree(x); |
| 176 | } |
| 177 | } |
| 178 | |
| 179 | |
| 180 | Cell *program(Node **a, int n) /* execute an awk program */ |
| 181 | { /* a[0] = BEGIN, a[1] = body, a[2] = END */ |
| 182 | Cell *x; |
| 183 | |
| 184 | if (setjmp(env) != 0) |
| 185 | goto ex; |
| 186 | if (a[0]) { /* BEGIN */ |
| 187 | x = execute(a[0]); |
| 188 | if (isexit(x)) |
| 189 | return(True); |
| 190 | if (isjump(x)) |
| 191 | FATAL("illegal break, continue, next or nextfile from BEGIN"); |
| 192 | tempfree(x); |
| 193 | } |
| 194 | if (a[1] || a[2]) |
Arnold D. Robbins | 108224b | 2019-11-10 21:19:18 +0200 | [diff] [blame] | 195 | while (getrec(&record, &recsize, true) > 0) { |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 196 | x = execute(a[1]); |
| 197 | if (isexit(x)) |
| 198 | break; |
| 199 | tempfree(x); |
| 200 | } |
| 201 | ex: |
| 202 | if (setjmp(env) != 0) /* handles exit within END */ |
| 203 | goto ex1; |
| 204 | if (a[2]) { /* END */ |
| 205 | x = execute(a[2]); |
| 206 | if (isbreak(x) || isnext(x) || iscont(x)) |
| 207 | FATAL("illegal break, continue, next or nextfile from END"); |
| 208 | tempfree(x); |
| 209 | } |
| 210 | ex1: |
| 211 | return(True); |
| 212 | } |
| 213 | |
| 214 | struct Frame { /* stack frame for awk function calls */ |
| 215 | int nargs; /* number of arguments in this call */ |
| 216 | Cell *fcncell; /* pointer to Cell for function */ |
| 217 | Cell **args; /* pointer to array of arguments after execute */ |
| 218 | Cell *retval; /* return value */ |
| 219 | }; |
| 220 | |
| 221 | #define NARGS 50 /* max args in a call */ |
| 222 | |
| 223 | struct Frame *frame = NULL; /* base of stack frames; dynamically allocated */ |
| 224 | int nframe = 0; /* number of frames allocated */ |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 225 | struct Frame *frp = NULL; /* frame pointer. bottom level unused */ |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 226 | |
| 227 | Cell *call(Node **a, int n) /* function call. very kludgy and fragile */ |
| 228 | { |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 229 | static const Cell newcopycell = { OCELL, CCOPY, 0, EMPTY, 0.0, NUM|STR|DONTFREE, NULL }; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 230 | int i, ncall, ndef; |
| 231 | int freed = 0; /* handles potential double freeing when fcn & param share a tempcell */ |
| 232 | Node *x; |
| 233 | Cell *args[NARGS], *oargs[NARGS]; /* BUG: fixed size arrays */ |
| 234 | Cell *y, *z, *fcn; |
| 235 | char *s; |
| 236 | |
| 237 | fcn = execute(a[0]); /* the function itself */ |
| 238 | s = fcn->nval; |
| 239 | if (!isfcn(fcn)) |
| 240 | FATAL("calling undefined function %s", s); |
| 241 | if (frame == NULL) { |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 242 | frp = frame = calloc(nframe += 100, sizeof(*frame)); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 243 | if (frame == NULL) |
| 244 | FATAL("out of space for stack frames calling %s", s); |
| 245 | } |
| 246 | for (ncall = 0, x = a[1]; x != NULL; x = x->nnext) /* args in call */ |
| 247 | ncall++; |
| 248 | ndef = (int) fcn->fval; /* args in defn */ |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 249 | dprintf( ("calling %s, %d args (%d in defn), frp=%d\n", s, ncall, ndef, (int) (frp-frame)) ); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 250 | if (ncall > ndef) |
| 251 | WARNING("function %s called with %d args, uses only %d", |
| 252 | s, ncall, ndef); |
| 253 | if (ncall + ndef > NARGS) |
| 254 | FATAL("function %s has %d arguments, limit %d", s, ncall+ndef, NARGS); |
| 255 | for (i = 0, x = a[1]; x != NULL; i++, x = x->nnext) { /* get call args */ |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 256 | dprintf( ("evaluate args[%d], frp=%d:\n", i, (int) (frp-frame)) ); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 257 | y = execute(x); |
| 258 | oargs[i] = y; |
| 259 | dprintf( ("args[%d]: %s %f <%s>, t=%o\n", |
| 260 | i, NN(y->nval), y->fval, isarr(y) ? "(array)" : NN(y->sval), y->tval) ); |
| 261 | if (isfcn(y)) |
| 262 | FATAL("can't use function %s as argument in %s", y->nval, s); |
| 263 | if (isarr(y)) |
| 264 | args[i] = y; /* arrays by ref */ |
| 265 | else |
| 266 | args[i] = copycell(y); |
| 267 | tempfree(y); |
| 268 | } |
| 269 | for ( ; i < ndef; i++) { /* add null args for ones not provided */ |
| 270 | args[i] = gettemp(); |
| 271 | *args[i] = newcopycell; |
| 272 | } |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 273 | frp++; /* now ok to up frame */ |
| 274 | if (frp >= frame + nframe) { |
| 275 | int dfp = frp - frame; /* old index */ |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 276 | frame = realloc(frame, (nframe += 100) * sizeof(*frame)); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 277 | if (frame == NULL) |
| 278 | FATAL("out of space for stack frames in %s", s); |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 279 | frp = frame + dfp; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 280 | } |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 281 | frp->fcncell = fcn; |
| 282 | frp->args = args; |
| 283 | frp->nargs = ndef; /* number defined with (excess are locals) */ |
| 284 | frp->retval = gettemp(); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 285 | |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 286 | dprintf( ("start exec of %s, frp=%d\n", s, (int) (frp-frame)) ); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 287 | y = execute((Node *)(fcn->sval)); /* execute body */ |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 288 | dprintf( ("finished exec of %s, frp=%d\n", s, (int) (frp-frame)) ); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 289 | |
| 290 | for (i = 0; i < ndef; i++) { |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 291 | Cell *t = frp->args[i]; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 292 | if (isarr(t)) { |
| 293 | if (t->csub == CCOPY) { |
| 294 | if (i >= ncall) { |
| 295 | freesymtab(t); |
| 296 | t->csub = CTEMP; |
| 297 | tempfree(t); |
| 298 | } else { |
| 299 | oargs[i]->tval = t->tval; |
| 300 | oargs[i]->tval &= ~(STR|NUM|DONTFREE); |
| 301 | oargs[i]->sval = t->sval; |
| 302 | tempfree(t); |
| 303 | } |
| 304 | } |
| 305 | } else if (t != y) { /* kludge to prevent freeing twice */ |
| 306 | t->csub = CTEMP; |
| 307 | tempfree(t); |
| 308 | } else if (t == y && t->csub == CCOPY) { |
| 309 | t->csub = CTEMP; |
| 310 | tempfree(t); |
| 311 | freed = 1; |
| 312 | } |
| 313 | } |
| 314 | tempfree(fcn); |
| 315 | if (isexit(y) || isnext(y)) |
| 316 | return y; |
| 317 | if (freed == 0) { |
| 318 | tempfree(y); /* don't free twice! */ |
| 319 | } |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 320 | z = frp->retval; /* return value */ |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 321 | dprintf( ("%s returns %g |%s| %o\n", s, getfval(z), getsval(z), z->tval) ); |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 322 | frp--; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 323 | return(z); |
| 324 | } |
| 325 | |
| 326 | Cell *copycell(Cell *x) /* make a copy of a cell in a temp */ |
| 327 | { |
| 328 | Cell *y; |
| 329 | |
Arnold D. Robbins | 32093f5 | 2018-08-22 20:40:26 +0300 | [diff] [blame] | 330 | /* copy is not constant or field */ |
| 331 | |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 332 | y = gettemp(); |
Arnold D. Robbins | 32093f5 | 2018-08-22 20:40:26 +0300 | [diff] [blame] | 333 | y->tval = x->tval & ~(CON|FLD|REC); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 334 | y->csub = CCOPY; /* prevents freeing until call is over */ |
| 335 | y->nval = x->nval; /* BUG? */ |
Arnold D. Robbins | 32093f5 | 2018-08-22 20:40:26 +0300 | [diff] [blame] | 336 | if (isstr(x) /* || x->ctype == OCELL */) { |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 337 | y->sval = tostring(x->sval); |
Arnold D. Robbins | 32093f5 | 2018-08-22 20:40:26 +0300 | [diff] [blame] | 338 | y->tval &= ~DONTFREE; |
| 339 | } else |
| 340 | y->tval |= DONTFREE; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 341 | y->fval = x->fval; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 342 | return y; |
| 343 | } |
| 344 | |
| 345 | Cell *arg(Node **a, int n) /* nth argument of a function */ |
| 346 | { |
| 347 | |
| 348 | n = ptoi(a[0]); /* argument number, counting from 0 */ |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 349 | dprintf( ("arg(%d), frp->nargs=%d\n", n, frp->nargs) ); |
| 350 | if (n+1 > frp->nargs) |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 351 | FATAL("argument #%d of function %s was not supplied", |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 352 | n+1, frp->fcncell->nval); |
| 353 | return frp->args[n]; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 354 | } |
| 355 | |
| 356 | Cell *jump(Node **a, int n) /* break, continue, next, nextfile, return */ |
| 357 | { |
| 358 | Cell *y; |
| 359 | |
| 360 | switch (n) { |
| 361 | case EXIT: |
| 362 | if (a[0] != NULL) { |
| 363 | y = execute(a[0]); |
| 364 | errorflag = (int) getfval(y); |
| 365 | tempfree(y); |
| 366 | } |
| 367 | longjmp(env, 1); |
| 368 | case RETURN: |
| 369 | if (a[0] != NULL) { |
| 370 | y = execute(a[0]); |
| 371 | if ((y->tval & (STR|NUM)) == (STR|NUM)) { |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 372 | setsval(frp->retval, getsval(y)); |
| 373 | frp->retval->fval = getfval(y); |
| 374 | frp->retval->tval |= NUM; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 375 | } |
