Annotation of /trunk/mkinitrd-magellan/busybox/coreutils/od_bloaty.c
Parent Directory | Revision Log
Revision 532 -
(hide annotations)
(download)
Sat Sep 1 22:45:15 2007 UTC (16 years, 8 months ago) by niro
File MIME type: text/plain
File size: 38809 byte(s)
Sat Sep 1 22:45:15 2007 UTC (16 years, 8 months ago) by niro
File MIME type: text/plain
File size: 38809 byte(s)
-import if magellan mkinitrd; it is a fork of redhats mkinitrd-5.0.8 with all magellan patches and features; deprecates magellan-src/mkinitrd
1 | niro | 532 | /* od -- dump files in octal and other formats |
2 | Copyright (C) 92, 1995-2004 Free Software Foundation, Inc. | ||
3 | |||
4 | This program is free software; you can redistribute it and/or modify | ||
5 | it under the terms of the GNU General Public License as published by | ||
6 | the Free Software Foundation; either version 2, or (at your option) | ||
7 | any later version. | ||
8 | |||
9 | This program is distributed in the hope that it will be useful, | ||
10 | but WITHOUT ANY WARRANTY; without even the implied warranty of | ||
11 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | ||
12 | GNU General Public License for more details. | ||
13 | |||
14 | You should have received a copy of the GNU General Public License | ||
15 | along with this program; if not, write to the Free Software Foundation, | ||
16 | Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ | ||
17 | |||
18 | /* Written by Jim Meyering. */ | ||
19 | |||
20 | /* Busyboxed by Denis Vlasenko | ||
21 | |||
22 | Based on od.c from coreutils-5.2.1 | ||
23 | Top bloat sources: | ||
24 | |||
25 | 00000073 t parse_old_offset | ||
26 | 0000007b t get_lcm | ||
27 | 00000090 r long_options | ||
28 | 00000092 t print_named_ascii | ||
29 | 000000bf t print_ascii | ||
30 | 00000168 t write_block | ||
31 | 00000366 t decode_format_string | ||
32 | 00000a71 T od_main | ||
33 | |||
34 | Tested for compat with coreutils 6.3 | ||
35 | using this script. Minor differences fixed. | ||
36 | |||
37 | #!/bin/sh | ||
38 | echo STD | ||
39 | time /path/to/coreutils/od \ | ||
40 | ...params... \ | ||
41 | >std | ||
42 | echo Exit code $? | ||
43 | echo BBOX | ||
44 | time ./busybox od \ | ||
45 | ...params... \ | ||
46 | >bbox | ||
47 | echo Exit code $? | ||
48 | diff -u -a std bbox >bbox.diff || { echo Different!; sleep 1; } | ||
49 | |||
50 | */ | ||
51 | |||
52 | |||
53 | #include "busybox.h" | ||
54 | #include <getopt.h> | ||
55 | |||
56 | #define assert(a) ((void)0) | ||
57 | |||
58 | /* Check for 0x7f is a coreutils 6.3 addition */ | ||
59 | #define ISPRINT(c) (((c)>=' ') && (c) != 0x7f) | ||
60 | |||
61 | typedef long double longdouble_t; | ||
62 | typedef unsigned long long ulonglong_t; | ||
63 | typedef long long llong; | ||
64 | |||
65 | #if ENABLE_LFS | ||
66 | # define xstrtooff_sfx xstrtoull_sfx | ||
67 | #else | ||
68 | # define xstrtooff_sfx xstrtoul_sfx | ||
69 | #endif | ||
70 | |||
71 | /* The default number of input bytes per output line. */ | ||
72 | #define DEFAULT_BYTES_PER_BLOCK 16 | ||
73 | |||
74 | /* The number of decimal digits of precision in a float. */ | ||
75 | #ifndef FLT_DIG | ||
76 | # define FLT_DIG 7 | ||
77 | #endif | ||
78 | |||
79 | /* The number of decimal digits of precision in a double. */ | ||
80 | #ifndef DBL_DIG | ||
81 | # define DBL_DIG 15 | ||
82 | #endif | ||
83 | |||
84 | /* The number of decimal digits of precision in a long double. */ | ||
85 | #ifndef LDBL_DIG | ||
86 | # define LDBL_DIG DBL_DIG | ||
87 | #endif | ||
88 | |||
89 | enum size_spec { | ||
90 | NO_SIZE, | ||
91 | CHAR, | ||
92 | SHORT, | ||
93 | INT, | ||
94 | LONG, | ||
95 | LONG_LONG, | ||
96 | FLOAT_SINGLE, | ||
97 | FLOAT_DOUBLE, | ||
98 | FLOAT_LONG_DOUBLE, | ||
99 | N_SIZE_SPECS | ||
100 | }; | ||
101 | |||
102 | enum output_format { | ||
103 | SIGNED_DECIMAL, | ||
104 | UNSIGNED_DECIMAL, | ||
105 | OCTAL, | ||
106 | HEXADECIMAL, | ||
107 | FLOATING_POINT, | ||
108 | NAMED_CHARACTER, | ||
109 | CHARACTER | ||
110 | }; | ||
111 | |||
112 | /* Each output format specification (from '-t spec' or from | ||
113 | old-style options) is represented by one of these structures. */ | ||
114 | struct tspec { | ||
115 | enum output_format fmt; | ||
116 | enum size_spec size; | ||
117 | void (*print_function) (size_t, const char *, const char *); | ||
118 | char *fmt_string; | ||
119 | int hexl_mode_trailer; | ||
120 | int field_width; | ||
121 | }; | ||
122 | |||
123 | /* Convert the number of 8-bit bytes of a binary representation to | ||
124 | the number of characters (digits + sign if the type is signed) | ||
125 | required to represent the same quantity in the specified base/type. | ||
126 | For example, a 32-bit (4-byte) quantity may require a field width | ||
127 | as wide as the following for these types: | ||
128 | 11 unsigned octal | ||
129 | 11 signed decimal | ||
130 | 10 unsigned decimal | ||
131 | 8 unsigned hexadecimal */ | ||
132 | |||
133 | static const uint8_t bytes_to_oct_digits[] = | ||
134 | {0, 3, 6, 8, 11, 14, 16, 19, 22, 25, 27, 30, 32, 35, 38, 41, 43}; | ||
135 | |||
136 | static const uint8_t bytes_to_signed_dec_digits[] = | ||
137 | {1, 4, 6, 8, 11, 13, 16, 18, 20, 23, 25, 28, 30, 33, 35, 37, 40}; | ||
138 | |||
139 | static const uint8_t bytes_to_unsigned_dec_digits[] = | ||
140 | {0, 3, 5, 8, 10, 13, 15, 17, 20, 22, 25, 27, 29, 32, 34, 37, 39}; | ||
141 | |||
142 | static const uint8_t bytes_to_hex_digits[] = | ||
143 | {0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30, 32}; | ||
144 | |||
145 | /* Convert enum size_spec to the size of the named type. */ | ||
146 | static const signed char width_bytes[] = { | ||
147 | -1, | ||
148 | sizeof(char), | ||
149 | sizeof(short), | ||
150 | sizeof(int), | ||
151 | sizeof(long), | ||
152 | sizeof(ulonglong_t), | ||
153 | sizeof(float), | ||
154 | sizeof(double), | ||
155 | sizeof(longdouble_t) | ||
156 | }; | ||
157 | |||
158 | /* Ensure that for each member of 'enum size_spec' there is an | ||
159 | initializer in the width_bytes array. */ | ||
160 | struct dummy { | ||
161 | int assert_width_bytes_matches_size_spec_decl | ||
162 | [sizeof width_bytes / sizeof width_bytes[0] == N_SIZE_SPECS ? 1 : -1]; | ||
163 | }; | ||
164 | |||
165 | static size_t string_min; | ||
166 | static int flag_dump_strings; | ||
167 | |||
168 | /* Non-zero if an old-style 'pseudo-address' was specified. */ | ||
169 | static int flag_pseudo_start; | ||
170 | |||
171 | /* The difference between the old-style pseudo starting address and | ||
172 | the number of bytes to skip. */ | ||
173 | static off_t pseudo_offset; | ||
174 | |||
