Print this page
4818 printf(1) should support n$ width and precision specifiers
4854 printf(1) doesn't support %b and \c properly
Reviewed by: Keith Wesolowski <keith.wesolowski@joyent.com>
Approved by: TBD
| Split |
Close |
| Expand all |
| Collapse all |
--- old/usr/src/cmd/printf/printf.c
+++ new/usr/src/cmd/printf/printf.c
1 1 /*
2 + * Copyright 2014 Garrett D'Amore <garrett@damore.org>
2 3 * Copyright 2010 Nexenta Systems, Inc. All rights reserved.
3 4 * Copyright (c) 1989, 1993
4 5 * The Regents of the University of California. All rights reserved.
5 6 *
6 7 * Redistribution and use in source and binary forms, with or without
7 8 * modification, are permitted provided that the following conditions
8 9 * are met:
9 10 * 1. Redistributions of source code must retain the above copyright
10 11 * notice, this list of conditions and the following disclaimer.
11 12 * 2. Redistributions in binary form must reproduce the above copyright
12 13 * notice, this list of conditions and the following disclaimer in the
13 14 * documentation and/or other materials provided with the distribution.
14 15 * 4. Neither the name of the University nor the names of its contributors
15 16 * may be used to endorse or promote products derived from this software
16 17 * without specific prior written permission.
17 18 *
18 19 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
19 20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
20 21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
21 22 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
22 23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
23 24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
24 25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
25 26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
26 27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
27 28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
28 29 * SUCH DAMAGE.
29 30 */
30 31
|
↓ open down ↓ |
19 lines elided |
↑ open up ↑ |
31 32 #include <sys/types.h>
32 33
33 34 #include <err.h>
34 35 #include <errno.h>
35 36 #include <inttypes.h>
36 37 #include <limits.h>
37 38 #include <stdio.h>
38 39 #include <stdlib.h>
39 40 #include <string.h>
40 41 #include <unistd.h>
42 +#include <alloca.h>
43 +#include <ctype.h>
41 44 #include <locale.h>
42 45 #include <note.h>
43 46
44 47 #define warnx1(a, b, c) warnx(a)
45 48 #define warnx2(a, b, c) warnx(a, b)
46 49 #define warnx3(a, b, c) warnx(a, b, c)
47 50
48 51 #define PTRDIFF(x, y) ((uintptr_t)(x) - (uintptr_t)(y))
49 52
50 53 #define _(x) gettext(x)
51 54
52 55 #define PF(f, func) do { \
53 56 char *b = NULL; \
54 - int dollar = 0; \
55 - if (*f == '$') { \
56 - dollar++; \
57 - *f = '%'; \
58 - } \
59 57 if (havewidth) \
60 58 if (haveprec) \
61 59 (void) asprintf(&b, f, fieldwidth, precision, func); \
62 60 else \
63 61 (void) asprintf(&b, f, fieldwidth, func); \
64 62 else if (haveprec) \
65 63 (void) asprintf(&b, f, precision, func); \
66 64 else \
67 65 (void) asprintf(&b, f, func); \
68 66 if (b) { \
69 67 (void) fputs(b, stdout); \
70 68 free(b); \
71 69 } \
72 - if (dollar) \
73 - *f = '$'; \
74 70 _NOTE(CONSTCOND) } while (0)
75 71
76 72 static int asciicode(void);
77 73 static char *doformat(char *, int *);
78 74 static int escape(char *, int, size_t *);
79 75 static int getchr(void);
80 76 static int getfloating(long double *, int);
81 77 static int getint(int *);
82 78 static int getnum(intmax_t *, uintmax_t *, int);
83 79 static const char
84 80 *getstr(void);
85 81 static char *mknum(char *, char);
86 82 static void usage(void);
87 83
84 +static const char digits[] = "0123456789";
85 +
88 86 static int myargc;
89 87 static char **myargv;
90 88 static char **gargv;
89 +static char **maxargv;
91 90
92 91 int
93 92 main(int argc, char *argv[])
94 93 {
95 94 size_t len;
96 - int chopped, end, rval;
95 + int end, rval;
97 96 char *format, *fmt, *start;
98 97
99 98 (void) setlocale(LC_ALL, "");
100 99
101 100 argv++;
102 101 argc--;
103 102
104 103 /*
105 104 * POSIX says: Standard utilities that do not accept options,
106 105 * but that do accept operands, shall recognize "--" as a
107 106 * first argument to be discarded.