| 376 | else if (y->tval & STR) |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 377 | setsval(frp->retval, getsval(y)); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 378 | else if (y->tval & NUM) |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 379 | setfval(frp->retval, getfval(y)); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 380 | else /* can't happen */ |
| 381 | FATAL("bad type variable %d", y->tval); |
| 382 | tempfree(y); |
| 383 | } |
| 384 | return(jret); |
| 385 | case NEXT: |
| 386 | return(jnext); |
| 387 | case NEXTFILE: |
| 388 | nextfile(); |
| 389 | return(jnextfile); |
| 390 | case BREAK: |
| 391 | return(jbreak); |
| 392 | case CONTINUE: |
| 393 | return(jcont); |
| 394 | default: /* can't happen */ |
| 395 | FATAL("illegal jump type %d", n); |
| 396 | } |
| 397 | return 0; /* not reached */ |
| 398 | } |
| 399 | |
| 400 | Cell *awkgetline(Node **a, int n) /* get next line from specific input */ |
| 401 | { /* a[0] is variable, a[1] is operator, a[2] is filename */ |
| 402 | Cell *r, *x; |
| 403 | extern Cell **fldtab; |
| 404 | FILE *fp; |
| 405 | char *buf; |
| 406 | int bufsize = recsize; |
| 407 | int mode; |
| 408 | |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 409 | if ((buf = malloc(bufsize)) == NULL) |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 410 | FATAL("out of memory in getline"); |
| 411 | |
| 412 | fflush(stdout); /* in case someone is waiting for a prompt */ |
| 413 | r = gettemp(); |
| 414 | if (a[1] != NULL) { /* getline < file */ |
| 415 | x = execute(a[2]); /* filename */ |
| 416 | mode = ptoi(a[1]); |
| 417 | if (mode == '|') /* input pipe */ |
| 418 | mode = LE; /* arbitrary flag */ |
| 419 | fp = openfile(mode, getsval(x)); |
| 420 | tempfree(x); |
| 421 | if (fp == NULL) |
| 422 | n = -1; |
| 423 | else |
| 424 | n = readrec(&buf, &bufsize, fp); |
| 425 | if (n <= 0) { |
| 426 | ; |
| 427 | } else if (a[0] != NULL) { /* getline var <file */ |
| 428 | x = execute(a[0]); |
| 429 | setsval(x, buf); |
Arnold D. Robbins | 9dbd1f1 | 2019-01-25 12:56:06 +0200 | [diff] [blame] | 430 | if (is_number(x->sval)) { |
| 431 | x->fval = atof(x->sval); |
| 432 | x->tval |= NUM; |
| 433 | } |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 434 | tempfree(x); |
| 435 | } else { /* getline <file */ |
| 436 | setsval(fldtab[0], buf); |
| 437 | if (is_number(fldtab[0]->sval)) { |
| 438 | fldtab[0]->fval = atof(fldtab[0]->sval); |
| 439 | fldtab[0]->tval |= NUM; |
| 440 | } |
| 441 | } |
| 442 | } else { /* bare getline; use current input */ |
| 443 | if (a[0] == NULL) /* getline */ |
Arnold D. Robbins | 108224b | 2019-11-10 21:19:18 +0200 | [diff] [blame] | 444 | n = getrec(&record, &recsize, true); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 445 | else { /* getline var */ |
Arnold D. Robbins | 108224b | 2019-11-10 21:19:18 +0200 | [diff] [blame] | 446 | n = getrec(&buf, &bufsize, false); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 447 | x = execute(a[0]); |
| 448 | setsval(x, buf); |
Arnold D. Robbins | 9dbd1f1 | 2019-01-25 12:56:06 +0200 | [diff] [blame] | 449 | if (is_number(x->sval)) { |
| 450 | x->fval = atof(x->sval); |
| 451 | x->tval |= NUM; |
| 452 | } |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 453 | tempfree(x); |
| 454 | } |
| 455 | } |
| 456 | setfval(r, (Awkfloat) n); |
| 457 | free(buf); |
| 458 | return r; |
| 459 | } |
| 460 | |
| 461 | Cell *getnf(Node **a, int n) /* get NF */ |
| 462 | { |
Arnold D. Robbins | 108224b | 2019-11-10 21:19:18 +0200 | [diff] [blame] | 463 | if (!donefld) |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 464 | fldbld(); |
| 465 | return (Cell *) a[0]; |
| 466 | } |
| 467 | |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 468 | static char * |
| 469 | makearraystring(Node *p, const char *func) |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 470 | { |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 471 | char *buf; |
| 472 | int bufsz = recsize; |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 473 | size_t blen, seplen; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 474 | |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 475 | if ((buf = malloc(bufsz)) == NULL) { |
| 476 | FATAL("%s: out of memory", func); |
| 477 | } |
| 478 | |
| 479 | blen = 0; |
| 480 | buf[blen] = '\0'; |
| 481 | seplen = strlen(getsval(subseploc)); |
| 482 | |
| 483 | for (; p; p = p->nnext) { |
| 484 | Cell *x = execute(p); /* expr */ |
| 485 | char *s = getsval(x); |
| 486 | size_t nsub = p->nnext ? seplen : 0; |
| 487 | size_t slen = strlen(s); |
| 488 | size_t tlen = blen + slen + nsub; |
| 489 | |
| 490 | if (!adjbuf(&buf, &bufsz, tlen + 1, recsize, 0, func)) { |
| 491 | FATAL("%s: out of memory %s[%s...]", |
| 492 | func, x->nval, buf); |
| 493 | } |
| 494 | memcpy(buf + blen, s, slen); |
| 495 | if (nsub) { |
| 496 | memcpy(buf + blen + slen, *SUBSEP, nsub); |
| 497 | } |
| 498 | buf[tlen] = '\0'; |
| 499 | blen = tlen; |
| 500 | tempfree(x); |
| 501 | } |
| 502 | return buf; |
| 503 | } |
| 504 | |
| 505 | Cell *array(Node **a, int n) /* a[0] is symtab, a[1] is list of subscripts */ |
| 506 | { |
| 507 | Cell *x, *z; |
| 508 | char *buf; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 509 | |
| 510 | x = execute(a[0]); /* Cell* for symbol table */ |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 511 | buf = makearraystring(a[1], __func__); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 512 | if (!isarr(x)) { |
| 513 | dprintf( ("making %s into an array\n", NN(x->nval)) ); |
| 514 | if (freeable(x)) |
| 515 | xfree(x->sval); |
| 516 | x->tval &= ~(STR|NUM|DONTFREE); |
| 517 | x->tval |= ARR; |
| 518 | x->sval = (char *) makesymtab(NSYMTAB); |
| 519 | } |
| 520 | z = setsymtab(buf, "", 0.0, STR|NUM, (Array *) x->sval); |
| 521 | z->ctype = OCELL; |
| 522 | z->csub = CVAR; |
| 523 | tempfree(x); |
| 524 | free(buf); |
| 525 | return(z); |
| 526 | } |
| 527 | |
| 528 | Cell *awkdelete(Node **a, int n) /* a[0] is symtab, a[1] is list of subscripts */ |
| 529 | { |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 530 | Cell *x; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 531 | |
| 532 | x = execute(a[0]); /* Cell* for symbol table */ |
Cody Mello | ae99b75 | 2019-06-17 10:08:54 -0900 | [diff] [blame] | 533 | if (x == symtabloc) { |
| 534 | FATAL("cannot delete SYMTAB or its elements"); |
| 535 | } |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 536 | if (!isarr(x)) |
| 537 | return True; |
pfg | 5242194 | 2016-06-03 21:23:11 +0000 | [diff] [blame] | 538 | if (a[1] == NULL) { /* delete the elements, not the table */ |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 539 | freesymtab(x); |
| 540 | x->tval &= ~STR; |
| 541 | x->tval |= ARR; |
| 542 | x->sval = (char *) makesymtab(NSYMTAB); |
| 543 | } else { |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 544 | char *buf = makearraystring(a[1], __func__); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 545 | freeelem(x, buf); |
| 546 | free(buf); |
| 547 | } |
| 548 | tempfree(x); |
| 549 | return True; |
| 550 | } |
| 551 | |
| 552 | Cell *intest(Node **a, int n) /* a[0] is index (list), a[1] is symtab */ |
| 553 | { |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 554 | Cell *ap, *k; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 555 | char *buf; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 556 | |
| 557 | ap = execute(a[1]); /* array name */ |
| 558 | if (!isarr(ap)) { |
| 559 | dprintf( ("making %s into an array\n", ap->nval) ); |
| 560 | if (freeable(ap)) |
| 561 | xfree(ap->sval); |
| 562 | ap->tval &= ~(STR|NUM|DONTFREE); |
| 563 | ap->tval |= ARR; |
| 564 | ap->sval = (char *) makesymtab(NSYMTAB); |
| 565 | } |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 566 | buf = makearraystring(a[0], __func__); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 567 | k = lookup(buf, (Array *) ap->sval); |
| 568 | tempfree(ap); |
| 569 | free(buf); |
| 570 | if (k == NULL) |
| 571 | return(False); |
| 572 | else |
| 573 | return(True); |
| 574 | } |
| 575 | |
| 576 | |
| 577 | Cell *matchop(Node **a, int n) /* ~ and match() */ |
| 578 | { |
| 579 | Cell *x, *y; |
| 580 | char *s, *t; |
| 581 | int i; |
| 582 | fa *pfa; |
| 583 | int (*mf)(fa *, const char *) = match, mode = 0; |
| 584 | |
| 585 | if (n == MATCHFCN) { |
| 586 | mf = pmatch; |
| 587 | mode = 1; |
| 588 | } |
| 589 | x = execute(a[1]); /* a[1] = target text */ |
| 590 | s = getsval(x); |
pfg | 5242194 | 2016-06-03 21:23:11 +0000 | [diff] [blame] | 591 | if (a[0] == NULL) /* a[1] == 0: already-compiled reg expr */ |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 592 | i = (*mf)((fa *) a[2], s); |
| 593 | else { |
| 594 | y = execute(a[2]); /* a[2] = regular expr */ |
| 595 | t = getsval(y); |
| 596 | pfa = makedfa(t, mode); |
| 597 | i = (*mf)(pfa, s); |
| 598 | tempfree(y); |
| 599 | } |
| 600 | tempfree(x); |
| 601 | if (n == MATCHFCN) { |
| 602 | int start = patbeg - s + 1; |
| 603 | if (patlen < 0) |
| 604 | start = 0; |
| 605 | setfval(rstartloc, (Awkfloat) start); |
| 606 | setfval(rlengthloc, (Awkfloat) patlen); |
| 607 | x = gettemp(); |
| 608 | x->tval = NUM; |
| 609 | x->fval = start; |
| 610 | return x; |
| 611 | } else if ((n == MATCH && i == 1) || (n == NOTMATCH && i == 0)) |
| 612 | return(True); |
| 613 | else |
| 614 | return(False); |
| 615 | } |
| 616 | |
| 617 | |
| 618 | Cell *boolop(Node **a, int n) /* a[0] || a[1], a[0] && a[1], !a[0] */ |
| 619 | { |
| 620 | Cell *x, *y; |
| 621 | int i; |
| 622 | |
| 623 | x = execute(a[0]); |
| 624 | i = istrue(x); |
| 625 | tempfree(x); |
| 626 | switch (n) { |
| 627 | case BOR: |