175 | /* Function that accepts an address and an optional following char, | ||
176 | and prints the address and char to stdout. */ | ||
177 | static void (*format_address) (off_t, char); | ||
178 | |||
179 | /* The number of input bytes to skip before formatting and writing. */ | ||
180 | static off_t n_bytes_to_skip; // = 0; | ||
181 | |||
182 | /* When zero, MAX_BYTES_TO_FORMAT and END_OFFSET are ignored, and all | ||
183 | input is formatted. */ | ||
184 | static int limit_bytes_to_format; // = 0; | ||
185 | |||
186 | /* The maximum number of bytes that will be formatted. */ | ||
187 | static off_t max_bytes_to_format; | ||
188 | |||
189 | /* The offset of the first byte after the last byte to be formatted. */ | ||
190 | static off_t end_offset; | ||
191 | |||
192 | /* When nonzero and two or more consecutive blocks are equal, format | ||
193 | only the first block and output an asterisk alone on the following | ||
194 | line to indicate that identical blocks have been elided. */ | ||
195 | static int abbreviate_duplicate_blocks = 1; | ||
196 | |||
197 | /* An array of specs describing how to format each input block. */ | ||
198 | static size_t n_specs; | ||
199 | static struct tspec *spec; | ||
200 | |||
201 | /* The number of input bytes formatted per output line. It must be | ||
202 | a multiple of the least common multiple of the sizes associated with | ||
203 | the specified output types. It should be as large as possible, but | ||
204 | no larger than 16 -- unless specified with the -w option. */ | ||
205 | static size_t bytes_per_block; | ||
206 | |||
207 | /* Human-readable representation of *file_list (for error messages). | ||
208 | It differs from *file_list only when *file_list is "-". */ | ||
209 | static char const *input_filename; | ||
210 | |||
211 | /* A NULL-terminated list of the file-arguments from the command line. */ | ||
212 | static char const *const *file_list; | ||
213 | |||
214 | /* Initializer for file_list if no file-arguments | ||
215 | were specified on the command line. */ | ||
216 | static char const *const default_file_list[] = { "-", NULL }; | ||
217 | |||
218 | /* The input stream associated with the current file. */ | ||
219 | static FILE *in_stream; | ||
220 | |||
221 | static int ioerror; | ||
222 | |||
223 | #define MAX_INTEGRAL_TYPE_SIZE sizeof(ulonglong_t) | ||
224 | static unsigned char integral_type_size[MAX_INTEGRAL_TYPE_SIZE + 1] = { | ||
225 | [sizeof(char)] = CHAR, | ||
226 | #if USHRT_MAX != UCHAR_MAX | ||
227 | [sizeof(short)] = SHORT, | ||
228 | #endif | ||
229 | #if UINT_MAX != USHRT_MAX | ||
230 | [sizeof(int)] = INT, | ||
231 | #endif | ||
232 | #if ULONG_MAX != UINT_MAX | ||
233 | [sizeof(long)] = LONG, | ||
234 | #endif | ||
235 | #if ULLONG_MAX != ULONG_MAX | ||
236 | [sizeof(ulonglong_t)] = LONG_LONG, | ||
237 | #endif | ||
238 | }; | ||
239 | |||
240 | #define MAX_FP_TYPE_SIZE sizeof(longdouble_t) | ||
241 | static unsigned char fp_type_size[MAX_FP_TYPE_SIZE + 1] = { | ||
242 | /* gcc seems to allow repeated indexes. Last one stays */ | ||
243 | [sizeof(longdouble_t)] = FLOAT_LONG_DOUBLE, | ||
244 | [sizeof(double)] = FLOAT_DOUBLE, | ||
245 | [sizeof(float)] = FLOAT_SINGLE, | ||
246 | }; | ||
247 | |||
248 | |||
249 | static unsigned | ||
250 | gcd(unsigned u, unsigned v) | ||
251 | { | ||
252 | unsigned t; | ||
253 | while (v != 0) { | ||
254 | t = u % v; | ||
255 | u = v; | ||
256 | v = t; | ||
257 | } | ||
258 | return u; | ||
259 | } | ||
260 | |||
261 | /* Compute the least common multiple of U and V. */ | ||
262 | static unsigned | ||
263 | lcm(unsigned u, unsigned v) { | ||
264 | unsigned t = gcd(u, v); | ||
265 | if (t == 0) | ||
266 | return 0; | ||
267 | return u * v / t; | ||
268 | } | ||
269 | |||
270 | static void | ||
271 | print_s_char(size_t n_bytes, const char *block, const char *fmt_string) | ||
272 | { | ||
273 | while (n_bytes--) { | ||
274 | int tmp = *(signed char *) block; | ||
275 | printf(fmt_string, tmp); | ||
276 | block += sizeof(unsigned char); | ||
277 | } | ||
278 | } | ||
279 | |||
280 | static void | ||
281 | print_char(size_t n_bytes, const char *block, const char *fmt_string) | ||
282 | { | ||
283 | while (n_bytes--) { | ||
284 | unsigned tmp = *(unsigned char *) block; | ||
285 | printf(fmt_string, tmp); | ||
286 | block += sizeof(unsigned char); | ||
287 | } | ||
288 | } | ||
289 | |||
290 | static void | ||
291 | print_s_short(size_t n_bytes, const char *block, const char *fmt_string) | ||
292 | { | ||
293 | n_bytes /= sizeof(signed short); | ||
294 | while (n_bytes--) { | ||
295 | int tmp = *(signed short *) block; | ||
296 | printf(fmt_string, tmp); | ||
297 | block += sizeof(unsigned short); | ||
298 | } | ||
299 | } | ||
300 | |||
301 | static void | ||
302 | print_short(size_t n_bytes, const char *block, const char *fmt_string) | ||
303 | { | ||
304 | n_bytes /= sizeof(unsigned short); | ||
305 | while (n_bytes--) { | ||
306 | unsigned tmp = *(unsigned short *) block; | ||
307 | printf(fmt_string, tmp); | ||
308 | block += sizeof(unsigned short); | ||
309 | } | ||
310 | } | ||
311 | |||
312 | static void | ||
313 | print_int(size_t n_bytes, const char *block, const char *fmt_string) | ||
314 | { | ||
315 | n_bytes /= sizeof(unsigned); | ||
316 | while (n_bytes--) { | ||
317 | unsigned tmp = *(unsigned *) block; | ||
318 | printf(fmt_string, tmp); | ||
319 | block += sizeof(unsigned); | ||
320 | } | ||
321 | } | ||
322 | |||
323 | #if UINT_MAX == ULONG_MAX | ||
324 | # define print_long print_int | ||
325 | #else | ||
326 | static void | ||
327 | print_long(size_t n_bytes, const char *block, const char *fmt_string) | ||
328 | { | ||
329 | n_bytes /= sizeof(unsigned long); | ||
330 | while (n_bytes--) { | ||
331 | unsigned long tmp = *(unsigned long *) block; | ||
332 | printf(fmt_string, tmp); | ||
333 | block += sizeof(unsigned long); | ||
334 | } | ||
335 | } | ||
336 | #endif | ||
337 | |||
338 | #if ULONG_MAX == ULLONG_MAX | ||
339 | # define print_long_long print_long | ||
340 | #else | ||
341 | static void | ||
342 | print_long_long(size_t n_bytes, const char *block, const char *fmt_string) | ||
343 | { | ||
344 | n_bytes /= sizeof(ulonglong_t); | ||
345 | while (n_bytes--) { | ||
346 | ulonglong_t tmp = *(ulonglong_t *) block; | ||
347 | printf(fmt_string, tmp); | ||
348 | block += sizeof(ulonglong_t); | ||
349 | } | ||
350 | } | ||
351 | #endif | ||
352 | |||
353 | static void | ||
354 | print_float(size_t n_bytes, const char *block, const char *fmt_string) | ||
355 | { | ||
356 | n_bytes /= sizeof(float); | ||
357 | while (n_bytes--) { | ||
358 | float tmp = *(float *) block; | ||
359 | printf(fmt_string, tmp); | ||