108 107 */
109 108 if (argc && strcmp(argv[0], "--") == 0) {
110 109 argc--;
111 110 argv++;
112 111 }
113 112
114 113 if (argc < 1) {
115 114 usage();
116 115 return (1);
117 116 }
|
↓ open down ↓ |
11 lines elided |
↑ open up ↑ |
118 117
119 118 /*
120 119 * Basic algorithm is to scan the format string for conversion
121 120 * specifications -- once one is found, find out if the field
122 121 * width or precision is a '*'; if it is, gather up value. Note,
123 122 * format strings are reused as necessary to use up the provided
124 123 * arguments, arguments of zero/null string are provided to use
125 124 * up the format string.
126 125 */
127 126 fmt = format = *argv;
128 - chopped = escape(fmt, 1, &len); /* backslash interpretation */
127 + (void) escape(fmt, 1, &len); /* backslash interpretation */
129 128 rval = end = 0;
130 129 gargv = ++argv;
131 130
132 131 for (;;) {
133 - char **maxargv = gargv;
132 + maxargv = gargv;
134 133
135 134 myargv = gargv;
136 135 for (myargc = 0; gargv[myargc]; myargc++)
137 136 /* nop */;
138 137 start = fmt;
139 138 while (fmt < format + len) {
140 139 if (fmt[0] == '%') {
141 140 (void) fwrite(start, 1, PTRDIFF(fmt, start),
142 141 stdout);
143 142 if (fmt[1] == '%') {
144 143 /* %% prints a % */
145 144 (void) putchar('%');
146 145 fmt += 2;
147 146 } else {
148 147 fmt = doformat(fmt, &rval);
149 148 if (fmt == NULL)
150 149 return (1);
151 150 end = 0;
152 151 }
153 152 start = fmt;
154 153 } else
155 154 fmt++;
|
↓ open down ↓ |
12 lines elided |
↑ open up ↑ |
156 155 if (gargv > maxargv)
157 156 maxargv = gargv;
158 157 }
159 158 gargv = maxargv;
160 159
161 160 if (end == 1) {
162 161 warnx1(_("missing format character"), NULL, NULL);
163 162 return (1);
164 163 }
165 164 (void) fwrite(start, 1, PTRDIFF(fmt, start), stdout);
166 - if (chopped || !*gargv)
165 + if (!*gargv)
167 166 return (rval);
168 167 /* Restart at the beginning of the format string. */
169 168 fmt = format;
170 169 end = 1;
171 170 }
172 171 /* NOTREACHED */
173 172 }
174 173
175 174
176 175 static char *
177 -doformat(char *start, int *rval)
176 +doformat(char *fmt, int *rval)
178 177 {
179 178 static const char skip1[] = "#'-+ 0";
180 - static const char skip2[] = "0123456789";
181 - char *fmt;
182 179 int fieldwidth, haveprec, havewidth, mod_ldbl, precision;
183 180 char convch, nextch;
181 + char *start;
182 + char **fargv;
183 + char *dptr;
184 + int l;
184 185
185 - fmt = start + 1;
186 + start = alloca(strlen(fmt) + 1);
186 187
188 + dptr = start;
189 + *dptr++ = '%';
190 + *dptr = 0;
191 +
192 + fmt++;
193 +
187 194 /* look for "n$" field index specifier */
188 - fmt += strspn(fmt, skip2);
189 - if ((*fmt == '$') && (fmt != (start + 1))) {
190 - int idx = atoi(start + 1);
195 + l = strspn(fmt, digits);
196 + if ((l > 0) && (fmt[l] == '$')) {
197 + int idx = atoi(fmt);
191 198 if (idx <= myargc) {
192 199 gargv = &myargv[idx - 1];
193 200 } else {
194 201 gargv = &myargv[myargc];
195 202 }
196 - start = fmt;
197 - fmt++;
203 + if (gargv > maxargv) {
204 + maxargv = gargv;
205 + }
206 + fmt += l + 1;
207 +
208 + /* save format argument */
209 + fargv = gargv;
198 210 } else {
199 - fmt = start + 1;
211 + fargv = NULL;
200 212 }
201 213
202 214 /* skip to field width */
203 - fmt += strspn(fmt, skip1);
215 + while (*fmt && strchr(skip1, *fmt) != NULL) {
216 + *dptr++ = *fmt++;
217 + *dptr = 0;
218 + }
219 +
220 +
204 221 if (*fmt == '*') {
222 +
223 + fmt++;
224 + l = strspn(fmt, digits);
225 + if ((l > 0) && (fmt[l] == '$')) {