| 628 | if (i) return(True); |
| 629 | y = execute(a[1]); |
| 630 | i = istrue(y); |
| 631 | tempfree(y); |
| 632 | if (i) return(True); |
| 633 | else return(False); |
| 634 | case AND: |
| 635 | if ( !i ) return(False); |
| 636 | y = execute(a[1]); |
| 637 | i = istrue(y); |
| 638 | tempfree(y); |
| 639 | if (i) return(True); |
| 640 | else return(False); |
| 641 | case NOT: |
| 642 | if (i) return(False); |
| 643 | else return(True); |
| 644 | default: /* can't happen */ |
| 645 | FATAL("unknown boolean operator %d", n); |
| 646 | } |
| 647 | return 0; /*NOTREACHED*/ |
| 648 | } |
| 649 | |
| 650 | Cell *relop(Node **a, int n) /* a[0 < a[1], etc. */ |
| 651 | { |
| 652 | int i; |
| 653 | Cell *x, *y; |
| 654 | Awkfloat j; |
| 655 | |
| 656 | x = execute(a[0]); |
| 657 | y = execute(a[1]); |
| 658 | if (x->tval&NUM && y->tval&NUM) { |
| 659 | j = x->fval - y->fval; |
| 660 | i = j<0? -1: (j>0? 1: 0); |
| 661 | } else { |
| 662 | i = strcmp(getsval(x), getsval(y)); |
| 663 | } |
| 664 | tempfree(x); |
| 665 | tempfree(y); |
| 666 | switch (n) { |
| 667 | case LT: if (i<0) return(True); |
| 668 | else return(False); |
| 669 | case LE: if (i<=0) return(True); |
| 670 | else return(False); |
| 671 | case NE: if (i!=0) return(True); |
| 672 | else return(False); |
| 673 | case EQ: if (i == 0) return(True); |
| 674 | else return(False); |
| 675 | case GE: if (i>=0) return(True); |
| 676 | else return(False); |
| 677 | case GT: if (i>0) return(True); |
| 678 | else return(False); |
| 679 | default: /* can't happen */ |
| 680 | FATAL("unknown relational operator %d", n); |
| 681 | } |
| 682 | return 0; /*NOTREACHED*/ |
| 683 | } |
| 684 | |
| 685 | void tfree(Cell *a) /* free a tempcell */ |
| 686 | { |
| 687 | if (freeable(a)) { |
| 688 | dprintf( ("freeing %s %s %o\n", NN(a->nval), NN(a->sval), a->tval) ); |
| 689 | xfree(a->sval); |
| 690 | } |
| 691 | if (a == tmps) |
| 692 | FATAL("tempcell list is curdled"); |
| 693 | a->cnext = tmps; |
| 694 | tmps = a; |
| 695 | } |
| 696 | |
| 697 | Cell *gettemp(void) /* get a tempcell */ |
| 698 | { int i; |
| 699 | Cell *x; |
| 700 | |
| 701 | if (!tmps) { |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 702 | tmps = calloc(100, sizeof(*tmps)); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 703 | if (!tmps) |
| 704 | FATAL("out of space for temporaries"); |
Arnold D. Robbins | 0ba1d03 | 2019-10-08 10:30:09 +0300 | [diff] [blame] | 705 | for (i = 1; i < 100; i++) |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 706 | tmps[i-1].cnext = &tmps[i]; |
pfg | 5242194 | 2016-06-03 21:23:11 +0000 | [diff] [blame] | 707 | tmps[i-1].cnext = NULL; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 708 | } |
| 709 | x = tmps; |
| 710 | tmps = x->cnext; |
| 711 | *x = tempcell; |
| 712 | return(x); |
| 713 | } |
| 714 | |
| 715 | Cell *indirect(Node **a, int n) /* $( a[0] ) */ |
| 716 | { |
| 717 | Awkfloat val; |
| 718 | Cell *x; |
| 719 | int m; |
| 720 | char *s; |
| 721 | |
| 722 | x = execute(a[0]); |
| 723 | val = getfval(x); /* freebsd: defend against super large field numbers */ |
| 724 | if ((Awkfloat)INT_MAX < val) |
| 725 | FATAL("trying to access out of range field %s", x->nval); |
| 726 | m = (int) val; |
| 727 | if (m == 0 && !is_number(s = getsval(x))) /* suspicion! */ |
| 728 | FATAL("illegal field $(%s), name \"%s\"", s, x->nval); |
| 729 | /* BUG: can x->nval ever be null??? */ |
| 730 | tempfree(x); |
| 731 | x = fieldadr(m); |
| 732 | x->ctype = OCELL; /* BUG? why are these needed? */ |
| 733 | x->csub = CFLD; |
| 734 | return(x); |
| 735 | } |
| 736 | |
| 737 | Cell *substr(Node **a, int nnn) /* substr(a[0], a[1], a[2]) */ |
| 738 | { |
| 739 | int k, m, n; |
| 740 | char *s; |
| 741 | int temp; |
pfg | 5242194 | 2016-06-03 21:23:11 +0000 | [diff] [blame] | 742 | Cell *x, *y, *z = NULL; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 743 | |
| 744 | x = execute(a[0]); |
| 745 | y = execute(a[1]); |
pfg | 5242194 | 2016-06-03 21:23:11 +0000 | [diff] [blame] | 746 | if (a[2] != NULL) |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 747 | z = execute(a[2]); |
| 748 | s = getsval(x); |
| 749 | k = strlen(s) + 1; |
| 750 | if (k <= 1) { |
| 751 | tempfree(x); |
| 752 | tempfree(y); |
pfg | 5242194 | 2016-06-03 21:23:11 +0000 | [diff] [blame] | 753 | if (a[2] != NULL) { |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 754 | tempfree(z); |
| 755 | } |
| 756 | x = gettemp(); |
| 757 | setsval(x, ""); |
| 758 | return(x); |
| 759 | } |
| 760 | m = (int) getfval(y); |
| 761 | if (m <= 0) |
| 762 | m = 1; |
| 763 | else if (m > k) |
| 764 | m = k; |
| 765 | tempfree(y); |
pfg | 5242194 | 2016-06-03 21:23:11 +0000 | [diff] [blame] | 766 | if (a[2] != NULL) { |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 767 | n = (int) getfval(z); |
| 768 | tempfree(z); |
| 769 | } else |
| 770 | n = k - 1; |
| 771 | if (n < 0) |
| 772 | n = 0; |
| 773 | else if (n > k - m) |
| 774 | n = k - m; |
| 775 | dprintf( ("substr: m=%d, n=%d, s=%s\n", m, n, s) ); |
| 776 | y = gettemp(); |
| 777 | temp = s[n+m-1]; /* with thanks to John Linderman */ |
| 778 | s[n+m-1] = '\0'; |
| 779 | setsval(y, s + m - 1); |
| 780 | s[n+m-1] = temp; |
| 781 | tempfree(x); |
| 782 | return(y); |
| 783 | } |
| 784 | |
| 785 | Cell *sindex(Node **a, int nnn) /* index(a[0], a[1]) */ |
| 786 | { |
| 787 | Cell *x, *y, *z; |
| 788 | char *s1, *s2, *p1, *p2, *q; |
| 789 | Awkfloat v = 0.0; |
| 790 | |
| 791 | x = execute(a[0]); |
| 792 | s1 = getsval(x); |
| 793 | y = execute(a[1]); |
| 794 | s2 = getsval(y); |
| 795 | |
| 796 | z = gettemp(); |
| 797 | for (p1 = s1; *p1 != '\0'; p1++) { |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 798 | for (q = p1, p2 = s2; *p2 != '\0' && *q == *p2; q++, p2++) |
| 799 | continue; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 800 | if (*p2 == '\0') { |
| 801 | v = (Awkfloat) (p1 - s1 + 1); /* origin 1 */ |
| 802 | break; |
| 803 | } |
| 804 | } |
| 805 | tempfree(x); |
| 806 | tempfree(y); |
| 807 | setfval(z, v); |
| 808 | return(z); |
| 809 | } |
| 810 | |
| 811 | #define MAXNUMSIZE 50 |
| 812 | |
| 813 | int format(char **pbuf, int *pbufsize, const char *s, Node *a) /* printf-like conversions */ |
| 814 | { |
| 815 | char *fmt; |
| 816 | char *p, *t; |
| 817 | const char *os; |
| 818 | Cell *x; |
| 819 | int flag = 0, n; |
| 820 | int fmtwd; /* format width */ |
| 821 | int fmtsz = recsize; |
| 822 | char *buf = *pbuf; |
| 823 | int bufsize = *pbufsize; |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 824 | #define FMTSZ(a) (fmtsz - ((a) - fmt)) |
| 825 | #define BUFSZ(a) (bufsize - ((a) - buf)) |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 826 | |
Arnold D. Robbins | 108224b | 2019-11-10 21:19:18 +0200 | [diff] [blame] | 827 | static bool first = true; |
| 828 | static bool have_a_format = false; |
Arnold D. Robbins | 32093f5 | 2018-08-22 20:40:26 +0300 | [diff] [blame] | 829 | |
| 830 | if (first) { |
| 831 | char buf[100]; |
| 832 | |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 833 | snprintf(buf, sizeof(buf), "%a", 42.0); |
Arnold D. Robbins | 32093f5 | 2018-08-22 20:40:26 +0300 | [diff] [blame] | 834 | have_a_format = (strcmp(buf, "0x1.5p+5") == 0); |
Arnold D. Robbins | 108224b | 2019-11-10 21:19:18 +0200 | [diff] [blame] | 835 | first = false; |
Arnold D. Robbins | 32093f5 | 2018-08-22 20:40:26 +0300 | [diff] [blame] | 836 | } |
| 837 | |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 838 | os = s; |
| 839 | p = buf; |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 840 | if ((fmt = malloc(fmtsz)) == NULL) |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 841 | FATAL("out of memory in format()"); |
| 842 | while (*s) { |
| 843 | adjbuf(&buf, &bufsize, MAXNUMSIZE+1+p-buf, recsize, &p, "format1"); |
| 844 | if (*s != '%') { |
| 845 | *p++ = *s++; |
| 846 | continue; |
| 847 | } |
| 848 | if (*(s+1) == '%') { |
| 849 | *p++ = '%'; |
| 850 | s += 2; |
| 851 | continue; |
| 852 | } |
| 853 | /* have to be real careful in case this is a huge number, eg, %100000d */ |
| 854 | fmtwd = atoi(s+1); |
| 855 | if (fmtwd < 0) |
| 856 | fmtwd = -fmtwd; |
| 857 | adjbuf(&buf, &bufsize, fmtwd+1+p-buf, recsize, &p, "format2"); |
| 858 | for (t = fmt; (*t++ = *s) != '\0'; s++) { |
| 859 | if (!adjbuf(&fmt, &fmtsz, MAXNUMSIZE+1+t-fmt, recsize, &t, "format3")) |
| 860 | FATAL("format item %.30s... ran format() out of memory", os); |
zoulasc | a96aebb | 2019-12-11 02:17:34 -0500 | [diff] [blame] | 861 | /* Ignore size specifiers */ |
Arnold D. Robbins | 0b82bc6 | 2019-12-11 09:24:38 +0200 | [diff] [blame] | 862 | if (strchr("hjLlqtz", *s) != NULL) { /* the ansi panoply */ |
zoulasc | a96aebb | 2019-12-11 02:17:34 -0500 | [diff] [blame] | 863 | t--; |
| 864 | continue; |
| 865 | } |
| 866 | if (isalpha((uschar)*s)) |
Arnold D. Robbins | 0b82bc6 | 2019-12-11 09:24:38 +0200 | [diff] [blame] | 867 | break; |
Brian Kernighan | 22aff9e | 2018-08-27 08:52:34 -0400 | [diff] [blame] | 868 | if (*s == '$') { |
| 869 | FATAL("'$' not permitted in awk formats"); |
| 870 | } |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 871 | if (*s == '*') { |
Cody Peter Mello | 6cf37e9 | 2018-09-14 17:29:06 -0700 | [diff] [blame] | 872 | if (a == NULL) { |
| 873 | FATAL("not enough args in printf(%s)", os); |
| 874 | } |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 875 | x = execute(a); |
| 876 | a = a->nnext; |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 877 | snprintf(t - 1, FMTSZ(t - 1), |
| 878 | "%d", fmtwd=(int) getfval(x)); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 879 | if (fmtwd < 0) |