360 | block += sizeof(float); | ||
361 | } | ||
362 | } | ||
363 | |||
364 | static void | ||
365 | print_double(size_t n_bytes, const char *block, const char *fmt_string) | ||
366 | { | ||
367 | n_bytes /= sizeof(double); | ||
368 | while (n_bytes--) { | ||
369 | double tmp = *(double *) block; | ||
370 | printf(fmt_string, tmp); | ||
371 | block += sizeof(double); | ||
372 | } | ||
373 | } | ||
374 | |||
375 | static void | ||
376 | print_long_double(size_t n_bytes, const char *block, const char *fmt_string) | ||
377 | { | ||
378 | n_bytes /= sizeof(longdouble_t); | ||
379 | while (n_bytes--) { | ||
380 | longdouble_t tmp = *(longdouble_t *) block; | ||
381 | printf(fmt_string, tmp); | ||
382 | block += sizeof(longdouble_t); | ||
383 | } | ||
384 | } | ||
385 | |||
386 | /* print_[named]_ascii are optimized for speed. | ||
387 | * Remember, someday you may want to pump gigabytes thru this thing. | ||
388 | * Saving a dozen of .text bytes here is counter-productive */ | ||
389 | |||
390 | static void | ||
391 | print_named_ascii(size_t n_bytes, const char *block, | ||
392 | const char *unused_fmt_string ATTRIBUTE_UNUSED) | ||
393 | { | ||
394 | /* Names for some non-printing characters. */ | ||
395 | static const char charname[33][3] = { | ||
396 | "nul", "soh", "stx", "etx", "eot", "enq", "ack", "bel", | ||
397 | " bs", " ht", " nl", " vt", " ff", " cr", " so", " si", | ||
398 | "dle", "dc1", "dc2", "dc3", "dc4", "nak", "syn", "etb", | ||
399 | "can", " em", "sub", "esc", " fs", " gs", " rs", " us", | ||
400 | " sp" | ||
401 | }; | ||
402 | // buf[N] pos: 01234 56789 | ||
403 | char buf[12] = " x\0 0xx\0"; | ||
404 | // actually " x\0 xxx\0", but I want to share the string with below. | ||
405 | // [12] because we take three 32bit stack slots anyway, and | ||
406 | // gcc is too dumb to initialize with constant stores, | ||
407 | // it copies initializer from rodata. Oh well. | ||
408 | |||
409 | while (n_bytes--) { | ||
410 | unsigned masked_c = *(unsigned char *) block++; | ||
411 | |||
412 | masked_c &= 0x7f; | ||
413 | if (masked_c == 0x7f) { | ||
414 | fputs(" del", stdout); | ||
415 | continue; | ||
416 | } | ||
417 | if (masked_c > ' ') { | ||
418 | buf[3] = masked_c; | ||
419 | fputs(buf, stdout); | ||
420 | continue; | ||
421 | } | ||
422 | /* Why? Because printf(" %3.3s") is much slower... */ | ||
423 | buf[6] = charname[masked_c][0]; | ||
424 | buf[7] = charname[masked_c][1]; | ||
425 | buf[8] = charname[masked_c][2]; | ||
426 | fputs(buf+5, stdout); | ||
427 | } | ||
428 | } | ||
429 | |||
430 | static void | ||
431 | print_ascii(size_t n_bytes, const char *block, | ||
432 | const char *unused_fmt_string ATTRIBUTE_UNUSED) | ||
433 | { | ||
434 | // buf[N] pos: 01234 56789 | ||
435 | char buf[12] = " x\0 0xx\0"; | ||
436 | |||
437 | while (n_bytes--) { | ||
438 | const char *s; | ||
439 | unsigned c = *(unsigned char *) block++; | ||
440 | |||
441 | if (ISPRINT(c)) { | ||
442 | buf[3] = c; | ||
443 | fputs(buf, stdout); | ||
444 | continue; | ||
445 | } | ||
446 | switch (c) { | ||
447 | case '\0': | ||
448 | s = " \\0"; | ||
449 | break; | ||
450 | case '\007': | ||
451 | s = " \\a"; | ||
452 | break; | ||
453 | case '\b': | ||
454 | s = " \\b"; | ||
455 | break; | ||
456 | case '\f': | ||
457 | s = " \\f"; | ||
458 | break; | ||
459 | case '\n': | ||
460 | s = " \\n"; | ||
461 | break; | ||
462 | case '\r': | ||
463 | s = " \\r"; | ||
464 | break; | ||
465 | case '\t': | ||
466 | s = " \\t"; | ||
467 | break; | ||
468 | case '\v': | ||
469 | s = " \\v"; | ||
470 | break; | ||
471 | case '\x7f': | ||
472 | s = " 177"; | ||
473 | break; | ||
474 | default: /* c is never larger than 040 */ | ||
475 | buf[7] = (c >> 3) + '0'; | ||
476 | buf[8] = (c & 7) + '0'; | ||
477 | s = buf + 5; | ||
478 | } | ||
479 | fputs(s, stdout); | ||
480 | } | ||
481 | } | ||
482 | |||
483 | /* Given a list of one or more input filenames FILE_LIST, set the global | ||
484 | file pointer IN_STREAM and the global string INPUT_FILENAME to the | ||
485 | first one that can be successfully opened. Modify FILE_LIST to | ||
486 | reference the next filename in the list. A file name of "-" is | ||
487 | interpreted as standard input. If any file open fails, give an error | ||
488 | message and return nonzero. */ | ||
489 | |||
490 | static void | ||
491 | open_next_file(void) | ||
492 | { | ||
493 | while (1) { | ||
494 | input_filename = *file_list; | ||
495 | if (!input_filename) | ||
496 | return; | ||
497 | file_list++; | ||
498 | in_stream = fopen_or_warn_stdin(input_filename); | ||
499 | if (in_stream) { | ||
500 | if (in_stream == stdin) | ||
501 | input_filename = bb_msg_standard_input; | ||
502 | break; | ||
503 | } | ||
504 | ioerror = 1; | ||
505 | } | ||
506 | |||
507 | if (limit_bytes_to_format && !flag_dump_strings) | ||
508 | setbuf(in_stream, NULL); | ||
509 | } | ||
510 | |||
511 | /* Test whether there have been errors on in_stream, and close it if | ||
512 | it is not standard input. Return nonzero if there has been an error | ||
513 | on in_stream or stdout; return zero otherwise. This function will | ||
514 | report more than one error only if both a read and a write error | ||
515 | have occurred. IN_ERRNO, if nonzero, is the error number | ||
516 | corresponding to the most recent action for IN_STREAM. */ | ||
517 | |||
518 | static void | ||
519 | check_and_close(void) | ||
520 | { | ||
521 | if (in_stream) { | ||
522 | if (ferror(in_stream)) { | ||
523 | bb_error_msg("%s: read error", input_filename); | ||
524 | ioerror = 1; | ||
525 | } | ||
526 | fclose_if_not_stdin(in_stream); | ||
527 | in_stream = NULL; | ||
528 | } | ||
529 | |||
530 | if (ferror(stdout)) { | ||
531 | bb_error_msg("write error"); | ||
532 | ioerror = 1; | ||
533 | } | ||
534 | } | ||
535 | |||
536 | /* If S points to a single valid modern od format string, put | ||
537 | a description of that format in *TSPEC, make *NEXT point at the | ||
538 | character following the just-decoded format (if *NEXT is non-NULL), | ||
539 | and return zero. For example, if S were "d4afL" | ||
540 | *NEXT would be set to "afL" and *TSPEC would be | ||
541 | { | ||
542 | fmt = SIGNED_DECIMAL; | ||
543 | size = INT or LONG; (whichever integral_type_size[4] resolves to) | ||
544 | print_function = print_int; (assuming size == INT) | ||
545 | fmt_string = "%011d%c"; | ||
546 | } | ||
547 | S_ORIG is solely for reporting errors. It should be the full format | ||
548 | string argument. */ | ||
549 | |||
550 | static void | ||
551 | decode_one_format(const char *s_orig, const char *s, const char **next, | ||
552 | struct tspec *tspec) | ||
553 | { | ||
554 | enum size_spec size_spec; | ||
555 | unsigned size; | ||
556 | enum output_format fmt; | ||
557 | const char *p; | ||
558 | char *end; | ||
559 | char *fmt_string = NULL; | ||
560 | void (*print_function) (size_t, const char *, const char *); | ||
561 | unsigned c; | ||
562 | unsigned field_width = 0; | ||
563 | int pos; | ||
564 | |||
565 | assert(tspec != NULL); | ||
566 | |||
567 | switch (*s) { | ||
568 | case 'd': | ||
569 | case 'o': | ||
570 | case 'u': | ||
571 | case 'x': { | ||
572 | static const char CSIL[] = "CSIL"; | ||
573 | |||
574 | c = *s++; | ||
575 | p = strchr(CSIL, *s); | ||
576 | if (!p) { | ||
577 | size = sizeof(int); | ||
578 | if (isdigit(s[0])) { | ||
579 | size = bb_strtou(s, &end, 0); | ||
580 | if (errno == ERANGE | ||
581 | || MAX_INTEGRAL_TYPE_SIZE < size | ||
582 | || integral_type_size[size] == NO_SIZE | ||
583 | ) { | ||
584 | bb_error_msg_and_die("invalid type string '%s'; " | ||
585 | "%u-byte %s type is not supported", | ||
586 | s_orig, size, "integral"); | ||
587 | } | ||
588 | s = end; | ||
589 | } | ||
590 | } else { | ||
591 | static const uint8_t CSIL_sizeof[] = { | ||
592 | sizeof(char), | ||
593 | sizeof(short), | ||
594 | sizeof(int), | ||
595 | sizeof(long), | ||
596 | }; | ||
597 | size = CSIL_sizeof[p - CSIL]; | ||
598 | } | ||
599 | |||
600 | #define ISPEC_TO_FORMAT(Spec, Min_format, Long_format, Max_format) \ | ||
601 | ((Spec) == LONG_LONG ? (Max_format) \ | ||
602 | : ((Spec) == LONG ? (Long_format) : (Min_format))) | ||
603 | |||
604 | #define FMT_BYTES_ALLOCATED 9 | ||
605 | size_spec = integral_type_size[size]; | ||
606 | |||
607 | { | ||
608 | static const char doux[] = "doux"; | ||
609 | static const char doux_fmt_letter[][4] = { | ||
610 | "lld", "llo", "llu", "llx" | ||
611 | }; | ||
612 | static const enum output_format doux_fmt[] = { | ||
613 | SIGNED_DECIMAL, | ||
614 | OCTAL, | ||
615 | UNSIGNED_DECIMAL, | ||
616 | HEXADECIMAL, | ||
617 | }; | ||
618 | static const uint8_t *const doux_bytes_to_XXX[] = { | ||
619 | bytes_to_signed_dec_digits, | ||
620 | bytes_to_oct_digits, | ||
621 | bytes_to_unsigned_dec_digits, | ||
622 | bytes_to_hex_digits, | ||
623 | }; | ||
624 | static const char doux_fmtstring[][sizeof(" %%0%u%s")] = { | ||
625 | " %%%u%s", | ||
626 | " %%0%u%s", | ||
627 | " %%%u%s", | ||
628 | " %%0%u%s", | ||
629 | }; | ||
630 | |||
631 | pos = strchr(doux, c) - doux; | ||
632 | fmt = doux_fmt[pos]; | ||
633 | field_width = doux_bytes_to_XXX[pos][size]; | ||
634 | p = doux_fmt_letter[pos] + 2; | ||
635 | if (size_spec == LONG) p--; | ||
636 | if (size_spec == LONG_LONG) p -= 2; | ||
637 | fmt_string = xasprintf(doux_fmtstring[pos], field_width, p); | ||
638 | } | ||
639 | |||
640 | switch (size_spec) { | ||
641 | case CHAR: | ||
642 | print_function = (fmt == SIGNED_DECIMAL | ||
643 | ? print_s_char | ||
644 | : print_char); | ||
645 | break; | ||
646 | case SHORT: | ||
647 | print_function = (fmt == SIGNED_DECIMAL | ||
648 | ? print_s_short | ||
649 | : print_short); | ||
650 | break; | ||
651 | case INT: | ||
652 | print_function = print_int; | ||
653 | break; | ||
654 | case LONG: | ||
655 | print_function = print_long; | ||
656 | break; | ||
657 | default: /* case LONG_LONG: */ | ||
658 | print_function = print_long_long; | ||
659 | break; | ||
660 | } | ||
661 | break; | ||
662 | } | ||
663 | |||
664 | case 'f': { | ||
665 | static const char FDL[] = "FDL"; | ||
666 | |||
667 | fmt = FLOATING_POINT; | ||
668 | ++s; | ||
669 | p = strchr(FDL, *s); | ||
670 | if (!p) { | ||
671 | size = sizeof(double); | ||
672 | if (isdigit(s[0])) { | ||
673 | size = bb_strtou(s, &end, 0); | ||
674 | if (errno == ERANGE || size > MAX_FP_TYPE_SIZE | ||
675 | || fp_type_size[size] == NO_SIZE | ||
676 | ) { | ||
677 | bb_error_msg_and_die("invalid type string '%s'; " | ||
678 | "%u-byte %s type is not supported", | ||
679 | s_orig, size, "floating point"); | ||
680 | } | ||
681 | s = end; | ||
682 | } | ||
683 | } else { | ||
684 | static const uint8_t FDL_sizeof[] = { | ||
685 | sizeof(float), | ||
686 | sizeof(double), | ||
687 | sizeof(longdouble_t), | ||
688 | }; | ||
689 | |||
690 | size = FDL_sizeof[p - FDL]; | ||
691 | } | ||
692 | |||
693 | size_spec = fp_type_size[size]; | ||
694 | |||
695 | switch (size_spec) { | ||
696 | case FLOAT_SINGLE: | ||
697 | print_function = print_float; | ||
698 | field_width = FLT_DIG + 8; | ||
699 | /* Don't use %#e; not all systems support it. */ | ||
700 | fmt_string = xasprintf(" %%%d.%de", field_width, FLT_DIG); | ||
701 | break; | ||
702 | case FLOAT_DOUBLE: | ||
703 | print_function = print_double; | ||
704 | field_width = DBL_DIG + 8; | ||
705 | fmt_string = xasprintf(" %%%d.%de", field_width, DBL_DIG); | ||
706 | break; | ||
707 | default: /* case FLOAT_LONG_DOUBLE: */ | ||
708 | print_function = print_long_double; | ||
709 | field_width = LDBL_DIG + 8; | ||
710 | fmt_string = xasprintf(" %%%d.%dLe", field_width, LDBL_DIG); | ||
711 | break; | ||
712 | } | ||
713 | break; | ||
714 | } | ||
715 | |||
716 | case 'a': | ||
717 | ++s; | ||
718 | fmt = NAMED_CHARACTER; | ||
719 | size_spec = CHAR; | ||
720 | print_function = print_named_ascii; | ||
721 | field_width = 3; | ||
722 | break; | ||
723 | case 'c': | ||
724 | ++s; | ||
725 | fmt = CHARACTER; | ||
726 | size_spec = CHAR; | ||
727 | print_function = print_ascii; | ||
728 | field_width = 3; | ||
729 | break; | ||
730 | default: | ||
731 | bb_error_msg_and_die("invalid character '%c' " | ||
732 | "in type string '%s'", *s, s_orig); | ||
733 | } | ||
734 | |||
735 | tspec->size = size_spec; | ||
736 | tspec->fmt = fmt; | ||
737 | tspec->print_function = print_function; | ||
738 | tspec->fmt_string = fmt_string; | ||
739 | |||
740 | tspec->field_width = field_width; | ||
741 | tspec->hexl_mode_trailer = (*s == 'z'); | ||
742 | if (tspec->hexl_mode_trailer) | ||
743 | s++; | ||
744 | |||
745 | if (next != NULL) | ||
746 | *next = s; | ||
747 | } | ||
748 | |||
749 | /* Decode the modern od format string S. Append the decoded | ||
750 | representation to the global array SPEC, reallocating SPEC if | ||
751 | necessary. Return zero if S is valid, nonzero otherwise. */ | ||
752 | |||
753 | static void | ||
754 | decode_format_string(const char *s) | ||
755 | { | ||
756 | const char *s_orig = s; | ||
757 | |||
758 | while (*s != '\0') { | ||
759 | struct tspec tspec; | ||
760 | const char *next; | ||
761 | |||
762 | decode_one_format(s_orig, s, &next, &tspec); | ||
763 | |||
764 | assert(s != next); | ||
765 | s = next; | ||
766 | n_specs++; | ||
767 | spec = xrealloc(spec, n_specs * sizeof(*spec)); | ||
768 | memcpy(&spec[n_specs-1], &tspec, sizeof *spec); | ||
769 | } | ||
770 | } | ||
771 | |||
772 | /* Given a list of one or more input filenames FILE_LIST, set the global | ||
773 | file pointer IN_STREAM to position N_SKIP in the concatenation of | ||
774 | those files. If any file operation fails or if there are fewer than | ||
775 | N_SKIP bytes in the combined input, give an error message and return | ||
776 | nonzero. When possible, use seek rather than read operations to | ||
777 | advance IN_STREAM. */ | ||
778 | |||
779 | static void | ||
780 | skip(off_t n_skip) | ||
781 | { | ||
782 | if (n_skip == 0) | ||
783 | return; | ||
784 | |||
785 | while (in_stream) { /* !EOF */ | ||
786 | struct stat file_stats; | ||
787 | |||
788 | /* First try seeking. For large offsets, this extra work is | ||
789 | worthwhile. If the offset is below some threshold it may be | ||
790 | more efficient to move the pointer by reading. There are two | ||
791 | issues when trying to seek: | ||
792 | - the file must be seekable. | ||
793 | - before seeking to the specified position, make sure | ||
794 | that the new position is in the current file. | ||
795 | Try to do that by getting file's size using fstat. | ||
796 | But that will work only for regular files. */ | ||
797 | |||
798 | /* The st_size field is valid only for regular files | ||
799 | (and for symbolic links, which cannot occur here). | ||
800 | If the number of bytes left to skip is at least | ||
801 | as large as the size of the current file, we can | ||
802 | decrement n_skip and go on to the next file. */ | ||
803 | if (fstat(fileno(in_stream), &file_stats) == 0 | ||
804 | && S_ISREG(file_stats.st_mode) && file_stats.st_size >= 0 | ||
805 | ) { | ||
806 | if (file_stats.st_size < n_skip) { | ||
807 | n_skip -= file_stats.st_size; | ||
808 | /* take check&close / open_next route */ | ||
809 | } else { | ||
810 | if (fseeko(in_stream, n_skip, SEEK_CUR) != 0) | ||
811 | ioerror = 1; | ||
812 | return; | ||
813 | } | ||
814 | } else { | ||
815 | /* If it's not a regular file with nonnegative size, | ||
816 | position the file pointer by reading. */ | ||
817 | char buf[BUFSIZ]; | ||
818 | size_t n_bytes_read, n_bytes_to_read = BUFSIZ; | ||
819 | |||
820 | while (n_skip > 0) { | ||
821 | if (n_skip < n_bytes_to_read) | ||
822 | n_bytes_to_read = n_skip; | ||
823 | n_bytes_read = fread(buf, 1, n_bytes_to_read, in_stream); | ||
824 | n_skip -= n_bytes_read; | ||
825 | if (n_bytes_read != n_bytes_to_read) | ||
826 | break; /* EOF on this file or error */ | ||
827 | } | ||
828 | } | ||
829 | if (n_skip == 0) | ||
830 | return; | ||
831 | |||
832 | check_and_close(); | ||
833 | open_next_file(); | ||
834 | } | ||
835 | |||
836 | if (n_skip) | ||
837 | bb_error_msg_and_die("cannot skip past end of combined input"); | ||
838 | } | ||
839 | |||
840 | |||
841 | typedef void FN_format_address(off_t address, char c); | ||
842 | |||
843 | static void | ||
844 | format_address_none(off_t address ATTRIBUTE_UNUSED, char c ATTRIBUTE_UNUSED) | ||
845 | { | ||
846 | } | ||
847 | |||
848 | static char address_fmt[] = "%0n"OFF_FMT"xc"; | ||
849 | /* Corresponds to 'x' above */ | ||
850 | #define address_base_char address_fmt[sizeof(address_fmt)-3] | ||
851 | /* Corresponds to 'n' above */ | ||
852 | #define address_pad_len_char address_fmt[2] | ||
853 | |||
854 | static void | ||
855 | format_address_std(off_t address, char c) | ||
856 | { | ||
857 | /* Corresponds to 'c' */ | ||
858 | address_fmt[sizeof(address_fmt)-2] = c; | ||
859 | printf(address_fmt, address); | ||
860 | } | ||
861 | |||
862 | #if ENABLE_GETOPT_LONG | ||
863 | /* only used with --traditional */ | ||
864 | static void | ||
865 | format_address_paren(off_t address, char c) | ||
866 | { | ||
867 | putchar('('); | ||
868 | format_address_std(address, ')'); | ||
869 | if (c) putchar(c); | ||
870 | } | ||
871 | |||
872 | static void | ||
873 | format_address_label(off_t address, char c) | ||
874 | { | ||
875 | format_address_std(address, ' '); | ||
876 | format_address_paren(address + pseudo_offset, c); | ||
877 | } | ||
878 | #endif | ||
879 | |||
880 | static void | ||
881 | dump_hexl_mode_trailer(size_t n_bytes, const char *block) | ||
882 | { | ||
883 | fputs(" >", stdout); | ||
884 | while (n_bytes--) { | ||
885 | unsigned c = *(unsigned char *) block++; | ||
886 | c = (ISPRINT(c) ? c : '.'); | ||
887 | putchar(c); | ||
888 | } | ||
889 | putchar('<'); | ||
890 | } | ||
891 | |||
892 | /* Write N_BYTES bytes from CURR_BLOCK to standard output once for each | ||
893 | of the N_SPEC format specs. CURRENT_OFFSET is the byte address of | ||
894 | CURR_BLOCK in the concatenation of input files, and it is printed | ||
895 | (optionally) only before the output line associated with the first | ||
896 | format spec. When duplicate blocks are being abbreviated, the output | ||
897 | for a sequence of identical input blocks is the output for the first | ||
898 | block followed by an asterisk alone on a line. It is valid to compare | ||
899 | the blocks PREV_BLOCK and CURR_BLOCK only when N_BYTES == BYTES_PER_BLOCK. | ||
900 | That condition may be false only for the last input block -- and then | ||
901 | only when it has not been padded to length BYTES_PER_BLOCK. */ | ||
902 | |||
903 | static void | ||
904 | write_block(off_t current_offset, size_t n_bytes, | ||
905 | const char *prev_block, const char *curr_block) | ||
906 | { | ||
907 | static char first = 1; | ||
908 | static char prev_pair_equal = 0; | ||
909 | size_t i; | ||
910 | |||
911 | if (abbreviate_duplicate_blocks | ||
912 | && !first | ||
913 | && n_bytes == bytes_per_block | ||
914 | && memcmp(prev_block, curr_block, bytes_per_block) == 0 | ||
915 | ) { | ||
916 | if (prev_pair_equal) { | ||
917 | /* The two preceding blocks were equal, and the current | ||
918 | block is the same as the last one, so print nothing. */ | ||
919 | } else { | ||
920 | puts("*"); | ||
921 | prev_pair_equal = 1; | ||
922 | } | ||
923 | } else { | ||
924 | first = 0; | ||
925 | prev_pair_equal = 0; | ||
926 | for (i = 0; i < n_specs; i++) { | ||
927 | if (i == 0) | ||
928 | format_address(current_offset, '\0'); | ||
929 | else | ||
930 | printf("%*s", address_pad_len_char - '0', ""); | ||
931 | (*spec[i].print_function) (n_bytes, curr_block, spec[i].fmt_string); | ||
932 | if (spec[i].hexl_mode_trailer) { | ||
933 | /* space-pad out to full line width, then dump the trailer */ | ||
934 | int datum_width = width_bytes[spec[i].size]; | ||
935 | int blank_fields = (bytes_per_block - n_bytes) / datum_width; | ||
936 | int field_width = spec[i].field_width + 1; | ||
937 | printf("%*s", blank_fields * field_width, ""); | ||
938 | dump_hexl_mode_trailer(n_bytes, curr_block); | ||
939 | } | ||
940 | putchar('\n'); | ||
941 | } | ||
942 | } | ||
943 | } | ||
944 | |||
945 | static void | ||
946 | read_block(size_t n, char *block, size_t *n_bytes_in_buffer) | ||
947 | { | ||
948 | assert(0 < n && n <= bytes_per_block); | ||
949 | |||
950 | *n_bytes_in_buffer = 0; | ||
951 | |||
952 | if (n == 0) | ||
953 | return; | ||
954 | |||
955 | while (in_stream != NULL) { /* EOF. */ | ||
956 | size_t n_needed; | ||
957 | size_t n_read; | ||
958 | |||
959 | n_needed = n - *n_bytes_in_buffer; | ||
960 | n_read = fread(block + *n_bytes_in_buffer, 1, n_needed, in_stream); | ||
961 | *n_bytes_in_buffer += n_read; | ||
962 | if (n_read == n_needed) | ||
963 | break; | ||
964 | /* error check is done in check_and_close */ | ||
965 | check_and_close(); | ||
966 | open_next_file(); | ||
967 | } | ||
968 | } | ||
969 | |||
970 | /* Return the least common multiple of the sizes associated | ||
971 | with the format specs. */ | ||
972 | |||
973 | static int | ||
974 | get_lcm(void) | ||
975 | { | ||
976 | size_t i; | ||
977 | int l_c_m = 1; | ||
978 | |||
979 | for (i = 0; i < n_specs; i++) | ||
980 | l_c_m = lcm(l_c_m, width_bytes[(int) spec[i].size]); | ||
981 | return l_c_m; | ||
982 | } | ||
983 | |||
984 | #if ENABLE_GETOPT_LONG | ||
985 | /* If S is a valid traditional offset specification with an optional | ||
986 | leading '+' return nonzero and set *OFFSET to the offset it denotes. */ | ||
987 | |||
988 | static int | ||
989 | parse_old_offset(const char *s, off_t *offset) | ||
990 | { | ||
991 | static const struct suffix_mult Bb[] = { | ||
992 | { "B", 1024 }, | ||
993 | { "b", 512 }, | ||
994 | { NULL, 0 } | ||
995 | }; | ||
996 | char *p; | ||
997 | int radix; | ||
998 | |||
999 | /* Skip over any leading '+'. */ | ||
1000 | if (s[0] == '+') ++s; | ||
1001 | |||
1002 | /* Determine the radix we'll use to interpret S. If there is a '.', | ||
1003 | * it's decimal, otherwise, if the string begins with '0X'or '0x', | ||
1004 | * it's hexadecimal, else octal. */ | ||
1005 | p = strchr(s, '.'); | ||
1006 | radix = 8; | ||
1007 | if (p) { | ||
1008 | p[0] = '\0'; /* cheating */ | ||
1009 | radix = 10; | ||
1010 | } else if (s[0] == '0' && (s[1] == 'x' || s[1] == 'X')) | ||
1011 | radix = 16; | ||
1012 | |||
1013 | *offset = xstrtooff_sfx(s, radix, Bb); | ||
1014 | if (p) p[0] = '.'; | ||
1015 | |||
1016 | return (*offset >= 0); | ||
1017 | } | ||
1018 | #endif | ||
1019 | |||
1020 | /* Read a chunk of size BYTES_PER_BLOCK from the input files, write the | ||
1021 | formatted block to standard output, and repeat until the specified | ||
1022 | maximum number of bytes has been read or until all input has been | ||
1023 | processed. If the last block read is smaller than BYTES_PER_BLOCK | ||
1024 | and its size is not a multiple of the size associated with a format | ||
1025 | spec, extend the input block with zero bytes until its length is a | ||
1026 | multiple of all format spec sizes. Write the final block. Finally, | ||
1027 | write on a line by itself the offset of the byte after the last byte | ||
1028 | read. Accumulate return values from calls to read_block and | ||
1029 | check_and_close, and if any was nonzero, return nonzero. | ||
1030 | Otherwise, return zero. */ | ||
1031 | |||
1032 | static void | ||
1033 | dump(void) | ||
1034 | { | ||
1035 | char *block[2]; | ||
1036 | off_t current_offset; | ||
1037 | int idx; | ||
1038 | size_t n_bytes_read; | ||
1039 | |||
1040 | block[0] = xmalloc(2*bytes_per_block); | ||
1041 | block[1] = block[0] + bytes_per_block; | ||
1042 | |||
1043 | current_offset = n_bytes_to_skip; | ||
1044 | |||
1045 | idx = 0; | ||
1046 | if (limit_bytes_to_format) { | ||
1047 | while (1) { | ||
1048 | size_t n_needed; | ||
1049 | if (current_offset >= end_offset) { | ||
1050 | n_bytes_read = 0; | ||
1051 | break; | ||
1052 | } | ||
1053 | n_needed = MIN(end_offset - current_offset, | ||
1054 | (off_t) bytes_per_block); | ||
1055 | read_block(n_needed, block[idx], &n_bytes_read); | ||
1056 | if (n_bytes_read < bytes_per_block) | ||
1057 | break; | ||
1058 | assert(n_bytes_read == bytes_per_block); | ||
1059 | write_block(current_offset, n_bytes_read, | ||
1060 | block[!idx], block[idx]); | ||
1061 | current_offset += n_bytes_read; | ||
1062 | idx = !idx; | ||
1063 | } | ||
1064 | } else { | ||
1065 | while (1) { | ||
1066 | read_block(bytes_per_block, block[idx], &n_bytes_read); | ||
1067 | if (n_bytes_read < bytes_per_block) | ||
1068 | break; | ||
1069 | assert(n_bytes_read == bytes_per_block); | ||
1070 | write_block(current_offset, n_bytes_read, | ||
1071 | block[!idx], block[idx]); | ||
1072 | current_offset += n_bytes_read; | ||
1073 | idx = !idx; | ||
1074 | } | ||
1075 | } | ||
1076 | |||
1077 | if (n_bytes_read > 0) { | ||
1078 | int l_c_m; | ||
1079 | size_t bytes_to_write; | ||
1080 | |||
1081 | l_c_m = get_lcm(); | ||
1082 | |||
1083 | /* Make bytes_to_write the smallest multiple of l_c_m that | ||
1084 | is at least as large as n_bytes_read. */ | ||
1085 | bytes_to_write = l_c_m * ((n_bytes_read + l_c_m - 1) / l_c_m); | ||
1086 | |||
1087 | memset(block[idx] + n_bytes_read, 0, bytes_to_write - n_bytes_read); | ||
1088 | write_block(current_offset, bytes_to_write, | ||
1089 | block[!idx], block[idx]); | ||
1090 | current_offset += n_bytes_read; | ||
1091 | } | ||
1092 | |||
1093 | format_address(current_offset, '\n'); | ||
1094 | |||
1095 | if (limit_bytes_to_format && current_offset >= end_offset) | ||
1096 | check_and_close(); | ||
1097 | |||
1098 | free(block[0]); | ||
1099 | } | ||
1100 | |||
1101 | /* Read a single byte into *C from the concatenation of the input files | ||
1102 | named in the global array FILE_LIST. On the first call to this | ||
1103 | function, the global variable IN_STREAM is expected to be an open | ||
1104 | stream associated with the input file INPUT_FILENAME. If IN_STREAM | ||
1105 | is at end-of-file, close it and update the global variables IN_STREAM | ||
1106 | and INPUT_FILENAME so they correspond to the next file in the list. | ||
1107 | Then try to read a byte from the newly opened file. Repeat if | ||
1108 | necessary until EOF is reached for the last file in FILE_LIST, then | ||
1109 | set *C to EOF and return. Subsequent calls do likewise. The return | ||
1110 | value is nonzero if any errors occured, zero otherwise. */ | ||
1111 | |||
1112 | static void | ||
1113 | read_char(int *c) | ||
1114 | { | ||
1115 | while (in_stream) { /* !EOF */ | ||
1116 | *c = fgetc(in_stream); | ||
1117 | if (*c != EOF) | ||
1118 | return; | ||
1119 | check_and_close(); | ||
1120 | open_next_file(); | ||
1121 | } | ||
1122 | *c = EOF; | ||
1123 | } | ||
1124 | |||
1125 | /* Read N bytes into BLOCK from the concatenation of the input files | ||
1126 | named in the global array FILE_LIST. On the first call to this | ||
1127 | function, the global variable IN_STREAM is expected to be an open | ||
1128 | stream associated with the input file INPUT_FILENAME. If all N | ||
1129 | bytes cannot be read from IN_STREAM, close IN_STREAM and update | ||
1130 | the global variables IN_STREAM and INPUT_FILENAME. Then try to | ||
1131 | read the remaining bytes from the newly opened file. Repeat if | ||
1132 | necessary until EOF is reached for the last file in FILE_LIST. | ||
1133 | On subsequent calls, don't modify BLOCK and return zero. Set | ||
1134 | *N_BYTES_IN_BUFFER to the number of bytes read. If an error occurs, | ||
1135 | it will be detected through ferror when the stream is about to be | ||
1136 | closed. If there is an error, give a message but continue reading | ||
1137 | as usual and return nonzero. Otherwise return zero. */ | ||
1138 | |||
1139 | /* STRINGS mode. Find each "string constant" in the input. | ||
1140 | A string constant is a run of at least 'string_min' ASCII | ||
1141 | graphic (or formatting) characters terminated by a null. | ||
1142 | Based on a function written by Richard Stallman for a | ||
1143 | traditional version of od. Return nonzero if an error | ||
1144 | occurs. Otherwise, return zero. */ | ||
1145 | |||
1146 | static void | ||
1147 | dump_strings(void) | ||
1148 | { | ||
1149 | size_t bufsize = MAX(100, string_min); | ||
1150 | char *buf = xmalloc(bufsize); | ||
1151 | off_t address = n_bytes_to_skip; | ||
1152 | |||
1153 | while (1) { | ||
1154 | size_t i; | ||
1155 | int c; | ||
1156 | |||
1157 | /* See if the next 'string_min' chars are all printing chars. */ | ||
1158 | tryline: | ||
1159 | if (limit_bytes_to_format && (end_offset - string_min <= address)) | ||
1160 | break; | ||
1161 | i = 0; | ||
1162 | while (!limit_bytes_to_format || address < end_offset) { | ||
1163 | if (i == bufsize) { | ||
1164 | bufsize += bufsize/8; | ||
1165 | buf = xrealloc(buf, bufsize); | ||
1166 | } | ||
1167 | read_char(&c); | ||
1168 | if (c < 0) { /* EOF */ | ||
1169 | free(buf); | ||
1170 | return; | ||
1171 | } | ||
1172 | address++; | ||
1173 | if (!c) | ||
1174 | break; | ||
1175 | if (!ISPRINT(c)) | ||
1176 | goto tryline; /* It isn't; give up on this string. */ | ||
1177 | buf[i++] = c; /* String continues; store it all. */ | ||
1178 | } | ||
1179 | |||
1180 | if (i < string_min) /* Too short! */ | ||
1181 | goto tryline; | ||
1182 | |||
1183 | /* If we get here, the string is all printable and null-terminated, | ||
1184 | * so print it. It is all in 'buf' and 'i' is its length. */ | ||
1185 | buf[i] = 0; | ||
1186 | format_address(address - i - 1, ' '); | ||
1187 | |||
1188 | for (i = 0; (c = buf[i]); i++) { | ||
1189 | switch (c) { | ||
1190 | case '\007': fputs("\\a", stdout); break; | ||
1191 | case '\b': fputs("\\b", stdout); break; | ||
1192 | case '\f': fputs("\\f", stdout); break; | ||
1193 | case '\n': fputs("\\n", stdout); break; | ||
1194 | case '\r': fputs("\\r", stdout); break; | ||
1195 | case '\t': fputs("\\t", stdout); break; | ||
1196 | case '\v': fputs("\\v", stdout); break; | ||
1197 | default: putc(c, stdout); | ||
1198 | } | ||
1199 | } | ||
1200 | putchar('\n'); | ||
1201 | } | ||
1202 | |||
1203 | /* We reach this point only if we search through | ||
1204 | (max_bytes_to_format - string_min) bytes before reaching EOF. */ | ||
1205 | free(buf); | ||
1206 | |||
1207 | check_and_close(); | ||
1208 | } | ||
1209 | |||
1210 | int | ||
1211 | od_main(int argc, char **argv) | ||
1212 | { | ||
1213 | static const struct suffix_mult bkm[] = { | ||
1214 | { "b", 512 }, | ||
1215 | { "k", 1024 }, | ||
1216 | { "m", 1024*1024 }, | ||
1217 | { NULL, 0 } | ||
1218 | }; | ||
1219 | unsigned opt; | ||
1220 | int l_c_m; | ||
1221 | /* The old-style 'pseudo starting address' to be printed in parentheses | ||
1222 | after any true address. */ | ||
1223 | off_t pseudo_start = 0; // only for gcc | ||
1224 | enum { | ||
1225 | OPT_A = 1 << 0, | ||
1226 | OPT_N = 1 << 1, | ||
1227 | OPT_a = 1 << 2, | ||
1228 | OPT_b = 1 << 3, | ||
1229 | OPT_c = 1 << 4, | ||
1230 | OPT_d = 1 << 5, | ||
1231 | OPT_f = 1 << 6, | ||
1232 | OPT_h = 1 << 7, | ||
1233 | OPT_i = 1 << 8, | ||
1234 | OPT_j = 1 << 9, | ||
1235 | OPT_l = 1 << 10, | ||
1236 | OPT_o = 1 << 11, | ||
1237 | OPT_t = 1 << 12, | ||
1238 | OPT_v = 1 << 13, | ||
1239 | OPT_x = 1 << 14, | ||
1240 | OPT_s = 1 << 15, | ||
1241 | OPT_S = 1 << 16, | ||
1242 | OPT_w = 1 << 17, | ||
1243 | OPT_traditional = (1 << 18) * ENABLE_GETOPT_LONG, | ||
1244 | }; | ||
1245 | #if ENABLE_GETOPT_LONG | ||
1246 | static const struct option long_options[] = { | ||
1247 | { "skip-bytes", required_argument, NULL, 'j' }, | ||
1248 | { "address-radix", required_argument, NULL, 'A' }, | ||
1249 | { "read-bytes", required_argument, NULL, 'N' }, | ||
1250 | { "format", required_argument, NULL, 't' }, | ||
1251 | { "output-duplicates", no_argument, NULL, 'v' }, | ||
1252 | { "strings", optional_argument, NULL, 'S' }, | ||
1253 | { "width", optional_argument, NULL, 'w' }, | ||
1254 | { "traditional", no_argument, NULL, 0xff }, | ||
1255 | { NULL, 0, NULL, 0 } | ||
1256 | }; | ||
1257 | #endif | ||
1258 | char *str_A, *str_N, *str_j, *str_S; | ||
1259 | char *str_w = NULL; | ||
1260 | llist_t *lst_t = NULL; | ||
1261 | |||
1262 | spec = NULL; | ||
1263 | format_address = format_address_std; | ||
1264 | address_base_char = 'o'; | ||
1265 | address_pad_len_char = '7'; | ||
1266 | flag_dump_strings = 0; | ||
1267 | |||
1268 | /* Parse command line */ | ||
1269 | opt_complementary = "t::"; // list | ||
1270 | #if ENABLE_GETOPT_LONG | ||
1271 | applet_long_options = long_options; | ||
1272 | #endif | ||
1273 | opt = getopt32(argc, argv, "A:N:abcdfhij:lot:vxsS:" | ||
1274 | "w::", // -w with optional param | ||
1275 | // -S was -s and also had optional parameter | ||
1276 | // but in coreutils 6.3 it was renamed and now has | ||
1277 | // _mandatory_ parameter | ||
1278 | &str_A, &str_N, &str_j, &lst_t, &str_S, &str_w); | ||
1279 | argc -= optind; | ||
1280 | argv += optind; | ||
1281 | if (opt & OPT_A) { | ||
1282 | static const char doxn[] = "doxn"; | ||
1283 | static const char doxn_address_base_char[] = { | ||
1284 | 'u', 'o', 'x', /* '?' fourth one is not important */ | ||
1285 | }; | ||
1286 | static const uint8_t doxn_address_pad_len_char[] = { | ||
1287 | '7', '7', '6', /* '?' */ | ||
1288 | }; | ||
1289 | char *p; | ||
1290 | int pos; | ||
1291 | p = strchr(doxn, str_A[0]); | ||
1292 | if (!p) | ||
1293 | bb_error_msg_and_die("bad output address radix " | ||
1294 | "'%c' (must be [doxn])", str_A[0]); | ||
1295 | pos = p - doxn; | ||
1296 | if (pos == 3) format_address = format_address_none; | ||
1297 | address_base_char = doxn_address_base_char[pos]; | ||
1298 | address_pad_len_char = doxn_address_pad_len_char[pos]; | ||
1299 | } | ||
1300 | if (opt & OPT_N) { | ||
1301 | limit_bytes_to_format = 1; | ||
1302 | max_bytes_to_format = xstrtooff_sfx(str_N, 0, bkm); | ||
1303 | } | ||
1304 | if (opt & OPT_a) decode_format_string("a"); | ||
1305 | if (opt & OPT_b) decode_format_string("oC"); | ||
1306 | if (opt & OPT_c) decode_format_string("c"); | ||
1307 | if (opt & OPT_d) decode_format_string("u2"); | ||
1308 | if (opt & OPT_f) decode_format_string("fF"); | ||
1309 | if (opt & OPT_h) decode_format_string("x2"); | ||
1310 | if (opt & OPT_i) decode_format_string("d2"); | ||
1311 | if (opt & OPT_j) n_bytes_to_skip = xstrtooff_sfx(str_j, 0, bkm); | ||
1312 | if (opt & OPT_l) decode_format_string("d4"); | ||
1313 | if (opt & OPT_o) decode_format_string("o2"); | ||
1314 | //if (opt & OPT_t)... | ||
1315 | lst_t = rev_llist(lst_t); | ||
1316 | while (lst_t) { | ||
1317 | decode_format_string(lst_t->data); | ||
1318 | lst_t = lst_t->link; | ||
1319 | } | ||
1320 | if (opt & OPT_v) abbreviate_duplicate_blocks = 0; | ||
1321 | if (opt & OPT_x) decode_format_string("x2"); | ||
1322 | if (opt & OPT_s) decode_format_string("d2"); | ||
1323 | if (opt & OPT_S) { | ||
1324 | string_min = 3; | ||
1325 | string_min = xstrtou_sfx(str_S, 0, bkm); | ||
1326 | flag_dump_strings = 1; | ||
1327 | } | ||
1328 | //if (opt & OPT_w)... | ||
1329 | //if (opt & OPT_traditional)... | ||
1330 | |||
1331 | if (flag_dump_strings && n_specs > 0) | ||
1332 | bb_error_msg_and_die("no type may be specified when dumping strings"); | ||
1333 | |||
1334 | /* If the --traditional option is used, there may be from | ||
1335 | * 0 to 3 remaining command line arguments; handle each case | ||
1336 | * separately. | ||
1337 | * od [file] [[+]offset[.][b] [[+]label[.][b]]] | ||
1338 | * The offset and pseudo_start have the same syntax. | ||
1339 | * | ||
1340 | * FIXME: POSIX 1003.1-2001 with XSI requires support for the | ||
1341 | * traditional syntax even if --traditional is not given. */ | ||
1342 | |||
1343 | #if ENABLE_GETOPT_LONG | ||
1344 | if (opt & OPT_traditional) { | ||
1345 | off_t o1, o2; | ||
1346 | |||
1347 | if (argc == 1) { | ||
1348 | if (parse_old_offset(argv[0], &o1)) { | ||
1349 | n_bytes_to_skip = o1; | ||
1350 | --argc; | ||
1351 | ++argv; | ||
1352 | } | ||
1353 | } else if (argc == 2) { | ||
1354 | if (parse_old_offset(argv[0], &o1) | ||
1355 | && parse_old_offset(argv[1], &o2) | ||
1356 | ) { | ||
1357 | n_bytes_to_skip = o1; | ||
1358 | flag_pseudo_start = 1; | ||
1359 | pseudo_start = o2; | ||
1360 | argv += 2; | ||
1361 | argc -= 2; | ||
1362 | } else if (parse_old_offset(argv[1], &o2)) { | ||
1363 | n_bytes_to_skip = o2; | ||
1364 | --argc; | ||
1365 | argv[1] = argv[0]; | ||
1366 | ++argv; | ||
1367 | } else { | ||
1368 | bb_error_msg_and_die("invalid second operand " | ||
1369 | "in compatibility mode '%s'", argv[1]); | ||
1370 | } | ||
1371 | } else if (argc == 3) { | ||
1372 | if (parse_old_offset(argv[1], &o1) | ||
1373 | && parse_old_offset(argv[2], &o2) | ||
1374 | ) { | ||
1375 | n_bytes_to_skip = o1; | ||
1376 | flag_pseudo_start = 1; | ||
1377 | pseudo_start = o2; | ||
1378 | argv[2] = argv[0]; | ||
1379 | argv += 2; | ||
1380 | argc -= 2; | ||
1381 | } else { | ||
1382 | bb_error_msg_and_die("in compatibility mode " | ||
1383 | "the last two arguments must be offsets"); | ||
1384 | } | ||
1385 | } else if (argc > 3) { | ||
1386 | bb_error_msg_and_die("compatibility mode supports " | ||
1387 | "at most three arguments"); | ||
1388 | } | ||
1389 | |||
1390 | if (flag_pseudo_start) { | ||
1391 | if (format_address == format_address_none) { | ||
1392 | address_base_char = 'o'; | ||
1393 | address_pad_len_char = '7'; | ||
1394 | format_address = format_address_paren; | ||
1395 | } else | ||
1396 | format_address = format_address_label; | ||
1397 | } | ||
1398 | } | ||
1399 | #endif | ||
1400 | |||
1401 | if (limit_bytes_to_format) { | ||
1402 | end_offset = n_bytes_to_skip + max_bytes_to_format; | ||
1403 | if (end_offset < n_bytes_to_skip) | ||
1404 | bb_error_msg_and_die("skip-bytes + read-bytes is too large"); | ||
1405 | } | ||
1406 | |||
1407 | if (n_specs == 0) { | ||
1408 | decode_format_string("o2"); | ||
1409 | n_specs = 1; | ||
1410 | } | ||
1411 | |||
1412 | /* If no files were listed on the command line, | ||
1413 | set the global pointer FILE_LIST so that it | ||
1414 | references the null-terminated list of one name: "-". */ | ||
1415 | file_list = default_file_list; | ||
1416 | if (argc > 0) { | ||
1417 | /* Set the global pointer FILE_LIST so that it | ||
1418 | references the first file-argument on the command-line. */ | ||
1419 | file_list = (char const *const *) argv; | ||
1420 | } | ||
1421 | |||
1422 | /* open the first input file */ | ||
1423 | open_next_file(); | ||
1424 | /* skip over any unwanted header bytes */ | ||
1425 | skip(n_bytes_to_skip); | ||
1426 | if (!in_stream) | ||
1427 | return 1; | ||
1428 | |||
1429 | pseudo_offset = (flag_pseudo_start ? pseudo_start - n_bytes_to_skip : 0); | ||
1430 | |||
1431 | /* Compute output block length. */ | ||
1432 | l_c_m = get_lcm(); | ||
1433 | |||
1434 | if (opt & OPT_w) { /* -w: width */ | ||
1435 | bytes_per_block = 32; | ||
1436 | if (str_w) | ||
1437 | bytes_per_block = xatou(str_w); | ||
1438 | if (!bytes_per_block || bytes_per_block % l_c_m != 0) { | ||
1439 | bb_error_msg("warning: invalid width %zu; using %d instead", | ||
1440 | bytes_per_block, l_c_m); | ||
1441 | bytes_per_block = l_c_m; | ||
1442 | } | ||
1443 | } else { | ||
1444 | bytes_per_block = l_c_m; | ||
1445 | if (l_c_m < DEFAULT_BYTES_PER_BLOCK) | ||
1446 | bytes_per_block *= DEFAULT_BYTES_PER_BLOCK / l_c_m; | ||
1447 | } | ||
1448 | |||
1449 | #ifdef DEBUG | ||
1450 | for (i = 0; i < n_specs; i++) { | ||
1451 | printf("%d: fmt=\"%s\" width=%d\n", | ||
1452 | i, spec[i].fmt_string, width_bytes[spec[i].size]); | ||
1453 | } | ||
1454 | #endif | ||
1455 | |||
1456 | if (flag_dump_strings) | ||
1457 | dump_strings(); | ||
1458 | else | ||
1459 | dump(); | ||
1460 | |||
1461 | if (fclose(stdin) == EOF) | ||
1462 | bb_perror_msg_and_die(bb_msg_standard_input); | ||
1463 | |||
1464 | return (ioerror != 0); /* err != 0 - return 1 (failure) */ | ||
1465 | } |