226 + int idx = atoi(fmt);
227 + if (fargv == NULL) {
228 + warnx1(_("incomplete use of n$"), NULL, NULL);
229 + return (NULL);
230 + }
231 + if (idx <= myargc) {
232 + gargv = &myargv[idx - 1];
233 + } else {
234 + gargv = &myargv[myargc];
235 + }
236 + fmt += l + 1;
237 + } else if (fargv != NULL) {
238 + warnx1(_("incomplete use of n$"), NULL, NULL);
239 + return (NULL);
240 + }
241 +
205 242 if (getint(&fieldwidth))
206 243 return (NULL);
244 + if (gargv > maxargv) {
245 + maxargv = gargv;
246 + }
207 247 havewidth = 1;
208 - ++fmt;
248 +
249 + *dptr++ = '*';
250 + *dptr = 0;
209 251 } else {
210 252 havewidth = 0;
211 253
212 254 /* skip to possible '.', get following precision */
213 - fmt += strspn(fmt, skip2);
255 + while (isdigit(*fmt)) {
256 + *dptr++ = *fmt++;
257 + *dptr = 0;
258 + }
214 259 }
260 +
215 261 if (*fmt == '.') {
216 262 /* precision present? */
217 - ++fmt;
263 + fmt++;
264 + *dptr++ = '.';
265 +
218 266 if (*fmt == '*') {
267 +
268 + fmt++;
269 + l = strspn(fmt, digits);
270 + if ((l > 0) && (fmt[l] == '$')) {
271 + int idx = atoi(fmt);
272 + if (fargv == NULL) {
273 + warnx1(_("incomplete use of n$"),
274 + NULL, NULL);
275 + return (NULL);
276 + }
277 + if (idx <= myargc) {
278 + gargv = &myargv[idx - 1];
279 + } else {
280 + gargv = &myargv[myargc];
281 + }
282 + fmt += l + 1;
283 + } else if (fargv != NULL) {
284 + warnx1(_("incomplete use of n$"), NULL, NULL);
285 + return (NULL);
286 + }
287 +
219 288 if (getint(&precision))
220 289 return (NULL);
290 + if (gargv > maxargv) {
291 + maxargv = gargv;
292 + }
221 293 haveprec = 1;
222 - ++fmt;
294 + *dptr++ = '*';
295 + *dptr = 0;
223 296 } else {
224 297 haveprec = 0;
225 298
226 299 /* skip to conversion char */
227 - fmt += strspn(fmt, skip2);
300 + while (isdigit(*fmt)) {
301 + *dptr++ = *fmt++;
302 + *dptr = 0;
303 + }
228 304 }
229 305 } else
230 306 haveprec = 0;
231 307 if (!*fmt) {
232 308 warnx1(_("missing format character"), NULL, NULL);
233 309 return (NULL);
234 310 }
311 + *dptr++ = *fmt;
312 + *dptr = 0;
235 313
236 314 /*
237 315 * Look for a length modifier. POSIX doesn't have these, so
238 316 * we only support them for floating-point conversions, which
239 317 * are extensions. This is useful because the L modifier can
240 318 * be used to gain extra range and precision, while omitting
241 319 * it is more likely to produce consistent results on different
242 320 * architectures. This is not so important for integers
243 321 * because overflow is the only bad thing that can happen to
244 322 * them, but consider the command printf %a 1.1
245 323 */
246 324 if (*fmt == 'L') {
|
↓ open down ↓ |
2 lines elided |
↑ open up ↑ |
247 325 mod_ldbl = 1;
248 326 fmt++;
249 327 if (!strchr("aAeEfFgG", *fmt)) {
250 328 warnx2(_("bad modifier L for %%%c"), *fmt, NULL);
251 329 return (NULL);
252 330 }
253 331 } else {
254 332 mod_ldbl = 0;
255 333 }
256 334
335 + /* save the current arg offset, and set to the format arg */
336 + if (fargv != NULL) {
337 + gargv = fargv;
338 + }
339 +
257 340 convch = *fmt;
258 341 nextch = *++fmt;
342 +
259 343 *fmt = '\0';
260 344 switch (convch) {
261 345 case 'b': {
262 346 size_t len;
263 347 char *p;
264 348 int getout;
265 349
266 350 p = strdup(getstr());
267 351 if (p == NULL) {
268 352 warnx2("%s", strerror(ENOMEM), NULL);
269 353 return (NULL);
270 354 }
271 355 getout = escape(p, 0, &len);
272 - *(fmt - 1) = 's';
273 - PF(start, p);
274 - *(fmt - 1) = 'b';
356 + (void) fputs(p, stdout);