| 880 | fmtwd = -fmtwd; |
| 881 | adjbuf(&buf, &bufsize, fmtwd+1+p-buf, recsize, &p, "format"); |
| 882 | t = fmt + strlen(fmt); |
| 883 | tempfree(x); |
| 884 | } |
| 885 | } |
| 886 | *t = '\0'; |
| 887 | if (fmtwd < 0) |
| 888 | fmtwd = -fmtwd; |
| 889 | adjbuf(&buf, &bufsize, fmtwd+1+p-buf, recsize, &p, "format4"); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 890 | switch (*s) { |
Arnold D. Robbins | 32093f5 | 2018-08-22 20:40:26 +0300 | [diff] [blame] | 891 | case 'a': case 'A': |
| 892 | if (have_a_format) |
| 893 | flag = *s; |
| 894 | else |
| 895 | flag = 'f'; |
| 896 | break; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 897 | case 'f': case 'e': case 'g': case 'E': case 'G': |
| 898 | flag = 'f'; |
| 899 | break; |
zoulasc | a96aebb | 2019-12-11 02:17:34 -0500 | [diff] [blame] | 900 | case 'd': case 'i': case 'o': case 'x': case 'X': case 'u': |
Arnold D. Robbins | 0b82bc6 | 2019-12-11 09:24:38 +0200 | [diff] [blame] | 901 | flag = (*s == 'd' || *s == 'i') ? 'd' : 'u'; |
zoulasc | ff5d676 | 2019-12-08 14:41:27 -0500 | [diff] [blame] | 902 | *(t-1) = 'j'; |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 903 | *t = *s; |
| 904 | *++t = '\0'; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 905 | break; |
| 906 | case 's': |
| 907 | flag = 's'; |
| 908 | break; |
| 909 | case 'c': |
| 910 | flag = 'c'; |
| 911 | break; |
| 912 | default: |
| 913 | WARNING("weird printf conversion %s", fmt); |
| 914 | flag = '?'; |
| 915 | break; |
| 916 | } |
| 917 | if (a == NULL) |
| 918 | FATAL("not enough args in printf(%s)", os); |
| 919 | x = execute(a); |
| 920 | a = a->nnext; |
| 921 | n = MAXNUMSIZE; |
| 922 | if (fmtwd > n) |
| 923 | n = fmtwd; |
| 924 | adjbuf(&buf, &bufsize, 1+n+p-buf, recsize, &p, "format5"); |
| 925 | switch (flag) { |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 926 | case '?': snprintf(p, BUFSZ(p), "%s", fmt); /* unknown, so dump it too */ |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 927 | t = getsval(x); |
| 928 | n = strlen(t); |
| 929 | if (fmtwd > n) |
| 930 | n = fmtwd; |
| 931 | adjbuf(&buf, &bufsize, 1+strlen(p)+n+p-buf, recsize, &p, "format6"); |
| 932 | p += strlen(p); |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 933 | snprintf(p, BUFSZ(p), "%s", t); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 934 | break; |
Arnold D. Robbins | 32093f5 | 2018-08-22 20:40:26 +0300 | [diff] [blame] | 935 | case 'a': |
| 936 | case 'A': |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 937 | case 'f': snprintf(p, BUFSZ(p), fmt, getfval(x)); break; |
zoulasc | ff5d676 | 2019-12-08 14:41:27 -0500 | [diff] [blame] | 938 | case 'd': snprintf(p, BUFSZ(p), fmt, (intmax_t) getfval(x)); break; |
| 939 | case 'u': snprintf(p, BUFSZ(p), fmt, (uintmax_t) getfval(x)); break; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 940 | case 's': |
| 941 | t = getsval(x); |
| 942 | n = strlen(t); |
| 943 | if (fmtwd > n) |
| 944 | n = fmtwd; |
| 945 | if (!adjbuf(&buf, &bufsize, 1+n+p-buf, recsize, &p, "format7")) |
| 946 | FATAL("huge string/format (%d chars) in printf %.30s... ran format() out of memory", n, t); |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 947 | snprintf(p, BUFSZ(p), fmt, t); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 948 | break; |
| 949 | case 'c': |
| 950 | if (isnum(x)) { |
M. Warner Losh | 966fdc9 | 2019-06-02 14:53:15 -0600 | [diff] [blame] | 951 | if ((int)getfval(x)) |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 952 | snprintf(p, BUFSZ(p), fmt, (int) getfval(x)); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 953 | else { |
| 954 | *p++ = '\0'; /* explicit null byte */ |
| 955 | *p = '\0'; /* next output will start here */ |
| 956 | } |
| 957 | } else |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 958 | snprintf(p, BUFSZ(p), fmt, getsval(x)[0]); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 959 | break; |
| 960 | default: |
| 961 | FATAL("can't happen: bad conversion %c in format()", flag); |
| 962 | } |
| 963 | tempfree(x); |
| 964 | p += strlen(p); |
| 965 | s++; |
| 966 | } |
| 967 | *p = '\0'; |
| 968 | free(fmt); |
| 969 | for ( ; a; a = a->nnext) /* evaluate any remaining args */ |
| 970 | execute(a); |
| 971 | *pbuf = buf; |
| 972 | *pbufsize = bufsize; |
| 973 | return p - buf; |
| 974 | } |
| 975 | |
| 976 | Cell *awksprintf(Node **a, int n) /* sprintf(a[0]) */ |
| 977 | { |
| 978 | Cell *x; |
| 979 | Node *y; |
| 980 | char *buf; |
| 981 | int bufsz=3*recsize; |
| 982 | |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 983 | if ((buf = malloc(bufsz)) == NULL) |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 984 | FATAL("out of memory in awksprintf"); |
| 985 | y = a[0]->nnext; |
| 986 | x = execute(a[0]); |
| 987 | if (format(&buf, &bufsz, getsval(x), y) == -1) |
| 988 | FATAL("sprintf string %.30s... too long. can't happen.", buf); |
| 989 | tempfree(x); |
| 990 | x = gettemp(); |
| 991 | x->sval = buf; |
| 992 | x->tval = STR; |
| 993 | return(x); |
| 994 | } |
| 995 | |
| 996 | Cell *awkprintf(Node **a, int n) /* printf */ |
| 997 | { /* a[0] is list of args, starting with format string */ |
| 998 | /* a[1] is redirection operator, a[2] is redirection file */ |
| 999 | FILE *fp; |
| 1000 | Cell *x; |
| 1001 | Node *y; |
| 1002 | char *buf; |
| 1003 | int len; |
| 1004 | int bufsz=3*recsize; |
| 1005 | |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 1006 | if ((buf = malloc(bufsz)) == NULL) |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1007 | FATAL("out of memory in awkprintf"); |
| 1008 | y = a[0]->nnext; |
| 1009 | x = execute(a[0]); |
| 1010 | if ((len = format(&buf, &bufsz, getsval(x), y)) == -1) |
| 1011 | FATAL("printf string %.30s... too long. can't happen.", buf); |
| 1012 | tempfree(x); |
| 1013 | if (a[1] == NULL) { |
| 1014 | /* fputs(buf, stdout); */ |
| 1015 | fwrite(buf, len, 1, stdout); |
| 1016 | if (ferror(stdout)) |
| 1017 | FATAL("write error on stdout"); |
| 1018 | } else { |
| 1019 | fp = redirect(ptoi(a[1]), a[2]); |
| 1020 | /* fputs(buf, fp); */ |
| 1021 | fwrite(buf, len, 1, fp); |
| 1022 | fflush(fp); |
| 1023 | if (ferror(fp)) |
| 1024 | FATAL("write error on %s", filename(fp)); |
| 1025 | } |
| 1026 | free(buf); |
| 1027 | return(True); |
| 1028 | } |
| 1029 | |
| 1030 | Cell *arith(Node **a, int n) /* a[0] + a[1], etc. also -a[0] */ |
| 1031 | { |
| 1032 | Awkfloat i, j = 0; |
| 1033 | double v; |
| 1034 | Cell *x, *y, *z; |
| 1035 | |
| 1036 | x = execute(a[0]); |
| 1037 | i = getfval(x); |
| 1038 | tempfree(x); |
Arnold D. Robbins | 32093f5 | 2018-08-22 20:40:26 +0300 | [diff] [blame] | 1039 | if (n != UMINUS && n != UPLUS) { |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1040 | y = execute(a[1]); |
| 1041 | j = getfval(y); |
| 1042 | tempfree(y); |
| 1043 | } |
| 1044 | z = gettemp(); |
| 1045 | switch (n) { |
| 1046 | case ADD: |
| 1047 | i += j; |
| 1048 | break; |
| 1049 | case MINUS: |
| 1050 | i -= j; |
| 1051 | break; |
| 1052 | case MULT: |
| 1053 | i *= j; |
| 1054 | break; |
| 1055 | case DIVIDE: |
| 1056 | if (j == 0) |
| 1057 | FATAL("division by zero"); |
| 1058 | i /= j; |
| 1059 | break; |
| 1060 | case MOD: |
| 1061 | if (j == 0) |
| 1062 | FATAL("division by zero in mod"); |
| 1063 | modf(i/j, &v); |
| 1064 | i = i - j * v; |
| 1065 | break; |
| 1066 | case UMINUS: |
| 1067 | i = -i; |
| 1068 | break; |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1069 | case UPLUS: /* handled by getfval(), above */ |
Arnold D. Robbins | 32093f5 | 2018-08-22 20:40:26 +0300 | [diff] [blame] | 1070 | break; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1071 | case POWER: |
| 1072 | if (j >= 0 && modf(j, &v) == 0.0) /* pos integer exponent */ |
| 1073 | i = ipow(i, (int) j); |
| 1074 | else |
| 1075 | i = errcheck(pow(i, j), "pow"); |
| 1076 | break; |
| 1077 | default: /* can't happen */ |
| 1078 | FATAL("illegal arithmetic operator %d", n); |
| 1079 | } |
| 1080 | setfval(z, i); |
| 1081 | return(z); |
| 1082 | } |
| 1083 | |
| 1084 | double ipow(double x, int n) /* x**n. ought to be done by pow, but isn't always */ |
| 1085 | { |
| 1086 | double v; |
| 1087 | |
| 1088 | if (n <= 0) |
| 1089 | return 1; |
| 1090 | v = ipow(x, n/2); |
| 1091 | if (n % 2 == 0) |
| 1092 | return v * v; |
| 1093 | else |
| 1094 | return x * v * v; |
| 1095 | } |
| 1096 | |
| 1097 | Cell *incrdecr(Node **a, int n) /* a[0]++, etc. */ |
| 1098 | { |
| 1099 | Cell *x, *z; |
| 1100 | int k; |
| 1101 | Awkfloat xf; |
| 1102 | |
| 1103 | x = execute(a[0]); |
| 1104 | xf = getfval(x); |
| 1105 | k = (n == PREINCR || n == POSTINCR) ? 1 : -1; |
| 1106 | if (n == PREINCR || n == PREDECR) { |
| 1107 | setfval(x, xf + k); |
| 1108 | return(x); |
| 1109 | } |
| 1110 | z = gettemp(); |
| 1111 | setfval(z, xf); |
| 1112 | setfval(x, xf + k); |
| 1113 | tempfree(x); |
| 1114 | return(z); |
| 1115 | } |
| 1116 | |
| 1117 | Cell *assign(Node **a, int n) /* a[0] = a[1], a[0] += a[1], etc. */ |
| 1118 | { /* this is subtle; don't muck with it. */ |
| 1119 | Cell *x, *y; |
| 1120 | Awkfloat xf, yf; |
| 1121 | double v; |
| 1122 | |
| 1123 | y = execute(a[1]); |
| 1124 | x = execute(a[0]); |
| 1125 | if (n == ASSIGN) { /* ordinary assignment */ |
Cody Peter Mello | 6315525 | 2018-09-23 17:59:52 -0700 | [diff] [blame] | 1126 | if (x == y && !(x->tval & (FLD|REC)) && x != nfloc) |
| 1127 | ; /* self-assignment: leave alone unless it's a field or NF */ |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1128 | else if ((y->tval & (STR|NUM)) == (STR|NUM)) { |
| 1129 | setsval(x, getsval(y)); |
| 1130 | x->fval = getfval(y); |
| 1131 | x->tval |= NUM; |
| 1132 | } |
| 1133 | else if (isstr(y)) |
| 1134 | setsval(x, getsval(y)); |
| 1135 | else if (isnum(y)) |
| 1136 | setfval(x, getfval(y)); |
| 1137 | else |
| 1138 | funnyvar(y, "read value of"); |
| 1139 | tempfree(y); |
| 1140 | return(x); |
| 1141 | } |
| 1142 | xf = getfval(x); |
| 1143 | yf = getfval(y); |
| 1144 | switch (n) { |
| 1145 | case ADDEQ: |
| 1146 | xf += yf; |
| 1147 | break; |
| 1148 | case SUBEQ: |
| 1149 | xf -= yf; |
| 1150 | break; |
| 1151 | case MULTEQ: |
| 1152 | xf *= yf; |
| 1153 | break; |
| 1154 | case DIVEQ: |
| 1155 | if (yf == 0) |
| 1156 | FATAL("division by zero in /="); |
| 1157 | xf /= yf; |
| 1158 | break; |
| 1159 | case MODEQ: |
| 1160 | if (yf == 0) |
| 1161 | FATAL("division by zero in %%="); |
| 1162 | modf(xf/yf, &v); |
| 1163 | xf = xf - yf * v; |
| 1164 | break; |
| 1165 | case POWEQ: |
| 1166 | if (yf >= 0 && modf(yf, &v) == 0.0) /* pos integer exponent */ |
| 1167 | xf = ipow(xf, (int) yf); |
| 1168 | else |
| 1169 | xf = errcheck(pow(xf, yf), "pow"); |
| 1170 | break; |
| 1171 | default: |
| 1172 | FATAL("illegal assignment operator %d", n); |
| 1173 | break; |
| 1174 | } |
| 1175 | tempfree(y); |
| 1176 | setfval(x, xf); |
| 1177 | return(x); |
| 1178 | } |
| 1179 | |
| 1180 | Cell *cat(Node **a, int q) /* a[0] cat a[1] */ |
| 1181 | { |
| 1182 | Cell *x, *y, *z; |
| 1183 | int n1, n2; |
Cody Peter Mello | e262374 | 2018-09-15 01:38:39 -0700 | [diff] [blame] | 1184 | char *s = NULL; |
| 1185 | int ssz = 0; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1186 | |
| 1187 | x = execute(a[0]); |
Cody Peter Mello | e262374 | 2018-09-15 01:38:39 -0700 | [diff] [blame] | 1188 | n1 = strlen(getsval(x)); |
Cody Peter Mello | e262374 | 2018-09-15 01:38:39 -0700 | [diff] [blame] | 1189 | |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1190 | y = execute(a[1]); |
Cody Peter Mello | e262374 | 2018-09-15 01:38:39 -0700 | [diff] [blame] | 1191 | n2 = strlen(getsval(y)); |
Arnold D. Robbins | 1d6ddfd | 2019-10-24 10:06:10 -0400 | [diff] [blame] | 1192 | |
| 1193 | adjbuf(&s, &ssz, n1 + n2 + 1, recsize, 0, "cat"); |
| 1194 | memcpy(s, x->sval, n1); |
| 1195 | memcpy(s + n1, y->sval, n2); |
| 1196 | s[n1 + n2] = '\0'; |
Cody Peter Mello | e262374 | 2018-09-15 01:38:39 -0700 | [diff] [blame] | 1197 | |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1198 | tempfree(x); |
| 1199 | tempfree(y); |
Cody Peter Mello | e262374 | 2018-09-15 01:38:39 -0700 | [diff] [blame] | 1200 | |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1201 | z = gettemp(); |
| 1202 | z->sval = s; |
| 1203 | z->tval = STR; |
Cody Peter Mello | e262374 | 2018-09-15 01:38:39 -0700 | [diff] [blame] | 1204 | |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1205 | return(z); |
| 1206 | } |
| 1207 | |
| 1208 | Cell *pastat(Node **a, int n) /* a[0] { a[1] } */ |
| 1209 | { |
| 1210 | Cell *x; |
| 1211 | |
pfg | 5242194 | 2016-06-03 21:23:11 +0000 | [diff] [blame] | 1212 | if (a[0] == NULL) |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1213 | x = execute(a[1]); |
| 1214 | else { |
| 1215 | x = execute(a[0]); |
| 1216 | if (istrue(x)) { |
| 1217 | tempfree(x); |
| 1218 | x = execute(a[1]); |
| 1219 | } |
| 1220 | } |
| 1221 | return x; |
| 1222 | } |
| 1223 | |
| 1224 | Cell *dopa2(Node **a, int n) /* a[0], a[1] { a[2] } */ |
| 1225 | { |
| 1226 | Cell *x; |
| 1227 | int pair; |
| 1228 | |
| 1229 | pair = ptoi(a[3]); |
| 1230 | if (pairstack[pair] == 0) { |
| 1231 | x = execute(a[0]); |
| 1232 | if (istrue(x)) |
| 1233 | pairstack[pair] = 1; |
| 1234 | tempfree(x); |
| 1235 | } |
| 1236 | if (pairstack[pair] == 1) { |
| 1237 | x = execute(a[1]); |
| 1238 | if (istrue(x)) |
| 1239 | pairstack[pair] = 0; |
| 1240 | tempfree(x); |
| 1241 | x = execute(a[2]); |
| 1242 | return(x); |
| 1243 | } |
| 1244 | return(False); |
| 1245 | } |
| 1246 | |
| 1247 | Cell *split(Node **a, int nnn) /* split(a[0], a[1], a[2]); a[3] is type */ |
| 1248 | { |
pfg | 5242194 | 2016-06-03 21:23:11 +0000 | [diff] [blame] | 1249 | Cell *x = NULL, *y, *ap; |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 1250 | const char *s, *origs, *t; |
zoulasc | 6a87709 | 2020-01-24 04:11:59 -0500 | [diff] [blame] | 1251 | const char *fs = NULL; |
| 1252 | char *origfs = NULL; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1253 | int sep; |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 1254 | char temp, num[50]; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1255 | int n, tempstat, arg3type; |
| 1256 | |
| 1257 | y = execute(a[0]); /* source string */ |
| 1258 | origs = s = strdup(getsval(y)); |
| 1259 | arg3type = ptoi(a[3]); |
pfg | 5242194 | 2016-06-03 21:23:11 +0000 | [diff] [blame] | 1260 | if (a[2] == NULL) /* fs string */ |
Cody Peter Mello | 52566c0 | 2018-09-18 15:45:55 -0700 | [diff] [blame] | 1261 | fs = getsval(fsloc); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1262 | else if (arg3type == STRING) { /* split(str,arr,"string") */ |
| 1263 | x = execute(a[2]); |
zoulasc | 6a87709 | 2020-01-24 04:11:59 -0500 | [diff] [blame] | 1264 | fs = origfs = strdup(getsval(x)); |
Cody Peter Mello | d45db5e | 2018-09-18 15:20:44 -0700 | [diff] [blame] | 1265 | tempfree(x); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1266 | } else if (arg3type == REGEXPR) |
| 1267 | fs = "(regexpr)"; /* split(str,arr,/regexpr/) */ |
| 1268 | else |
| 1269 | FATAL("illegal type of split"); |
| 1270 | sep = *fs; |
| 1271 | ap = execute(a[1]); /* array name */ |
| 1272 | freesymtab(ap); |
| 1273 | dprintf( ("split: s=|%s|, a=%s, sep=|%s|\n", s, NN(ap->nval), fs) ); |
| 1274 | ap->tval &= ~STR; |
| 1275 | ap->tval |= ARR; |
| 1276 | ap->sval = (char *) makesymtab(NSYMTAB); |
| 1277 | |
| 1278 | n = 0; |
| 1279 | if (arg3type == REGEXPR && strlen((char*)((fa*)a[2])->restr) == 0) { |
| 1280 | /* split(s, a, //); have to arrange that it looks like empty sep */ |
| 1281 | arg3type = 0; |
| 1282 | fs = ""; |
| 1283 | sep = 0; |
| 1284 | } |
| 1285 | if (*s != '\0' && (strlen(fs) > 1 || arg3type == REGEXPR)) { /* reg expr */ |
| 1286 | fa *pfa; |
| 1287 | if (arg3type == REGEXPR) { /* it's ready already */ |
| 1288 | pfa = (fa *) a[2]; |
| 1289 | } else { |
| 1290 | pfa = makedfa(fs, 1); |
| 1291 | } |
| 1292 | if (nematch(pfa,s)) { |
| 1293 | tempstat = pfa->initstat; |
| 1294 | pfa->initstat = 2; |
| 1295 | do { |
| 1296 | n++; |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 1297 | snprintf(num, sizeof(num), "%d", n); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1298 | temp = *patbeg; |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 1299 | setptr(patbeg, '\0'); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1300 | if (is_number(s)) |
| 1301 | setsymtab(num, s, atof(s), STR|NUM, (Array *) ap->sval); |
| 1302 | else |
| 1303 | setsymtab(num, s, 0.0, STR, (Array *) ap->sval); |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 1304 | setptr(patbeg, temp); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1305 | s = patbeg + patlen; |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1306 | if (*(patbeg+patlen-1) == '\0' || *s == '\0') { |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1307 | n++; |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 1308 | snprintf(num, sizeof(num), "%d", n); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1309 | setsymtab(num, "", 0.0, STR, (Array *) ap->sval); |
| 1310 | pfa->initstat = tempstat; |
| 1311 | goto spdone; |
| 1312 | } |
| 1313 | } while (nematch(pfa,s)); |
| 1314 | pfa->initstat = tempstat; /* bwk: has to be here to reset */ |
| 1315 | /* cf gsub and refldbld */ |
| 1316 | } |
| 1317 | n++; |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 1318 | snprintf(num, sizeof(num), "%d", n); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1319 | if (is_number(s)) |
| 1320 | setsymtab(num, s, atof(s), STR|NUM, (Array *) ap->sval); |
| 1321 | else |
| 1322 | setsymtab(num, s, 0.0, STR, (Array *) ap->sval); |
| 1323 | spdone: |
| 1324 | pfa = NULL; |
| 1325 | } else if (sep == ' ') { |
| 1326 | for (n = 0; ; ) { |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1327 | #define ISWS(c) ((c) == ' ' || (c) == '\t' || (c) == '\n') |
| 1328 | while (ISWS(*s)) |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1329 | s++; |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1330 | if (*s == '\0') |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1331 | break; |
| 1332 | n++; |
| 1333 | t = s; |
| 1334 | do |
| 1335 | s++; |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1336 | while (*s != '\0' && !ISWS(*s)); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1337 | temp = *s; |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 1338 | setptr(s, '\0'); |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 1339 | snprintf(num, sizeof(num), "%d", n); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1340 | if (is_number(t)) |
| 1341 | setsymtab(num, t, atof(t), STR|NUM, (Array *) ap->sval); |
| 1342 | else |
| 1343 | setsymtab(num, t, 0.0, STR, (Array *) ap->sval); |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 1344 | setptr(s, temp); |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1345 | if (*s != '\0') |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1346 | s++; |
| 1347 | } |
| 1348 | } else if (sep == 0) { /* new: split(s, a, "") => 1 char/elem */ |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1349 | for (n = 0; *s != '\0'; s++) { |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1350 | char buf[2]; |
| 1351 | n++; |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 1352 | snprintf(num, sizeof(num), "%d", n); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1353 | buf[0] = *s; |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1354 | buf[1] = '\0'; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1355 | if (isdigit((uschar)buf[0])) |
| 1356 | setsymtab(num, buf, atof(buf), STR|NUM, (Array *) ap->sval); |
| 1357 | else |
| 1358 | setsymtab(num, buf, 0.0, STR, (Array *) ap->sval); |
| 1359 | } |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1360 | } else if (*s != '\0') { |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1361 | for (;;) { |
| 1362 | n++; |
| 1363 | t = s; |
| 1364 | while (*s != sep && *s != '\n' && *s != '\0') |
| 1365 | s++; |
| 1366 | temp = *s; |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 1367 | setptr(s, '\0'); |
zoulasc | 0d8778b | 2019-10-25 10:59:10 -0400 | [diff] [blame] | 1368 | snprintf(num, sizeof(num), "%d", n); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1369 | if (is_number(t)) |
| 1370 | setsymtab(num, t, atof(t), STR|NUM, (Array *) ap->sval); |
| 1371 | else |
| 1372 | setsymtab(num, t, 0.0, STR, (Array *) ap->sval); |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 1373 | setptr(s, temp); |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1374 | if (*s++ == '\0') |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1375 | break; |
| 1376 | } |
| 1377 | } |
| 1378 | tempfree(ap); |
| 1379 | tempfree(y); |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 1380 | xfree(origs); |
| 1381 | xfree(origfs); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1382 | x = gettemp(); |
| 1383 | x->tval = NUM; |
| 1384 | x->fval = n; |
| 1385 | return(x); |
| 1386 | } |
| 1387 | |
| 1388 | Cell *condexpr(Node **a, int n) /* a[0] ? a[1] : a[2] */ |
| 1389 | { |
| 1390 | Cell *x; |
| 1391 | |
| 1392 | x = execute(a[0]); |
| 1393 | if (istrue(x)) { |
| 1394 | tempfree(x); |
| 1395 | x = execute(a[1]); |
| 1396 | } else { |
| 1397 | tempfree(x); |
| 1398 | x = execute(a[2]); |
| 1399 | } |
| 1400 | return(x); |
| 1401 | } |
| 1402 | |
| 1403 | Cell *ifstat(Node **a, int n) /* if (a[0]) a[1]; else a[2] */ |
| 1404 | { |
| 1405 | Cell *x; |
| 1406 | |
| 1407 | x = execute(a[0]); |
| 1408 | if (istrue(x)) { |
| 1409 | tempfree(x); |
| 1410 | x = execute(a[1]); |
pfg | 5242194 | 2016-06-03 21:23:11 +0000 | [diff] [blame] | 1411 | } else if (a[2] != NULL) { |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1412 | tempfree(x); |
| 1413 | x = execute(a[2]); |
| 1414 | } |
| 1415 | return(x); |
| 1416 | } |
| 1417 | |
| 1418 | Cell *whilestat(Node **a, int n) /* while (a[0]) a[1] */ |
| 1419 | { |
| 1420 | Cell *x; |
| 1421 | |
| 1422 | for (;;) { |
| 1423 | x = execute(a[0]); |
| 1424 | if (!istrue(x)) |
| 1425 | return(x); |
| 1426 | tempfree(x); |
| 1427 | x = execute(a[1]); |
| 1428 | if (isbreak(x)) { |
| 1429 | x = True; |
| 1430 | return(x); |
| 1431 | } |
| 1432 | if (isnext(x) || isexit(x) || isret(x)) |
| 1433 | return(x); |
| 1434 | tempfree(x); |
| 1435 | } |
| 1436 | } |
| 1437 | |
| 1438 | Cell *dostat(Node **a, int n) /* do a[0]; while(a[1]) */ |
| 1439 | { |
| 1440 | Cell *x; |
| 1441 | |
| 1442 | for (;;) { |
| 1443 | x = execute(a[0]); |
| 1444 | if (isbreak(x)) |
| 1445 | return True; |
| 1446 | if (isnext(x) || isexit(x) || isret(x)) |
| 1447 | return(x); |
| 1448 | tempfree(x); |
| 1449 | x = execute(a[1]); |
| 1450 | if (!istrue(x)) |
| 1451 | return(x); |
| 1452 | tempfree(x); |
| 1453 | } |
| 1454 | } |
| 1455 | |
| 1456 | Cell *forstat(Node **a, int n) /* for (a[0]; a[1]; a[2]) a[3] */ |
| 1457 | { |
| 1458 | Cell *x; |
| 1459 | |
| 1460 | x = execute(a[0]); |
| 1461 | tempfree(x); |
| 1462 | for (;;) { |
pfg | 5242194 | 2016-06-03 21:23:11 +0000 | [diff] [blame] | 1463 | if (a[1]!=NULL) { |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1464 | x = execute(a[1]); |
| 1465 | if (!istrue(x)) return(x); |
| 1466 | else tempfree(x); |
| 1467 | } |
| 1468 | x = execute(a[3]); |
| 1469 | if (isbreak(x)) /* turn off break */ |
| 1470 | return True; |
| 1471 | if (isnext(x) || isexit(x) || isret(x)) |
| 1472 | return(x); |
| 1473 | tempfree(x); |
| 1474 | x = execute(a[2]); |
| 1475 | tempfree(x); |
| 1476 | } |
| 1477 | } |
| 1478 | |
| 1479 | Cell *instat(Node **a, int n) /* for (a[0] in a[1]) a[2] */ |
| 1480 | { |
| 1481 | Cell *x, *vp, *arrayp, *cp, *ncp; |
| 1482 | Array *tp; |
| 1483 | int i; |
| 1484 | |
| 1485 | vp = execute(a[0]); |
| 1486 | arrayp = execute(a[1]); |
| 1487 | if (!isarr(arrayp)) { |
| 1488 | return True; |
| 1489 | } |
| 1490 | tp = (Array *) arrayp->sval; |
| 1491 | tempfree(arrayp); |
| 1492 | for (i = 0; i < tp->size; i++) { /* this routine knows too much */ |
| 1493 | for (cp = tp->tab[i]; cp != NULL; cp = ncp) { |
| 1494 | setsval(vp, cp->nval); |
| 1495 | ncp = cp->cnext; |
| 1496 | x = execute(a[2]); |
| 1497 | if (isbreak(x)) { |
| 1498 | tempfree(vp); |
| 1499 | return True; |
| 1500 | } |
| 1501 | if (isnext(x) || isexit(x) || isret(x)) { |
| 1502 | tempfree(vp); |
| 1503 | return(x); |
| 1504 | } |
| 1505 | tempfree(x); |
| 1506 | } |
| 1507 | } |
| 1508 | return True; |
| 1509 | } |
| 1510 | |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1511 | static char *nawk_convert(const char *s, int (*fun_c)(int), |
| 1512 | wint_t (*fun_wc)(wint_t)) |
| 1513 | { |
| 1514 | char *buf = NULL; |
| 1515 | char *pbuf = NULL; |
| 1516 | const char *ps = NULL; |
| 1517 | size_t n = 0; |
| 1518 | mbstate_t mbs, mbs2; |
| 1519 | wchar_t wc; |
| 1520 | size_t sz = MB_CUR_MAX; |
| 1521 | |
| 1522 | if (sz == 1) { |
| 1523 | buf = tostring(s); |
| 1524 | |
| 1525 | for (pbuf = buf; *pbuf; pbuf++) |
| 1526 | *pbuf = fun_c((uschar)*pbuf); |
| 1527 | |
| 1528 | return buf; |
| 1529 | } else { |
| 1530 | /* upper/lower character may be shorter/longer */ |
| 1531 | buf = tostringN(s, strlen(s) * sz + 1); |
| 1532 | |
| 1533 | memset(&mbs, 0, sizeof(mbs)); |
| 1534 | memset(&mbs2, 0, sizeof(mbs2)); |
| 1535 | |
| 1536 | ps = s; |
| 1537 | pbuf = buf; |
| 1538 | while (n = mbrtowc(&wc, ps, sz, &mbs), |
| 1539 | n > 0 && n != (size_t)-1 && n != (size_t)-2) |
| 1540 | { |
| 1541 | ps += n; |
| 1542 | |
| 1543 | n = wcrtomb(pbuf, fun_wc(wc), &mbs2); |
| 1544 | if (n == (size_t)-1) |
| 1545 | FATAL("illegal wide character %s", s); |
| 1546 | |
| 1547 | pbuf += n; |
| 1548 | } |
| 1549 | |
| 1550 | *pbuf = '\0'; |
| 1551 | |
| 1552 | if (n) |
| 1553 | FATAL("illegal byte sequence %s", s); |
| 1554 | |
| 1555 | return buf; |
| 1556 | } |
| 1557 | } |
| 1558 | |
| 1559 | static char *nawk_toupper(const char *s) |
| 1560 | { |
| 1561 | return nawk_convert(s, toupper, towupper); |
| 1562 | } |
| 1563 | |
| 1564 | static char *nawk_tolower(const char *s) |
| 1565 | { |
| 1566 | return nawk_convert(s, tolower, towlower); |
| 1567 | } |
| 1568 | |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1569 | Cell *bltin(Node **a, int n) /* builtin functions. a[0] is type, a[1] is arg list */ |
| 1570 | { |
| 1571 | Cell *x, *y; |
| 1572 | Awkfloat u; |
| 1573 | int t; |
| 1574 | Awkfloat tmp; |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1575 | char *buf; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1576 | Node *nextarg; |
| 1577 | FILE *fp; |
Arnold D. Robbins | 32093f5 | 2018-08-22 20:40:26 +0300 | [diff] [blame] | 1578 | int status = 0; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1579 | |
| 1580 | t = ptoi(a[0]); |
| 1581 | x = execute(a[1]); |
| 1582 | nextarg = a[1]->nnext; |
| 1583 | switch (t) { |
| 1584 | case FLENGTH: |
| 1585 | if (isarr(x)) |
| 1586 | u = ((Array *) x->sval)->nelem; /* GROT. should be function*/ |
| 1587 | else |
| 1588 | u = strlen(getsval(x)); |
| 1589 | break; |
| 1590 | case FLOG: |
| 1591 | u = errcheck(log(getfval(x)), "log"); break; |
| 1592 | case FINT: |
| 1593 | modf(getfval(x), &u); break; |
| 1594 | case FEXP: |
| 1595 | u = errcheck(exp(getfval(x)), "exp"); break; |
| 1596 | case FSQRT: |
| 1597 | u = errcheck(sqrt(getfval(x)), "sqrt"); break; |
| 1598 | case FSIN: |
| 1599 | u = sin(getfval(x)); break; |
| 1600 | case FCOS: |
| 1601 | u = cos(getfval(x)); break; |
| 1602 | case FATAN: |
pfg | 5242194 | 2016-06-03 21:23:11 +0000 | [diff] [blame] | 1603 | if (nextarg == NULL) { |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1604 | WARNING("atan2 requires two arguments; returning 1.0"); |
| 1605 | u = 1.0; |
| 1606 | } else { |
| 1607 | y = execute(a[1]->nnext); |
| 1608 | u = atan2(getfval(x), getfval(y)); |
| 1609 | tempfree(y); |
| 1610 | nextarg = nextarg->nnext; |
| 1611 | } |
| 1612 | break; |
| 1613 | case FSYSTEM: |
| 1614 | fflush(stdout); /* in case something is buffered already */ |
Arnold D. Robbins | 32093f5 | 2018-08-22 20:40:26 +0300 | [diff] [blame] | 1615 | status = system(getsval(x)); |
| 1616 | u = status; |
| 1617 | if (status != -1) { |
| 1618 | if (WIFEXITED(status)) { |
| 1619 | u = WEXITSTATUS(status); |
| 1620 | } else if (WIFSIGNALED(status)) { |
| 1621 | u = WTERMSIG(status) + 256; |
| 1622 | #ifdef WCOREDUMP |
| 1623 | if (WCOREDUMP(status)) |
| 1624 | u += 256; |
| 1625 | #endif |
| 1626 | } else /* something else?!? */ |
| 1627 | u = 0; |
| 1628 | } |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1629 | break; |
| 1630 | case FRAND: |
pfg | c70b9fe | 2014-09-19 18:24:02 +0000 | [diff] [blame] | 1631 | /* random() returns numbers in [0..2^31-1] |
| 1632 | * in order to get a number in [0, 1), divide it by 2^31 |
| 1633 | */ |
| 1634 | u = (Awkfloat) random() / (0x7fffffffL + 0x1UL); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1635 | break; |
| 1636 | case FSRAND: |
| 1637 | if (isrec(x)) /* no argument provided */ |
| 1638 | u = time((time_t *)0); |
| 1639 | else |
| 1640 | u = getfval(x); |
| 1641 | tmp = u; |
pfg | c70b9fe | 2014-09-19 18:24:02 +0000 | [diff] [blame] | 1642 | srandom((unsigned long) u); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1643 | u = srand_seed; |
| 1644 | srand_seed = tmp; |
| 1645 | break; |
| 1646 | case FTOUPPER: |
| 1647 | case FTOLOWER: |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1648 | if (t == FTOUPPER) |
| 1649 | buf = nawk_toupper(getsval(x)); |
| 1650 | else |
| 1651 | buf = nawk_tolower(getsval(x)); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1652 | tempfree(x); |
| 1653 | x = gettemp(); |
| 1654 | setsval(x, buf); |
| 1655 | free(buf); |
| 1656 | return x; |
| 1657 | case FFLUSH: |
| 1658 | if (isrec(x) || strlen(getsval(x)) == 0) { |
| 1659 | flush_all(); /* fflush() or fflush("") -> all */ |
| 1660 | u = 0; |
| 1661 | } else if ((fp = openfile(FFLUSH, getsval(x))) == NULL) |
| 1662 | u = EOF; |
| 1663 | else |
| 1664 | u = fflush(fp); |
| 1665 | break; |
| 1666 | default: /* can't happen */ |
| 1667 | FATAL("illegal function type %d", t); |
| 1668 | break; |
| 1669 | } |
| 1670 | tempfree(x); |
| 1671 | x = gettemp(); |
| 1672 | setfval(x, u); |
pfg | 5242194 | 2016-06-03 21:23:11 +0000 | [diff] [blame] | 1673 | if (nextarg != NULL) { |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1674 | WARNING("warning: function has too many arguments"); |
| 1675 | for ( ; nextarg; nextarg = nextarg->nnext) |
| 1676 | execute(nextarg); |
| 1677 | } |
| 1678 | return(x); |
| 1679 | } |
| 1680 | |
| 1681 | Cell *printstat(Node **a, int n) /* print a[0] */ |
| 1682 | { |
| 1683 | Node *x; |
| 1684 | Cell *y; |
| 1685 | FILE *fp; |
| 1686 | |
pfg | 5242194 | 2016-06-03 21:23:11 +0000 | [diff] [blame] | 1687 | if (a[1] == NULL) /* a[1] is redirection operator, a[2] is file */ |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1688 | fp = stdout; |
| 1689 | else |
| 1690 | fp = redirect(ptoi(a[1]), a[2]); |
| 1691 | for (x = a[0]; x != NULL; x = x->nnext) { |
| 1692 | y = execute(x); |
| 1693 | fputs(getpssval(y), fp); |
| 1694 | tempfree(y); |
| 1695 | if (x->nnext == NULL) |
Cody Peter Mello | 52566c0 | 2018-09-18 15:45:55 -0700 | [diff] [blame] | 1696 | fputs(getsval(orsloc), fp); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1697 | else |
Cody Peter Mello | 52566c0 | 2018-09-18 15:45:55 -0700 | [diff] [blame] | 1698 | fputs(getsval(ofsloc), fp); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1699 | } |
pfg | 5242194 | 2016-06-03 21:23:11 +0000 | [diff] [blame] | 1700 | if (a[1] != NULL) |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1701 | fflush(fp); |
| 1702 | if (ferror(fp)) |
| 1703 | FATAL("write error on %s", filename(fp)); |
| 1704 | return(True); |
| 1705 | } |
| 1706 | |
| 1707 | Cell *nullproc(Node **a, int n) |
| 1708 | { |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1709 | return 0; |
| 1710 | } |
| 1711 | |
| 1712 | |
| 1713 | FILE *redirect(int a, Node *b) /* set up all i/o redirections */ |
| 1714 | { |
| 1715 | FILE *fp; |
| 1716 | Cell *x; |
| 1717 | char *fname; |
| 1718 | |
| 1719 | x = execute(b); |
| 1720 | fname = getsval(x); |
| 1721 | fp = openfile(a, fname); |
| 1722 | if (fp == NULL) |
| 1723 | FATAL("can't open file %s", fname); |
| 1724 | tempfree(x); |
| 1725 | return fp; |
| 1726 | } |
| 1727 | |
| 1728 | struct files { |
| 1729 | FILE *fp; |
| 1730 | const char *fname; |
| 1731 | int mode; /* '|', 'a', 'w' => LE/LT, GT */ |
| 1732 | } *files; |
| 1733 | |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1734 | size_t nfiles; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1735 | |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1736 | static void stdinit(void) /* in case stdin, etc., are not constants */ |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1737 | { |
| 1738 | nfiles = FOPEN_MAX; |
| 1739 | files = calloc(nfiles, sizeof(*files)); |
| 1740 | if (files == NULL) |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1741 | FATAL("can't allocate file memory for %zu files", nfiles); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1742 | files[0].fp = stdin; |
| 1743 | files[0].fname = "/dev/stdin"; |
| 1744 | files[0].mode = LT; |
| 1745 | files[1].fp = stdout; |
| 1746 | files[1].fname = "/dev/stdout"; |
| 1747 | files[1].mode = GT; |
| 1748 | files[2].fp = stderr; |
| 1749 | files[2].fname = "/dev/stderr"; |
| 1750 | files[2].mode = GT; |
| 1751 | } |
| 1752 | |
| 1753 | FILE *openfile(int a, const char *us) |
| 1754 | { |
| 1755 | const char *s = us; |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1756 | size_t i; |
| 1757 | int m; |
pfg | 5242194 | 2016-06-03 21:23:11 +0000 | [diff] [blame] | 1758 | FILE *fp = NULL; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1759 | |
| 1760 | if (*s == '\0') |
| 1761 | FATAL("null file name in print or getline"); |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1762 | for (i = 0; i < nfiles; i++) |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1763 | if (files[i].fname && strcmp(s, files[i].fname) == 0) { |
| 1764 | if (a == files[i].mode || (a==APPEND && files[i].mode==GT)) |
| 1765 | return files[i].fp; |
| 1766 | if (a == FFLUSH) |
| 1767 | return files[i].fp; |
| 1768 | } |
| 1769 | if (a == FFLUSH) /* didn't find it, so don't create it! */ |
| 1770 | return NULL; |
| 1771 | |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1772 | for (i = 0; i < nfiles; i++) |
pfg | 5242194 | 2016-06-03 21:23:11 +0000 | [diff] [blame] | 1773 | if (files[i].fp == NULL) |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1774 | break; |
| 1775 | if (i >= nfiles) { |
| 1776 | struct files *nf; |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1777 | size_t nnf = nfiles + FOPEN_MAX; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1778 | nf = realloc(files, nnf * sizeof(*nf)); |
| 1779 | if (nf == NULL) |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1780 | FATAL("cannot grow files for %s and %zu files", s, nnf); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1781 | memset(&nf[nfiles], 0, FOPEN_MAX * sizeof(*nf)); |
| 1782 | nfiles = nnf; |
| 1783 | files = nf; |
| 1784 | } |
| 1785 | fflush(stdout); /* force a semblance of order */ |
| 1786 | m = a; |
| 1787 | if (a == GT) { |
| 1788 | fp = fopen(s, "w"); |
| 1789 | } else if (a == APPEND) { |
| 1790 | fp = fopen(s, "a"); |
| 1791 | m = GT; /* so can mix > and >> */ |
| 1792 | } else if (a == '|') { /* output pipe */ |
| 1793 | fp = popen(s, "w"); |
| 1794 | } else if (a == LE) { /* input pipe */ |
| 1795 | fp = popen(s, "r"); |
| 1796 | } else if (a == LT) { /* getline <file */ |
| 1797 | fp = strcmp(s, "-") == 0 ? stdin : fopen(s, "r"); /* "-" is stdin */ |
| 1798 | } else /* can't happen */ |
| 1799 | FATAL("illegal redirection %d", a); |
| 1800 | if (fp != NULL) { |
| 1801 | files[i].fname = tostring(s); |
| 1802 | files[i].fp = fp; |
| 1803 | files[i].mode = m; |
Arnold D. Robbins | 5a18f63 | 2020-01-22 02:10:59 -0700 | [diff] [blame] | 1804 | if (fp != stdin && fp != stdout && fp != stderr) |
| 1805 | (void) fcntl(fileno(fp), F_SETFD, FD_CLOEXEC); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1806 | } |
| 1807 | return fp; |
| 1808 | } |
| 1809 | |
| 1810 | const char *filename(FILE *fp) |
| 1811 | { |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1812 | size_t i; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1813 | |
| 1814 | for (i = 0; i < nfiles; i++) |
| 1815 | if (fp == files[i].fp) |
| 1816 | return files[i].fname; |
| 1817 | return "???"; |
| 1818 | } |
| 1819 | |
Arnold D. Robbins | cd55211 | 2020-02-06 22:32:55 +0200 | [diff] [blame] | 1820 | Cell *closefile(Node **a, int n) |
| 1821 | { |
| 1822 | Cell *x; |
| 1823 | size_t i; |
| 1824 | bool stat; |
| 1825 | |
| 1826 | x = execute(a[0]); |
| 1827 | getsval(x); |
| 1828 | stat = true; |
| 1829 | for (i = 0; i < nfiles; i++) { |
| 1830 | if (!files[i].fname || strcmp(x->sval, files[i].fname) != 0) |
| 1831 | continue; |
| 1832 | if (ferror(files[i].fp)) |
| 1833 | FATAL("i/o error occurred on %s", files[i].fname); |
| 1834 | if (files[i].mode == '|' || files[i].mode == LE) |
| 1835 | stat = pclose(files[i].fp) == -1; |
| 1836 | else |
| 1837 | stat = fclose(files[i].fp) == EOF; |
| 1838 | if (stat) |
| 1839 | FATAL("i/o error occurred closing %s", files[i].fname); |
| 1840 | if (i > 2) /* don't do /dev/std... */ |
| 1841 | xfree(files[i].fname); |
| 1842 | files[i].fname = NULL; /* watch out for ref thru this */ |
| 1843 | files[i].fp = NULL; |
| 1844 | } |
| 1845 | tempfree(x); |
| 1846 | x = gettemp(); |
| 1847 | setfval(x, (Awkfloat) (stat ? -1 : 0)); |
| 1848 | return(x); |
| 1849 | } |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1850 | |
| 1851 | void closeall(void) |
| 1852 | { |
Arnold D. Robbins | e6fe674 | 2020-02-06 22:38:30 +0200 | [diff] [blame] | 1853 | size_t i; |
| 1854 | bool stat = false; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1855 | |
Arnold D. Robbins | e6fe674 | 2020-02-06 22:38:30 +0200 | [diff] [blame] | 1856 | for (i = 0; i < nfiles; i++) { |
| 1857 | if (! files[i].fp) |
| 1858 | continue; |
| 1859 | if (ferror(files[i].fp)) |
| 1860 | FATAL( "i/o error occurred on %s", files[i].fname ); |
| 1861 | if (files[i].mode == '|' || files[i].mode == LE) |
| 1862 | stat = pclose(files[i].fp) == -1; |
| 1863 | else |
| 1864 | stat = fclose(files[i].fp) == EOF; |
| 1865 | if (stat) |
| 1866 | FATAL( "i/o error occurred while closing %s", files[i].fname ); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1867 | } |
| 1868 | } |
| 1869 | |
Arnold D. Robbins | e6fe674 | 2020-02-06 22:38:30 +0200 | [diff] [blame] | 1870 | static void flush_all(void) |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1871 | { |
| 1872 | int i; |
| 1873 | |
| 1874 | for (i = 0; i < nfiles; i++) |
| 1875 | if (files[i].fp) |
| 1876 | fflush(files[i].fp); |
| 1877 | } |
| 1878 | |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 1879 | void backsub(char **pb_ptr, const char **sptr_ptr); |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1880 | |
| 1881 | Cell *sub(Node **a, int nnn) /* substitute command */ |
| 1882 | { |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 1883 | const char *sptr, *q; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1884 | Cell *x, *y, *result; |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 1885 | char *t, *buf, *pb; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1886 | fa *pfa; |
| 1887 | int bufsz = recsize; |
| 1888 | |
Arnold D. Robbins | 938b26c | 2019-10-25 11:01:17 -0400 | [diff] [blame] | 1889 | if ((buf = malloc(bufsz)) == NULL) |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1890 | FATAL("out of memory in sub"); |
| 1891 | x = execute(a[3]); /* target string */ |
| 1892 | t = getsval(x); |
pfg | 5242194 | 2016-06-03 21:23:11 +0000 | [diff] [blame] | 1893 | if (a[0] == NULL) /* 0 => a[1] is already-compiled regexpr */ |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1894 | pfa = (fa *) a[1]; /* regular expression */ |
| 1895 | else { |
| 1896 | y = execute(a[1]); |
| 1897 | pfa = makedfa(getsval(y), 1); |
| 1898 | tempfree(y); |
| 1899 | } |
| 1900 | y = execute(a[2]); /* replacement string */ |
| 1901 | result = False; |
| 1902 | if (pmatch(pfa, t)) { |
| 1903 | sptr = t; |
| 1904 | adjbuf(&buf, &bufsz, 1+patbeg-sptr, recsize, 0, "sub"); |
| 1905 | pb = buf; |
| 1906 | while (sptr < patbeg) |
| 1907 | *pb++ = *sptr++; |
| 1908 | sptr = getsval(y); |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1909 | while (*sptr != '\0') { |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1910 | adjbuf(&buf, &bufsz, 5+pb-buf, recsize, &pb, "sub"); |
| 1911 | if (*sptr == '\\') { |
| 1912 | backsub(&pb, &sptr); |
| 1913 | } else if (*sptr == '&') { |
| 1914 | sptr++; |
| 1915 | adjbuf(&buf, &bufsz, 1+patlen+pb-buf, recsize, &pb, "sub"); |
| 1916 | for (q = patbeg; q < patbeg+patlen; ) |
| 1917 | *pb++ = *q++; |
| 1918 | } else |
| 1919 | *pb++ = *sptr++; |
| 1920 | } |
| 1921 | *pb = '\0'; |
| 1922 | if (pb > buf + bufsz) |
| 1923 | FATAL("sub result1 %.30s too big; can't happen", buf); |
| 1924 | sptr = patbeg + patlen; |
| 1925 | if ((patlen == 0 && *patbeg) || (patlen && *(sptr-1))) { |
| 1926 | adjbuf(&buf, &bufsz, 1+strlen(sptr)+pb-buf, 0, &pb, "sub"); |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1927 | while ((*pb++ = *sptr++) != '\0') |
| 1928 | continue; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1929 | } |
| 1930 | if (pb > buf + bufsz) |
| 1931 | FATAL("sub result2 %.30s too big; can't happen", buf); |
| 1932 | setsval(x, buf); /* BUG: should be able to avoid copy */ |
pfg | 5242194 | 2016-06-03 21:23:11 +0000 | [diff] [blame] | 1933 | result = True; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1934 | } |
| 1935 | tempfree(x); |
| 1936 | tempfree(y); |
| 1937 | free(buf); |
| 1938 | return result; |
| 1939 | } |
| 1940 | |
| 1941 | Cell *gsub(Node **a, int nnn) /* global substitute */ |
| 1942 | { |
| 1943 | Cell *x, *y; |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 1944 | char *rptr, *pb; |
| 1945 | const char *q, *t, *sptr; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1946 | char *buf; |
| 1947 | fa *pfa; |
| 1948 | int mflag, tempstat, num; |
| 1949 | int bufsz = recsize; |
| 1950 | |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 1951 | if ((buf = malloc(bufsz)) == NULL) |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1952 | FATAL("out of memory in gsub"); |
| 1953 | mflag = 0; /* if mflag == 0, can replace empty string */ |
| 1954 | num = 0; |
| 1955 | x = execute(a[3]); /* target string */ |
| 1956 | t = getsval(x); |
pfg | 5242194 | 2016-06-03 21:23:11 +0000 | [diff] [blame] | 1957 | if (a[0] == NULL) /* 0 => a[1] is already-compiled regexpr */ |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1958 | pfa = (fa *) a[1]; /* regular expression */ |
| 1959 | else { |
| 1960 | y = execute(a[1]); |
| 1961 | pfa = makedfa(getsval(y), 1); |
| 1962 | tempfree(y); |
| 1963 | } |
| 1964 | y = execute(a[2]); /* replacement string */ |
| 1965 | if (pmatch(pfa, t)) { |
| 1966 | tempstat = pfa->initstat; |
| 1967 | pfa->initstat = 2; |
| 1968 | pb = buf; |
| 1969 | rptr = getsval(y); |
| 1970 | do { |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1971 | if (patlen == 0 && *patbeg != '\0') { /* matched empty string */ |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1972 | if (mflag == 0) { /* can replace empty */ |
| 1973 | num++; |
| 1974 | sptr = rptr; |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1975 | while (*sptr != '\0') { |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1976 | adjbuf(&buf, &bufsz, 5+pb-buf, recsize, &pb, "gsub"); |
| 1977 | if (*sptr == '\\') { |
| 1978 | backsub(&pb, &sptr); |
| 1979 | } else if (*sptr == '&') { |
| 1980 | sptr++; |
| 1981 | adjbuf(&buf, &bufsz, 1+patlen+pb-buf, recsize, &pb, "gsub"); |
| 1982 | for (q = patbeg; q < patbeg+patlen; ) |
| 1983 | *pb++ = *q++; |
| 1984 | } else |
| 1985 | *pb++ = *sptr++; |
| 1986 | } |
| 1987 | } |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 1988 | if (*t == '\0') /* at end */ |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 1989 | goto done; |
| 1990 | adjbuf(&buf, &bufsz, 2+pb-buf, recsize, &pb, "gsub"); |
| 1991 | *pb++ = *t++; |
| 1992 | if (pb > buf + bufsz) /* BUG: not sure of this test */ |
| 1993 | FATAL("gsub result0 %.30s too big; can't happen", buf); |
| 1994 | mflag = 0; |
| 1995 | } |
| 1996 | else { /* matched nonempty string */ |
| 1997 | num++; |
| 1998 | sptr = t; |
| 1999 | adjbuf(&buf, &bufsz, 1+(patbeg-sptr)+pb-buf, recsize, &pb, "gsub"); |
| 2000 | while (sptr < patbeg) |
| 2001 | *pb++ = *sptr++; |
| 2002 | sptr = rptr; |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 2003 | while (*sptr != '\0') { |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 2004 | adjbuf(&buf, &bufsz, 5+pb-buf, recsize, &pb, "gsub"); |
| 2005 | if (*sptr == '\\') { |
| 2006 | backsub(&pb, &sptr); |
| 2007 | } else if (*sptr == '&') { |
| 2008 | sptr++; |
| 2009 | adjbuf(&buf, &bufsz, 1+patlen+pb-buf, recsize, &pb, "gsub"); |
| 2010 | for (q = patbeg; q < patbeg+patlen; ) |
| 2011 | *pb++ = *q++; |
| 2012 | } else |
| 2013 | *pb++ = *sptr++; |
| 2014 | } |
| 2015 | t = patbeg + patlen; |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 2016 | if (patlen == 0 || *t == '\0' || *(t-1) == '\0') |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 2017 | goto done; |
| 2018 | if (pb > buf + bufsz) |
| 2019 | FATAL("gsub result1 %.30s too big; can't happen", buf); |
| 2020 | mflag = 1; |
| 2021 | } |
| 2022 | } while (pmatch(pfa,t)); |
| 2023 | sptr = t; |
| 2024 | adjbuf(&buf, &bufsz, 1+strlen(sptr)+pb-buf, 0, &pb, "gsub"); |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 2025 | while ((*pb++ = *sptr++) != '\0') |
| 2026 | continue; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 2027 | done: if (pb < buf + bufsz) |
| 2028 | *pb = '\0'; |
| 2029 | else if (*(pb-1) != '\0') |
| 2030 | FATAL("gsub result2 %.30s truncated; can't happen", buf); |
| 2031 | setsval(x, buf); /* BUG: should be able to avoid copy + free */ |
| 2032 | pfa->initstat = tempstat; |
| 2033 | } |
| 2034 | tempfree(x); |
| 2035 | tempfree(y); |
| 2036 | x = gettemp(); |
| 2037 | x->tval = NUM; |
| 2038 | x->fval = num; |
| 2039 | free(buf); |
| 2040 | return(x); |
| 2041 | } |
| 2042 | |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 2043 | void backsub(char **pb_ptr, const char **sptr_ptr) /* handle \\& variations */ |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 2044 | { /* sptr[0] == '\\' */ |
zoulasc | 6589208 | 2019-10-24 09:40:15 -0400 | [diff] [blame] | 2045 | char *pb = *pb_ptr; |
| 2046 | const char *sptr = *sptr_ptr; |
Arnold D. Robbins | de6284e | 2020-01-19 20:37:33 +0200 | [diff] [blame] | 2047 | static bool first = true; |
| 2048 | static bool do_posix = false; |
| 2049 | |
| 2050 | if (first) { |
| 2051 | first = false; |
| 2052 | do_posix = (getenv("POSIXLY_CORRECT") != NULL); |
| 2053 | } |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 2054 | |
| 2055 | if (sptr[1] == '\\') { |
| 2056 | if (sptr[2] == '\\' && sptr[3] == '&') { /* \\\& -> \& */ |
| 2057 | *pb++ = '\\'; |
| 2058 | *pb++ = '&'; |
| 2059 | sptr += 4; |
| 2060 | } else if (sptr[2] == '&') { /* \\& -> \ + matched */ |
| 2061 | *pb++ = '\\'; |
| 2062 | sptr += 2; |
Arnold D. Robbins | de6284e | 2020-01-19 20:37:33 +0200 | [diff] [blame] | 2063 | } else if (do_posix) { /* \\x -> \x */ |
| 2064 | sptr++; |
| 2065 | *pb++ = *sptr++; |
Brian Kernighan | 87b9493 | 2012-12-22 10:35:39 -0500 | [diff] [blame] | 2066 | } else { /* \\x -> \\x */ |
| 2067 | *pb++ = *sptr++; |
| 2068 | *pb++ = *sptr++; |
| 2069 | } |
| 2070 | } else if (sptr[1] == '&') { /* literal & */ |
| 2071 | sptr++; |
| 2072 | *pb++ = *sptr++; |
| 2073 | } else /* literal \ */ |
| 2074 | *pb++ = *sptr++; |
| 2075 | |
| 2076 | *pb_ptr = pb; |
| 2077 | *sptr_ptr = sptr; |
| 2078 | } |