275 357 free(p);
276 358
277 359 if (getout)
278 - return (fmt);
360 + exit(*rval);
279 361 break;
280 362 }
281 363 case 'c': {
282 364 char p;
283 365
284 366 p = getchr();
285 367 PF(start, p);
286 368 break;
287 369 }
288 370 case 's': {
289 371 const char *p;
290 372
291 373 p = getstr();
292 374 PF(start, p);
293 375 break;
294 376 }
295 377 case 'd': case 'i': case 'o': case 'u': case 'x': case 'X': {
296 378 char *f;
297 379 intmax_t val;
298 380 uintmax_t uval;
299 381 int signedconv;
300 382
301 383 signedconv = (convch == 'd' || convch == 'i');
302 384 if ((f = mknum(start, convch)) == NULL)
303 385 return (NULL);
304 386 if (getnum(&val, &uval, signedconv))
305 387 *rval = 1;
306 388 if (signedconv)
307 389 PF(f, val);
308 390 else
309 391 PF(f, uval);
310 392 break;
311 393 }
312 394 case 'e': case 'E':
313 395 case 'f': case 'F':
314 396 case 'g': case 'G':
315 397 case 'a': case 'A': {
316 398 long double p;
317 399
318 400 if (getfloating(&p, mod_ldbl))
319 401 *rval = 1;
320 402 if (mod_ldbl)
|
↓ open down ↓ |
32 lines elided |
↑ open up ↑ |
321 403 PF(start, p);
322 404 else
323 405 PF(start, (double)p);
324 406 break;
325 407 }
326 408 default:
327 409 warnx2(_("illegal format character %c"), convch, NULL);
328 410 return (NULL);
329 411 }
330 412 *fmt = nextch;
413 +
414 + /* return the gargv to the next element */
331 415 return (fmt);
332 416 }
333 417
334 418 static char *
335 419 mknum(char *str, char ch)
336 420 {
337 421 static char *copy;
338 422 static size_t copy_size;
339 423 char *newcopy;
340 424 size_t len, newlen;
341 425
342 426 len = strlen(str) + 2;
343 427 if (len > copy_size) {
344 428 newlen = ((len + 1023) >> 10) << 10;
345 429 if ((newcopy = realloc(copy, newlen)) == NULL) {
346 430 warnx2("%s", strerror(ENOMEM), NULL);
347 431 return (NULL);
348 432 }
349 433 copy = newcopy;
350 434 copy_size = newlen;
351 435 }
352 436
353 437 (void) memmove(copy, str, len - 3);
354 438 copy[len - 3] = 'j';
355 439 copy[len - 2] = ch;
356 440 copy[len - 1] = '\0';
357 441 return (copy);
358 442 }
359 443
360 444 static int
361 445 escape(char *fmt, int percent, size_t *len)
362 446 {
363 447 char *save, *store, c;
364 448 int value;
365 449
366 450 for (save = store = fmt; ((c = *fmt) != 0); ++fmt, ++store) {
367 451 if (c != '\\') {
368 452 *store = c;
369 453 continue;
370 454 }
371 455 switch (*++fmt) {
372 456 case '\0': /* EOS, user error */
373 457 *store = '\\';
374 458 *++store = '\0';
375 459 *len = PTRDIFF(store, save);
376 460 return (0);
377 461 case '\\': /* backslash */
|
↓ open down ↓ |
37 lines elided |
↑ open up ↑ |
378 462 case '\'': /* single quote */
379 463 *store = *fmt;
380 464 break;
381 465 case 'a': /* bell/alert */
382 466 *store = '\a';
383 467 break;
384 468 case 'b': /* backspace */
385 469 *store = '\b';
386 470 break;
387 471 case 'c':
388 - *store = '\0';
389 - *len = PTRDIFF(store, save);
390 - return (1);
472 + if (!percent) {
473 + *store = '\0';
474 + *len = PTRDIFF(store, save);
475 + return (1);
476 + }
477 + *store = 'c';
478 + break;
391 479 case 'f': /* form-feed */
392 480 *store = '\f';
393 481 break;
394 482 case 'n': /* newline */
395 483 *store = '\n';
396 484 break;
397 485 case 'r': /* carriage-return */
398 486 *store = '\r';
399 487 break;
400 488 case 't': /* horizontal tab */
401 489 *store = '\t';
402 490 break;
403 491 case 'v': /* vertical tab */
404 492 *store = '\v';
405 493 break;
406 494 /* octal constant */
407 495 case '0': case '1': case '2': case '3':
408 496 case '4': case '5': case '6': case '7':
409 497 c = (!percent && *fmt == '0') ? 4 : 3;
410 498 for (value = 0;
411 499 c-- && *fmt >= '0' && *fmt <= '7'; ++fmt) {
412 500 value <<= 3;
413 501 value += *fmt - '0';
414 502 }
415 503 --fmt;
416 504 if (percent && value == '%') {
417 505 *store++ = '%';
418 506 *store = '%';
419 507 } else
420 508 *store = (char)value;
421 509 break;
422 510 default:
423 511 *store = *fmt;
424 512 break;
425 513 }
426 514 }
427 515 *store = '\0';
428 516 *len = PTRDIFF(store, save);
429 517 return (0);
430 518 }
431 519
432 520 static int
433 521 getchr(void)
434 522 {
435 523 if (!*gargv)
436 524 return ('\0');
437 525 return ((int)**gargv++);
438 526 }
439 527
440 528 static const char *
441 529 getstr(void)
442 530 {
443 531 if (!*gargv)
444 532 return ("");
445 533 return (*gargv++);
446 534 }
447 535
448 536 static int
449 537 getint(int *ip)
450 538 {
451 539 intmax_t val;
452 540 uintmax_t uval;
453 541 int rval;
454 542
455 543 if (getnum(&val, &uval, 1))
456 544 return (1);
457 545 rval = 0;
458 546 if (val < INT_MIN || val > INT_MAX) {
459 547 warnx3("%s: %s", *gargv, strerror(ERANGE));
460 548 rval = 1;
461 549 }
462 550 *ip = (int)val;
463 551 return (rval);
464 552 }
465 553
466 554 static int
467 555 getnum(intmax_t *ip, uintmax_t *uip, int signedconv)
468 556 {
469 557 char *ep;
470 558 int rval;
471 559
472 560 if (!*gargv) {
473 561 *ip = 0;
474 562 return (0);
475 563 }
476 564 if (**gargv == '"' || **gargv == '\'') {
477 565 if (signedconv)
478 566 *ip = asciicode();
479 567 else
480 568 *uip = asciicode();
481 569 return (0);
482 570 }
483 571 rval = 0;
484 572 errno = 0;
485 573 if (signedconv)
486 574 *ip = strtoimax(*gargv, &ep, 0);
487 575 else
488 576 *uip = strtoumax(*gargv, &ep, 0);
489 577 if (ep == *gargv) {
490 578 warnx2(_("%s: expected numeric value"), *gargv, NULL);
491 579 rval = 1;
492 580 } else if (*ep != '\0') {
493 581 warnx2(_("%s: not completely converted"), *gargv, NULL);
494 582 rval = 1;
495 583 }
496 584 if (errno == ERANGE) {
497 585 warnx3("%s: %s", *gargv, strerror(ERANGE));
498 586 rval = 1;
499 587 }
500 588 ++gargv;
501 589 return (rval);
502 590 }
503 591
504 592 static int
505 593 getfloating(long double *dp, int mod_ldbl)
506 594 {
507 595 char *ep;
508 596 int rval;
509 597
510 598 if (!*gargv) {
511 599 *dp = 0.0;
512 600 return (0);
513 601 }
514 602 if (**gargv == '"' || **gargv == '\'') {
515 603 *dp = asciicode();
516 604 return (0);
517 605 }
518 606 rval = 0;
519 607 errno = 0;
520 608 if (mod_ldbl)
521 609 *dp = strtold(*gargv, &ep);
522 610 else
523 611 *dp = strtod(*gargv, &ep);
524 612 if (ep == *gargv) {
525 613 warnx2(_("%s: expected numeric value"), *gargv, NULL);
526 614 rval = 1;
527 615 } else if (*ep != '\0') {
528 616 warnx2(_("%s: not completely converted"), *gargv, NULL);
529 617 rval = 1;
530 618 }
531 619 if (errno == ERANGE) {
532 620 warnx3("%s: %s", *gargv, strerror(ERANGE));
533 621 rval = 1;
534 622 }
535 623 ++gargv;
536 624 return (rval);
537 625 }
538 626
539 627 static int
540 628 asciicode(void)
541 629 {
542 630 int ch;
543 631
544 632 ch = **gargv;
545 633 if (ch == '\'' || ch == '"')
546 634 ch = (*gargv)[1];
547 635 ++gargv;
548 636 return (ch);
549 637 }
550 638
551 639 static void
552 640 usage(void)
553 641 {
554 642 (void) fprintf(stderr, _("usage: printf format [arguments ...]\n"));
555 643 }
|
↓ open down ↓ |
155 lines elided |
↑ open up ↑ |
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX