]>
Commit | Line | Data |
---|---|---|
7cf3a80b ML |
1 | Submitted by: DJ Lucas (dj_AT_linuxfromscratch_DOT_org) |
2 | Date: 2016-02-09 | |
3 | Initial Package Version: 8.25 | |
4 | Upstream Status: Rejected | |
5 | Origin: Based on Suse's i18n patches at https://build.opensuse.org/package/view_file/Base:System/coreutils/coreutils-i18n.patch | |
6 | Description: Fixes several i18n issues with various Coreutils programs | |
7 | ||
8 | diff -Naurp coreutils-8.25-orig/lib/linebuffer.h coreutils-8.25/lib/linebuffer.h | |
9 | --- coreutils-8.25-orig/lib/linebuffer.h 2016-01-01 07:45:55.000000000 -0600 | |
10 | +++ coreutils-8.25/lib/linebuffer.h 2016-02-08 19:07:10.298944609 -0600 | |
11 | @@ -21,6 +21,11 @@ | |
12 | ||
13 | # include <stdio.h> | |
14 | ||
15 | +/* Get mbstate_t. */ | |
16 | +# if HAVE_WCHAR_H | |
17 | +# include <wchar.h> | |
18 | +# endif | |
19 | + | |
20 | /* A 'struct linebuffer' holds a line of text. */ | |
21 | ||
22 | struct linebuffer | |
23 | @@ -28,6 +33,9 @@ struct linebuffer | |
24 | size_t size; /* Allocated. */ | |
25 | size_t length; /* Used. */ | |
26 | char *buffer; | |
27 | +# if HAVE_WCHAR_H | |
28 | + mbstate_t state; | |
29 | +# endif | |
30 | }; | |
31 | ||
32 | /* Initialize linebuffer LINEBUFFER for use. */ | |
33 | diff -Naurp coreutils-8.25-orig/src/cut.c coreutils-8.25/src/cut.c | |
34 | --- coreutils-8.25-orig/src/cut.c 2016-01-13 05:08:59.000000000 -0600 | |
35 | +++ coreutils-8.25/src/cut.c 2016-02-08 19:07:10.300944616 -0600 | |
36 | @@ -28,6 +28,11 @@ | |
37 | #include <assert.h> | |
38 | #include <getopt.h> | |
39 | #include <sys/types.h> | |
40 | + | |
41 | +/* Get mbstate_t, mbrtowc(). */ | |
42 | +#if HAVE_WCHAR_H | |
43 | +# include <wchar.h> | |
44 | +#endif | |
45 | #include "system.h" | |
46 | ||
47 | #include "error.h" | |
48 | @@ -38,6 +43,18 @@ | |
49 | ||
50 | #include "set-fields.h" | |
51 | ||
52 | +/* MB_LEN_MAX is incorrectly defined to be 1 in at least one GCC | |
53 | + installation; work around this configuration error. */ | |
54 | +#if !defined MB_LEN_MAX || MB_LEN_MAX < 2 | |
55 | +# undef MB_LEN_MAX | |
56 | +# define MB_LEN_MAX 16 | |
57 | +#endif | |
58 | + | |
59 | +/* Some systems, like BeOS, have multibyte encodings but lack mbstate_t. */ | |
60 | +#if HAVE_MBRTOWC && defined mbstate_t | |
61 | +# define mbrtowc(pwc, s, n, ps) (mbrtowc) (pwc, s, n, 0) | |
62 | +#endif | |
63 | + | |
64 | /* The official name of this program (e.g., no 'g' prefix). */ | |
65 | #define PROGRAM_NAME "cut" | |
66 | ||
67 | @@ -54,6 +71,52 @@ | |
68 | } \ | |
69 | while (0) | |
70 | ||
71 | +/* Refill the buffer BUF to get a multibyte character. */ | |
72 | +#define REFILL_BUFFER(BUF, BUFPOS, BUFLEN, STREAM) \ | |
73 | + do \ | |
74 | + { \ | |
75 | + if (BUFLEN < MB_LEN_MAX && !feof (STREAM) && !ferror (STREAM)) \ | |
76 | + { \ | |
77 | + memmove (BUF, BUFPOS, BUFLEN); \ | |
78 | + BUFLEN += fread (BUF + BUFLEN, sizeof(char), BUFSIZ, STREAM); \ | |
79 | + BUFPOS = BUF; \ | |
80 | + } \ | |
81 | + } \ | |
82 | + while (0) | |
83 | + | |
84 | +/* Get wide character on BUFPOS. BUFPOS is not included after that. | |
85 | + If byte sequence is not valid as a character, CONVFAIL is true. Otherwise false. */ | |
86 | +#define GET_NEXT_WC_FROM_BUFFER(WC, BUFPOS, BUFLEN, MBLENGTH, STATE, CONVFAIL) \ | |
87 | + do \ | |
88 | + { \ | |
89 | + mbstate_t state_bak; \ | |
90 | + \ | |
91 | + if (BUFLEN < 1) \ | |
92 | + { \ | |
93 | + WC = WEOF; \ | |
94 | + break; \ | |
95 | + } \ | |
96 | + \ | |
97 | + /* Get a wide character. */ \ | |
98 | + CONVFAIL = false; \ | |
99 | + state_bak = STATE; \ | |
100 | + MBLENGTH = mbrtowc ((wchar_t *)&WC, BUFPOS, BUFLEN, &STATE); \ | |
101 | + \ | |
102 | + switch (MBLENGTH) \ | |
103 | + { \ | |
104 | + case (size_t)-1: \ | |
105 | + case (size_t)-2: \ | |
106 | + CONVFAIL = true; \ | |
107 | + STATE = state_bak; \ | |
108 | + /* Fall througn. */ \ | |
109 | + \ | |
110 | + case 0: \ | |
111 | + MBLENGTH = 1; \ | |
112 | + break; \ | |
113 | + } \ | |
114 | + } \ | |
115 | + while (0) | |
116 | + | |
117 | ||
118 | /* Pointer inside RP. When checking if a byte or field is selected | |
119 | by a finite range, we check if it is between CURRENT_RP.LO | |
120 | @@ -61,6 +124,9 @@ | |
121 | CURRENT_RP.HI then we make CURRENT_RP to point to the next range pair. */ | |
122 | static struct field_range_pair *current_rp; | |
123 | ||
124 | +/* Length of the delimiter given as argument to -d. */ | |
125 | +size_t delimlen; | |
126 | + | |
127 | /* This buffer is used to support the semantics of the -s option | |
128 | (or lack of same) when the specified field list includes (does | |
129 | not include) the first field. In both of those cases, the entire | |
130 | @@ -77,15 +143,25 @@ enum operating_mode | |
131 | { | |
132 | undefined_mode, | |
133 | ||
134 | - /* Output characters that are in the given bytes. */ | |
135 | + /* Output bytes that are at the given positions. */ | |
136 | byte_mode, | |
137 | ||
138 | + /* Output characters that are at the given positions. */ | |
139 | + character_mode, | |
140 | + | |
141 | /* Output the given delimiter-separated fields. */ | |
142 | field_mode | |
143 | }; | |
144 | ||
145 | static enum operating_mode operating_mode; | |
146 | ||
147 | +/* If nonzero, when in byte mode, don't split multibyte characters. */ | |
148 | +static int byte_mode_character_aware; | |
149 | + | |
150 | +/* If nonzero, the function for single byte locale is work | |
151 | + if this program runs on multibyte locale. */ | |
152 | +static int force_singlebyte_mode; | |
153 | + | |
154 | /* If true do not output lines containing no delimiter characters. | |
155 | Otherwise, all such lines are printed. This option is valid only | |
156 | with field mode. */ | |
157 | @@ -97,6 +173,9 @@ static bool complement; | |
158 | ||
159 | /* The delimiter character for field mode. */ | |
160 | static unsigned char delim; | |
161 | +#if HAVE_WCHAR_H | |
162 | +static wchar_t wcdelim; | |
163 | +#endif | |
164 | ||
165 | /* The delimiter for each line/record. */ | |
166 | static unsigned char line_delim = '\n'; | |
167 | @@ -164,7 +243,7 @@ Print selected parts of lines from each | |
168 | -f, --fields=LIST select only these fields; also print any line\n\ | |
169 | that contains no delimiter character, unless\n\ | |
170 | the -s option is specified\n\ | |
171 | - -n (ignored)\n\ | |
172 | + -n with -b: don't split multibyte characters\n\ | |
173 | "), stdout); | |
174 | fputs (_("\ | |
175 | --complement complement the set of selected bytes, characters\n\ | |
176 | @@ -280,6 +359,82 @@ cut_bytes (FILE *stream) | |
177 | } | |
178 | } | |
179 | ||
180 | +#if HAVE_MBRTOWC | |
181 | +/* This function is in use for the following case. | |
182 | + | |
183 | + 1. Read from the stream STREAM, printing to standard output any selected | |
184 | + characters. | |
185 | + | |
186 | + 2. Read from stream STREAM, printing to standard output any selected bytes, | |
187 | + without splitting multibyte characters. */ | |
188 | + | |
189 | +static void | |
190 | +cut_characters_or_cut_bytes_no_split (FILE *stream) | |
191 | +{ | |
192 | + size_t idx; /* number of bytes or characters in the line so far. */ | |
193 | + char buf[MB_LEN_MAX + BUFSIZ]; /* For spooling a read byte sequence. */ | |
194 | + char *bufpos; /* Next read position of BUF. */ | |
195 | + size_t buflen; /* The length of the byte sequence in buf. */ | |
196 | + wint_t wc; /* A gotten wide character. */ | |
197 | + size_t mblength; /* The byte size of a multibyte character which shows | |
198 | + as same character as WC. */ | |
199 | + mbstate_t state; /* State of the stream. */ | |
200 | + bool convfail = false; /* true, when conversion failed. Otherwise false. */ | |
201 | + /* Whether to begin printing delimiters between ranges for the current line. | |
202 | + Set after we've begun printing data corresponding to the first range. */ | |
203 | + bool print_delimiter = false; | |
204 | + | |
205 | + idx = 0; | |
206 | + buflen = 0; | |
207 | + bufpos = buf; | |
208 | + memset (&state, '\0', sizeof(mbstate_t)); | |
209 | + | |
210 | + current_rp = frp; | |
211 | + | |
212 | + while (1) | |
213 | + { | |
214 | + REFILL_BUFFER (buf, bufpos, buflen, stream); | |
215 | + | |
216 | + GET_NEXT_WC_FROM_BUFFER (wc, bufpos, buflen, mblength, state, convfail); | |
217 | + (void) convfail; /* ignore unused */ | |
218 | + | |
219 | + if (wc == WEOF) | |
220 | + { | |
221 | + if (idx > 0) | |
222 | + putchar (line_delim); | |
223 | + break; | |
224 | + } | |
225 | + else if (wc == line_delim) | |
226 | + { | |
227 | + putchar (line_delim); | |
228 | + idx = 0; | |
229 | + print_delimiter = false; | |
230 | + current_rp = frp; | |
231 | + } | |
232 | + else | |
233 | + { | |
234 | + next_item (&idx); | |
235 | + if (print_kth (idx)) | |
236 | + { | |
237 | + if (output_delimiter_specified) | |
238 | + { | |
239 | + if (print_delimiter && is_range_start_index (idx)) | |
240 | + { | |
241 | + fwrite (output_delimiter_string, sizeof (char), | |
242 | + output_delimiter_length, stdout); | |
243 | + } | |
244 | + print_delimiter = true; | |
245 | + } | |
246 | + fwrite (bufpos, mblength, sizeof(char), stdout); | |
247 | + } | |
248 | + } | |
249 | + | |
250 | + buflen -= mblength; | |
251 | + bufpos += mblength; | |
252 | + } | |
253 | +} | |
254 | +#endif | |
255 | + | |
256 | /* Read from stream STREAM, printing to standard output any selected fields. */ | |
257 | ||
258 | static void | |
259 | @@ -425,13 +580,211 @@ cut_fields (FILE *stream) | |
260 | } | |
261 | } | |
262 | ||
263 | +#if HAVE_MBRTOWC | |
264 | +static void | |
265 | +cut_fields_mb (FILE *stream) | |
266 | +{ | |
267 | + int c; | |
268 | + size_t field_idx; | |
269 | + int found_any_selected_field; | |
270 | + int buffer_first_field; | |
271 | + int empty_input; | |
272 | + char buf[MB_LEN_MAX + BUFSIZ]; /* For spooling a read byte sequence. */ | |
273 | + char *bufpos; /* Next read position of BUF. */ | |
274 | + size_t buflen; /* The length of the byte sequence in buf. */ | |
275 | + wint_t wc = 0; /* A gotten wide character. */ | |
276 | + size_t mblength; /* The byte size of a multibyte character which shows | |
277 | + as same character as WC. */ | |
278 | + mbstate_t state; /* State of the stream. */ | |
279 | + bool convfail = false; /* true, when conversion failed. Otherwise false. */ | |
280 | + | |
281 | + current_rp = frp; | |
282 | + | |
283 | + found_any_selected_field = 0; | |
284 | + field_idx = 1; | |
285 | + bufpos = buf; | |
286 | + buflen = 0; | |
287 | + memset (&state, '\0', sizeof(mbstate_t)); | |
288 | + | |
289 | + c = getc (stream); | |
290 | + empty_input = (c == EOF); | |
291 | + if (c != EOF) | |
292 | + { | |
293 | + ungetc (c, stream); | |
294 | + wc = 0; | |
295 | + } | |
296 | + else | |
297 | + wc = WEOF; | |
298 | + | |
299 | + /* To support the semantics of the -s flag, we may have to buffer | |
300 | + all of the first field to determine whether it is `delimited.' | |
301 | + But that is unnecessary if all non-delimited lines must be printed | |
302 | + and the first field has been selected, or if non-delimited lines | |
303 | + must be suppressed and the first field has *not* been selected. | |
304 | + That is because a non-delimited line has exactly one field. */ | |
305 | + buffer_first_field = (suppress_non_delimited ^ !print_kth (1)); | |
306 | + | |
307 | + while (1) | |
308 | + { | |
309 | + if (field_idx == 1 && buffer_first_field) | |
310 | + { | |
311 | + int len = 0; | |
312 | + | |
313 | + while (1) | |
314 | + { | |
315 | + REFILL_BUFFER (buf, bufpos, buflen, stream); | |
316 | + | |
317 | + GET_NEXT_WC_FROM_BUFFER | |
318 | + (wc, bufpos, buflen, mblength, state, convfail); | |
319 | + | |
320 | + if (wc == WEOF) | |
321 | + break; | |
322 | + | |
323 | + field_1_buffer = xrealloc (field_1_buffer, len + mblength); | |
324 | + memcpy (field_1_buffer + len, bufpos, mblength); | |
325 | + len += mblength; | |
326 | + buflen -= mblength; | |
327 | + bufpos += mblength; | |
328 | + | |
329 | + if (!convfail && (wc == line_delim || wc == wcdelim)) | |
330 | + break; | |
331 | + } | |
332 | + | |
333 | + if (len <= 0 && wc == WEOF) | |
334 | + break; | |
335 | + | |
336 | + /* If the first field extends to the end of line (it is not | |
337 | + delimited) and we are printing all non-delimited lines, | |
338 | + print this one. */ | |
339 | + if (convfail || (!convfail && wc != wcdelim)) | |
340 | + { | |
341 | + if (suppress_non_delimited) | |
342 | + { | |
343 | + /* Empty. */ | |
344 | + } | |
345 | + else | |
346 | + { | |
347 | + fwrite (field_1_buffer, sizeof (char), len, stdout); | |
348 | + /* Make sure the output line is newline terminated. */ | |
349 | + if (convfail || (!convfail && wc != line_delim)) | |
350 | + putchar (line_delim); | |
351 | + } | |
352 | + continue; | |
353 | + } | |
354 | + | |
355 | + if (print_kth (1)) | |
356 | + { | |
357 | + /* Print the field, but not the trailing delimiter. */ | |
358 | + fwrite (field_1_buffer, sizeof (char), len - 1, stdout); | |
359 | + found_any_selected_field = 1; | |
360 | + } | |
361 | + next_item (&field_idx); | |
362 | + } | |
363 | + | |
364 | + if (wc != WEOF) | |
365 | + { | |
366 | + if (print_kth (field_idx)) | |
367 | + { | |
368 | + if (found_any_selected_field) | |
369 | + { | |
370 | + fwrite (output_delimiter_string, sizeof (char), | |
371 | + output_delimiter_length, stdout); | |
372 | + } | |
373 | + found_any_selected_field = 1; | |
374 | + } | |
375 | + | |
376 | + while (1) | |
377 | + { | |
378 | + REFILL_BUFFER (buf, bufpos, buflen, stream); | |
379 | + | |
380 | + GET_NEXT_WC_FROM_BUFFER | |
381 | + (wc, bufpos, buflen, mblength, state, convfail); | |
382 | + | |
383 | + if (wc == WEOF) | |
384 | + break; | |
385 | + else if (!convfail && (wc == wcdelim || wc == line_delim)) | |
386 | + { | |
387 | + buflen -= mblength; | |
388 | + bufpos += mblength; | |
389 | + break; | |
390 | + } | |
391 | + | |
392 | + if (print_kth (field_idx)) | |
393 | + fwrite (bufpos, mblength, sizeof(char), stdout); | |
394 | + | |
395 | + buflen -= mblength; | |
396 | + bufpos += mblength; | |
397 | + } | |
398 | + } | |
399 | + | |
400 | + if ((!convfail || wc == line_delim) && buflen < 1) | |
401 | + wc = WEOF; | |
402 | + | |
403 | + if (!convfail && wc == wcdelim) | |
404 | + next_item (&field_idx); | |
405 | + else if (wc == WEOF || (!convfail && wc == line_delim)) | |
406 | + { | |
407 | + if (found_any_selected_field | |
408 | + || (!empty_input && !(suppress_non_delimited && field_idx == 1))) | |
409 | + putchar (line_delim); | |
410 | + if (wc == WEOF) | |
411 | + break; | |
412 | + field_idx = 1; | |
413 | + current_rp = frp; | |
414 | + found_any_selected_field = 0; | |
415 | + } | |
416 | + } | |
417 | +} | |
418 | +#endif | |
419 | + | |
420 | static void | |
421 | cut_stream (FILE *stream) | |
422 | { | |
423 | - if (operating_mode == byte_mode) | |
424 | - cut_bytes (stream); | |
425 | +#if HAVE_MBRTOWC | |
426 | + if (MB_CUR_MAX > 1 && !force_singlebyte_mode) | |
427 | + { | |
428 | + switch (operating_mode) | |
429 | + { | |
430 | + case byte_mode: | |
431 | + if (byte_mode_character_aware) | |
432 | + cut_characters_or_cut_bytes_no_split (stream); | |
433 | + else | |
434 | + cut_bytes (stream); | |
435 | + break; | |
436 | + | |
437 | + case character_mode: | |
438 | + cut_characters_or_cut_bytes_no_split (stream); | |
439 | + break; | |
440 | + | |
441 | + case field_mode: | |
442 | + if (delimlen == 1) | |
443 | + { | |
444 | + /* Check if we have utf8 multibyte locale, so we can use this | |
445 | + optimization because of uniqueness of characters, which is | |
446 | + not true for e.g. SJIS */ | |
447 | + char * loc = setlocale(LC_CTYPE, NULL); | |
448 | + if (loc && (strstr (loc, "UTF-8") || strstr (loc, "utf-8") || | |
449 | + strstr (loc, "UTF8") || strstr (loc, "utf8"))) | |
450 | + { | |
451 | + cut_fields (stream); | |
452 | + break; | |
453 | + } | |
454 | + } | |
455 | + cut_fields_mb (stream); | |
456 | + break; | |
457 | + | |
458 | + default: | |
459 | + abort (); | |
460 | + } | |
461 | + } | |
462 | else | |
463 | - cut_fields (stream); | |
464 | +#endif | |
465 | + { | |
466 | + if (operating_mode == field_mode) | |
467 | + cut_fields (stream); | |
468 | + else | |
469 | + cut_bytes (stream); | |
470 | + } | |
471 | } | |
472 | ||
473 | /* Process file FILE to standard output. | |
474 | @@ -483,6 +836,7 @@ main (int argc, char **argv) | |
475 | bool ok; | |
476 | bool delim_specified = false; | |
477 | char *spec_list_string IF_LINT ( = NULL); | |
478 | + char mbdelim[MB_LEN_MAX + 1]; | |
479 | ||
480 | initialize_main (&argc, &argv); | |
481 | set_program_name (argv[0]); | |
482 | @@ -505,7 +859,6 @@ main (int argc, char **argv) | |
483 | switch (optc) | |
484 | { | |
485 | case 'b': | |
486 | - case 'c': | |
487 | /* Build the byte list. */ | |
488 | if (operating_mode != undefined_mode) | |
489 | FATAL_ERROR (_("only one type of list may be specified")); | |
490 | @@ -513,6 +866,14 @@ main (int argc, char **argv) | |
491 | spec_list_string = optarg; | |
492 | break; | |
493 | ||
494 | + case 'c': | |
495 | + /* Build the character list. */ | |
496 | + if (operating_mode != undefined_mode) | |
497 | + FATAL_ERROR (_("only one type of list may be specified")); | |
498 | + operating_mode = character_mode; | |
499 | + spec_list_string = optarg; | |
500 | + break; | |
501 | + | |
502 | case 'f': | |
503 | /* Build the field list. */ | |
504 | if (operating_mode != undefined_mode) | |
505 | @@ -524,10 +885,38 @@ main (int argc, char **argv) | |
506 | case 'd': | |
507 | /* New delimiter. */ | |
508 | /* Interpret -d '' to mean 'use the NUL byte as the delimiter.' */ | |
509 | - if (optarg[0] != '\0' && optarg[1] != '\0') | |
510 | - FATAL_ERROR (_("the delimiter must be a single character")); | |
511 | - delim = optarg[0]; | |
512 | - delim_specified = true; | |
513 | + { | |
514 | +#if HAVE_MBRTOWC | |
515 | + if(MB_CUR_MAX > 1) | |
516 | + { | |
517 | + mbstate_t state; | |
518 | + | |
519 | + memset (&state, '\0', sizeof(mbstate_t)); | |
520 | + delimlen = mbrtowc (&wcdelim, optarg, strnlen(optarg, MB_LEN_MAX), &state); | |
521 | + | |
522 | + if (delimlen == (size_t)-1 || delimlen == (size_t)-2) | |
523 | + ++force_singlebyte_mode; | |
524 | + else | |
525 | + { | |
526 | + delimlen = (delimlen < 1) ? 1 : delimlen; | |
527 | + if (wcdelim != L'\0' && *(optarg + delimlen) != '\0') | |
528 | + FATAL_ERROR (_("the delimiter must be a single character")); | |
529 | + memcpy (mbdelim, optarg, delimlen); | |
530 | + mbdelim[delimlen] = '\0'; | |
531 | + if (delimlen == 1) | |
532 | + delim = *optarg; | |
533 | + } | |
534 | + } | |
535 | + | |
536 | + if (MB_CUR_MAX <= 1 || force_singlebyte_mode) | |
537 | +#endif | |
538 | + { | |
539 | + if (optarg[0] != '\0' && optarg[1] != '\0') | |
540 | + FATAL_ERROR (_("the delimiter must be a single character")); | |
541 | + delim = (unsigned char) optarg[0]; | |
542 | + } | |
543 | + delim_specified = true; | |
544 | + } | |
545 | break; | |
546 | ||
547 | case OUTPUT_DELIMITER_OPTION: | |
548 | @@ -540,6 +929,7 @@ main (int argc, char **argv) | |
549 | break; | |
550 | ||
551 | case 'n': | |
552 | + byte_mode_character_aware = 1; | |
553 | break; | |
554 | ||
555 | case 's': | |
556 | @@ -579,15 +969,34 @@ main (int argc, char **argv) | |
557 | | (complement ? SETFLD_COMPLEMENT : 0) ); | |
558 | ||
559 | if (!delim_specified) | |
560 | - delim = '\t'; | |
561 | + { | |
562 | + delim = '\t'; | |
563 | +#ifdef HAVE_MBRTOWC | |
564 | + wcdelim = L'\t'; | |
565 | + mbdelim[0] = '\t'; | |
566 | + mbdelim[1] = '\0'; | |
567 | + delimlen = 1; | |
568 | +#endif | |
569 | + } | |
570 | ||
571 | if (output_delimiter_string == NULL) | |
572 | { | |
573 | - static char dummy[2]; | |
574 | - dummy[0] = delim; | |
575 | - dummy[1] = '\0'; | |
576 | - output_delimiter_string = dummy; | |
577 | - output_delimiter_length = 1; | |
578 | +#ifdef HAVE_MBRTOWC | |
579 | + if (MB_CUR_MAX > 1 && !force_singlebyte_mode) | |
580 | + { | |
581 | + output_delimiter_string = xstrdup(mbdelim); | |
582 | + output_delimiter_length = delimlen; | |
583 | + } | |
584 | + | |
585 | + if (MB_CUR_MAX <= 1 || force_singlebyte_mode) | |
586 | +#endif | |
587 | + { | |
588 | + static char dummy[2]; | |
589 | + dummy[0] = delim; | |
590 | + dummy[1] = '\0'; | |
591 | + output_delimiter_string = dummy; | |
592 | + output_delimiter_length = 1; | |
593 | + } | |
594 | } | |
595 | ||
596 | if (optind == argc) | |
597 | diff -Naurp coreutils-8.25-orig/src/expand.c coreutils-8.25/src/expand.c | |
598 | --- coreutils-8.25-orig/src/expand.c 2016-01-01 07:48:50.000000000 -0600 | |
599 | +++ coreutils-8.25/src/expand.c 2016-02-08 19:07:10.301944619 -0600 | |
600 | @@ -37,12 +37,34 @@ | |
601 | #include <stdio.h> | |
602 | #include <getopt.h> | |
603 | #include <sys/types.h> | |
604 | + | |
605 | +/* Get mbstate_t, mbrtowc(), wcwidth(). */ | |
606 | +#if HAVE_WCHAR_H | |
607 | +# include <wchar.h> | |
608 | +#endif | |
609 | + | |
610 | +/* Get iswblank(). */ | |
611 | +#if HAVE_WCTYPE_H | |
612 | +# include <wctype.h> | |
613 | +#endif | |
614 | + | |
615 | #include "system.h" | |
616 | #include "error.h" | |
617 | #include "fadvise.h" | |
618 | #include "quote.h" | |
619 | #include "xstrndup.h" | |
620 | ||
621 | +/* MB_LEN_MAX is incorrectly defined to be 1 in at least one GCC | |
622 | + installation; work around this configuration error. */ | |
623 | +#if !defined MB_LEN_MAX || MB_LEN_MAX < 2 | |
624 | +# define MB_LEN_MAX 16 | |
625 | +#endif | |
626 | + | |
627 | +/* Some systems, like BeOS, have multibyte encodings but lack mbstate_t. */ | |
628 | +#if HAVE_MBRTOWC && defined mbstate_t | |
629 | +# define mbrtowc(pwc, s, n, ps) (mbrtowc) (pwc, s, n, 0) | |
630 | +#endif | |
631 | + | |
632 | /* The official name of this program (e.g., no 'g' prefix). */ | |
633 | #define PROGRAM_NAME "expand" | |
634 | ||
635 | @@ -357,6 +379,142 @@ expand (void) | |
636 | } | |
637 | } | |
638 | ||
639 | +#if HAVE_MBRTOWC | |
640 | +static void | |
641 | +expand_multibyte (void) | |
642 | +{ | |
643 | + FILE *fp; /* Input strem. */ | |
644 | + mbstate_t i_state; /* Current shift state of the input stream. */ | |
645 | + mbstate_t i_state_bak; /* Back up the I_STATE. */ | |
646 | + mbstate_t o_state; /* Current shift state of the output stream. */ | |
647 | + char buf[MB_LEN_MAX + BUFSIZ]; /* For spooling a read byte sequence. */ | |
648 | + char *bufpos = buf; /* Next read position of BUF. */ | |
649 | + size_t buflen = 0; /* The length of the byte sequence in buf. */ | |
650 | + wchar_t wc; /* A gotten wide character. */ | |
651 | + size_t mblength; /* The byte size of a multibyte character | |
652 | + which shows as same character as WC. */ | |
653 | + int tab_index = 0; /* Index in `tab_list' of next tabstop. */ | |
654 | + int column = 0; /* Column on screen of the next char. */ | |
655 | + int next_tab_column; /* Column the next tab stop is on. */ | |
656 | + int convert = 1; /* If nonzero, perform translations. */ | |
657 | + | |
658 | + fp = next_file ((FILE *) NULL); | |
659 | + if (fp == NULL) | |
660 | + return; | |
661 | + | |
662 | + memset (&o_state, '\0', sizeof(mbstate_t)); | |
663 | + memset (&i_state, '\0', sizeof(mbstate_t)); | |
664 | + | |
665 | + for (;;) | |
666 | + { | |
667 | + /* Refill the buffer BUF. */ | |
668 | + if (buflen < MB_LEN_MAX && !feof(fp) && !ferror(fp)) | |
669 | + { | |
670 | + memmove (buf, bufpos, buflen); | |
671 | + buflen += fread (buf + buflen, sizeof(char), BUFSIZ, fp); | |
672 | + bufpos = buf; | |
673 | + } | |
674 | + | |
675 | + /* No character is left in BUF. */ | |
676 | + if (buflen < 1) | |
677 | + { | |
678 | + fp = next_file (fp); | |
679 | + | |
680 | + if (fp == NULL) | |
681 | + break; /* No more files. */ | |
682 | + else | |
683 | + { | |
684 | + memset (&i_state, '\0', sizeof(mbstate_t)); | |
685 | + continue; | |
686 | + } | |
687 | + } | |
688 | + | |
689 | + /* Get a wide character. */ | |
690 | + i_state_bak = i_state; | |
691 | + mblength = mbrtowc (&wc, bufpos, buflen, &i_state); | |
692 | + | |
693 | + switch (mblength) | |
694 | + { | |
695 | + case (size_t)-1: /* illegal byte sequence. */ | |
696 | + case (size_t)-2: | |
697 | + mblength = 1; | |
698 | + i_state = i_state_bak; | |
699 | + if (convert) | |
700 | + { | |
701 | + ++column; | |
702 | + if (convert_entire_line == 0 && !isblank(*bufpos)) | |
703 | + convert = 0; | |
704 | + } | |
705 | + putchar (*bufpos); | |
706 | + break; | |
707 | + | |
708 | + case 0: /* null. */ | |
709 | + mblength = 1; | |
710 | + if (convert && convert_entire_line == 0) | |
711 | + convert = 0; | |
712 | + putchar ('\0'); | |
713 | + break; | |
714 | + | |
715 | + default: | |
716 | + if (wc == L'\n') /* LF. */ | |
717 | + { | |
718 | + tab_index = 0; | |
719 | + column = 0; | |
720 | + convert = 1; | |
721 | + putchar ('\n'); | |
722 | + } | |
723 | + else if (wc == L'\t' && convert) /* Tab. */ | |
724 | + { | |
725 | + if (tab_size == 0) | |
726 | + { | |
727 | + /* Do not let tab_index == first_free_tab; | |
728 | + stop when it is 1 less. */ | |
729 | + while (tab_index < first_free_tab - 1 | |
730 | + && column >= tab_list[tab_index]) | |
731 | + tab_index++; | |
732 | + next_tab_column = tab_list[tab_index]; | |
733 | + if (tab_index < first_free_tab - 1) | |
734 | + tab_index++; | |
735 | + if (column >= next_tab_column) | |
736 | + next_tab_column = column + 1; | |
737 | + } | |
738 | + else | |
739 | + next_tab_column = column + tab_size - column % tab_size; | |
740 | + | |
741 | + while (column < next_tab_column) | |
742 | + { | |
743 | + putchar (' '); | |
744 | + ++column; | |
745 | + } | |
746 | + } | |
747 | + else /* Others. */ | |
748 | + { | |
749 | + if (convert) | |
750 | + { | |
751 | + if (wc == L'\b') | |
752 | + { | |
753 | + if (column > 0) | |
754 | + --column; | |
755 | + } | |
756 | + else | |
757 | + { | |
758 | + int width; /* The width of WC. */ | |
759 | + | |
760 | + width = wcwidth (wc); | |
761 | + column += (width > 0) ? width : 0; | |
762 | + if (convert_entire_line == 0 && !iswblank(wc)) | |
763 | + convert = 0; | |
764 | + } | |
765 | + } | |
766 | + fwrite (bufpos, sizeof(char), mblength, stdout); | |
767 | + } | |
768 | + } | |
769 | + buflen -= mblength; | |
770 | + bufpos += mblength; | |
771 | + } | |
772 | +} | |
773 | +#endif | |
774 | + | |
775 | int | |
776 | main (int argc, char **argv) | |
777 | { | |
778 | @@ -421,7 +579,12 @@ main (int argc, char **argv) | |
779 | ||
780 | file_list = (optind < argc ? &argv[optind] : stdin_argv); | |
781 | ||
782 | - expand (); | |
783 | +#if HAVE_MBRTOWC | |
784 | + if (MB_CUR_MAX > 1) | |
785 | + expand_multibyte (); | |
786 | + else | |
787 | +#endif | |
788 | + expand (); | |
789 | ||
790 | if (have_read_stdin && fclose (stdin) != 0) | |
791 | error (EXIT_FAILURE, errno, "-"); | |
792 | diff -Naurp coreutils-8.25-orig/src/fold.c coreutils-8.25/src/fold.c | |
793 | --- coreutils-8.25-orig/src/fold.c 2016-01-01 07:48:50.000000000 -0600 | |
794 | +++ coreutils-8.25/src/fold.c 2016-02-08 19:07:10.302944622 -0600 | |
795 | @@ -22,11 +22,33 @@ | |
796 | #include <getopt.h> | |
797 | #include <sys/types.h> | |
798 | ||
799 | +/* Get mbstate_t, mbrtowc(), wcwidth(). */ | |
800 | +#if HAVE_WCHAR_H | |
801 | +# include <wchar.h> | |
802 | +#endif | |
803 | + | |
804 | +/* Get iswprint(), iswblank(), wcwidth(). */ | |
805 | +#if HAVE_WCTYPE_H | |
806 | +# include <wctype.h> | |
807 | +#endif | |
808 | + | |
809 | #include "system.h" | |
810 | #include "error.h" | |
811 | #include "fadvise.h" | |
812 | #include "xdectoint.h" | |
813 | ||
814 | +/* MB_LEN_MAX is incorrectly defined to be 1 in at least one GCC | |
815 | + installation; work around this configuration error. */ | |
816 | +#if !defined MB_LEN_MAX || MB_LEN_MAX < 2 | |
817 | +# undef MB_LEN_MAX | |
818 | +# define MB_LEN_MAX 16 | |
819 | +#endif | |
820 | + | |
821 | +/* Some systems, like BeOS, have multibyte encodings but lack mbstate_t. */ | |
822 | +#if HAVE_MBRTOWC && defined mbstate_t | |
823 | +# define mbrtowc(pwc, s, n, ps) (mbrtowc) (pwc, s, n, 0) | |
824 | +#endif | |
825 | + | |
826 | #define TAB_WIDTH 8 | |
827 | ||
828 | /* The official name of this program (e.g., no 'g' prefix). */ | |
829 | @@ -34,20 +56,41 @@ | |
830 | ||
831 | #define AUTHORS proper_name ("David MacKenzie") | |
832 | ||
833 | +#define FATAL_ERROR(Message) \ | |
834 | + do \ | |
835 | + { \ | |
836 | + error (0, 0, (Message)); \ | |
837 | + usage (2); \ | |
838 | + } \ | |
839 | + while (0) | |
840 | + | |
841 | +enum operating_mode | |
842 | +{ | |
843 | + /* Fold texts by columns that are at the given positions. */ | |
844 | + column_mode, | |
845 | + | |
846 | + /* Fold texts by bytes that are at the given positions. */ | |
847 | + byte_mode, | |
848 | + | |
849 | + /* Fold texts by characters that are at the given positions. */ | |
850 | + character_mode, | |
851 | +}; | |
852 | + | |
853 | +/* The argument shows current mode. (Default: column_mode) */ | |
854 | +static enum operating_mode operating_mode; | |
855 | + | |
856 | /* If nonzero, try to break on whitespace. */ | |
857 | static bool break_spaces; | |
858 | ||
859 | -/* If nonzero, count bytes, not column positions. */ | |
860 | -static bool count_bytes; | |
861 | - | |
862 | /* If nonzero, at least one of the files we read was standard input. */ | |
863 | static bool have_read_stdin; | |
864 | ||
865 | -static char const shortopts[] = "bsw:0::1::2::3::4::5::6::7::8::9::"; | |
866 | +static char const shortopts[] = "bcsw:0::1::2::3::4::5::6::7::8::9::"; | |
867 | ||
868 | static struct option const longopts[] = | |
869 | { | |
870 | {"bytes", no_argument, NULL, 'b'}, | |
871 | + {"characters", no_argument, NULL, 'c'}, | |
872 | {"spaces", no_argument, NULL, 's'}, | |
873 | {"width", required_argument, NULL, 'w'}, | |
874 | {GETOPT_HELP_OPTION_DECL}, | |
875 | @@ -75,6 +118,7 @@ Wrap input lines in each FILE, writing t | |
876 | ||
877 | fputs (_("\ | |
878 | -b, --bytes count bytes rather than columns\n\ | |
879 | + -c, --characters count characters rather than columns\n\ | |
880 | -s, --spaces break at spaces\n\ | |
881 | -w, --width=WIDTH use WIDTH columns instead of 80\n\ | |
882 | "), stdout); | |
883 | @@ -92,7 +136,7 @@ Wrap input lines in each FILE, writing t | |
884 | static size_t | |
885 | adjust_column (size_t column, char c) | |
886 | { | |
887 | - if (!count_bytes) | |
888 | + if (operating_mode != byte_mode) | |
889 | { | |
890 | if (c == '\b') | |
891 | { | |
892 | @@ -115,30 +159,14 @@ adjust_column (size_t column, char c) | |
893 | to stdout, with maximum line length WIDTH. | |
894 | Return true if successful. */ | |
895 | ||
896 | -static bool | |
897 | -fold_file (char const *filename, size_t width) | |
898 | +static void | |
899 | +fold_text (FILE *istream, size_t width, int *saved_errno) | |
900 | { | |
901 | - FILE *istream; | |
902 | int c; | |
903 | size_t column = 0; /* Screen column where next char will go. */ | |
904 | size_t offset_out = 0; /* Index in 'line_out' for next char. */ | |
905 | static char *line_out = NULL; | |
906 | static size_t allocated_out = 0; | |
907 | - int saved_errno; | |
908 | - | |
909 | - if (STREQ (filename, "-")) | |
910 | - { | |
911 | - istream = stdin; | |
912 | - have_read_stdin = true; | |
913 | - } | |
914 | - else | |
915 | - istream = fopen (filename, "r"); | |
916 | - | |
917 | - if (istream == NULL) | |
918 | - { | |
919 | - error (0, errno, "%s", quotef (filename)); | |
920 | - return false; | |
921 | - } | |
922 | ||
923 | fadvise (istream, FADVISE_SEQUENTIAL); | |
924 | ||
925 | @@ -168,6 +196,15 @@ fold_file (char const *filename, size_t | |
926 | bool found_blank = false; | |
927 | size_t logical_end = offset_out; | |
928 | ||
929 | + /* If LINE_OUT has no wide character, | |
930 | + put a new wide character in LINE_OUT | |
931 | + if column is bigger than width. */ | |
932 | + if (offset_out == 0) | |
933 | + { | |
934 | + line_out[offset_out++] = c; | |
935 | + continue; | |
936 | + } | |
937 | + | |
938 | /* Look for the last blank. */ | |
939 | while (logical_end) | |
940 | { | |
941 | @@ -214,11 +251,221 @@ fold_file (char const *filename, size_t | |
942 | line_out[offset_out++] = c; | |
943 | } | |
944 | ||
945 | - saved_errno = errno; | |
946 | + *saved_errno = errno; | |
947 | + | |
948 | + if (offset_out) | |
949 | + fwrite (line_out, sizeof (char), (size_t) offset_out, stdout); | |
950 | + | |
951 | +} | |
952 | + | |
953 | +#if HAVE_MBRTOWC | |
954 | +static void | |
955 | +fold_multibyte_text (FILE *istream, size_t width, int *saved_errno) | |
956 | +{ | |
957 | + char buf[MB_LEN_MAX + BUFSIZ]; /* For spooling a read byte sequence. */ | |
958 | + size_t buflen = 0; /* The length of the byte sequence in buf. */ | |
959 | + char *bufpos = buf; /* Next read position of BUF. */ | |
960 | + wint_t wc; /* A gotten wide character. */ | |
961 | + size_t mblength; /* The byte size of a multibyte character which shows | |
962 | + as same character as WC. */ | |
963 | + mbstate_t state, state_bak; /* State of the stream. */ | |
964 | + int convfail = 0; /* 1, when conversion is failed. Otherwise 0. */ | |
965 | + | |
966 | + static char *line_out = NULL; | |
967 | + size_t offset_out = 0; /* Index in `line_out' for next char. */ | |
968 | + static size_t allocated_out = 0; | |
969 | + | |
970 | + int increment; | |
971 | + size_t column = 0; | |
972 | + | |
973 | + size_t last_blank_pos; | |
974 | + size_t last_blank_column; | |
975 | + int is_blank_seen; | |
976 | + int last_blank_increment = 0; | |
977 | + int is_bs_following_last_blank; | |
978 | + size_t bs_following_last_blank_num; | |
979 | + int is_cr_after_last_blank; | |
980 | + | |
981 | +#define CLEAR_FLAGS \ | |
982 | + do \ | |
983 | + { \ | |
984 | + last_blank_pos = 0; \ | |
985 | + last_blank_column = 0; \ | |
986 | + is_blank_seen = 0; \ | |
987 | + is_bs_following_last_blank = 0; \ | |
988 | + bs_following_last_blank_num = 0; \ | |
989 | + is_cr_after_last_blank = 0; \ | |
990 | + } \ | |
991 | + while (0) | |
992 | + | |
993 | +#define START_NEW_LINE \ | |
994 | + do \ | |
995 | + { \ | |
996 | + putchar ('\n'); \ | |
997 | + column = 0; \ | |
998 | + offset_out = 0; \ | |
999 | + CLEAR_FLAGS; \ | |
1000 | + } \ | |
1001 | + while (0) | |
1002 | + | |
1003 | + CLEAR_FLAGS; | |
1004 | + memset (&state, '\0', sizeof(mbstate_t)); | |
1005 | + | |
1006 | + for (;; bufpos += mblength, buflen -= mblength) | |
1007 | + { | |
1008 | + if (buflen < MB_LEN_MAX && !feof (istream) && !ferror (istream)) | |
1009 | + { | |
1010 | + memmove (buf, bufpos, buflen); | |
1011 | + buflen += fread (buf + buflen, sizeof(char), BUFSIZ, istream); | |
1012 | + bufpos = buf; | |
1013 | + } | |
1014 | + | |
1015 | + if (buflen < 1) | |
1016 | + break; | |
1017 | + | |
1018 | + /* Get a wide character. */ | |
1019 | + state_bak = state; | |
1020 | + mblength = mbrtowc ((wchar_t *)&wc, bufpos, buflen, &state); | |
1021 | + | |
1022 | + switch (mblength) | |
1023 | + { | |
1024 | + case (size_t)-1: | |
1025 | + case (size_t)-2: | |
1026 | + convfail++; | |
1027 | + state = state_bak; | |
1028 | + /* Fall through. */ | |
1029 | + | |
1030 | + case 0: | |
1031 | + mblength = 1; | |
1032 | + break; | |
1033 | + } | |
1034 | + | |
1035 | +rescan: | |
1036 | + if (operating_mode == byte_mode) /* byte mode */ | |
1037 | + increment = mblength; | |
1038 | + else if (operating_mode == character_mode) /* character mode */ | |
1039 | + increment = 1; | |
1040 | + else /* column mode */ | |
1041 | + { | |
1042 | + if (convfail) | |
1043 | + increment = 1; | |
1044 | + else | |
1045 | + { | |
1046 | + switch (wc) | |
1047 | + { | |
1048 | + case L'\n': | |
1049 | + fwrite (line_out, sizeof(char), offset_out, stdout); | |
1050 | + START_NEW_LINE; | |
1051 | + continue; | |
1052 | + | |
1053 | + case L'\b': | |
1054 | + increment = (column > 0) ? -1 : 0; | |
1055 | + break; | |
1056 | + | |
1057 | + case L'\r': | |
1058 | + increment = -1 * column; | |
1059 | + break; | |
1060 | + | |
1061 | + case L'\t': | |
1062 | + increment = 8 - column % 8; | |
1063 | + break; | |
1064 | + | |
1065 | + default: | |
1066 | + increment = wcwidth (wc); | |
1067 | + increment = (increment < 0) ? 0 : increment; | |
1068 | + } | |
1069 | + } | |
1070 | + } | |
1071 | + | |
1072 | + if (column + increment > width && break_spaces && last_blank_pos) | |
1073 | + { | |
1074 | + fwrite (line_out, sizeof(char), last_blank_pos, stdout); | |
1075 | + putchar ('\n'); | |
1076 | + | |
1077 | + offset_out = offset_out - last_blank_pos; | |
1078 | + column = column - last_blank_column + ((is_cr_after_last_blank) | |
1079 | + ? last_blank_increment : bs_following_last_blank_num); | |
1080 | + memmove (line_out, line_out + last_blank_pos, offset_out); | |
1081 | + CLEAR_FLAGS; | |
1082 | + goto rescan; | |
1083 | + } | |
1084 | + | |
1085 | + if (column + increment > width && column != 0) | |
1086 | + { | |
1087 | + fwrite (line_out, sizeof(char), offset_out, stdout); | |
1088 | + START_NEW_LINE; | |
1089 | + goto rescan; | |
1090 | + } | |
1091 | + | |
1092 | + if (allocated_out < offset_out + mblength) | |
1093 | + { | |
1094 | + line_out = X2REALLOC (line_out, &allocated_out); | |
1095 | + } | |
1096 | + | |
1097 | + memcpy (line_out + offset_out, bufpos, mblength); | |
1098 | + offset_out += mblength; | |
1099 | + column += increment; | |
1100 | + | |
1101 | + if (is_blank_seen && !convfail && wc == L'\r') | |
1102 | + is_cr_after_last_blank = 1; | |
1103 | + | |
1104 | + if (is_bs_following_last_blank && !convfail && wc == L'\b') | |
1105 | + ++bs_following_last_blank_num; | |
1106 | + else | |
1107 | + is_bs_following_last_blank = 0; | |
1108 | + | |
1109 | + if (break_spaces && !convfail && iswblank (wc)) | |
1110 | + { | |
1111 | + last_blank_pos = offset_out; | |
1112 | + last_blank_column = column; | |
1113 | + is_blank_seen = 1; | |
1114 | + last_blank_increment = increment; | |
1115 | + is_bs_following_last_blank = 1; | |
1116 | + bs_following_last_blank_num = 0; | |
1117 | + is_cr_after_last_blank = 0; | |
1118 | + } | |
1119 | + } | |
1120 | + | |
1121 | + *saved_errno = errno; | |
1122 | ||
1123 | if (offset_out) | |
1124 | fwrite (line_out, sizeof (char), (size_t) offset_out, stdout); | |
1125 | ||
1126 | +} | |
1127 | +#endif | |
1128 | + | |
1129 | +/* Fold file FILENAME, or standard input if FILENAME is "-", | |
1130 | + to stdout, with maximum line length WIDTH. | |
1131 | + Return 0 if successful, 1 if an error occurs. */ | |
1132 | + | |
1133 | +static bool | |
1134 | +fold_file (char const *filename, size_t width) | |
1135 | +{ | |
1136 | + FILE *istream; | |
1137 | + int saved_errno; | |
1138 | + | |
1139 | + if (STREQ (filename, "-")) | |
1140 | + { | |
1141 | + istream = stdin; | |
1142 | + have_read_stdin = 1; | |
1143 | + } | |
1144 | + else | |
1145 | + istream = fopen (filename, "r"); | |
1146 | + | |
1147 | + if (istream == NULL) | |
1148 | + { | |
1149 | + error (0, errno, "%s", quotef (filename)); | |
1150 | + return 1; | |
1151 | + } | |
1152 | + | |
1153 | + /* Define how ISTREAM is being folded. */ | |
1154 | +#if HAVE_MBRTOWC | |
1155 | + if (MB_CUR_MAX > 1) | |
1156 | + fold_multibyte_text (istream, width, &saved_errno); | |
1157 | + else | |
1158 | +#endif | |
1159 | + fold_text (istream, width, &saved_errno); | |
1160 | + | |
1161 | if (ferror (istream)) | |
1162 | { | |
1163 | error (0, saved_errno, "%s", quotef (filename)); | |
1164 | @@ -251,7 +498,8 @@ main (int argc, char **argv) | |
1165 | ||
1166 | atexit (close_stdout); | |
1167 | ||
1168 | - break_spaces = count_bytes = have_read_stdin = false; | |
1169 | + operating_mode = column_mode; | |
1170 | + break_spaces = have_read_stdin = false; | |
1171 | ||
1172 | while ((optc = getopt_long (argc, argv, shortopts, longopts, NULL)) != -1) | |
1173 | { | |
1174 | @@ -260,7 +508,15 @@ main (int argc, char **argv) | |
1175 | switch (optc) | |
1176 | { | |
1177 | case 'b': /* Count bytes rather than columns. */ | |
1178 | - count_bytes = true; | |
1179 | + if (operating_mode != column_mode) | |
1180 | + FATAL_ERROR (_("only one way of folding may be specified")); | |
1181 | + operating_mode = byte_mode; | |
1182 | + break; | |
1183 | + | |
1184 | + case 'c': | |
1185 | + if (operating_mode != column_mode) | |
1186 | + FATAL_ERROR (_("only one way of folding may be specified")); | |
1187 | + operating_mode = character_mode; | |
1188 | break; | |
1189 | ||
1190 | case 's': /* Break at word boundaries. */ | |
1191 | diff -Naurp coreutils-8.25-orig/src/join.c coreutils-8.25/src/join.c | |
1192 | --- coreutils-8.25-orig/src/join.c 2016-01-13 05:08:59.000000000 -0600 | |
1193 | +++ coreutils-8.25/src/join.c 2016-02-08 19:07:10.303944625 -0600 | |
1194 | @@ -22,18 +22,32 @@ | |
1195 | #include <sys/types.h> | |
1196 | #include <getopt.h> | |
1197 | ||
1198 | +/* Get mbstate_t, mbrtowc(), mbrtowc(), wcwidth(). */ | |
1199 | +#if HAVE_WCHAR_H | |
1200 | +# include <wchar.h> | |
1201 | +#endif | |
1202 | + | |
1203 | +/* Get iswblank(), towupper. */ | |
1204 | +#if HAVE_WCTYPE_H | |
1205 | +# include <wctype.h> | |
1206 | +#endif | |
1207 | + | |
1208 | #include "system.h" | |
1209 | #include "error.h" | |
1210 | #include "fadvise.h" | |
1211 | #include "hard-locale.h" | |
1212 | #include "linebuffer.h" | |
1213 | -#include "memcasecmp.h" | |
1214 | #include "quote.h" | |
1215 | #include "stdio--.h" | |
1216 | #include "xmemcoll.h" | |
1217 | #include "xstrtol.h" | |
1218 | #include "argmatch.h" | |
1219 | ||
1220 | +/* Some systems, like BeOS, have multibyte encodings but lack mbstate_t. */ | |
1221 | +#if HAVE_MBRTOWC && defined mbstate_t | |
1222 | +# define mbrtowc(pwc, s, n, ps) (mbrtowc) (pwc, s, n, 0) | |
1223 | +#endif | |
1224 | + | |
1225 | /* The official name of this program (e.g., no 'g' prefix). */ | |
1226 | #define PROGRAM_NAME "join" | |
1227 | ||
1228 | @@ -135,10 +149,12 @@ static struct outlist outlist_head; | |
1229 | /* Last element in 'outlist', where a new element can be added. */ | |
1230 | static struct outlist *outlist_end = &outlist_head; | |
1231 | ||
1232 | -/* Tab character separating fields. If negative, fields are separated | |
1233 | - by any nonempty string of blanks, otherwise by exactly one | |
1234 | - tab character whose value (when cast to unsigned char) equals TAB. */ | |
1235 | -static int tab = -1; | |
1236 | +/* Tab character separating fields. If NULL, fields are separated | |
1237 | + by any nonempty string of blanks. */ | |
1238 | +static char *tab = NULL; | |
1239 | + | |
1240 | +/* The number of bytes used for tab. */ | |
1241 | +static size_t tablen = 0; | |
1242 | ||
1243 | /* If nonzero, check that the input is correctly ordered. */ | |
1244 | static enum | |
1245 | @@ -275,13 +291,14 @@ xfields (struct line *line) | |
1246 | if (ptr == lim) | |
1247 | return; | |
1248 | ||
1249 | - if (0 <= tab && tab != '\n') | |
1250 | + if (tab != NULL) | |
1251 | { | |
1252 | + unsigned char t = tab[0]; | |
1253 | char *sep; | |
1254 | - for (; (sep = memchr (ptr, tab, lim - ptr)) != NULL; ptr = sep + 1) | |
1255 | + for (; (sep = memchr (ptr, t, lim - ptr)) != NULL; ptr = sep + 1) | |
1256 | extract_field (line, ptr, sep - ptr); | |
1257 | } | |
1258 | - else if (tab < 0) | |
1259 | + else | |
1260 | { | |
1261 | /* Skip leading blanks before the first field. */ | |
1262 | while (field_sep (*ptr)) | |
1263 | @@ -305,6 +322,147 @@ xfields (struct line *line) | |
1264 | extract_field (line, ptr, lim - ptr); | |
1265 | } | |
1266 | ||
1267 | +#if HAVE_MBRTOWC | |
1268 | +static void | |
1269 | +xfields_multibyte (struct line *line) | |
1270 | +{ | |
1271 | + char *ptr = line->buf.buffer; | |
1272 | + char const *lim = ptr + line->buf.length - 1; | |
1273 | + wchar_t wc = 0; | |
1274 | + size_t mblength = 1; | |
1275 | + mbstate_t state, state_bak; | |
1276 | + | |
1277 | + memset (&state, 0, sizeof (mbstate_t)); | |
1278 | + | |
1279 | + if (ptr >= lim) | |
1280 | + return; | |
1281 | + | |
1282 | + if (tab != NULL) | |
1283 | + { | |
1284 | + char *sep = ptr; | |
1285 | + for (; ptr < lim; ptr = sep + mblength) | |
1286 | + { | |
1287 | + sep = ptr; | |
1288 | + while (sep < lim) | |
1289 | + { | |
1290 | + state_bak = state; | |
1291 | + mblength = mbrtowc (&wc, sep, lim - sep + 1, &state); | |
1292 | + | |
1293 | + if (mblength == (size_t)-1 || mblength == (size_t)-2) | |
1294 | + { | |
1295 | + mblength = 1; | |
1296 | + state = state_bak; | |
1297 | + } | |
1298 | + mblength = (mblength < 1) ? 1 : mblength; | |
1299 | + | |
1300 | + if (mblength == tablen && !memcmp (sep, tab, mblength)) | |
1301 | + break; | |
1302 | + else | |
1303 | + { | |
1304 | + sep += mblength; | |
1305 | + continue; | |
1306 | + } | |
1307 | + } | |
1308 | + | |
1309 | + if (sep >= lim) | |
1310 | + break; | |
1311 | + | |
1312 | + extract_field (line, ptr, sep - ptr); | |
1313 | + } | |
1314 | + } | |
1315 | + else | |
1316 | + { | |
1317 | + /* Skip leading blanks before the first field. */ | |
1318 | + while(ptr < lim) | |
1319 | + { | |
1320 | + state_bak = state; | |
1321 | + mblength = mbrtowc (&wc, ptr, lim - ptr + 1, &state); | |
1322 | + | |
1323 | + if (mblength == (size_t)-1 || mblength == (size_t)-2) | |
1324 | + { | |
1325 | + mblength = 1; | |
1326 | + state = state_bak; | |
1327 | + break; | |
1328 | + } | |
1329 | + mblength = (mblength < 1) ? 1 : mblength; | |
1330 | + | |
1331 | + if (!iswblank(wc) && wc != '\n') | |
1332 | + break; | |
1333 | + ptr += mblength; | |
1334 | + } | |
1335 | + | |
1336 | + do | |
1337 | + { | |
1338 | + char *sep; | |
1339 | + state_bak = state; | |
1340 | + mblength = mbrtowc (&wc, ptr, lim - ptr + 1, &state); | |
1341 | + if (mblength == (size_t)-1 || mblength == (size_t)-2) | |
1342 | + { | |
1343 | + mblength = 1; | |
1344 | + state = state_bak; | |
1345 | + break; | |
1346 | + } | |
1347 | + mblength = (mblength < 1) ? 1 : mblength; | |
1348 | + | |
1349 | + sep = ptr + mblength; | |
1350 | + while (sep < lim) | |
1351 | + { | |
1352 | + state_bak = state; | |
1353 | + mblength = mbrtowc (&wc, sep, lim - sep + 1, &state); | |
1354 | + if (mblength == (size_t)-1 || mblength == (size_t)-2) | |
1355 | + { | |
1356 | + mblength = 1; | |
1357 | + state = state_bak; | |
1358 | + break; | |
1359 | + } | |
1360 | + mblength = (mblength < 1) ? 1 : mblength; | |
1361 | + | |
1362 | + if (iswblank (wc) || wc == '\n') | |
1363 | + break; | |
1364 | + | |
1365 | + sep += mblength; | |
1366 | + } | |
1367 | + | |
1368 | + extract_field (line, ptr, sep - ptr); | |
1369 | + if (sep >= lim) | |
1370 | + return; | |
1371 | + | |
1372 | + state_bak = state; | |
1373 | + mblength = mbrtowc (&wc, sep, lim - sep + 1, &state); | |
1374 | + if (mblength == (size_t)-1 || mblength == (size_t)-2) | |
1375 | + { | |
1376 | + mblength = 1; | |
1377 | + state = state_bak; | |
1378 | + break; | |
1379 | + } | |
1380 | + mblength = (mblength < 1) ? 1 : mblength; | |
1381 | + | |
1382 | + ptr = sep + mblength; | |
1383 | + while (ptr < lim) | |
1384 | + { | |
1385 | + state_bak = state; | |
1386 | + mblength = mbrtowc (&wc, ptr, lim - ptr + 1, &state); | |
1387 | + if (mblength == (size_t)-1 || mblength == (size_t)-2) | |
1388 | + { | |
1389 | + mblength = 1; | |
1390 | + state = state_bak; | |
1391 | + break; | |
1392 | + } | |
1393 | + mblength = (mblength < 1) ? 1 : mblength; | |
1394 | + | |
1395 | + if (!iswblank (wc) && wc != '\n') | |
1396 | + break; | |
1397 | + | |
1398 | + ptr += mblength; | |
1399 | + } | |
1400 | + } | |
1401 | + while (ptr < lim); | |
1402 | + } | |
1403 | + | |
1404 | + extract_field (line, ptr, lim - ptr); | |
1405 | +} | |
1406 | +#endif | |
1407 | + | |
1408 | static void | |
1409 | freeline (struct line *line) | |
1410 | { | |
1411 | @@ -326,56 +484,133 @@ keycmp (struct line const *line1, struct | |
1412 | size_t jf_1, size_t jf_2) | |
1413 | { | |
1414 | /* Start of field to compare in each file. */ | |
1415 | - char *beg1; | |
1416 | - char *beg2; | |
1417 | - | |
1418 | - size_t len1; | |
1419 | - size_t len2; /* Length of fields to compare. */ | |
1420 | + char *beg[2]; | |
1421 | + char *copy[2]; | |
1422 | + size_t len[2]; /* Length of fields to compare. */ | |
1423 | int diff; | |
1424 | + int i, j; | |
1425 | + int mallocd = 0; | |
1426 | ||
1427 | if (jf_1 < line1->nfields) | |
1428 | { | |
1429 | - beg1 = line1->fields[jf_1].beg; | |
1430 | - len1 = line1->fields[jf_1].len; | |
1431 | + beg[0] = line1->fields[jf_1].beg; | |
1432 | + len[0] = line1->fields[jf_1].len; | |
1433 | } | |
1434 | else | |
1435 | { | |
1436 | - beg1 = NULL; | |
1437 | - len1 = 0; | |
1438 | + beg[0] = NULL; | |
1439 | + len[0] = 0; | |
1440 | } | |
1441 | ||
1442 | if (jf_2 < line2->nfields) | |
1443 | { | |
1444 | - beg2 = line2->fields[jf_2].beg; | |
1445 | - len2 = line2->fields[jf_2].len; | |
1446 | + beg[1] = line2->fields[jf_2].beg; | |
1447 | + len[1] = line2->fields[jf_2].len; | |
1448 | } | |
1449 | else | |
1450 | { | |
1451 | - beg2 = NULL; | |
1452 | - len2 = 0; | |
1453 | + beg[1] = NULL; | |
1454 | + len[1] = 0; | |
1455 | } | |
1456 | ||
1457 | - if (len1 == 0) | |
1458 | - return len2 == 0 ? 0 : -1; | |
1459 | - if (len2 == 0) | |
1460 | + if (len[0] == 0) | |
1461 | + return len[1] == 0 ? 0 : -1; | |
1462 | + if (len[1] == 0) | |
1463 | return 1; | |
1464 | ||
1465 | if (ignore_case) | |
1466 | { | |
1467 | - /* FIXME: ignore_case does not work with NLS (in particular, | |
1468 | - with multibyte chars). */ | |
1469 | - diff = memcasecmp (beg1, beg2, MIN (len1, len2)); | |
1470 | +#ifdef HAVE_MBRTOWC | |
1471 | + if (MB_CUR_MAX > 1) | |
1472 | + { | |
1473 | + size_t mblength; | |
1474 | + wchar_t wc, uwc; | |
1475 | + mbstate_t state, state_bak; | |
1476 | + | |
1477 | + memset (&state, '\0', sizeof (mbstate_t)); | |
1478 | + | |
1479 | + for (i = 0; i < 2; i++) | |
1480 | + { | |
1481 | + mallocd = 1; | |
1482 | + copy[i] = xmalloc (len[i] + 1); | |
1483 | + memset (copy[i], '\0',len[i] + 1); | |
1484 | + | |
1485 | + for (j = 0; j < MIN (len[0], len[1]);) | |
1486 | + { | |
1487 | + state_bak = state; | |
1488 | + mblength = mbrtowc (&wc, beg[i] + j, len[i] - j, &state); | |
1489 | + | |
1490 | + switch (mblength) | |
1491 | + { | |
1492 | + case (size_t) -1: | |
1493 | + case (size_t) -2: | |
1494 | + state = state_bak; | |
1495 | + /* Fall through */ | |
1496 | + case 0: | |
1497 | + mblength = 1; | |
1498 | + break; | |
1499 | + | |
1500 | + default: | |
1501 | + uwc = towupper (wc); | |
1502 | + | |
1503 | + if (uwc != wc) | |
1504 | + { | |
1505 | + mbstate_t state_wc; | |
1506 | + size_t mblen; | |
1507 | + | |
1508 | + memset (&state_wc, '\0', sizeof (mbstate_t)); | |
1509 | + mblen = wcrtomb (copy[i] + j, uwc, &state_wc); | |
1510 | + assert (mblen != (size_t)-1); | |
1511 | + } | |
1512 | + else | |
1513 | + memcpy (copy[i] + j, beg[i] + j, mblength); | |
1514 | + } | |
1515 | + j += mblength; | |
1516 | + } | |
1517 | + copy[i][j] = '\0'; | |
1518 | + } | |
1519 | + } | |
1520 | + else | |
1521 | +#endif | |
1522 | + { | |
1523 | + for (i = 0; i < 2; i++) | |
1524 | + { | |
1525 | + mallocd = 1; | |
1526 | + copy[i] = xmalloc (len[i] + 1); | |
1527 | + | |
1528 | + for (j = 0; j < MIN (len[0], len[1]); j++) | |
1529 | + copy[i][j] = toupper (beg[i][j]); | |
1530 | + | |
1531 | + copy[i][j] = '\0'; | |
1532 | + } | |
1533 | + } | |
1534 | } | |
1535 | else | |
1536 | { | |
1537 | - if (hard_LC_COLLATE) | |
1538 | - return xmemcoll (beg1, len1, beg2, len2); | |
1539 | - diff = memcmp (beg1, beg2, MIN (len1, len2)); | |
1540 | + copy[0] = beg[0]; | |
1541 | + copy[1] = beg[1]; | |
1542 | + } | |
1543 | + | |
1544 | + if (hard_LC_COLLATE) | |
1545 | + { | |
1546 | + diff = xmemcoll ((char *) copy[0], len[0], (char *) copy[1], len[1]); | |
1547 | + | |
1548 | + if (mallocd) | |
1549 | + for (i = 0; i < 2; i++) | |
1550 | + free (copy[i]); | |
1551 | + | |
1552 | + return diff; | |
1553 | } | |
1554 | + diff = memcmp (copy[0], copy[1], MIN (len[0], len[1])); | |
1555 | + | |
1556 | + if (mallocd) | |
1557 | + for (i = 0; i < 2; i++) | |
1558 | + free (copy[i]); | |
1559 | + | |
1560 | ||
1561 | if (diff) | |
1562 | return diff; | |
1563 | - return len1 < len2 ? -1 : len1 != len2; | |
1564 | + return len[0] - len[1]; | |
1565 | } | |
1566 | ||
1567 | /* Check that successive input lines PREV and CURRENT from input file | |
1568 | @@ -467,6 +702,11 @@ get_line (FILE *fp, struct line **linep, | |
1569 | } | |
1570 | ++line_no[which - 1]; | |
1571 | ||
1572 | +#if HAVE_MBRTOWC | |
1573 | + if (MB_CUR_MAX > 1) | |
1574 | + xfields_multibyte (line); | |
1575 | + else | |
1576 | +#endif | |
1577 | xfields (line); | |
1578 | ||
1579 | if (prevline[which - 1]) | |
1580 | @@ -566,21 +806,28 @@ prfield (size_t n, struct line const *li | |
1581 | ||
1582 | /* Output all the fields in line, other than the join field. */ | |
1583 | ||
1584 | +#define PUT_TAB_CHAR \ | |
1585 | + do \ | |
1586 | + { \ | |
1587 | + (tab != NULL) ? \ | |
1588 | + fwrite(tab, sizeof(char), tablen, stdout) : putchar (' '); \ | |
1589 | + } \ | |
1590 | + while (0) | |
1591 | + | |
1592 | static void | |
1593 | prfields (struct line const *line, size_t join_field, size_t autocount) | |
1594 | { | |
1595 | size_t i; | |
1596 | size_t nfields = autoformat ? autocount : line->nfields; | |
1597 | - char output_separator = tab < 0 ? ' ' : tab; | |
1598 | ||
1599 | for (i = 0; i < join_field && i < nfields; ++i) | |
1600 | { | |
1601 | - putchar (output_separator); | |
1602 | + PUT_TAB_CHAR; | |
1603 | prfield (i, line); | |
1604 | } | |
1605 | for (i = join_field + 1; i < nfields; ++i) | |
1606 | { | |
1607 | - putchar (output_separator); | |
1608 | + PUT_TAB_CHAR; | |
1609 | prfield (i, line); | |
1610 | } | |
1611 | } | |
1612 | @@ -591,7 +838,6 @@ static void | |
1613 | prjoin (struct line const *line1, struct line const *line2) | |
1614 | { | |
1615 | const struct outlist *outlist; | |
1616 | - char output_separator = tab < 0 ? ' ' : tab; | |
1617 | size_t field; | |
1618 | struct line const *line; | |
1619 | ||
1620 | @@ -625,7 +871,7 @@ prjoin (struct line const *line1, struct | |
1621 | o = o->next; | |
1622 | if (o == NULL) | |
1623 | break; | |
1624 | - putchar (output_separator); | |
1625 | + PUT_TAB_CHAR; | |
1626 | } | |
1627 | putchar (eolchar); | |
1628 | } | |
1629 | @@ -1103,21 +1349,46 @@ main (int argc, char **argv) | |
1630 | ||
1631 | case 't': | |
1632 | { | |
1633 | - unsigned char newtab = optarg[0]; | |
1634 | + char *newtab = NULL; | |
1635 | + size_t newtablen; | |
1636 | + newtab = xstrdup (optarg); | |
1637 | +#if HAVE_MBRTOWC | |
1638 | + if (MB_CUR_MAX > 1) | |
1639 | + { | |
1640 | + mbstate_t state; | |
1641 | + | |
1642 | + memset (&state, 0, sizeof (mbstate_t)); | |
1643 | + newtablen = mbrtowc (NULL, newtab, | |
1644 | + strnlen (newtab, MB_LEN_MAX), | |
1645 | + &state); | |
1646 | + if (newtablen == (size_t) 0 | |
1647 | + || newtablen == (size_t) -1 | |
1648 | + || newtablen == (size_t) -2) | |
1649 | + newtablen = 1; | |
1650 | + } | |
1651 | + else | |
1652 | +#endif | |
1653 | + newtablen = 1; | |
1654 | if (! newtab) | |
1655 | - newtab = '\n'; /* '' => process the whole line. */ | |
1656 | + { | |
1657 | + newtab = (char*)"\n"; /* '' => process the whole line. */ | |
1658 | + } | |
1659 | else if (optarg[1]) | |
1660 | { | |
1661 | - if (STREQ (optarg, "\\0")) | |
1662 | - newtab = '\0'; | |
1663 | - else | |
1664 | - error (EXIT_FAILURE, 0, _("multi-character tab %s"), | |
1665 | - quote (optarg)); | |
1666 | + if (newtablen == 1 && newtab[1]) | |
1667 | + { | |
1668 | + if (STREQ (newtab, "\\0")) | |
1669 | + newtab[0] = '\0'; | |
1670 | + } | |
1671 | + } | |
1672 | + if (tab != NULL && strcmp (tab, newtab)) | |
1673 | + { | |
1674 | + free (newtab); | |
1675 | + error (EXIT_FAILURE, 0, _("incompatible tabs")); | |
1676 | } | |
1677 | - if (0 <= tab && tab != newtab) | |
1678 | - error (EXIT_FAILURE, 0, _("incompatible tabs")); | |
1679 | tab = newtab; | |
1680 | - } | |
1681 | + tablen = newtablen; | |
1682 | + } | |
1683 | break; | |
1684 | ||
1685 | case 'z': | |
1686 | diff -Naurp coreutils-8.25-orig/src/pr.c coreutils-8.25/src/pr.c | |
1687 | --- coreutils-8.25-orig/src/pr.c 2016-01-01 07:48:50.000000000 -0600 | |
1688 | +++ coreutils-8.25/src/pr.c 2016-02-08 19:07:10.306944635 -0600 | |
1689 | @@ -311,6 +311,24 @@ | |
1690 | ||
1691 | #include <getopt.h> | |
1692 | #include <sys/types.h> | |
1693 | + | |
1694 | +/* Get MB_LEN_MAX. */ | |
1695 | +#include <limits.h> | |
1696 | +/* MB_LEN_MAX is incorrectly defined to be 1 in at least one GCC | |
1697 | + installation; work around this configuration error. */ | |
1698 | +#if !defined MB_LEN_MAX || MB_LEN_MAX == 1 | |
1699 | +# define MB_LEN_MAX 16 | |
1700 | +#endif | |
1701 | + | |
1702 | +/* Get MB_CUR_MAX. */ | |
1703 | +#include <stdlib.h> | |
1704 | + | |
1705 | +/* Solaris 2.5 has a bug: <wchar.h> must be included before <wctype.h>. */ | |
1706 | +/* Get mbstate_t, mbrtowc(), wcwidth(). */ | |
1707 | +#if HAVE_WCHAR_H | |
1708 | +# include <wchar.h> | |
1709 | +#endif | |
1710 | + | |
1711 | #include "system.h" | |
1712 | #include "error.h" | |
1713 | #include "fadvise.h" | |
1714 | @@ -323,6 +341,18 @@ | |
1715 | #include "xstrtol.h" | |
1716 | #include "xdectoint.h" | |
1717 | ||
1718 | +/* Some systems, like BeOS, have multibyte encodings but lack mbstate_t. */ | |
1719 | +#if HAVE_MBRTOWC && defined mbstate_t | |
1720 | +# define mbrtowc(pwc, s, n, ps) (mbrtowc) (pwc, s, n, 0) | |
1721 | +#endif | |
1722 | + | |
1723 | +#ifndef HAVE_DECL_WCWIDTH | |
1724 | +"this configure-time declaration test was not run" | |
1725 | +#endif | |
1726 | +#if !HAVE_DECL_WCWIDTH | |
1727 | +extern int wcwidth (); | |
1728 | +#endif | |
1729 | + | |
1730 | /* The official name of this program (e.g., no 'g' prefix). */ | |
1731 | #define PROGRAM_NAME "pr" | |
1732 | ||
1733 | @@ -415,7 +445,20 @@ struct COLUMN | |
1734 | ||
1735 | typedef struct COLUMN COLUMN; | |
1736 | ||
1737 | -static int char_to_clump (char c); | |
1738 | +/* Funtion pointers to switch functions for single byte locale or for | |
1739 | + multibyte locale. If multibyte functions do not exist in your sysytem, | |
1740 | + these pointers always point the function for single byte locale. */ | |
1741 | +static void (*print_char) (char c); | |
1742 | +static int (*char_to_clump) (char c); | |
1743 | + | |
1744 | +/* Functions for single byte locale. */ | |
1745 | +static void print_char_single (char c); | |
1746 | +static int char_to_clump_single (char c); | |
1747 | + | |
1748 | +/* Functions for multibyte locale. */ | |
1749 | +static void print_char_multi (char c); | |
1750 | +static int char_to_clump_multi (char c); | |
1751 | + | |
1752 | static bool read_line (COLUMN *p); | |
1753 | static bool print_page (void); | |
1754 | static bool print_stored (COLUMN *p); | |
1755 | @@ -427,6 +470,7 @@ static void add_line_number (COLUMN *p); | |
1756 | static void getoptnum (const char *n_str, int min, int *num, | |
1757 | const char *errfmt); | |
1758 | static void getoptarg (char *arg, char switch_char, char *character, | |
1759 | + int *character_length, int *character_width, | |
1760 | int *number); | |
1761 | static void print_files (int number_of_files, char **av); | |
1762 | static void init_parameters (int number_of_files); | |
1763 | @@ -440,7 +484,6 @@ static void store_char (char c); | |
1764 | static void pad_down (unsigned int lines); | |
1765 | static void read_rest_of_line (COLUMN *p); | |
1766 | static void skip_read (COLUMN *p, int column_number); | |
1767 | -static void print_char (char c); | |
1768 | static void cleanup (void); | |
1769 | static void print_sep_string (void); | |
1770 | static void separator_string (const char *optarg_S); | |
1771 | @@ -452,7 +495,7 @@ static COLUMN *column_vector; | |
1772 | we store the leftmost columns contiguously in buff. | |
1773 | To print a line from buff, get the index of the first character | |
1774 | from line_vector[i], and print up to line_vector[i + 1]. */ | |
1775 | -static char *buff; | |
1776 | +static unsigned char *buff; | |
1777 | ||
1778 | /* Index of the position in buff where the next character | |
1779 | will be stored. */ | |
1780 | @@ -556,7 +599,7 @@ static int chars_per_column; | |
1781 | static bool untabify_input = false; | |
1782 | ||
1783 | /* (-e) The input tab character. */ | |
1784 | -static char input_tab_char = '\t'; | |
1785 | +static char input_tab_char[MB_LEN_MAX] = "\t"; | |
1786 | ||
1787 | /* (-e) Tabstops are at chars_per_tab, 2*chars_per_tab, 3*chars_per_tab, ... | |
1788 | where the leftmost column is 1. */ | |
1789 | @@ -566,7 +609,10 @@ static int chars_per_input_tab = 8; | |
1790 | static bool tabify_output = false; | |
1791 | ||
1792 | /* (-i) The output tab character. */ | |
1793 | -static char output_tab_char = '\t'; | |
1794 | +static char output_tab_char[MB_LEN_MAX] = "\t"; | |
1795 | + | |
1796 | +/* (-i) The byte length of output tab character. */ | |
1797 | +static int output_tab_char_length = 1; | |
1798 | ||
1799 | /* (-i) The width of the output tab. */ | |
1800 | static int chars_per_output_tab = 8; | |
1801 | @@ -636,7 +682,13 @@ static int line_number; | |
1802 | static bool numbered_lines = false; | |
1803 | ||
1804 | /* (-n) Character which follows each line number. */ | |
1805 | -static char number_separator = '\t'; | |
1806 | +static char number_separator[MB_LEN_MAX] = "\t"; | |
1807 | + | |
1808 | +/* (-n) The byte length of the character which follows each line number. */ | |
1809 | +static int number_separator_length = 1; | |
1810 | + | |
1811 | +/* (-n) The character width of the character which follows each line number. */ | |
1812 | +static int number_separator_width = 0; | |
1813 | ||
1814 | /* (-n) line counting starts with 1st line of input file (not with 1st | |
1815 | line of 1st page printed). */ | |
1816 | @@ -689,6 +741,7 @@ static bool use_col_separator = false; | |
1817 | -a|COLUMN|-m is a 'space' and with the -J option a 'tab'. */ | |
1818 | static char *col_sep_string = (char *) ""; | |
1819 | static int col_sep_length = 0; | |
1820 | +static int col_sep_width = 0; | |
1821 | static char *column_separator = (char *) " "; | |
1822 | static char *line_separator = (char *) "\t"; | |
1823 | ||
1824 | @@ -839,6 +892,13 @@ separator_string (const char *optarg_S) | |
1825 | col_sep_length = (int) strlen (optarg_S); | |
1826 | col_sep_string = xmalloc (col_sep_length + 1); | |
1827 | strcpy (col_sep_string, optarg_S); | |
1828 | + | |
1829 | +#if HAVE_MBRTOWC | |
1830 | + if (MB_CUR_MAX > 1) | |
1831 | + col_sep_width = mbswidth (col_sep_string, 0); | |
1832 | + else | |
1833 | +#endif | |
1834 | + col_sep_width = col_sep_length; | |
1835 | } | |
1836 | ||
1837 | int | |
1838 | @@ -863,6 +923,21 @@ main (int argc, char **argv) | |
1839 | ||
1840 | atexit (close_stdout); | |
1841 | ||
1842 | +/* Define which functions are used, the ones for single byte locale or the ones | |
1843 | + for multibyte locale. */ | |
1844 | +#if HAVE_MBRTOWC | |
1845 | + if (MB_CUR_MAX > 1) | |
1846 | + { | |
1847 | + print_char = print_char_multi; | |
1848 | + char_to_clump = char_to_clump_multi; | |
1849 | + } | |
1850 | + else | |
1851 | +#endif | |
1852 | + { | |
1853 | + print_char = print_char_single; | |
1854 | + char_to_clump = char_to_clump_single; | |
1855 | + } | |
1856 | + | |
1857 | n_files = 0; | |
1858 | file_names = (argc > 1 | |
1859 | ? xmalloc ((argc - 1) * sizeof (char *)) | |
1860 | @@ -939,8 +1014,12 @@ main (int argc, char **argv) | |
1861 | break; | |
1862 | case 'e': | |
1863 | if (optarg) | |
1864 | - getoptarg (optarg, 'e', &input_tab_char, | |
1865 | - &chars_per_input_tab); | |
1866 | + { | |
1867 | + int dummy_length, dummy_width; | |
1868 | + | |
1869 | + getoptarg (optarg, 'e', input_tab_char, &dummy_length, | |
1870 | + &dummy_width, &chars_per_input_tab); | |
1871 | + } | |
1872 | /* Could check tab width > 0. */ | |
1873 | untabify_input = true; | |
1874 | break; | |
1875 | @@ -953,8 +1032,12 @@ main (int argc, char **argv) | |
1876 | break; | |
1877 | case 'i': | |
1878 | if (optarg) | |
1879 | - getoptarg (optarg, 'i', &output_tab_char, | |
1880 | - &chars_per_output_tab); | |
1881 | + { | |
1882 | + int dummy_width; | |
1883 | + | |
1884 | + getoptarg (optarg, 'i', output_tab_char, &output_tab_char_length, | |
1885 | + &dummy_width, &chars_per_output_tab); | |
1886 | + } | |
1887 | /* Could check tab width > 0. */ | |
1888 | tabify_output = true; | |
1889 | break; | |
1890 | @@ -972,8 +1055,8 @@ main (int argc, char **argv) | |
1891 | case 'n': | |
1892 | numbered_lines = true; | |
1893 | if (optarg) | |
1894 | - getoptarg (optarg, 'n', &number_separator, | |
1895 | - &chars_per_number); | |
1896 | + getoptarg (optarg, 'n', number_separator, &number_separator_length, | |
1897 | + &number_separator_width, &chars_per_number); | |
1898 | break; | |
1899 | case 'N': | |
1900 | skip_count = false; | |
1901 | @@ -997,7 +1080,7 @@ main (int argc, char **argv) | |
1902 | old_s = false; | |
1903 | /* Reset an additional input of -s, -S dominates -s */ | |
1904 | col_sep_string = bad_cast (""); | |
1905 | - col_sep_length = 0; | |
1906 | + col_sep_length = col_sep_width = 0; | |
1907 | use_col_separator = true; | |
1908 | if (optarg) | |
1909 | separator_string (optarg); | |
1910 | @@ -1152,10 +1235,45 @@ getoptnum (const char *n_str, int min, i | |
1911 | a number. */ | |
1912 | ||
1913 | static void | |
1914 | -getoptarg (char *arg, char switch_char, char *character, int *number) | |
1915 | +getoptarg (char *arg, char switch_char, char *character, int *character_length, | |
1916 | + int *character_width, int *number) | |
1917 | { | |
1918 | if (!ISDIGIT (*arg)) | |
1919 | - *character = *arg++; | |
1920 | + { | |
1921 | +#ifdef HAVE_MBRTOWC | |
1922 | + if (MB_CUR_MAX > 1) /* for multibyte locale. */ | |
1923 | + { | |
1924 | + wchar_t wc; | |
1925 | + size_t mblength; | |
1926 | + int width; | |
1927 | + mbstate_t state = {'\0'}; | |
1928 | + | |
1929 | + mblength = mbrtowc (&wc, arg, strnlen(arg, MB_LEN_MAX), &state); | |
1930 | + | |
1931 | + if (mblength == (size_t)-1 || mblength == (size_t)-2) | |
1932 | + { | |
1933 | + *character_length = 1; | |
1934 | + *character_width = 1; | |
1935 | + } | |
1936 | + else | |
1937 | + { | |
1938 | + *character_length = (mblength < 1) ? 1 : mblength; | |
1939 | + width = wcwidth (wc); | |
1940 | + *character_width = (width < 0) ? 0 : width; | |
1941 | + } | |
1942 | + | |
1943 | + strncpy (character, arg, *character_length); | |
1944 | + arg += *character_length; | |
1945 | + } | |
1946 | + else /* for single byte locale. */ | |
1947 | +#endif | |
1948 | + { | |
1949 | + *character = *arg++; | |
1950 | + *character_length = 1; | |
1951 | + *character_width = 1; | |
1952 | + } | |
1953 | + } | |
1954 | + | |
1955 | if (*arg) | |
1956 | { | |
1957 | long int tmp_long; | |
1958 | @@ -1177,6 +1295,11 @@ static void | |
1959 | init_parameters (int number_of_files) | |
1960 | { | |
1961 | int chars_used_by_number = 0; | |
1962 | + int mb_len = 1; | |
1963 | +#if HAVE_MBRTOWC | |
1964 | + if (MB_CUR_MAX > 1) | |
1965 | + mb_len = MB_LEN_MAX; | |
1966 | +#endif | |
1967 | ||
1968 | lines_per_body = lines_per_page - lines_per_header - lines_per_footer; | |
1969 | if (lines_per_body <= 0) | |
1970 | @@ -1214,7 +1337,7 @@ init_parameters (int number_of_files) | |
1971 | else | |
1972 | col_sep_string = column_separator; | |
1973 | ||
1974 | - col_sep_length = 1; | |
1975 | + col_sep_length = col_sep_width = 1; | |
1976 | use_col_separator = true; | |
1977 | } | |
1978 | /* It's rather pointless to define a TAB separator with column | |
1979 | @@ -1244,11 +1367,11 @@ init_parameters (int number_of_files) | |
1980 | + TAB_WIDTH (chars_per_input_tab, chars_per_number); */ | |
1981 | ||
1982 | /* Estimate chars_per_text without any margin and keep it constant. */ | |
1983 | - if (number_separator == '\t') | |
1984 | + if (number_separator[0] == '\t') | |
1985 | number_width = (chars_per_number | |
1986 | + TAB_WIDTH (chars_per_default_tab, chars_per_number)); | |
1987 | else | |
1988 | - number_width = chars_per_number + 1; | |
1989 | + number_width = chars_per_number + number_separator_width; | |
1990 | ||
1991 | /* The number is part of the column width unless we are | |
1992 | printing files in parallel. */ | |
1993 | @@ -1257,7 +1380,7 @@ init_parameters (int number_of_files) | |
1994 | } | |
1995 | ||
1996 | chars_per_column = (chars_per_line - chars_used_by_number | |
1997 | - - (columns - 1) * col_sep_length) / columns; | |
1998 | + - (columns - 1) * col_sep_width) / columns; | |
1999 | ||
2000 | if (chars_per_column < 1) | |
2001 | error (EXIT_FAILURE, 0, _("page width too narrow")); | |
2002 | @@ -1275,7 +1398,7 @@ init_parameters (int number_of_files) | |
2003 | We've to use 8 as the lower limit, if we use chars_per_default_tab = 8 | |
2004 | to expand a tab which is not an input_tab-char. */ | |
2005 | free (clump_buff); | |
2006 | - clump_buff = xmalloc (MAX (8, chars_per_input_tab)); | |
2007 | + clump_buff = xmalloc (mb_len * MAX (8, chars_per_input_tab)); | |
2008 | } | |
2009 | ||
2010 | /* Open the necessary files, | |
2011 | @@ -1383,7 +1506,7 @@ init_funcs (void) | |
2012 | ||
2013 | /* Enlarge p->start_position of first column to use the same form of | |
2014 | padding_not_printed with all columns. */ | |
2015 | - h = h + col_sep_length; | |
2016 | + h = h + col_sep_width; | |
2017 | ||
2018 | /* This loop takes care of all but the rightmost column. */ | |
2019 | ||
2020 | @@ -1417,7 +1540,7 @@ init_funcs (void) | |
2021 | } | |
2022 | else | |
2023 | { | |
2024 | - h = h_next + col_sep_length; | |
2025 | + h = h_next + col_sep_width; | |
2026 | h_next = h + chars_per_column; | |
2027 | } | |
2028 | } | |
2029 | @@ -1708,9 +1831,9 @@ static void | |
2030 | align_column (COLUMN *p) | |
2031 | { | |
2032 | padding_not_printed = p->start_position; | |
2033 | - if (padding_not_printed - col_sep_length > 0) | |
2034 | + if (padding_not_printed - col_sep_width > 0) | |
2035 | { | |
2036 | - pad_across_to (padding_not_printed - col_sep_length); | |
2037 | + pad_across_to (padding_not_printed - col_sep_width); | |
2038 | padding_not_printed = ANYWHERE; | |
2039 | } | |
2040 | ||
2041 | @@ -1981,13 +2104,13 @@ store_char (char c) | |
2042 | /* May be too generous. */ | |
2043 | buff = X2REALLOC (buff, &buff_allocated); | |
2044 | } | |
2045 | - buff[buff_current++] = c; | |
2046 | + buff[buff_current++] = (unsigned char) c; | |
2047 | } | |
2048 | ||
2049 | static void | |
2050 | add_line_number (COLUMN *p) | |
2051 | { | |
2052 | - int i; | |
2053 | + int i, j; | |
2054 | char *s; | |
2055 | int num_width; | |
2056 | ||
2057 | @@ -2004,22 +2127,24 @@ add_line_number (COLUMN *p) | |
2058 | /* Tabification is assumed for multiple columns, also for n-separators, | |
2059 | but 'default n-separator = TAB' hasn't been given priority over | |
2060 | equal column_width also specified by POSIX. */ | |
2061 | - if (number_separator == '\t') | |
2062 | + if (number_separator[0] == '\t') | |
2063 | { | |
2064 | i = number_width - chars_per_number; | |
2065 | while (i-- > 0) | |
2066 | (p->char_func) (' '); | |
2067 | } | |
2068 | else | |
2069 | - (p->char_func) (number_separator); | |
2070 | + for (j = 0; j < number_separator_length; j++) | |
2071 | + (p->char_func) (number_separator[j]); | |
2072 | } | |
2073 | else | |
2074 | /* To comply with POSIX, we avoid any expansion of default TAB | |
2075 | separator with a single column output. No column_width requirement | |
2076 | has to be considered. */ | |
2077 | { | |
2078 | - (p->char_func) (number_separator); | |
2079 | - if (number_separator == '\t') | |
2080 | + for (j = 0; j < number_separator_length; j++) | |
2081 | + (p->char_func) (number_separator[j]); | |
2082 | + if (number_separator[0] == '\t') | |
2083 | output_position = POS_AFTER_TAB (chars_per_output_tab, | |
2084 | output_position); | |
2085 | } | |
2086 | @@ -2180,7 +2305,7 @@ print_white_space (void) | |
2087 | while (goal - h_old > 1 | |
2088 | && (h_new = POS_AFTER_TAB (chars_per_output_tab, h_old)) <= goal) | |
2089 | { | |
2090 | - putchar (output_tab_char); | |
2091 | + fwrite (output_tab_char, sizeof(char), output_tab_char_length, stdout); | |
2092 | h_old = h_new; | |
2093 | } | |
2094 | while (++h_old <= goal) | |
2095 | @@ -2200,6 +2325,7 @@ print_sep_string (void) | |
2096 | { | |
2097 | char *s; | |
2098 | int l = col_sep_length; | |
2099 | + int not_space_flag; | |
2100 | ||
2101 | s = col_sep_string; | |
2102 | ||
2103 | @@ -2213,6 +2339,7 @@ print_sep_string (void) | |
2104 | { | |
2105 | for (; separators_not_printed > 0; --separators_not_printed) | |
2106 | { | |
2107 | + not_space_flag = 0; | |
2108 | while (l-- > 0) | |
2109 | { | |
2110 | /* 3 types of sep_strings: spaces only, spaces and chars, | |
2111 | @@ -2226,12 +2353,15 @@ print_sep_string (void) | |
2112 | } | |
2113 | else | |
2114 | { | |
2115 | + not_space_flag = 1; | |
2116 | if (spaces_not_printed > 0) | |
2117 | print_white_space (); | |
2118 | putchar (*s++); | |
2119 | - ++output_position; | |
2120 | } | |
2121 | } | |
2122 | + if (not_space_flag) | |
2123 | + output_position += col_sep_width; | |
2124 | + | |
2125 | /* sep_string ends with some spaces */ | |
2126 | if (spaces_not_printed > 0) | |
2127 | print_white_space (); | |
2128 | @@ -2259,7 +2389,7 @@ print_clump (COLUMN *p, int n, char *clu | |
2129 | required number of tabs and spaces. */ | |
2130 | ||
2131 | static void | |
2132 | -print_char (char c) | |
2133 | +print_char_single (char c) | |
2134 | { | |
2135 | if (tabify_output) | |
2136 | { | |
2137 | @@ -2283,6 +2413,74 @@ print_char (char c) | |
2138 | putchar (c); | |
2139 | } | |
2140 | ||
2141 | +#ifdef HAVE_MBRTOWC | |
2142 | +static void | |
2143 | +print_char_multi (char c) | |
2144 | +{ | |
2145 | + static size_t mbc_pos = 0; | |
2146 | + static char mbc[MB_LEN_MAX] = {'\0'}; | |
2147 | + static mbstate_t state = {'\0'}; | |
2148 | + mbstate_t state_bak; | |
2149 | + wchar_t wc; | |
2150 | + size_t mblength; | |
2151 | + int width; | |
2152 | + | |
2153 | + if (tabify_output) | |
2154 | + { | |
2155 | + state_bak = state; | |
2156 | + mbc[mbc_pos++] = c; | |
2157 | + mblength = mbrtowc (&wc, mbc, mbc_pos, &state); | |
2158 | + | |
2159 | + while (mbc_pos > 0) | |
2160 | + { | |
2161 | + switch (mblength) | |
2162 | + { | |
2163 | + case (size_t)-2: | |
2164 | + state = state_bak; | |
2165 | + return; | |
2166 | + | |
2167 | + case (size_t)-1: | |
2168 | + state = state_bak; | |
2169 | + ++output_position; | |
2170 | + putchar (mbc[0]); | |
2171 | + memmove (mbc, mbc + 1, MB_CUR_MAX - 1); | |
2172 | + --mbc_pos; | |
2173 | + break; | |
2174 | + | |
2175 | + case 0: | |
2176 | + mblength = 1; | |
2177 | + | |
2178 | + default: | |
2179 | + if (wc == L' ') | |
2180 | + { | |
2181 | + memmove (mbc, mbc + mblength, MB_CUR_MAX - mblength); | |
2182 | + --mbc_pos; | |
2183 | + ++spaces_not_printed; | |
2184 | + return; | |
2185 | + } | |
2186 | + else if (spaces_not_printed > 0) | |
2187 | + print_white_space (); | |
2188 | + | |
2189 | + /* Nonprintables are assumed to have width 0, except L'\b'. */ | |
2190 | + if ((width = wcwidth (wc)) < 1) | |
2191 | + { | |
2192 | + if (wc == L'\b') | |
2193 | + --output_position; | |
2194 | + } | |
2195 | + else | |
2196 | + output_position += width; | |
2197 | + | |
2198 | + fwrite (mbc, sizeof(char), mblength, stdout); | |
2199 | + memmove (mbc, mbc + mblength, MB_CUR_MAX - mblength); | |
2200 | + mbc_pos -= mblength; | |
2201 | + } | |
2202 | + } | |
2203 | + return; | |
2204 | + } | |
2205 | + putchar (c); | |
2206 | +} | |
2207 | +#endif | |
2208 | + | |
2209 | /* Skip to page PAGE before printing. | |
2210 | PAGE may be larger than total number of pages. */ | |
2211 | ||
2212 | @@ -2462,9 +2660,9 @@ read_line (COLUMN *p) | |
2213 | align_empty_cols = false; | |
2214 | } | |
2215 | ||
2216 | - if (padding_not_printed - col_sep_length > 0) | |
2217 | + if (padding_not_printed - col_sep_width > 0) | |
2218 | { | |
2219 | - pad_across_to (padding_not_printed - col_sep_length); | |
2220 | + pad_across_to (padding_not_printed - col_sep_width); | |
2221 | padding_not_printed = ANYWHERE; | |
2222 | } | |
2223 | ||
2224 | @@ -2534,7 +2732,7 @@ print_stored (COLUMN *p) | |
2225 | int i; | |
2226 | ||
2227 | int line = p->current_line++; | |
2228 | - char *first = &buff[line_vector[line]]; | |
2229 | + unsigned char *first = &buff[line_vector[line]]; | |
2230 | /* FIXME | |
2231 | UMR: Uninitialized memory read: | |
2232 | * This is occurring while in: | |
2233 | @@ -2546,7 +2744,7 @@ print_stored (COLUMN *p) | |
2234 | xmalloc [xmalloc.c:94] | |
2235 | init_store_cols [pr.c:1648] | |
2236 | */ | |
2237 | - char *last = &buff[line_vector[line + 1]]; | |
2238 | + unsigned char *last = &buff[line_vector[line + 1]]; | |
2239 | ||
2240 | pad_vertically = true; | |
2241 | ||
2242 | @@ -2565,9 +2763,9 @@ print_stored (COLUMN *p) | |
2243 | } | |
2244 | } | |
2245 | ||
2246 | - if (padding_not_printed - col_sep_length > 0) | |
2247 | + if (padding_not_printed - col_sep_width > 0) | |
2248 | { | |
2249 | - pad_across_to (padding_not_printed - col_sep_length); | |
2250 | + pad_across_to (padding_not_printed - col_sep_width); | |
2251 | padding_not_printed = ANYWHERE; | |
2252 | } | |
2253 | ||
2254 | @@ -2580,8 +2778,8 @@ print_stored (COLUMN *p) | |
2255 | if (spaces_not_printed == 0) | |
2256 | { | |
2257 | output_position = p->start_position + end_vector[line]; | |
2258 | - if (p->start_position - col_sep_length == chars_per_margin) | |
2259 | - output_position -= col_sep_length; | |
2260 | + if (p->start_position - col_sep_width == chars_per_margin) | |
2261 | + output_position -= col_sep_width; | |
2262 | } | |
2263 | ||
2264 | return true; | |
2265 | @@ -2600,7 +2798,7 @@ print_stored (COLUMN *p) | |
2266 | number of characters is 1.) */ | |
2267 | ||
2268 | static int | |
2269 | -char_to_clump (char c) | |
2270 | +char_to_clump_single (char c) | |
2271 | { | |
2272 | unsigned char uc = c; | |
2273 | char *s = clump_buff; | |
2274 | @@ -2610,10 +2808,10 @@ char_to_clump (char c) | |
2275 | int chars; | |
2276 | int chars_per_c = 8; | |
2277 | ||
2278 | - if (c == input_tab_char) | |
2279 | + if (c == input_tab_char[0]) | |
2280 | chars_per_c = chars_per_input_tab; | |
2281 | ||
2282 | - if (c == input_tab_char || c == '\t') | |
2283 | + if (c == input_tab_char[0] || c == '\t') | |
2284 | { | |
2285 | width = TAB_WIDTH (chars_per_c, input_position); | |
2286 | ||
2287 | @@ -2694,6 +2892,164 @@ char_to_clump (char c) | |
2288 | return chars; | |
2289 | } | |
2290 | ||
2291 | +#ifdef HAVE_MBRTOWC | |
2292 | +static int | |
2293 | +char_to_clump_multi (char c) | |
2294 | +{ | |
2295 | + static size_t mbc_pos = 0; | |
2296 | + static char mbc[MB_LEN_MAX] = {'\0'}; | |
2297 | + static mbstate_t state = {'\0'}; | |
2298 | + mbstate_t state_bak; | |
2299 | + wchar_t wc; | |
2300 | + size_t mblength; | |
2301 | + int wc_width; | |
2302 | + register char *s = clump_buff; | |
2303 | + register int i, j; | |
2304 | + char esc_buff[4]; | |
2305 | + int width; | |
2306 | + int chars; | |
2307 | + int chars_per_c = 8; | |
2308 | + | |
2309 | + state_bak = state; | |
2310 | + mbc[mbc_pos++] = c; | |
2311 | + mblength = mbrtowc (&wc, mbc, mbc_pos, &state); | |
2312 | + | |
2313 | + width = 0; | |
2314 | + chars = 0; | |
2315 | + while (mbc_pos > 0) | |
2316 | + { | |
2317 | + switch (mblength) | |
2318 | + { | |
2319 | + case (size_t)-2: | |
2320 | + state = state_bak; | |
2321 | + return 0; | |
2322 | + | |
2323 | + case (size_t)-1: | |
2324 | + state = state_bak; | |
2325 | + mblength = 1; | |
2326 | + | |
2327 | + if (use_esc_sequence || use_cntrl_prefix) | |
2328 | + { | |
2329 | + width = +4; | |
2330 | + chars = +4; | |
2331 | + *s++ = '\\'; | |
2332 | + sprintf (esc_buff, "%03o", (unsigned char) mbc[0]); | |
2333 | + for (i = 0; i <= 2; ++i) | |
2334 | + *s++ = (int) esc_buff[i]; | |
2335 | + } | |
2336 | + else | |
2337 | + { | |
2338 | + width += 1; | |
2339 | + chars += 1; | |
2340 | + *s++ = mbc[0]; | |
2341 | + } | |
2342 | + break; | |
2343 | + | |
2344 | + case 0: | |
2345 | + mblength = 1; | |
2346 | + /* Fall through */ | |
2347 | + | |
2348 | + default: | |
2349 | + if (memcmp (mbc, input_tab_char, mblength) == 0) | |
2350 | + chars_per_c = chars_per_input_tab; | |
2351 | + | |
2352 | + if (memcmp (mbc, input_tab_char, mblength) == 0 || c == '\t') | |
2353 | + { | |
2354 | + int width_inc; | |
2355 | + | |
2356 | + width_inc = TAB_WIDTH (chars_per_c, input_position); | |
2357 | + width += width_inc; | |
2358 | + | |
2359 | + if (untabify_input) | |
2360 | + { | |
2361 | + for (i = width_inc; i; --i) | |
2362 | + *s++ = ' '; | |
2363 | + chars += width_inc; | |
2364 | + } | |
2365 | + else | |
2366 | + { | |
2367 | + for (i = 0; i < mblength; i++) | |
2368 | + *s++ = mbc[i]; | |
2369 | + chars += mblength; | |
2370 | + } | |
2371 | + } | |
2372 | + else if ((wc_width = wcwidth (wc)) < 1) | |
2373 | + { | |
2374 | + if (use_esc_sequence) | |
2375 | + { | |
2376 | + for (i = 0; i < mblength; i++) | |
2377 | + { | |
2378 | + width += 4; | |
2379 | + chars += 4; | |
2380 | + *s++ = '\\'; | |
2381 | + sprintf (esc_buff, "%03o", (unsigned char) mbc[i]); | |
2382 | + for (j = 0; j <= 2; ++j) | |
2383 | + *s++ = (int) esc_buff[j]; | |
2384 | + } | |
2385 | + } | |
2386 | + else if (use_cntrl_prefix) | |
2387 | + { | |
2388 | + if (wc < 0200) | |
2389 | + { | |
2390 | + width += 2; | |
2391 | + chars += 2; | |
2392 | + *s++ = '^'; | |
2393 | + *s++ = wc ^ 0100; | |
2394 | + } | |
2395 | + else | |
2396 | + { | |
2397 | + for (i = 0; i < mblength; i++) | |
2398 | + { | |
2399 | + width += 4; | |
2400 | + chars += 4; | |
2401 | + *s++ = '\\'; | |
2402 | + sprintf (esc_buff, "%03o", (unsigned char) mbc[i]); | |
2403 | + for (j = 0; j <= 2; ++j) | |
2404 | + *s++ = (int) esc_buff[j]; | |
2405 | + } | |
2406 | + } | |
2407 | + } | |
2408 | + else if (wc == L'\b') | |
2409 | + { | |
2410 | + width += -1; | |
2411 | + chars += 1; | |
2412 | + *s++ = c; | |
2413 | + } | |
2414 | + else | |
2415 | + { | |
2416 | + width += 0; | |
2417 | + chars += mblength; | |
2418 | + for (i = 0; i < mblength; i++) | |
2419 | + *s++ = mbc[i]; | |
2420 | + } | |
2421 | + } | |
2422 | + else | |
2423 | + { | |
2424 | + width += wc_width; | |
2425 | + chars += mblength; | |
2426 | + for (i = 0; i < mblength; i++) | |
2427 | + *s++ = mbc[i]; | |
2428 | + } | |
2429 | + } | |
2430 | + memmove (mbc, mbc + mblength, MB_CUR_MAX - mblength); | |
2431 | + mbc_pos -= mblength; | |
2432 | + } | |
2433 | + | |
2434 | + /* Too many backspaces must put us in position 0 -- never negative. */ | |
2435 | + if (width < 0 && input_position == 0) | |
2436 | + { | |
2437 | + chars = 0; | |
2438 | + input_position = 0; | |
2439 | + } | |
2440 | + else if (width < 0 && input_position <= -width) | |
2441 | + input_position = 0; | |
2442 | + else | |
2443 | + input_position += width; | |
2444 | + | |
2445 | + return chars; | |
2446 | +} | |
2447 | +#endif | |
2448 | + | |
2449 | /* We've just printed some files and need to clean up things before | |
2450 | looking for more options and printing the next batch of files. | |
2451 | ||
2452 | diff -Naurp coreutils-8.25-orig/src/sort.c coreutils-8.25/src/sort.c | |
2453 | --- coreutils-8.25-orig/src/sort.c 2016-01-16 13:09:33.000000000 -0600 | |
2454 | +++ coreutils-8.25/src/sort.c 2016-02-08 19:07:10.310944648 -0600 | |
2455 | @@ -29,6 +29,14 @@ | |
2456 | #include <sys/wait.h> | |
2457 | #include <signal.h> | |
2458 | #include <assert.h> | |
2459 | +#if HAVE_WCHAR_H | |
2460 | +# include <wchar.h> | |
2461 | +#endif | |
2462 | +/* Get isw* functions. */ | |
2463 | +#if HAVE_WCTYPE_H | |
2464 | +# include <wctype.h> | |
2465 | +#endif | |
2466 | + | |
2467 | #include "system.h" | |
2468 | #include "argmatch.h" | |
2469 | #include "error.h" | |
2470 | @@ -163,14 +171,39 @@ static int decimal_point; | |
2471 | /* Thousands separator; if -1, then there isn't one. */ | |
2472 | static int thousands_sep; | |
2473 | ||
2474 | +/* True if -f is specified. */ | |
2475 | +static bool folding; | |
2476 | + | |
2477 | /* Nonzero if the corresponding locales are hard. */ | |
2478 | static bool hard_LC_COLLATE; | |
2479 | -#if HAVE_NL_LANGINFO | |
2480 | +#if HAVE_LANGINFO_CODESET | |
2481 | static bool hard_LC_TIME; | |
2482 | #endif | |
2483 | ||
2484 | #define NONZERO(x) ((x) != 0) | |
2485 | ||
2486 | +/* get a multibyte character's byte length. */ | |
2487 | +#define GET_BYTELEN_OF_CHAR(LIM, PTR, MBLENGTH, STATE) \ | |
2488 | + do \ | |
2489 | + { \ | |
2490 | + wchar_t wc; \ | |
2491 | + mbstate_t state_bak; \ | |
2492 | + \ | |
2493 | + state_bak = STATE; \ | |
2494 | + mblength = mbrtowc (&wc, PTR, LIM - PTR, &STATE); \ | |
2495 | + \ | |
2496 | + switch (MBLENGTH) \ | |
2497 | + { \ | |
2498 | + case (size_t)-1: \ | |
2499 | + case (size_t)-2: \ | |
2500 | + STATE = state_bak; \ | |
2501 | + /* Fall through. */ \ | |
2502 | + case 0: \ | |
2503 | + MBLENGTH = 1; \ | |
2504 | + } \ | |
2505 | + } \ | |
2506 | + while (0) | |
2507 | + | |
2508 | /* The kind of blanks for '-b' to skip in various options. */ | |
2509 | enum blanktype { bl_start, bl_end, bl_both }; | |
2510 | ||
2511 | @@ -344,13 +377,11 @@ static bool reverse; | |
2512 | they were read if all keys compare equal. */ | |
2513 | static bool stable; | |
2514 | ||
2515 | -/* If TAB has this value, blanks separate fields. */ | |
2516 | -enum { TAB_DEFAULT = CHAR_MAX + 1 }; | |
2517 | - | |
2518 | -/* Tab character separating fields. If TAB_DEFAULT, then fields are | |
2519 | +/* Tab character separating fields. If tab_length is 0, then fields are | |
2520 | separated by the empty string between a non-blank character and a blank | |
2521 | character. */ | |
2522 | -static int tab = TAB_DEFAULT; | |
2523 | +static char tab[MB_LEN_MAX + 1]; | |
2524 | +static size_t tab_length = 0; | |
2525 | ||
2526 | /* Flag to remove consecutive duplicate lines from the output. | |
2527 | Only the last of a sequence of equal lines will be output. */ | |
2528 | @@ -810,6 +841,46 @@ reap_all (void) | |
2529 | reap (-1); | |
2530 | } | |
2531 | ||
2532 | +/* Function pointers. */ | |
2533 | +static void | |
2534 | +(*inittables) (void); | |
2535 | +static char * | |
2536 | +(*begfield) (const struct line*, const struct keyfield *); | |
2537 | +static char * | |
2538 | +(*limfield) (const struct line*, const struct keyfield *); | |
2539 | +static void | |
2540 | +(*skipblanks) (char **ptr, char *lim); | |
2541 | +static int | |
2542 | +(*getmonth) (char const *, size_t, char **); | |
2543 | +static int | |
2544 | +(*keycompare) (const struct line *, const struct line *); | |
2545 | +static int | |
2546 | +(*numcompare) (const char *, const char *); | |
2547 | + | |
2548 | +/* Test for white space multibyte character. | |
2549 | + Set LENGTH the byte length of investigated multibyte character. */ | |
2550 | +#if HAVE_MBRTOWC | |
2551 | +static int | |
2552 | +ismbblank (const char *str, size_t len, size_t *length) | |
2553 | +{ | |
2554 | + size_t mblength; | |
2555 | + wchar_t wc; | |
2556 | + mbstate_t state; | |
2557 | + | |
2558 | + memset (&state, '\0', sizeof(mbstate_t)); | |
2559 | + mblength = mbrtowc (&wc, str, len, &state); | |
2560 | + | |
2561 | + if (mblength == (size_t)-1 || mblength == (size_t)-2) | |
2562 | + { | |
2563 | + *length = 1; | |
2564 | + return 0; | |
2565 | + } | |
2566 | + | |
2567 | + *length = (mblength < 1) ? 1 : mblength; | |
2568 | + return iswblank (wc) || wc == '\n'; | |
2569 | +} | |
2570 | +#endif | |
2571 | + | |
2572 | /* Clean up any remaining temporary files. */ | |
2573 | ||
2574 | static void | |
2575 | @@ -1254,7 +1325,7 @@ zaptemp (char const *name) | |
2576 | free (node); | |
2577 | } | |
2578 | ||
2579 | -#if HAVE_NL_LANGINFO | |
2580 | +#if HAVE_LANGINFO_CODESET | |
2581 | ||
2582 | static int | |
2583 | struct_month_cmp (void const *m1, void const *m2) | |
2584 | @@ -1269,7 +1340,7 @@ struct_month_cmp (void const *m1, void c | |
2585 | /* Initialize the character class tables. */ | |
2586 | ||
2587 | static void | |
2588 | -inittables (void) | |
2589 | +inittables_uni (void) | |
2590 | { | |
2591 | size_t i; | |
2592 | ||
2593 | @@ -1281,7 +1352,7 @@ inittables (void) | |
2594 | fold_toupper[i] = toupper (i); | |
2595 | } | |
2596 | ||
2597 | -#if HAVE_NL_LANGINFO | |
2598 | +#if HAVE_LANGINFO_CODESET | |
2599 | /* If we're not in the "C" locale, read different names for months. */ | |
2600 | if (hard_LC_TIME) | |
2601 | { | |
2602 | @@ -1363,6 +1434,84 @@ specify_nmerge (int oi, char c, char con | |
2603 | xstrtol_fatal (e, oi, c, long_options, s); | |
2604 | } | |
2605 | ||
2606 | +#if HAVE_MBRTOWC | |
2607 | +static void | |
2608 | +inittables_mb (void) | |
2609 | +{ | |
2610 | + int i, j, k, l; | |
2611 | + char *name, *s, *lc_time, *lc_ctype; | |
2612 | + size_t s_len, mblength; | |
2613 | + char mbc[MB_LEN_MAX]; | |
2614 | + wchar_t wc, pwc; | |
2615 | + mbstate_t state_mb, state_wc; | |
2616 | + | |
2617 | + lc_time = setlocale (LC_TIME, ""); | |
2618 | + if (lc_time) | |
2619 | + lc_time = xstrdup (lc_time); | |
2620 | + | |
2621 | + lc_ctype = setlocale (LC_CTYPE, ""); | |
2622 | + if (lc_ctype) | |
2623 | + lc_ctype = xstrdup (lc_ctype); | |
2624 | + | |
2625 | + if (lc_time && lc_ctype) | |
2626 | + /* temporarily set LC_CTYPE to match LC_TIME, so that we can convert | |
2627 | + * the names of months to upper case */ | |
2628 | + setlocale (LC_CTYPE, lc_time); | |
2629 | + | |
2630 | + for (i = 0; i < MONTHS_PER_YEAR; i++) | |
2631 | + { | |
2632 | + s = (char *) nl_langinfo (ABMON_1 + i); | |
2633 | + s_len = strlen (s); | |
2634 | + monthtab[i].name = name = (char *) xmalloc (s_len + 1); | |
2635 | + monthtab[i].val = i + 1; | |
2636 | + | |
2637 | + memset (&state_mb, '\0', sizeof (mbstate_t)); | |
2638 | + memset (&state_wc, '\0', sizeof (mbstate_t)); | |
2639 | + | |
2640 | + for (j = 0; j < s_len;) | |
2641 | + { | |
2642 | + if (!ismbblank (s + j, s_len - j, &mblength)) | |
2643 | + break; | |
2644 | + j += mblength; | |
2645 | + } | |
2646 | + | |
2647 | + for (k = 0; j < s_len;) | |
2648 | + { | |
2649 | + mblength = mbrtowc (&wc, (s + j), (s_len - j), &state_mb); | |
2650 | + assert (mblength != (size_t)-1 && mblength != (size_t)-2); | |
2651 | + if (mblength == 0) | |
2652 | + break; | |
2653 | + | |
2654 | + pwc = towupper (wc); | |
2655 | + if (pwc == wc) | |
2656 | + { | |
2657 | + memcpy (mbc, s + j, mblength); | |
2658 | + j += mblength; | |
2659 | + } | |
2660 | + else | |
2661 | + { | |
2662 | + j += mblength; | |
2663 | + mblength = wcrtomb (mbc, pwc, &state_wc); | |
2664 | + assert (mblength != (size_t)0 && mblength != (size_t)-1); | |
2665 | + } | |
2666 | + | |
2667 | + for (l = 0; l < mblength; l++) | |
2668 | + name[k++] = mbc[l]; | |
2669 | + } | |
2670 | + name[k] = '\0'; | |
2671 | + } | |
2672 | + qsort ((void *) monthtab, MONTHS_PER_YEAR, | |
2673 | + sizeof (struct month), struct_month_cmp); | |
2674 | + | |
2675 | + if (lc_time && lc_ctype) | |
2676 | + /* restore the original locales */ | |
2677 | + setlocale (LC_CTYPE, lc_ctype); | |
2678 | + | |
2679 | + free (lc_ctype); | |
2680 | + free (lc_time); | |
2681 | +} | |
2682 | +#endif | |
2683 | + | |
2684 | /* Specify the amount of main memory to use when sorting. */ | |
2685 | static void | |
2686 | specify_sort_size (int oi, char c, char const *s) | |
2687 | @@ -1596,7 +1745,7 @@ buffer_linelim (struct buffer const *buf | |
2688 | by KEY in LINE. */ | |
2689 | ||
2690 | static char * | |
2691 | -begfield (struct line const *line, struct keyfield const *key) | |
2692 | +begfield_uni (const struct line *line, const struct keyfield *key) | |
2693 | { | |
2694 | char *ptr = line->text, *lim = ptr + line->length - 1; | |
2695 | size_t sword = key->sword; | |
2696 | @@ -1605,10 +1754,10 @@ begfield (struct line const *line, struc | |
2697 | /* The leading field separator itself is included in a field when -t | |
2698 | is absent. */ | |
2699 | ||
2700 | - if (tab != TAB_DEFAULT) | |
2701 | + if (tab_length) | |
2702 | while (ptr < lim && sword--) | |
2703 | { | |
2704 | - while (ptr < lim && *ptr != tab) | |
2705 | + while (ptr < lim && *ptr != tab[0]) | |
2706 | ++ptr; | |
2707 | if (ptr < lim) | |
2708 | ++ptr; | |
2709 | @@ -1634,11 +1783,70 @@ begfield (struct line const *line, struc | |
2710 | return ptr; | |
2711 | } | |
2712 | ||
2713 | +#if HAVE_MBRTOWC | |
2714 | +static char * | |
2715 | +begfield_mb (const struct line *line, const struct keyfield *key) | |
2716 | +{ | |
2717 | + int i; | |
2718 | + char *ptr = line->text, *lim = ptr + line->length - 1; | |
2719 | + size_t sword = key->sword; | |
2720 | + size_t schar = key->schar; | |
2721 | + size_t mblength; | |
2722 | + mbstate_t state; | |
2723 | + | |
2724 | + memset (&state, '\0', sizeof(mbstate_t)); | |
2725 | + | |
2726 | + if (tab_length) | |
2727 | + while (ptr < lim && sword--) | |
2728 | + { | |
2729 | + while (ptr < lim && memcmp (ptr, tab, tab_length) != 0) | |
2730 | + { | |
2731 | + GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); | |
2732 | + ptr += mblength; | |
2733 | + } | |
2734 | + if (ptr < lim) | |
2735 | + { | |
2736 | + GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); | |
2737 | + ptr += mblength; | |
2738 | + } | |
2739 | + } | |
2740 | + else | |
2741 | + while (ptr < lim && sword--) | |
2742 | + { | |
2743 | + while (ptr < lim && ismbblank (ptr, lim - ptr, &mblength)) | |
2744 | + ptr += mblength; | |
2745 | + if (ptr < lim) | |
2746 | + { | |
2747 | + GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); | |
2748 | + ptr += mblength; | |
2749 | + } | |
2750 | + while (ptr < lim && !ismbblank (ptr, lim - ptr, &mblength)) | |
2751 | + ptr += mblength; | |
2752 | + } | |
2753 | + | |
2754 | + if (key->skipsblanks) | |
2755 | + while (ptr < lim && ismbblank (ptr, lim - ptr, &mblength)) | |
2756 | + ptr += mblength; | |
2757 | + | |
2758 | + for (i = 0; i < schar; i++) | |
2759 | + { | |
2760 | + GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); | |
2761 | + | |
2762 | + if (ptr + mblength > lim) | |
2763 | + break; | |
2764 | + else | |
2765 | + ptr += mblength; | |
2766 | + } | |
2767 | + | |
2768 | + return ptr; | |
2769 | +} | |
2770 | +#endif | |
2771 | + | |
2772 | /* Return the limit of (a pointer to the first character after) the field | |
2773 | in LINE specified by KEY. */ | |
2774 | ||
2775 | static char * | |
2776 | -limfield (struct line const *line, struct keyfield const *key) | |
2777 | +limfield_uni (const struct line *line, const struct keyfield *key) | |
2778 | { | |
2779 | char *ptr = line->text, *lim = ptr + line->length - 1; | |
2780 | size_t eword = key->eword, echar = key->echar; | |
2781 | @@ -1653,10 +1861,10 @@ limfield (struct line const *line, struc | |
2782 | 'beginning' is the first character following the delimiting TAB. | |
2783 | Otherwise, leave PTR pointing at the first 'blank' character after | |
2784 | the preceding field. */ | |
2785 | - if (tab != TAB_DEFAULT) | |
2786 | + if (tab_length) | |
2787 | while (ptr < lim && eword--) | |
2788 | { | |
2789 | - while (ptr < lim && *ptr != tab) | |
2790 | + while (ptr < lim && *ptr != tab[0]) | |
2791 | ++ptr; | |
2792 | if (ptr < lim && (eword || echar)) | |
2793 | ++ptr; | |
2794 | @@ -1702,10 +1910,10 @@ limfield (struct line const *line, struc | |
2795 | */ | |
2796 | ||
2797 | /* Make LIM point to the end of (one byte past) the current field. */ | |
2798 | - if (tab != TAB_DEFAULT) | |
2799 | + if (tab_length) | |
2800 | { | |
2801 | char *newlim; | |
2802 | - newlim = memchr (ptr, tab, lim - ptr); | |
2803 | + newlim = memchr (ptr, tab[0], lim - ptr); | |
2804 | if (newlim) | |
2805 | lim = newlim; | |
2806 | } | |
2807 | @@ -1736,6 +1944,130 @@ limfield (struct line const *line, struc | |
2808 | return ptr; | |
2809 | } | |
2810 | ||
2811 | +#if HAVE_MBRTOWC | |
2812 | +static char * | |
2813 | +limfield_mb (const struct line *line, const struct keyfield *key) | |
2814 | +{ | |
2815 | + char *ptr = line->text, *lim = ptr + line->length - 1; | |
2816 | + size_t eword = key->eword, echar = key->echar; | |
2817 | + int i; | |
2818 | + size_t mblength; | |
2819 | + mbstate_t state; | |
2820 | + | |
2821 | + if (echar == 0) | |
2822 | + eword++; /* skip all of end field. */ | |
2823 | + | |
2824 | + memset (&state, '\0', sizeof(mbstate_t)); | |
2825 | + | |
2826 | + if (tab_length) | |
2827 | + while (ptr < lim && eword--) | |
2828 | + { | |
2829 | + while (ptr < lim && memcmp (ptr, tab, tab_length) != 0) | |
2830 | + { | |
2831 | + GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); | |
2832 | + ptr += mblength; | |
2833 | + } | |
2834 | + if (ptr < lim && (eword | echar)) | |
2835 | + { | |
2836 | + GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); | |
2837 | + ptr += mblength; | |
2838 | + } | |
2839 | + } | |
2840 | + else | |
2841 | + while (ptr < lim && eword--) | |
2842 | + { | |
2843 | + while (ptr < lim && ismbblank (ptr, lim - ptr, &mblength)) | |
2844 | + ptr += mblength; | |
2845 | + if (ptr < lim) | |
2846 | + { | |
2847 | + GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); | |
2848 | + ptr += mblength; | |
2849 | + } | |
2850 | + while (ptr < lim && !ismbblank (ptr, lim - ptr, &mblength)) | |
2851 | + ptr += mblength; | |
2852 | + } | |
2853 | + | |
2854 | + | |
2855 | +# ifdef POSIX_UNSPECIFIED | |
2856 | + /* Make LIM point to the end of (one byte past) the current field. */ | |
2857 | + if (tab_length) | |
2858 | + { | |
2859 | + char *newlim, *p; | |
2860 | + | |
2861 | + newlim = NULL; | |
2862 | + for (p = ptr; p < lim;) | |
2863 | + { | |
2864 | + if (memcmp (p, tab, tab_length) == 0) | |
2865 | + { | |
2866 | + newlim = p; | |
2867 | + break; | |
2868 | + } | |
2869 | + | |
2870 | + GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); | |
2871 | + p += mblength; | |
2872 | + } | |
2873 | + } | |
2874 | + else | |
2875 | + { | |
2876 | + char *newlim; | |
2877 | + newlim = ptr; | |
2878 | + | |
2879 | + while (newlim < lim && ismbblank (newlim, lim - newlim, &mblength)) | |
2880 | + newlim += mblength; | |
2881 | + if (ptr < lim) | |
2882 | + { | |
2883 | + GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); | |
2884 | + ptr += mblength; | |
2885 | + } | |
2886 | + while (newlim < lim && !ismbblank (newlim, lim - newlim, &mblength)) | |
2887 | + newlim += mblength; | |
2888 | + lim = newlim; | |
2889 | + } | |
2890 | +# endif | |
2891 | + | |
2892 | + if (echar != 0) | |
2893 | + { | |
2894 | + /* If we're skipping leading blanks, don't start counting characters | |
2895 | + * until after skipping past any leading blanks. */ | |
2896 | + if (key->skipeblanks) | |
2897 | + while (ptr < lim && ismbblank (ptr, lim - ptr, &mblength)) | |
2898 | + ptr += mblength; | |
2899 | + | |
2900 | + memset (&state, '\0', sizeof(mbstate_t)); | |
2901 | + | |
2902 | + /* Advance PTR by ECHAR (if possible), but no further than LIM. */ | |
2903 | + for (i = 0; i < echar; i++) | |
2904 | + { | |
2905 | + GET_BYTELEN_OF_CHAR (lim, ptr, mblength, state); | |
2906 | + | |
2907 | + if (ptr + mblength > lim) | |
2908 | + break; | |
2909 | + else | |
2910 | + ptr += mblength; | |
2911 | + } | |
2912 | + } | |
2913 | + | |
2914 | + return ptr; | |
2915 | +} | |
2916 | +#endif | |
2917 | + | |
2918 | +static void | |
2919 | +skipblanks_uni (char **ptr, char *lim) | |
2920 | +{ | |
2921 | + while (*ptr < lim && blanks[to_uchar (**ptr)]) | |
2922 | + ++(*ptr); | |
2923 | +} | |
2924 | + | |
2925 | +#if HAVE_MBRTOWC | |
2926 | +static void | |
2927 | +skipblanks_mb (char **ptr, char *lim) | |
2928 | +{ | |
2929 | + size_t mblength; | |
2930 | + while (*ptr < lim && ismbblank (*ptr, lim - *ptr, &mblength)) | |
2931 | + (*ptr) += mblength; | |
2932 | +} | |
2933 | +#endif | |
2934 | + | |
2935 | /* Fill BUF reading from FP, moving buf->left bytes from the end | |
2936 | of buf->buf to the beginning first. If EOF is reached and the | |
2937 | file wasn't terminated by a newline, supply one. Set up BUF's line | |
2938 | @@ -1822,8 +2154,22 @@ fillbuf (struct buffer *buf, FILE *fp, c | |
2939 | else | |
2940 | { | |
2941 | if (key->skipsblanks) | |
2942 | - while (blanks[to_uchar (*line_start)]) | |
2943 | - line_start++; | |
2944 | + { | |
2945 | +#if HAVE_MBRTOWC | |
2946 | + if (MB_CUR_MAX > 1) | |
2947 | + { | |
2948 | + size_t mblength; | |
2949 | + while (line_start < line->keylim && | |
2950 | + ismbblank (line_start, | |
2951 | + line->keylim - line_start, | |
2952 | + &mblength)) | |
2953 | + line_start += mblength; | |
2954 | + } | |
2955 | + else | |
2956 | +#endif | |
2957 | + while (blanks[to_uchar (*line_start)]) | |
2958 | + line_start++; | |
2959 | + } | |
2960 | line->keybeg = line_start; | |
2961 | } | |
2962 | } | |
2963 | @@ -1944,7 +2290,7 @@ human_numcompare (char const *a, char co | |
2964 | hideously fast. */ | |
2965 | ||
2966 | static int | |
2967 | -numcompare (char const *a, char const *b) | |
2968 | +numcompare_uni (const char *a, const char *b) | |
2969 | { | |
2970 | while (blanks[to_uchar (*a)]) | |
2971 | a++; | |
2972 | @@ -1954,6 +2300,25 @@ numcompare (char const *a, char const *b | |
2973 | return strnumcmp (a, b, decimal_point, thousands_sep); | |
2974 | } | |
2975 | ||
2976 | +#if HAVE_MBRTOWC | |
2977 | +static int | |
2978 | +numcompare_mb (const char *a, const char *b) | |
2979 | +{ | |
2980 | + size_t mblength, len; | |
2981 | + len = strlen (a); /* okay for UTF-8 */ | |
2982 | + while (*a && ismbblank (a, len > MB_CUR_MAX ? MB_CUR_MAX : len, &mblength)) | |
2983 | + { | |
2984 | + a += mblength; | |
2985 | + len -= mblength; | |
2986 | + } | |
2987 | + len = strlen (b); /* okay for UTF-8 */ | |
2988 | + while (*b && ismbblank (b, len > MB_CUR_MAX ? MB_CUR_MAX : len, &mblength)) | |
2989 | + b += mblength; | |
2990 | + | |
2991 | + return strnumcmp (a, b, decimal_point, thousands_sep); | |
2992 | +} | |
2993 | +#endif /* HAV_EMBRTOWC */ | |
2994 | + | |
2995 | /* Work around a problem whereby the long double value returned by glibc's | |
2996 | strtold ("NaN", ...) contains uninitialized bits: clear all bytes of | |
2997 | A and B before calling strtold. FIXME: remove this function once | |
2998 | @@ -2004,7 +2369,7 @@ general_numcompare (char const *sa, char | |
2999 | Return 0 if the name in S is not recognized. */ | |
3000 | ||
3001 | static int | |
3002 | -getmonth (char const *month, char **ea) | |
3003 | +getmonth_uni (char const *month, size_t len, char **ea) | |
3004 | { | |
3005 | size_t lo = 0; | |
3006 | size_t hi = MONTHS_PER_YEAR; | |
3007 | @@ -2280,15 +2645,14 @@ debug_key (struct line const *line, stru | |
3008 | char saved = *lim; | |
3009 | *lim = '\0'; | |
3010 | ||
3011 | - while (blanks[to_uchar (*beg)]) | |
3012 | - beg++; | |
3013 | + skipblanks (&beg, lim); | |
3014 | ||
3015 | char *tighter_lim = beg; | |
3016 | ||
3017 | if (lim < beg) | |
3018 | tighter_lim = lim; | |
3019 | else if (key->month) | |
3020 | - getmonth (beg, &tighter_lim); | |
3021 | + getmonth (beg, lim-beg, &tighter_lim); | |
3022 | else if (key->general_numeric) | |
3023 | ignore_value (strtold (beg, &tighter_lim)); | |
3024 | else if (key->numeric || key->human_numeric) | |
3025 | @@ -2432,7 +2796,7 @@ key_warnings (struct keyfield const *gke | |
3026 | bool maybe_space_aligned = !hard_LC_COLLATE && default_key_compare (key) | |
3027 | && !(key->schar || key->echar); | |
3028 | bool line_offset = key->eword == 0 && key->echar != 0; /* -k1.x,1.y */ | |
3029 | - if (!gkey_only && tab == TAB_DEFAULT && !line_offset | |
3030 | + if (!gkey_only && !tab_length && !line_offset | |
3031 | && ((!key->skipsblanks && !(implicit_skip || maybe_space_aligned)) | |
3032 | || (!key->skipsblanks && key->schar) | |
3033 | || (!key->skipeblanks && key->echar))) | |
3034 | @@ -2490,11 +2854,87 @@ key_warnings (struct keyfield const *gke | |
3035 | error (0, 0, _("option '-r' only applies to last-resort comparison")); | |
3036 | } | |
3037 | ||
3038 | +#if HAVE_MBRTOWC | |
3039 | +static int | |
3040 | +getmonth_mb (const char *s, size_t len, char **ea) | |
3041 | +{ | |
3042 | + char *month; | |
3043 | + register size_t i; | |
3044 | + register int lo = 0, hi = MONTHS_PER_YEAR, result; | |
3045 | + char *tmp; | |
3046 | + size_t wclength, mblength; | |
3047 | + const char *pp; | |
3048 | + const wchar_t *wpp; | |
3049 | + wchar_t *month_wcs; | |
3050 | + mbstate_t state; | |
3051 | + | |
3052 | + while (len > 0 && ismbblank (s, len, &mblength)) | |
3053 | + { | |
3054 | + s += mblength; | |
3055 | + len -= mblength; | |
3056 | + } | |
3057 | + | |
3058 | + if (len == 0) | |
3059 | + return 0; | |
3060 | + | |
3061 | + if (SIZE_MAX - len < 1) | |
3062 | + xalloc_die (); | |
3063 | + | |
3064 | + month = (char *) xnmalloc (len + 1, MB_CUR_MAX); | |
3065 | + | |
3066 | + pp = tmp = (char *) xnmalloc (len + 1, MB_CUR_MAX); | |
3067 | + memcpy (tmp, s, len); | |
3068 | + tmp[len] = '\0'; | |
3069 | + wpp = month_wcs = (wchar_t *) xnmalloc (len + 1, sizeof (wchar_t)); | |
3070 | + memset (&state, '\0', sizeof (mbstate_t)); | |
3071 | + | |
3072 | + wclength = mbsrtowcs (month_wcs, &pp, len + 1, &state); | |
3073 | + if (wclength == (size_t)-1 || pp != NULL) | |
3074 | + error (SORT_FAILURE, 0, _("Invalid multibyte input %s."), quote(s)); | |
3075 | + | |
3076 | + for (i = 0; i < wclength; i++) | |
3077 | + { | |
3078 | + month_wcs[i] = towupper(month_wcs[i]); | |
3079 | + if (iswblank (month_wcs[i])) | |
3080 | + { | |
3081 | + month_wcs[i] = L'\0'; | |
3082 | + break; | |
3083 | + } | |
3084 | + } | |
3085 | + | |
3086 | + mblength = wcsrtombs (month, &wpp, (len + 1) * MB_CUR_MAX, &state); | |
3087 | + assert (mblength != (-1) && wpp == NULL); | |
3088 | + | |
3089 | + do | |
3090 | + { | |
3091 | + int ix = (lo + hi) / 2; | |
3092 | + | |
3093 | + if (strncmp (month, monthtab[ix].name, strlen (monthtab[ix].name)) < 0) | |
3094 | + hi = ix; | |
3095 | + else | |
3096 | + lo = ix; | |
3097 | + } | |
3098 | + while (hi - lo > 1); | |
3099 | + | |
3100 | + result = (!strncmp (month, monthtab[lo].name, strlen (monthtab[lo].name)) | |
3101 | + ? monthtab[lo].val : 0); | |
3102 | + | |
3103 | + if (ea && result) | |
3104 | + *ea = (char*) s + strlen (monthtab[lo].name); | |
3105 | + | |
3106 | + free (month); | |
3107 | + free (tmp); | |
3108 | + free (month_wcs); | |
3109 | + | |
3110 | + return result; | |
3111 | +} | |
3112 | +#endif | |
3113 | + | |
3114 | /* Compare two lines A and B trying every key in sequence until there | |
3115 | are no more keys or a difference is found. */ | |
3116 | ||
3117 | static int | |
3118 | -keycompare (struct line const *a, struct line const *b) | |
3119 | +keycompare_uni (const struct line *a, const struct line *b) | |
3120 | { | |
3121 | struct keyfield *key = keylist; | |
3122 | ||
3123 | @@ -2579,7 +3019,7 @@ keycompare (struct line const *a, struct | |
3124 | else if (key->human_numeric) | |
3125 | diff = human_numcompare (ta, tb); | |
3126 | else if (key->month) | |
3127 | - diff = getmonth (ta, NULL) - getmonth (tb, NULL); | |
3128 | + diff = getmonth (ta, tlena, NULL) - getmonth (tb, tlenb, NULL); | |
3129 | else if (key->random) | |
3130 | diff = compare_random (ta, tlena, tb, tlenb); | |
3131 | else if (key->version) | |
3132 | @@ -2695,6 +3135,211 @@ keycompare (struct line const *a, struct | |
3133 | return key->reverse ? -diff : diff; | |
3134 | } | |
3135 | ||
3136 | +#if HAVE_MBRTOWC | |
3137 | +static int | |
3138 | +keycompare_mb (const struct line *a, const struct line *b) | |
3139 | +{ | |
3140 | + struct keyfield *key = keylist; | |
3141 | + | |
3142 | + /* For the first iteration only, the key positions have been | |
3143 | + precomputed for us. */ | |
3144 | + char *texta = a->keybeg; | |
3145 | + char *textb = b->keybeg; | |
3146 | + char *lima = a->keylim; | |
3147 | + char *limb = b->keylim; | |
3148 | + | |
3149 | + size_t mblength_a, mblength_b; | |
3150 | + wchar_t wc_a, wc_b; | |
3151 | + mbstate_t state_a, state_b; | |
3152 | + | |
3153 | + int diff = 0; | |
3154 | + | |
3155 | + memset (&state_a, '\0', sizeof(mbstate_t)); | |
3156 | + memset (&state_b, '\0', sizeof(mbstate_t)); | |
3157 | + /* Ignore keys with start after end. */ | |
3158 | + if (a->keybeg - a->keylim > 0) | |
3159 | + return 0; | |
3160 | + | |
3161 | + | |
3162 | + /* Ignore and/or translate chars before comparing. */ | |
3163 | +# define IGNORE_CHARS(NEW_LEN, LEN, TEXT, COPY, WC, MBLENGTH, STATE) \ | |
3164 | + do \ | |
3165 | + { \ | |
3166 | + wchar_t uwc; \ | |
3167 | + char mbc[MB_LEN_MAX]; \ | |
3168 | + mbstate_t state_wc; \ | |
3169 | + \ | |
3170 | + for (NEW_LEN = i = 0; i < LEN;) \ | |
3171 | + { \ | |
3172 | + mbstate_t state_bak; \ | |
3173 | + \ | |
3174 | + state_bak = STATE; \ | |
3175 | + MBLENGTH = mbrtowc (&WC, TEXT + i, LEN - i, &STATE); \ | |
3176 | + \ | |
3177 | + if (MBLENGTH == (size_t)-2 || MBLENGTH == (size_t)-1 \ | |
3178 | + || MBLENGTH == 0) \ | |
3179 | + { \ | |
3180 | + if (MBLENGTH == (size_t)-2 || MBLENGTH == (size_t)-1) \ | |
3181 | + STATE = state_bak; \ | |
3182 | + if (!ignore) \ | |
3183 | + COPY[NEW_LEN++] = TEXT[i]; \ | |
3184 | + i++; \ | |
3185 | + continue; \ | |
3186 | + } \ | |
3187 | + \ | |
3188 | + if (ignore) \ | |
3189 | + { \ | |
3190 | + if ((ignore == nonprinting && !iswprint (WC)) \ | |
3191 | + || (ignore == nondictionary \ | |
3192 | + && !iswalnum (WC) && !iswblank (WC))) \ | |
3193 | + { \ | |
3194 | + i += MBLENGTH; \ | |
3195 | + continue; \ | |
3196 | + } \ | |
3197 | + } \ | |
3198 | + \ | |
3199 | + if (translate) \ | |
3200 | + { \ | |
3201 | + \ | |
3202 | + uwc = towupper(WC); \ | |
3203 | + if (WC == uwc) \ | |
3204 | + { \ | |
3205 | + memcpy (mbc, TEXT + i, MBLENGTH); \ | |
3206 | + i += MBLENGTH; \ | |
3207 | + } \ | |
3208 | + else \ | |
3209 | + { \ | |
3210 | + i += MBLENGTH; \ | |
3211 | + WC = uwc; \ | |
3212 | + memset (&state_wc, '\0', sizeof (mbstate_t)); \ | |
3213 | + \ | |
3214 | + MBLENGTH = wcrtomb (mbc, WC, &state_wc); \ | |
3215 | + assert (MBLENGTH != (size_t)-1 && MBLENGTH != 0); \ | |
3216 | + } \ | |
3217 | + \ | |
3218 | + for (j = 0; j < MBLENGTH; j++) \ | |
3219 | + COPY[NEW_LEN++] = mbc[j]; \ | |
3220 | + } \ | |
3221 | + else \ | |
3222 | + for (j = 0; j < MBLENGTH; j++) \ | |
3223 | + COPY[NEW_LEN++] = TEXT[i++]; \ | |
3224 | + } \ | |
3225 | + COPY[NEW_LEN] = '\0'; \ | |
3226 | + } \ | |
3227 | + while (0) | |
3228 | + | |
3229 | + /* Actually compare the fields. */ | |
3230 | + | |
3231 | + for (;;) | |
3232 | + { | |
3233 | + /* Find the lengths. */ | |
3234 | + size_t lena = lima <= texta ? 0 : lima - texta; | |
3235 | + size_t lenb = limb <= textb ? 0 : limb - textb; | |
3236 | + | |
3237 | + char enda IF_LINT (= 0); | |
3238 | + char endb IF_LINT (= 0); | |
3239 | + | |
3240 | + char const *translate = key->translate; | |
3241 | + bool const *ignore = key->ignore; | |
3242 | + | |
3243 | + if (ignore || translate) | |
3244 | + { | |
3245 | + if (SIZE_MAX - lenb - 2 < lena) | |
3246 | + xalloc_die (); | |
3247 | + char *copy_a = (char *) xnmalloc (lena + lenb + 2, MB_CUR_MAX); | |
3248 | + char *copy_b = copy_a + lena * MB_CUR_MAX + 1; | |
3249 | + size_t new_len_a, new_len_b; | |
3250 | + size_t i, j; | |
3251 | + | |
3252 | + IGNORE_CHARS (new_len_a, lena, texta, copy_a, | |
3253 | + wc_a, mblength_a, state_a); | |
3254 | + IGNORE_CHARS (new_len_b, lenb, textb, copy_b, | |
3255 | + wc_b, mblength_b, state_b); | |
3256 | + texta = copy_a; textb = copy_b; | |
3257 | + lena = new_len_a; lenb = new_len_b; | |
3258 | + } | |
3259 | + else | |
3260 | + { | |
3261 | + /* Use the keys in-place, temporarily null-terminated. */ | |
3262 | + enda = texta[lena]; texta[lena] = '\0'; | |
3263 | + endb = textb[lenb]; textb[lenb] = '\0'; | |
3264 | + } | |
3265 | + | |
3266 | + if (key->random) | |
3267 | + diff = compare_random (texta, lena, textb, lenb); | |
3268 | + else if (key->numeric | key->general_numeric | key->human_numeric) | |
3269 | + { | |
3270 | + char savea = *lima, saveb = *limb; | |
3271 | + | |
3272 | + *lima = *limb = '\0'; | |
3273 | + diff = (key->numeric ? numcompare (texta, textb) | |
3274 | + : key->general_numeric ? general_numcompare (texta, textb) | |
3275 | + : human_numcompare (texta, textb)); | |
3276 | + *lima = savea, *limb = saveb; | |
3277 | + } | |
3278 | + else if (key->version) | |
3279 | + diff = filevercmp (texta, textb); | |
3280 | + else if (key->month) | |
3281 | + diff = getmonth (texta, lena, NULL) - getmonth (textb, lenb, NULL); | |
3282 | + else if (lena == 0) | |
3283 | + diff = - NONZERO (lenb); | |
3284 | + else if (lenb == 0) | |
3285 | + diff = 1; | |
3286 | + else if (hard_LC_COLLATE && !folding) | |
3287 | + { | |
3288 | + diff = xmemcoll0 (texta, lena + 1, textb, lenb + 1); | |
3289 | + } | |
3290 | + else | |
3291 | + { | |
3292 | + diff = memcmp (texta, textb, MIN (lena, lenb)); | |
3293 | + if (diff == 0) | |
3294 | + diff = lena < lenb ? -1 : lena != lenb; | |
3295 | + } | |
3296 | + | |
3297 | + if (ignore || translate) | |
3298 | + free (texta); | |
3299 | + else | |
3300 | + { | |
3301 | + texta[lena] = enda; | |
3302 | + textb[lenb] = endb; | |
3303 | + } | |
3304 | + | |
3305 | + if (diff) | |
3306 | + goto not_equal; | |
3307 | + | |
3308 | + key = key->next; | |
3309 | + if (! key) | |
3310 | + break; | |
3311 | + | |
3312 | + /* Find the beginning and limit of the next field. */ | |
3313 | + if (key->eword != -1) | |
3314 | + lima = limfield (a, key), limb = limfield (b, key); | |
3315 | + else | |
3316 | + lima = a->text + a->length - 1, limb = b->text + b->length - 1; | |
3317 | + | |
3318 | + if (key->sword != -1) | |
3319 | + texta = begfield (a, key), textb = begfield (b, key); | |
3320 | + else | |
3321 | + { | |
3322 | + texta = a->text, textb = b->text; | |
3323 | + if (key->skipsblanks) | |
3324 | + { | |
3325 | + while (texta < lima && ismbblank (texta, lima - texta, &mblength_a)) | |
3326 | + texta += mblength_a; | |
3327 | + while (textb < limb && ismbblank (textb, limb - textb, &mblength_b)) | |
3328 | + textb += mblength_b; | |
3329 | + } | |
3330 | + } | |
3331 | + } | |
3332 | + | |
3333 | +not_equal: | |
3334 | + if (key && key->reverse) | |
3335 | + return -diff; | |
3336 | + else | |
3337 | + return diff; | |
3338 | +} | |
3339 | +#endif | |
3340 | + | |
3341 | /* Compare two lines A and B, returning negative, zero, or positive | |
3342 | depending on whether A compares less than, equal to, or greater than B. */ | |
3343 | ||
3344 | @@ -2722,7 +3367,7 @@ compare (struct line const *a, struct li | |
3345 | diff = - NONZERO (blen); | |
3346 | else if (blen == 0) | |
3347 | diff = 1; | |
3348 | - else if (hard_LC_COLLATE) | |
3349 | + else if (hard_LC_COLLATE && !folding) | |
3350 | { | |
3351 | /* Note xmemcoll0 is a performance enhancement as | |
3352 | it will not unconditionally write '\0' after the | |
3353 | @@ -4121,6 +4766,7 @@ set_ordering (char const *s, struct keyf | |
3354 | break; | |
3355 | case 'f': | |
3356 | key->translate = fold_toupper; | |
3357 | + folding = true; | |
3358 | break; | |
3359 | case 'g': | |
3360 | key->general_numeric = true; | |
3361 | @@ -4199,7 +4845,7 @@ main (int argc, char **argv) | |
3362 | initialize_exit_failure (SORT_FAILURE); | |
3363 | ||
3364 | hard_LC_COLLATE = hard_locale (LC_COLLATE); | |
3365 | -#if HAVE_NL_LANGINFO | |
3366 | +#if HAVE_LANGINFO_CODESET | |
3367 | hard_LC_TIME = hard_locale (LC_TIME); | |
3368 | #endif | |
3369 | ||
3370 | @@ -4220,6 +4866,29 @@ main (int argc, char **argv) | |
3371 | thousands_sep = -1; | |
3372 | } | |
3373 | ||
3374 | +#if HAVE_MBRTOWC | |
3375 | + if (MB_CUR_MAX > 1) | |
3376 | + { | |
3377 | + inittables = inittables_mb; | |
3378 | + begfield = begfield_mb; | |
3379 | + limfield = limfield_mb; | |
3380 | + skipblanks = skipblanks_mb; | |
3381 | + getmonth = getmonth_mb; | |
3382 | + keycompare = keycompare_mb; | |
3383 | + numcompare = numcompare_mb; | |
3384 | + } | |
3385 | + else | |
3386 | +#endif | |
3387 | + { | |
3388 | + inittables = inittables_uni; | |
3389 | + begfield = begfield_uni; | |
3390 | + limfield = limfield_uni; | |
3391 | + skipblanks = skipblanks_uni; | |
3392 | + getmonth = getmonth_uni; | |
3393 | + keycompare = keycompare_uni; | |
3394 | + numcompare = numcompare_uni; | |
3395 | + } | |
3396 | + | |
3397 | have_read_stdin = false; | |
3398 | inittables (); | |
3399 | ||
3400 | @@ -4494,13 +5163,34 @@ main (int argc, char **argv) | |
3401 | ||
3402 | case 't': | |
3403 | { | |
3404 | - char newtab = optarg[0]; | |
3405 | - if (! newtab) | |
3406 | + char newtab[MB_LEN_MAX + 1]; | |
3407 | + size_t newtab_length = 1; | |
3408 | + strncpy (newtab, optarg, MB_LEN_MAX); | |
3409 | + if (! newtab[0]) | |
3410 | error (SORT_FAILURE, 0, _("empty tab")); | |
3411 | - if (optarg[1]) | |
3412 | +#if HAVE_MBRTOWC | |
3413 | + if (MB_CUR_MAX > 1) | |
3414 | + { | |
3415 | + wchar_t wc; | |
3416 | + mbstate_t state; | |
3417 | + | |
3418 | + memset (&state, '\0', sizeof (mbstate_t)); | |
3419 | + newtab_length = mbrtowc (&wc, newtab, strnlen (newtab, | |
3420 | + MB_LEN_MAX), | |
3421 | + &state); | |
3422 | + switch (newtab_length) | |
3423 | + { | |
3424 | + case (size_t) -1: | |
3425 | + case (size_t) -2: | |
3426 | + case 0: | |
3427 | + newtab_length = 1; | |
3428 | + } | |
3429 | + } | |
3430 | +#endif | |
3431 | + if (newtab_length == 1 && optarg[1]) | |
3432 | { | |
3433 | if (STREQ (optarg, "\\0")) | |
3434 | - newtab = '\0'; | |
3435 | + newtab[0] = '\0'; | |
3436 | else | |
3437 | { | |
3438 | /* Provoke with 'sort -txx'. Complain about | |
3439 | @@ -4511,9 +5201,12 @@ main (int argc, char **argv) | |
3440 | quote (optarg)); | |
3441 | } | |
3442 | } | |
3443 | - if (tab != TAB_DEFAULT && tab != newtab) | |
3444 | + if (tab_length | |
3445 | + && (tab_length != newtab_length | |
3446 | + || memcmp (tab, newtab, tab_length) != 0)) | |
3447 | error (SORT_FAILURE, 0, _("incompatible tabs")); | |
3448 | - tab = newtab; | |
3449 | + memcpy (tab, newtab, newtab_length); | |
3450 | + tab_length = newtab_length; | |
3451 | } | |
3452 | break; | |
3453 | ||
3454 | @@ -4751,12 +5444,10 @@ main (int argc, char **argv) | |
3455 | sort (files, nfiles, outfile, nthreads); | |
3456 | } | |
3457 | ||
3458 | -#ifdef lint | |
3459 | if (files_from) | |
3460 | readtokens0_free (&tok); | |
3461 | else | |
3462 | free (files); | |
3463 | -#endif | |
3464 | ||
3465 | if (have_read_stdin && fclose (stdin) == EOF) | |
3466 | die (_("close failed"), "-"); | |
3467 | diff -Naurp coreutils-8.25-orig/src/unexpand.c coreutils-8.25/src/unexpand.c | |
3468 | --- coreutils-8.25-orig/src/unexpand.c 2016-01-01 07:48:50.000000000 -0600 | |
3469 | +++ coreutils-8.25/src/unexpand.c 2016-02-08 19:07:10.311944651 -0600 | |
3470 | @@ -38,12 +38,29 @@ | |
3471 | #include <stdio.h> | |
3472 | #include <getopt.h> | |
3473 | #include <sys/types.h> | |
3474 | + | |
3475 | +/* Get mbstate_t, mbrtowc(), wcwidth(). */ | |
3476 | +#if HAVE_WCHAR_H | |
3477 | +# include <wchar.h> | |
3478 | +#endif | |
3479 | + | |
3480 | #include "system.h" | |
3481 | #include "error.h" | |
3482 | #include "fadvise.h" | |
3483 | #include "quote.h" | |
3484 | #include "xstrndup.h" | |
3485 | ||
3486 | +/* MB_LEN_MAX is incorrectly defined to be 1 in at least one GCC | |
3487 | + installation; work around this configuration error. */ | |
3488 | +#if !defined MB_LEN_MAX || MB_LEN_MAX < 2 | |
3489 | +# define MB_LEN_MAX 16 | |
3490 | +#endif | |
3491 | + | |
3492 | +/* Some systems, like BeOS, have multibyte encodings but lack mbstate_t. */ | |
3493 | +#if HAVE_MBRTOWC && defined mbstate_t | |
3494 | +# define mbrtowc(pwc, s, n, ps) (mbrtowc) (pwc, s, n, 0) | |
3495 | +#endif | |
3496 | + | |
3497 | /* The official name of this program (e.g., no 'g' prefix). */ | |
3498 | #define PROGRAM_NAME "unexpand" | |
3499 | ||
3500 | @@ -103,6 +120,210 @@ static struct option const longopts[] = | |
3501 | {NULL, 0, NULL, 0} | |
3502 | }; | |
3503 | ||
3504 | +static FILE *next_file (FILE *fp); | |
3505 | + | |
3506 | +#if HAVE_MBRTOWC | |
3507 | +static void | |
3508 | +unexpand_multibyte (void) | |
3509 | +{ | |
3510 | + FILE *fp; /* Input stream. */ | |
3511 | + mbstate_t i_state; /* Current shift state of the input stream. */ | |
3512 | + mbstate_t i_state_bak; /* Back up the I_STATE. */ | |
3513 | + mbstate_t o_state; /* Current shift state of the output stream. */ | |
3514 | + char buf[MB_LEN_MAX + BUFSIZ]; /* For spooling a read byte sequence. */ | |
3515 | + char *bufpos = buf; /* Next read position of BUF. */ | |
3516 | + size_t buflen = 0; /* The length of the byte sequence in buf. */ | |
3517 | + wint_t wc; /* A gotten wide character. */ | |
3518 | + size_t mblength; /* The byte size of a multibyte character | |
3519 | + which shows as same character as WC. */ | |
3520 | + bool prev_tab = false; | |
3521 | + | |
3522 | + /* Index in `tab_list' of next tabstop: */ | |
3523 | + int tab_index = 0; /* For calculating width of pending tabs. */ | |
3524 | + int print_tab_index = 0; /* For printing as many tabs as possible. */ | |
3525 | + unsigned int column = 0; /* Column on screen of next char. */ | |
3526 | + int next_tab_column; /* Column the next tab stop is on. */ | |
3527 | + int convert = 1; /* If nonzero, perform translations. */ | |
3528 | + unsigned int pending = 0; /* Pending columns of blanks. */ | |
3529 | + | |
3530 | + fp = next_file ((FILE *) NULL); | |
3531 | + if (fp == NULL) | |
3532 | + return; | |
3533 | + | |
3534 | + memset (&o_state, '\0', sizeof(mbstate_t)); | |
3535 | + memset (&i_state, '\0', sizeof(mbstate_t)); | |
3536 | + | |
3537 | + for (;;) | |
3538 | + { | |
3539 | + if (buflen < MB_LEN_MAX && !feof(fp) && !ferror(fp)) | |
3540 | + { | |
3541 | + memmove (buf, bufpos, buflen); | |
3542 | + buflen += fread (buf + buflen, sizeof(char), BUFSIZ, fp); | |
3543 | + bufpos = buf; | |
3544 | + } | |
3545 | + | |
3546 | + /* Get a wide character. */ | |
3547 | + if (buflen < 1) | |
3548 | + { | |
3549 | + mblength = 1; | |
3550 | + wc = WEOF; | |
3551 | + } | |
3552 | + else | |
3553 | + { | |
3554 | + i_state_bak = i_state; | |
3555 | + mblength = mbrtowc ((wchar_t *)&wc, bufpos, buflen, &i_state); | |
3556 | + } | |
3557 | + | |
3558 | + if (mblength == (size_t)-1 || mblength == (size_t)-2) | |
3559 | + { | |
3560 | + i_state = i_state_bak; | |
3561 | + wc = L'\0'; | |
3562 | + } | |
3563 | + | |
3564 | + if (wc == L' ' && convert && column < INT_MAX) | |
3565 | + { | |
3566 | + ++pending; | |
3567 | + ++column; | |
3568 | + } | |
3569 | + else if (wc == L'\t' && convert) | |
3570 | + { | |
3571 | + if (tab_size == 0) | |
3572 | + { | |
3573 | + /* Do not let tab_index == first_free_tab; | |
3574 | + stop when it is 1 less. */ | |
3575 | + while (tab_index < first_free_tab - 1 | |
3576 | + && column >= tab_list[tab_index]) | |
3577 | + tab_index++; | |
3578 | + next_tab_column = tab_list[tab_index]; | |
3579 | + if (tab_index < first_free_tab - 1) | |
3580 | + tab_index++; | |
3581 | + if (column >= next_tab_column) | |
3582 | + { | |
3583 | + convert = 0; /* Ran out of tab stops. */ | |
3584 | + goto flush_pend_mb; | |
3585 | + } | |
3586 | + } | |
3587 | + else | |
3588 | + { | |
3589 | + next_tab_column = column + tab_size - column % tab_size; | |
3590 | + } | |
3591 | + pending += next_tab_column - column; | |
3592 | + column = next_tab_column; | |
3593 | + } | |
3594 | + else | |
3595 | + { | |
3596 | +flush_pend_mb: | |
3597 | + /* Flush pending spaces. Print as many tabs as possible, | |
3598 | + then print the rest as spaces. */ | |
3599 | + if (pending == 1 && column != 1 && !prev_tab) | |
3600 | + { | |
3601 | + putchar (' '); | |
3602 | + pending = 0; | |
3603 | + } | |
3604 | + column -= pending; | |
3605 | + while (pending > 0) | |
3606 | + { | |
3607 | + if (tab_size == 0) | |
3608 | + { | |
3609 | + /* Do not let print_tab_index == first_free_tab; | |
3610 | + stop when it is 1 less. */ | |
3611 | + while (print_tab_index < first_free_tab - 1 | |
3612 | + && column >= tab_list[print_tab_index]) | |
3613 | + print_tab_index++; | |
3614 | + next_tab_column = tab_list[print_tab_index]; | |
3615 | + if (print_tab_index < first_free_tab - 1) | |
3616 | + print_tab_index++; | |
3617 | + } | |
3618 | + else | |
3619 | + { | |
3620 | + next_tab_column = | |
3621 | + column + tab_size - column % tab_size; | |
3622 | + } | |
3623 | + if (next_tab_column - column <= pending) | |
3624 | + { | |
3625 | + putchar ('\t'); | |
3626 | + pending -= next_tab_column - column; | |
3627 | + column = next_tab_column; | |
3628 | + } | |
3629 | + else | |
3630 | + { | |
3631 | + --print_tab_index; | |
3632 | + column += pending; | |
3633 | + while (pending != 0) | |
3634 | + { | |
3635 | + putchar (' '); | |
3636 | + pending--; | |
3637 | + } | |
3638 | + } | |
3639 | + } | |
3640 | + | |
3641 | + if (wc == WEOF) | |
3642 | + { | |
3643 | + fp = next_file (fp); | |
3644 | + if (fp == NULL) | |
3645 | + break; /* No more files. */ | |
3646 | + else | |
3647 | + { | |
3648 | + memset (&i_state, '\0', sizeof(mbstate_t)); | |
3649 | + continue; | |
3650 | + } | |
3651 | + } | |
3652 | + | |
3653 | + if (mblength == (size_t)-1 || mblength == (size_t)-2) | |
3654 | + { | |
3655 | + if (convert) | |
3656 | + { | |
3657 | + ++column; | |
3658 | + if (convert_entire_line == 0) | |
3659 | + convert = 0; | |
3660 | + } | |
3661 | + mblength = 1; | |
3662 | + putchar (buf[0]); | |
3663 | + } | |
3664 | + else if (mblength == 0) | |
3665 | + { | |
3666 | + if (convert && convert_entire_line == 0) | |
3667 | + convert = 0; | |
3668 | + mblength = 1; | |
3669 | + putchar ('\0'); | |
3670 | + } | |
3671 | + else | |
3672 | + { | |
3673 | + if (convert) | |
3674 | + { | |
3675 | + if (wc == L'\b') | |
3676 | + { | |
3677 | + if (column > 0) | |
3678 | + --column; | |
3679 | + } | |
3680 | + else | |
3681 | + { | |
3682 | + int width; /* The width of WC. */ | |
3683 | + | |
3684 | + width = wcwidth (wc); | |
3685 | + column += (width > 0) ? width : 0; | |
3686 | + if (convert_entire_line == 0) | |
3687 | + convert = 0; | |
3688 | + } | |
3689 | + } | |
3690 | + | |
3691 | + if (wc == L'\n') | |
3692 | + { | |
3693 | + tab_index = print_tab_index = 0; | |
3694 | + column = pending = 0; | |
3695 | + convert = 1; | |
3696 | + } | |
3697 | + fwrite (bufpos, sizeof(char), mblength, stdout); | |
3698 | + } | |
3699 | + } | |
3700 | + prev_tab = wc == L'\t'; | |
3701 | + buflen -= mblength; | |
3702 | + bufpos += mblength; | |
3703 | + } | |
3704 | +} | |
3705 | +#endif | |
3706 | + | |
3707 | + | |
3708 | void | |
3709 | usage (int status) | |
3710 | { | |
3711 | @@ -523,7 +744,12 @@ main (int argc, char **argv) | |
3712 | ||
3713 | file_list = (optind < argc ? &argv[optind] : stdin_argv); | |
3714 | ||
3715 | - unexpand (); | |
3716 | +#if HAVE_MBRTOWC | |
3717 | + if (MB_CUR_MAX > 1) | |
3718 | + unexpand_multibyte (); | |
3719 | + else | |
3720 | +#endif | |
3721 | + unexpand (); | |
3722 | ||
3723 | if (have_read_stdin && fclose (stdin) != 0) | |
3724 | error (EXIT_FAILURE, errno, "-"); | |
3725 | diff -Naurp coreutils-8.25-orig/src/uniq.c coreutils-8.25/src/uniq.c | |
3726 | --- coreutils-8.25-orig/src/uniq.c 2016-01-13 05:08:59.000000000 -0600 | |
3727 | +++ coreutils-8.25/src/uniq.c 2016-02-08 19:07:10.312944654 -0600 | |
3728 | @@ -21,6 +21,17 @@ | |
3729 | #include <getopt.h> | |
3730 | #include <sys/types.h> | |
3731 | ||
3732 | +/* Get mbstate_t, mbrtowc(). */ | |
3733 | +#if HAVE_WCHAR_H | |
3734 | +# include <wchar.h> | |
3735 | +#endif | |
3736 | + | |
3737 | +/* Get isw* functions. */ | |
3738 | +#if HAVE_WCTYPE_H | |
3739 | +# include <wctype.h> | |
3740 | +#endif | |
3741 | +#include <assert.h> | |
3742 | + | |
3743 | #include "system.h" | |
3744 | #include "argmatch.h" | |
3745 | #include "linebuffer.h" | |
3746 | @@ -33,6 +44,18 @@ | |
3747 | #include "xstrtol.h" | |
3748 | #include "memcasecmp.h" | |
3749 | #include "quote.h" | |
3750 | +#include "xmemcoll.h" | |
3751 | + | |
3752 | +/* MB_LEN_MAX is incorrectly defined to be 1 in at least one GCC | |
3753 | + installation; work around this configuration error. */ | |
3754 | +#if !defined MB_LEN_MAX || MB_LEN_MAX < 2 | |
3755 | +# define MB_LEN_MAX 16 | |
3756 | +#endif | |
3757 | + | |
3758 | +/* Some systems, like BeOS, have multibyte encodings but lack mbstate_t. */ | |
3759 | +#if HAVE_MBRTOWC && defined mbstate_t | |
3760 | +# define mbrtowc(pwc, s, n, ps) (mbrtowc) (pwc, s, n, 0) | |
3761 | +#endif | |
3762 | ||
3763 | /* The official name of this program (e.g., no 'g' prefix). */ | |
3764 | #define PROGRAM_NAME "uniq" | |
3765 | @@ -143,6 +166,10 @@ enum | |
3766 | GROUP_OPTION = CHAR_MAX + 1 | |
3767 | }; | |
3768 | ||
3769 | +/* Function pointers. */ | |
3770 | +static char * | |
3771 | +(*find_field) (struct linebuffer *line); | |
3772 | + | |
3773 | static struct option const longopts[] = | |
3774 | { | |
3775 | {"count", no_argument, NULL, 'c'}, | |
3776 | @@ -252,7 +279,7 @@ size_opt (char const *opt, char const *m | |
3777 | return a pointer to the beginning of the line's field to be compared. */ | |
3778 | ||
3779 | static char * _GL_ATTRIBUTE_PURE | |
3780 | -find_field (struct linebuffer const *line) | |
3781 | +find_field_uni (struct linebuffer *line) | |
3782 | { | |
3783 | size_t count; | |
3784 | char const *lp = line->buffer; | |
3785 | @@ -272,6 +299,83 @@ find_field (struct linebuffer const *lin | |
3786 | return line->buffer + i; | |
3787 | } | |
3788 | ||
3789 | +#if HAVE_MBRTOWC | |
3790 | + | |
3791 | +# define MBCHAR_TO_WCHAR(WC, MBLENGTH, LP, POS, SIZE, STATEP, CONVFAIL) \ | |
3792 | + do \ | |
3793 | + { \ | |
3794 | + mbstate_t state_bak; \ | |
3795 | + \ | |
3796 | + CONVFAIL = 0; \ | |
3797 | + state_bak = *STATEP; \ | |
3798 | + \ | |
3799 | + MBLENGTH = mbrtowc (&WC, LP + POS, SIZE - POS, STATEP); \ | |
3800 | + \ | |
3801 | + switch (MBLENGTH) \ | |
3802 | + { \ | |
3803 | + case (size_t)-2: \ | |
3804 | + case (size_t)-1: \ | |
3805 | + *STATEP = state_bak; \ | |
3806 | + CONVFAIL++; \ | |
3807 | + /* Fall through */ \ | |
3808 | + case 0: \ | |
3809 | + MBLENGTH = 1; \ | |
3810 | + } \ | |
3811 | + } \ | |
3812 | + while (0) | |
3813 | + | |
3814 | +static char * | |
3815 | +find_field_multi (struct linebuffer *line) | |
3816 | +{ | |
3817 | + size_t count; | |
3818 | + char *lp = line->buffer; | |
3819 | + size_t size = line->length - 1; | |
3820 | + size_t pos; | |
3821 | + size_t mblength; | |
3822 | + wchar_t wc; | |
3823 | + mbstate_t *statep; | |
3824 | + int convfail = 0; | |
3825 | + | |
3826 | + pos = 0; | |
3827 | + statep = &(line->state); | |
3828 | + | |
3829 | + /* skip fields. */ | |
3830 | + for (count = 0; count < skip_fields && pos < size; count++) | |
3831 | + { | |
3832 | + while (pos < size) | |
3833 | + { | |
3834 | + MBCHAR_TO_WCHAR (wc, mblength, lp, pos, size, statep, convfail); | |
3835 | + | |
3836 | + if (convfail || !(iswblank (wc) || wc == '\n')) | |
3837 | + { | |
3838 | + pos += mblength; | |
3839 | + break; | |
3840 | + } | |
3841 | + pos += mblength; | |
3842 | + } | |
3843 | + | |
3844 | + while (pos < size) | |
3845 | + { | |
3846 | + MBCHAR_TO_WCHAR (wc, mblength, lp, pos, size, statep, convfail); | |
3847 | + | |
3848 | + if (!convfail && (iswblank (wc) || wc == '\n')) | |
3849 | + break; | |
3850 | + | |
3851 | + pos += mblength; | |
3852 | + } | |
3853 | + } | |
3854 | + | |
3855 | + /* skip fields. */ | |
3856 | + for (count = 0; count < skip_chars && pos < size; count++) | |
3857 | + { | |
3858 | + MBCHAR_TO_WCHAR (wc, mblength, lp, pos, size, statep, convfail); | |
3859 | + pos += mblength; | |
3860 | + } | |
3861 | + | |
3862 | + return lp + pos; | |
3863 | +} | |
3864 | +#endif | |
3865 | + | |
3866 | /* Return false if two strings OLD and NEW match, true if not. | |
3867 | OLD and NEW point not to the beginnings of the lines | |
3868 | but rather to the beginnings of the fields to compare. | |
3869 | @@ -280,6 +384,8 @@ find_field (struct linebuffer const *lin | |
3870 | static bool | |
3871 | different (char *old, char *new, size_t oldlen, size_t newlen) | |
3872 | { | |
3873 | + char *copy_old, *copy_new; | |
3874 | + | |
3875 | if (check_chars < oldlen) | |
3876 | oldlen = check_chars; | |
3877 | if (check_chars < newlen) | |
3878 | @@ -287,15 +393,104 @@ different (char *old, char *new, size_t | |
3879 | ||
3880 | if (ignore_case) | |
3881 | { | |
3882 | - /* FIXME: This should invoke strcoll somehow. */ | |
3883 | - return oldlen != newlen || memcasecmp (old, new, oldlen); | |
3884 | + size_t i; | |
3885 | + | |
3886 | + copy_old = xmalloc (oldlen + 1); | |
3887 | + copy_new = xmalloc (oldlen + 1); | |
3888 | + | |
3889 | + for (i = 0; i < oldlen; i++) | |
3890 | + { | |
3891 | + copy_old[i] = toupper (old[i]); | |
3892 | + copy_new[i] = toupper (new[i]); | |
3893 | + } | |
3894 | + bool rc = xmemcoll (copy_old, oldlen, copy_new, newlen); | |
3895 | + free (copy_old); | |
3896 | + free (copy_new); | |
3897 | + return rc; | |
3898 | } | |
3899 | - else if (hard_LC_COLLATE) | |
3900 | - return xmemcoll (old, oldlen, new, newlen) != 0; | |
3901 | else | |
3902 | - return oldlen != newlen || memcmp (old, new, oldlen); | |
3903 | + { | |
3904 | + copy_old = (char *)old; | |
3905 | + copy_new = (char *)new; | |
3906 | + } | |
3907 | + | |
3908 | + return xmemcoll (copy_old, oldlen, copy_new, newlen); | |
3909 | + | |
3910 | } | |
3911 | ||
3912 | +#if HAVE_MBRTOWC | |
3913 | +static int | |
3914 | +different_multi (const char *old, const char *new, size_t oldlen, size_t newlen, mbstate_t oldstate, mbstate_t newstate) | |
3915 | +{ | |
3916 | + size_t i, j, chars; | |
3917 | + const char *str[2]; | |
3918 | + char *copy[2]; | |
3919 | + size_t len[2]; | |
3920 | + mbstate_t state[2]; | |
3921 | + size_t mblength; | |
3922 | + wchar_t wc, uwc; | |
3923 | + mbstate_t state_bak; | |
3924 | + | |
3925 | + str[0] = old; | |
3926 | + str[1] = new; | |
3927 | + len[0] = oldlen; | |
3928 | + len[1] = newlen; | |
3929 | + state[0] = oldstate; | |
3930 | + state[1] = newstate; | |
3931 | + | |
3932 | + for (i = 0; i < 2; i++) | |
3933 | + { | |
3934 | + copy[i] = xmalloc (len[i] + 1); | |
3935 | + memset (copy[i], '\0', len[i] + 1); | |
3936 | + | |
3937 | + for (j = 0, chars = 0; j < len[i] && chars < check_chars; chars++) | |
3938 | + { | |
3939 | + state_bak = state[i]; | |
3940 | + mblength = mbrtowc (&wc, str[i] + j, len[i] - j, &(state[i])); | |
3941 | + | |
3942 | + switch (mblength) | |
3943 | + { | |
3944 | + case (size_t)-1: | |
3945 | + case (size_t)-2: | |
3946 | + state[i] = state_bak; | |
3947 | + /* Fall through */ | |
3948 | + case 0: | |
3949 | + mblength = 1; | |
3950 | + break; | |
3951 | + | |
3952 | + default: | |
3953 | + if (ignore_case) | |
3954 | + { | |
3955 | + uwc = towupper (wc); | |
3956 | + | |
3957 | + if (uwc != wc) | |
3958 | + { | |
3959 | + mbstate_t state_wc; | |
3960 | + size_t mblen; | |
3961 | + | |
3962 | + memset (&state_wc, '\0', sizeof(mbstate_t)); | |
3963 | + mblen = wcrtomb (copy[i] + j, uwc, &state_wc); | |
3964 | + assert (mblen != (size_t)-1); | |
3965 | + } | |
3966 | + else | |
3967 | + memcpy (copy[i] + j, str[i] + j, mblength); | |
3968 | + } | |
3969 | + else | |
3970 | + memcpy (copy[i] + j, str[i] + j, mblength); | |
3971 | + } | |
3972 | + j += mblength; | |
3973 | + } | |
3974 | + copy[i][j] = '\0'; | |
3975 | + len[i] = j; | |
3976 | + } | |
3977 | + int rc = xmemcoll (copy[0], len[0], copy[1], len[1]); | |
3978 | + free (copy[0]); | |
3979 | + free (copy[1]); | |
3980 | + return rc; | |
3981 | + | |
3982 | +} | |
3983 | +#endif | |
3984 | + | |
3985 | /* Output the line in linebuffer LINE to standard output | |
3986 | provided that the switches say it should be output. | |
3987 | MATCH is true if the line matches the previous line. | |
3988 | @@ -359,19 +554,38 @@ check_file (const char *infile, const ch | |
3989 | char *prevfield IF_LINT ( = NULL); | |
3990 | size_t prevlen IF_LINT ( = 0); | |
3991 | bool first_group_printed = false; | |
3992 | +#if HAVE_MBRTOWC | |
3993 | + mbstate_t prevstate; | |
3994 | + | |
3995 | + memset (&prevstate, '\0', sizeof (mbstate_t)); | |
3996 | +#endif | |
3997 | ||
3998 | while (!feof (stdin)) | |
3999 | { | |
4000 | char *thisfield; | |
4001 | size_t thislen; | |
4002 | bool new_group; | |
4003 | +#if HAVE_MBRTOWC | |
4004 | + mbstate_t thisstate; | |
4005 | +#endif | |
4006 | ||
4007 | if (readlinebuffer_delim (thisline, stdin, delimiter) == 0) | |
4008 | break; | |
4009 | ||
4010 | thisfield = find_field (thisline); | |
4011 | thislen = thisline->length - 1 - (thisfield - thisline->buffer); | |
4012 | +#if HAVE_MBRTOWC | |
4013 | + if (MB_CUR_MAX > 1) | |
4014 | + { | |
4015 | + thisstate = thisline->state; | |
4016 | ||
4017 | + new_group = (prevline->length == 0 | |
4018 | + || different_multi (thisfield, prevfield, | |
4019 | + thislen, prevlen, | |
4020 | + thisstate, prevstate)); | |
4021 | + } | |
4022 | + else | |
4023 | +#endif | |
4024 | new_group = (prevline->length == 0 | |
4025 | || different (thisfield, prevfield, thislen, prevlen)); | |
4026 | ||
4027 | @@ -389,6 +603,10 @@ check_file (const char *infile, const ch | |
4028 | SWAP_LINES (prevline, thisline); | |
4029 | prevfield = thisfield; | |
4030 | prevlen = thislen; | |
4031 | +#if HAVE_MBRTOWC | |
4032 | + if (MB_CUR_MAX > 1) | |
4033 | + prevstate = thisstate; | |
4034 | +#endif | |
4035 | first_group_printed = true; | |
4036 | } | |
4037 | } | |
4038 | @@ -401,17 +619,26 @@ check_file (const char *infile, const ch | |
4039 | size_t prevlen; | |
4040 | uintmax_t match_count = 0; | |
4041 | bool first_delimiter = true; | |
4042 | +#if HAVE_MBRTOWC | |
4043 | + mbstate_t prevstate; | |
4044 | +#endif | |
4045 | ||
4046 | if (readlinebuffer_delim (prevline, stdin, delimiter) == 0) | |
4047 | goto closefiles; | |
4048 | prevfield = find_field (prevline); | |
4049 | prevlen = prevline->length - 1 - (prevfield - prevline->buffer); | |
4050 | +#if HAVE_MBRTOWC | |
4051 | + prevstate = prevline->state; | |
4052 | +#endif | |
4053 | ||
4054 | while (!feof (stdin)) | |
4055 | { | |
4056 | bool match; | |
4057 | char *thisfield; | |
4058 | size_t thislen; | |
4059 | +#if HAVE_MBRTOWC | |
4060 | + mbstate_t thisstate = thisline->state; | |
4061 | +#endif | |
4062 | if (readlinebuffer_delim (thisline, stdin, delimiter) == 0) | |
4063 | { | |
4064 | if (ferror (stdin)) | |
4065 | @@ -420,6 +647,14 @@ check_file (const char *infile, const ch | |
4066 | } | |
4067 | thisfield = find_field (thisline); | |
4068 | thislen = thisline->length - 1 - (thisfield - thisline->buffer); | |
4069 | +#if HAVE_MBRTOWC | |
4070 | + if (MB_CUR_MAX > 1) | |
4071 | + { | |
4072 | + match = !different_multi (thisfield, prevfield, | |
4073 | + thislen, prevlen, thisstate, prevstate); | |
4074 | + } | |
4075 | + else | |
4076 | +#endif | |
4077 | match = !different (thisfield, prevfield, thislen, prevlen); | |
4078 | match_count += match; | |
4079 | ||
4080 | @@ -452,6 +687,9 @@ check_file (const char *infile, const ch | |
4081 | SWAP_LINES (prevline, thisline); | |
4082 | prevfield = thisfield; | |
4083 | prevlen = thislen; | |
4084 | +#if HAVE_MBRTOWC | |
4085 | + prevstate = thisstate; | |
4086 | +#endif | |
4087 | if (!match) | |
4088 | match_count = 0; | |
4089 | } | |
4090 | @@ -498,6 +736,19 @@ main (int argc, char **argv) | |
4091 | ||
4092 | atexit (close_stdout); | |
4093 | ||
4094 | +#if HAVE_MBRTOWC | |
4095 | + if (MB_CUR_MAX > 1) | |
4096 | + { | |
4097 | + find_field = find_field_multi; | |
4098 | + } | |
4099 | + else | |
4100 | +#endif | |
4101 | + { | |
4102 | + find_field = find_field_uni; | |
4103 | + } | |
4104 | + | |
4105 | + | |
4106 | + | |
4107 | skip_chars = 0; | |
4108 | skip_fields = 0; | |
4109 | check_chars = SIZE_MAX; | |
4110 | diff -Naurp coreutils-8.25-orig/tests/i18n/sort-month.sh coreutils-8.25/tests/i18n/sort-month.sh | |
4111 | --- coreutils-8.25-orig/tests/i18n/sort-month.sh 1969-12-31 18:00:00.000000000 -0600 | |
4112 | +++ coreutils-8.25/tests/i18n/sort-month.sh 2016-02-08 19:07:10.312944654 -0600 | |
4113 | @@ -0,0 +1,34 @@ | |
4114 | +#!/bin/sh | |
4115 | +# Verify sort -M multi-byte support. | |
4116 | + | |
4117 | +. "${srcdir=.}/tests/init.sh"; path_prepend_ ./src | |
4118 | +print_ver_ sort | |
4119 | +require_valgrind_ | |
4120 | + | |
4121 | +# Skip this test if some deallocations are | |
4122 | +# avoided at process end. | |
4123 | +grep '^#define lint 1' $CONFIG_HEADER > /dev/null || | |
4124 | + skip_ 'Allocation checks only work reliably in "lint" mode' | |
4125 | + | |
4126 | +export LC_ALL=en_US.UTF-8 | |
4127 | +locale -k LC_CTYPE | grep -q "charmap.*UTF-8" \ | |
4128 | + || skip_ "No UTF-8 locale available" | |
4129 | + | |
4130 | +# Note the use of ɑ here which expands to | |
4131 | +# a wider representation upon case conversion | |
4132 | +# which triggered an assertion in sort -M | |
4133 | +cat <<EOF > exp | |
4134 | +. | |
4135 | +ɑ | |
4136 | +EOF | |
4137 | + | |
4138 | + | |
4139 | +# check large mem leak with --month-sort | |
4140 | +# https://bugzilla.redhat.com/show_bug.cgi?id=1259942 | |
4141 | +valgrind --leak-check=full \ | |
4142 | + --error-exitcode=1 --errors-for-leak-kinds=definite \ | |
4143 | + sort -M < exp > out || fail=1 | |
4144 | +compare exp out || { fail=1; cat out; } | |
4145 | + | |
4146 | + | |
4147 | +Exit $fail | |
4148 | diff -Naurp coreutils-8.25-orig/tests/i18n/sort.sh coreutils-8.25/tests/i18n/sort.sh | |
4149 | --- coreutils-8.25-orig/tests/i18n/sort.sh 1969-12-31 18:00:00.000000000 -0600 | |
4150 | +++ coreutils-8.25/tests/i18n/sort.sh 2016-02-08 19:07:10.312944654 -0600 | |
4151 | @@ -0,0 +1,29 @@ | |
4152 | +#!/bin/sh | |
4153 | +# Verify sort's multi-byte support. | |
4154 | + | |
4155 | +. "${srcdir=.}/tests/init.sh"; path_prepend_ ./src | |
4156 | +print_ver_ sort | |
4157 | + | |
4158 | +export LC_ALL=en_US.UTF-8 | |
4159 | +locale -k LC_CTYPE | grep -q "charmap.*UTF-8" \ | |
4160 | + || skip_ "No UTF-8 locale available" | |
4161 | + | |
4162 | +# Enable heap consistency checkng on older systems | |
4163 | +export MALLOC_CHECK_=2 | |
4164 | + | |
4165 | + | |
4166 | +# check buffer overflow issue due to | |
4167 | +# expanding multi-byte representation due to case conversion | |
4168 | +# https://bugzilla.suse.com/show_bug.cgi?id=928749 | |
4169 | +cat <<EOF > exp | |
4170 | +. | |
4171 | +ɑ | |
4172 | +EOF | |
4173 | +cat <<EOF | sort -f > out || fail=1 | |
4174 | +. | |
4175 | +ɑ | |
4176 | +EOF | |
4177 | +compare exp out || { fail=1; cat out; } | |
4178 | + | |
4179 | + | |
4180 | +Exit $fail | |
4181 | diff -Naurp coreutils-8.25-orig/tests/local.mk coreutils-8.25/tests/local.mk | |
4182 | --- coreutils-8.25-orig/tests/local.mk 2016-01-16 12:18:13.000000000 -0600 | |
4183 | +++ coreutils-8.25/tests/local.mk 2016-02-08 19:07:10.313944658 -0600 | |
4184 | @@ -344,6 +344,9 @@ all_tests = \ | |
4185 | tests/misc/sort-discrim.sh \ | |
4186 | tests/misc/sort-files0-from.pl \ | |
4187 | tests/misc/sort-float.sh \ | |
4188 | + tests/misc/sort-mb-tests.sh \ | |
4189 | + tests/i18n/sort.sh \ | |
4190 | + tests/i18n/sort-month.sh \ | |
4191 | tests/misc/sort-merge.pl \ | |
4192 | tests/misc/sort-merge-fdlimit.sh \ | |
4193 | tests/misc/sort-month.sh \ | |
4194 | diff -Naurp coreutils-8.25-orig/tests/misc/cut.pl coreutils-8.25/tests/misc/cut.pl | |
4195 | --- coreutils-8.25-orig/tests/misc/cut.pl 2016-01-16 12:18:13.000000000 -0600 | |
4196 | +++ coreutils-8.25/tests/misc/cut.pl 2016-02-08 19:07:10.314944661 -0600 | |
4197 | @@ -23,9 +23,11 @@ use strict; | |
4198 | # Turn off localization of executable's output. | |
4199 | @ENV{qw(LANGUAGE LANG LC_ALL)} = ('C') x 3; | |
4200 | ||
4201 | -my $mb_locale = $ENV{LOCALE_FR_UTF8}; | |
4202 | +my $mb_locale; | |
4203 | +# uncommented enable multibyte paths | |
4204 | +$mb_locale = $ENV{LOCALE_FR_UTF8}; | |
4205 | ! defined $mb_locale || $mb_locale eq 'none' | |
4206 | - and $mb_locale = 'C'; | |
4207 | + and $mb_locale = 'C'; | |
4208 | ||
4209 | my $prog = 'cut'; | |
4210 | my $try = "Try '$prog --help' for more information.\n"; | |
4211 | @@ -240,6 +242,7 @@ if ($mb_locale ne 'C') | |
4212 | my @new_t = @$t; | |
4213 | my $test_name = shift @new_t; | |
4214 | ||
4215 | + next if ($test_name =~ "newline-[12][0-9]"); | |
4216 | push @new, ["$test_name-mb", @new_t, {ENV => "LC_ALL=$mb_locale"}]; | |
4217 | } | |
4218 | push @Tests, @new; | |
4219 | diff -Naurp coreutils-8.25-orig/tests/misc/expand.pl coreutils-8.25/tests/misc/expand.pl | |
4220 | --- coreutils-8.25-orig/tests/misc/expand.pl 2016-01-16 12:18:13.000000000 -0600 | |
4221 | +++ coreutils-8.25/tests/misc/expand.pl 2016-02-08 19:07:10.314944661 -0600 | |
4222 | @@ -23,6 +23,15 @@ use strict; | |
4223 | # Turn off localization of executable's output. | |
4224 | @ENV{qw(LANGUAGE LANG LC_ALL)} = ('C') x 3; | |
4225 | ||
4226 | +#comment out next line to disable multibyte tests | |
4227 | +my $mb_locale = $ENV{LOCALE_FR_UTF8}; | |
4228 | +! defined $mb_locale || $mb_locale eq 'none' | |
4229 | + and $mb_locale = 'C'; | |
4230 | + | |
4231 | +my $prog = 'expand'; | |
4232 | +my $try = "Try \`$prog --help' for more information.\n"; | |
4233 | +my $inval = "$prog: invalid byte, character or field list\n$try"; | |
4234 | + | |
4235 | my @Tests = | |
4236 | ( | |
4237 | ['t1', '--tabs=3', {IN=>"a\tb"}, {OUT=>"a b"}], | |
4238 | @@ -31,6 +40,37 @@ my @Tests = | |
4239 | ['i2', '--tabs=3 -i', {IN=>" \ta\tb"}, {OUT=>" a\tb"}], | |
4240 | ); | |
4241 | ||
4242 | +if ($mb_locale ne 'C') | |
4243 | + { | |
4244 | + # Duplicate each test vector, appending "-mb" to the test name and | |
4245 | + # inserting {ENV => "LC_ALL=$mb_locale"} in the copy, so that we | |
4246 | + # provide coverage for the distro-added multi-byte code paths. | |
4247 | + my @new; | |
4248 | + foreach my $t (@Tests) | |
4249 | + { | |
4250 | + my @new_t = @$t; | |
4251 | + my $test_name = shift @new_t; | |
4252 | + | |
4253 | + # Depending on whether expand is multi-byte-patched, | |
4254 | + # it emits different diagnostics: | |
4255 | + # non-MB: invalid byte or field list | |
4256 | + # MB: invalid byte, character or field list | |
4257 | + # Adjust the expected error output accordingly. | |
4258 | + if (grep {ref $_ eq 'HASH' && exists $_->{ERR} && $_->{ERR} eq $inval} | |
4259 | + (@new_t)) | |
4260 | + { | |
4261 | + my $sub = {ERR_SUBST => 's/, character//'}; | |
4262 | + push @new_t, $sub; | |
4263 | + push @$t, $sub; | |
4264 | + } | |
4265 | + push @new, ["$test_name-mb", @new_t, {ENV => "LC_ALL=$mb_locale"}]; | |
4266 | + } | |
4267 | + push @Tests, @new; | |
4268 | + } | |
4269 | + | |
4270 | + | |
4271 | +@Tests = triple_test \@Tests; | |
4272 | + | |
4273 | my $save_temps = $ENV{DEBUG}; | |
4274 | my $verbose = $ENV{VERBOSE}; | |
4275 | ||
4276 | diff -Naurp coreutils-8.25-orig/tests/misc/fold.pl coreutils-8.25/tests/misc/fold.pl | |
4277 | --- coreutils-8.25-orig/tests/misc/fold.pl 2016-01-16 12:18:13.000000000 -0600 | |
4278 | +++ coreutils-8.25/tests/misc/fold.pl 2016-02-08 19:07:10.314944661 -0600 | |
4279 | @@ -20,9 +20,18 @@ use strict; | |
4280 | ||
4281 | (my $program_name = $0) =~ s|.*/||; | |
4282 | ||
4283 | +my $prog = 'fold'; | |
4284 | +my $try = "Try \`$prog --help' for more information.\n"; | |
4285 | +my $inval = "$prog: invalid byte, character or field list\n$try"; | |
4286 | + | |
4287 | # Turn off localization of executable's output. | |
4288 | @ENV{qw(LANGUAGE LANG LC_ALL)} = ('C') x 3; | |
4289 | ||
4290 | +# uncommented to enable multibyte paths | |
4291 | +my $mb_locale = $ENV{LOCALE_FR_UTF8}; | |
4292 | +! defined $mb_locale || $mb_locale eq 'none' | |
4293 | + and $mb_locale = 'C'; | |
4294 | + | |
4295 | my @Tests = | |
4296 | ( | |
4297 | ['s1', '-w2 -s', {IN=>"a\t"}, {OUT=>"a\n\t"}], | |
4298 | @@ -31,9 +40,48 @@ my @Tests = | |
4299 | ['s4', '-w4 -s', {IN=>"abc ef\n"}, {OUT=>"abc \nef\n"}], | |
4300 | ); | |
4301 | ||
4302 | +# Add _POSIX2_VERSION=199209 to the environment of each test | |
4303 | +# that uses an old-style option like +1. | |
4304 | +if ($mb_locale ne 'C') | |
4305 | + { | |
4306 | + # Duplicate each test vector, appending "-mb" to the test name and | |
4307 | + # inserting {ENV => "LC_ALL=$mb_locale"} in the copy, so that we | |
4308 | + # provide coverage for the distro-added multi-byte code paths. | |
4309 | + my @new; | |
4310 | + foreach my $t (@Tests) | |
4311 | + { | |
4312 | + my @new_t = @$t; | |
4313 | + my $test_name = shift @new_t; | |
4314 | + | |
4315 | + # Depending on whether fold is multi-byte-patched, | |
4316 | + # it emits different diagnostics: | |
4317 | + # non-MB: invalid byte or field list | |
4318 | + # MB: invalid byte, character or field list | |
4319 | + # Adjust the expected error output accordingly. | |
4320 | + if (grep {ref $_ eq 'HASH' && exists $_->{ERR} && $_->{ERR} eq $inval} | |
4321 | + (@new_t)) | |
4322 | + { | |
4323 | + my $sub = {ERR_SUBST => 's/, character//'}; | |
4324 | + push @new_t, $sub; | |
4325 | + push @$t, $sub; | |
4326 | + } | |
4327 | + push @new, ["$test_name-mb", @new_t, {ENV => "LC_ALL=$mb_locale"}]; | |
4328 | + } | |
4329 | + push @Tests, @new; | |
4330 | + } | |
4331 | + | |
4332 | +@Tests = triple_test \@Tests; | |
4333 | + | |
4334 | +# Remember that triple_test creates from each test with exactly one "IN" | |
4335 | +# file two more tests (.p and .r suffix on name) corresponding to reading | |
4336 | +# input from a file and from a pipe. The pipe-reading test would fail | |
4337 | +# due to a race condition about 1 in 20 times. | |
4338 | +# Remove the IN_PIPE version of the "output-is-input" test above. | |
4339 | +# The others aren't susceptible because they have three inputs each. | |
4340 | +@Tests = grep {$_->[0] ne 'output-is-input.p'} @Tests; | |
4341 | + | |
4342 | my $save_temps = $ENV{DEBUG}; | |
4343 | my $verbose = $ENV{VERBOSE}; | |
4344 | ||
4345 | -my $prog = 'fold'; | |
4346 | my $fail = run_tests ($program_name, $prog, \@Tests, $save_temps, $verbose); | |
4347 | exit $fail; | |
4348 | diff -Naurp coreutils-8.25-orig/tests/misc/join.pl coreutils-8.25/tests/misc/join.pl | |
4349 | --- coreutils-8.25-orig/tests/misc/join.pl 2016-01-16 12:18:13.000000000 -0600 | |
4350 | +++ coreutils-8.25/tests/misc/join.pl 2016-02-08 19:07:10.315944664 -0600 | |
4351 | @@ -25,6 +25,15 @@ my $limits = getlimits (); | |
4352 | ||
4353 | my $prog = 'join'; | |
4354 | ||
4355 | +my $try = "Try \`$prog --help' for more information.\n"; | |
4356 | +my $inval = "$prog: invalid byte, character or field list\n$try"; | |
4357 | + | |
4358 | +my $mb_locale; | |
4359 | +#Comment out next line to disable multibyte tests | |
4360 | +$mb_locale = $ENV{LOCALE_FR_UTF8}; | |
4361 | +! defined $mb_locale || $mb_locale eq 'none' | |
4362 | + and $mb_locale = 'C'; | |
4363 | + | |
4364 | my $delim = chr 0247; | |
4365 | sub t_subst ($) | |
4366 | { | |
4367 | @@ -329,8 +338,49 @@ foreach my $t (@tv) | |
4368 | push @Tests, $new_ent; | |
4369 | } | |
4370 | ||
4371 | +# Add _POSIX2_VERSION=199209 to the environment of each test | |
4372 | +# that uses an old-style option like +1. | |
4373 | +if ($mb_locale ne 'C') | |
4374 | + { | |
4375 | + # Duplicate each test vector, appending "-mb" to the test name and | |
4376 | + # inserting {ENV => "LC_ALL=$mb_locale"} in the copy, so that we | |
4377 | + # provide coverage for the distro-added multi-byte code paths. | |
4378 | + my @new; | |
4379 | + foreach my $t (@Tests) | |
4380 | + { | |
4381 | + my @new_t = @$t; | |
4382 | + my $test_name = shift @new_t; | |
4383 | + | |
4384 | + # Depending on whether join is multi-byte-patched, | |
4385 | + # it emits different diagnostics: | |
4386 | + # non-MB: invalid byte or field list | |
4387 | + # MB: invalid byte, character or field list | |
4388 | + # Adjust the expected error output accordingly. | |
4389 | + if (grep {ref $_ eq 'HASH' && exists $_->{ERR} && $_->{ERR} eq $inval} | |
4390 | + (@new_t)) | |
4391 | + { | |
4392 | + my $sub = {ERR_SUBST => 's/, character//'}; | |
4393 | + push @new_t, $sub; | |
4394 | + push @$t, $sub; | |
4395 | + } | |
4396 | + #Adjust the output some error messages including test_name for mb | |
4397 | + if (grep {ref $_ eq 'HASH' && exists $_->{ERR}} | |
4398 | + (@new_t)) | |
4399 | + { | |
4400 | + my $sub2 = {ERR_SUBST => "s/$test_name-mb/$test_name/"}; | |
4401 | + push @new_t, $sub2; | |
4402 | + push @$t, $sub2; | |
4403 | + } | |
4404 | + push @new, ["$test_name-mb", @new_t, {ENV => "LC_ALL=$mb_locale"}]; | |
4405 | + } | |
4406 | + push @Tests, @new; | |
4407 | + } | |
4408 | + | |
4409 | @Tests = triple_test \@Tests; | |
4410 | ||
4411 | +#skip invalid-j-mb test, it is failing because of the format | |
4412 | +@Tests = grep {$_->[0] ne 'invalid-j-mb'} @Tests; | |
4413 | + | |
4414 | my $save_temps = $ENV{DEBUG}; | |
4415 | my $verbose = $ENV{VERBOSE}; | |
4416 | ||
4417 | diff -Naurp coreutils-8.25-orig/tests/misc/sort-mb-tests.sh coreutils-8.25/tests/misc/sort-mb-tests.sh | |
4418 | --- coreutils-8.25-orig/tests/misc/sort-mb-tests.sh 1969-12-31 18:00:00.000000000 -0600 | |
4419 | +++ coreutils-8.25/tests/misc/sort-mb-tests.sh 2016-02-08 19:07:10.315944664 -0600 | |
4420 | @@ -0,0 +1,45 @@ | |
4421 | +#!/bin/sh | |
4422 | +# Verify sort's multi-byte support. | |
4423 | + | |
4424 | +. "${srcdir=.}/tests/init.sh"; path_prepend_ ./src | |
4425 | +print_ver_ sort | |
4426 | + | |
4427 | +export LC_ALL=en_US.UTF-8 | |
4428 | +locale -k LC_CTYPE | grep -q "charmap.*UTF-8" \ | |
4429 | + || skip_ "No UTF-8 locale available" | |
4430 | + | |
4431 | + | |
4432 | +cat <<EOF > exp | |
4433 | +Banana@5 | |
4434 | +Apple@10 | |
4435 | +Citrus@20 | |
4436 | +Cherry@30 | |
4437 | +EOF | |
4438 | + | |
4439 | +cat <<EOF | sort -t @ -k2 -n > out || fail=1 | |
4440 | +Apple@10 | |
4441 | +Banana@5 | |
4442 | +Citrus@20 | |
4443 | +Cherry@30 | |
4444 | +EOF | |
4445 | + | |
4446 | +compare exp out || { fail=1; cat out; } | |
4447 | + | |
4448 | + | |
4449 | +cat <<EOF > exp | |
4450 | +Citrus@AA20@@5 | |
4451 | +Cherry@AA30@@10 | |
4452 | +Apple@AA10@@20 | |
4453 | +Banana@AA5@@30 | |
4454 | +EOF | |
4455 | + | |
4456 | +cat <<EOF | sort -t @ -k4 -n > out || fail=1 | |
4457 | +Apple@AA10@@20 | |
4458 | +Banana@AA5@@30 | |
4459 | +Citrus@AA20@@5 | |
4460 | +Cherry@AA30@@10 | |
4461 | +EOF | |
4462 | + | |
4463 | +compare exp out || { fail=1; cat out; } | |
4464 | + | |
4465 | +Exit $fail | |
4466 | diff -Naurp coreutils-8.25-orig/tests/misc/sort-merge.pl coreutils-8.25/tests/misc/sort-merge.pl | |
4467 | --- coreutils-8.25-orig/tests/misc/sort-merge.pl 2016-01-16 12:18:14.000000000 -0600 | |
4468 | +++ coreutils-8.25/tests/misc/sort-merge.pl 2016-02-08 19:07:10.316944667 -0600 | |
4469 | @@ -26,6 +26,15 @@ my $prog = 'sort'; | |
4470 | # Turn off localization of executable's output. | |
4471 | @ENV{qw(LANGUAGE LANG LC_ALL)} = ('C') x 3; | |
4472 | ||
4473 | +my $mb_locale; | |
4474 | +# uncommented according to upstream commit enabling multibyte paths | |
4475 | +$mb_locale = $ENV{LOCALE_FR_UTF8}; | |
4476 | +! defined $mb_locale || $mb_locale eq 'none' | |
4477 | + and $mb_locale = 'C'; | |
4478 | + | |
4479 | +my $try = "Try \`$prog --help' for more information.\n"; | |
4480 | +my $inval = "$prog: invalid byte, character or field list\n$try"; | |
4481 | + | |
4482 | # three empty files and one that says 'foo' | |
4483 | my @inputs = (+(map{{IN=> {"empty$_"=> ''}}}1..3), {IN=> {foo=> "foo\n"}}); | |
4484 | ||
4485 | @@ -77,6 +86,39 @@ my @Tests = | |
4486 | {OUT=>$big_input}], | |
4487 | ); | |
4488 | ||
4489 | +# Add _POSIX2_VERSION=199209 to the environment of each test | |
4490 | +# that uses an old-style option like +1. | |
4491 | +if ($mb_locale ne 'C') | |
4492 | + { | |
4493 | + # Duplicate each test vector, appending "-mb" to the test name and | |
4494 | + # inserting {ENV => "LC_ALL=$mb_locale"} in the copy, so that we | |
4495 | + # provide coverage for the distro-added multi-byte code paths. | |
4496 | + my @new; | |
4497 | + foreach my $t (@Tests) | |
4498 | + { | |
4499 | + my @new_t = @$t; | |
4500 | + my $test_name = shift @new_t; | |
4501 | + | |
4502 | + # Depending on whether sort is multi-byte-patched, | |
4503 | + # it emits different diagnostics: | |
4504 | + # non-MB: invalid byte or field list | |
4505 | + # MB: invalid byte, character or field list | |
4506 | + # Adjust the expected error output accordingly. | |
4507 | + if (grep {ref $_ eq 'HASH' && exists $_->{ERR} && $_->{ERR} eq $inval} | |
4508 | + (@new_t)) | |
4509 | + { | |
4510 | + my $sub = {ERR_SUBST => 's/, character//'}; | |
4511 | + push @new_t, $sub; | |
4512 | + push @$t, $sub; | |
4513 | + } | |
4514 | + next if ($test_name =~ "nmerge-."); | |
4515 | + push @new, ["$test_name-mb", @new_t, {ENV => "LC_ALL=$mb_locale"}]; | |
4516 | + } | |
4517 | + push @Tests, @new; | |
4518 | + } | |
4519 | + | |
4520 | +@Tests = triple_test \@Tests; | |
4521 | + | |
4522 | my $save_temps = $ENV{DEBUG}; | |
4523 | my $verbose = $ENV{VERBOSE}; | |
4524 | ||
4525 | diff -Naurp coreutils-8.25-orig/tests/misc/sort.pl coreutils-8.25/tests/misc/sort.pl | |
4526 | --- coreutils-8.25-orig/tests/misc/sort.pl 2016-01-16 12:18:14.000000000 -0600 | |
4527 | +++ coreutils-8.25/tests/misc/sort.pl 2016-02-08 19:07:10.316944667 -0600 | |
4528 | @@ -24,10 +24,15 @@ my $prog = 'sort'; | |
4529 | # Turn off localization of executable's output. | |
4530 | @ENV{qw(LANGUAGE LANG LC_ALL)} = ('C') x 3; | |
4531 | ||
4532 | -my $mb_locale = $ENV{LOCALE_FR_UTF8}; | |
4533 | +my $mb_locale; | |
4534 | +#Comment out next line to disable multibyte tests | |
4535 | +$mb_locale = $ENV{LOCALE_FR_UTF8}; | |
4536 | ! defined $mb_locale || $mb_locale eq 'none' | |
4537 | and $mb_locale = 'C'; | |
4538 | ||
4539 | +my $try = "Try \`$prog --help' for more information.\n"; | |
4540 | +my $inval = "$prog: invalid byte, character or field list\n$try"; | |
4541 | + | |
4542 | # Since each test is run with a file name and with redirected stdin, | |
4543 | # the name in the diagnostic is either the file name or "-". | |
4544 | # Normalize each diagnostic to use '-'. | |
4545 | @@ -424,6 +429,38 @@ foreach my $t (@Tests) | |
4546 | } | |
4547 | } | |
4548 | ||
4549 | +if ($mb_locale ne 'C') | |
4550 | + { | |
4551 | + # Duplicate each test vector, appending "-mb" to the test name and | |
4552 | + # inserting {ENV => "LC_ALL=$mb_locale"} in the copy, so that we | |
4553 | + # provide coverage for the distro-added multi-byte code paths. | |
4554 | + my @new; | |
4555 | + foreach my $t (@Tests) | |
4556 | + { | |
4557 | + my @new_t = @$t; | |
4558 | + my $test_name = shift @new_t; | |
4559 | + | |
4560 | + # Depending on whether sort is multi-byte-patched, | |
4561 | + # it emits different diagnostics: | |
4562 | + # non-MB: invalid byte or field list | |
4563 | + # MB: invalid byte, character or field list | |
4564 | + # Adjust the expected error output accordingly. | |
4565 | + if (grep {ref $_ eq 'HASH' && exists $_->{ERR} && $_->{ERR} eq $inval} | |
4566 | + (@new_t)) | |
4567 | + { | |
4568 | + my $sub = {ERR_SUBST => 's/, character//'}; | |
4569 | + push @new_t, $sub; | |
4570 | + push @$t, $sub; | |
4571 | + } | |
4572 | + #disable several failing tests until investigation, disable all tests with envvars set | |
4573 | + next if (grep {ref $_ eq 'HASH' && exists $_->{ENV}} (@new_t)); | |
4574 | + next if ($test_name =~ "18g" or $test_name =~ "sort-numeric" or $test_name =~ "08[ab]" or $test_name =~ "03[def]" or $test_name =~ "h4" or $test_name =~ "n1" or $test_name =~ "2[01]a"); | |
4575 | + next if ($test_name =~ "11[ab]"); # avoid FP: expected result differs to MB result due to collation rules. | |
4576 | + push @new, ["$test_name-mb", @new_t, {ENV => "LC_ALL=$mb_locale"}]; | |
4577 | + } | |
4578 | + push @Tests, @new; | |
4579 | + } | |
4580 | + | |
4581 | @Tests = triple_test \@Tests; | |
4582 | ||
4583 | # Remember that triple_test creates from each test with exactly one "IN" | |
4584 | @@ -433,6 +470,7 @@ foreach my $t (@Tests) | |
4585 | # Remove the IN_PIPE version of the "output-is-input" test above. | |
4586 | # The others aren't susceptible because they have three inputs each. | |
4587 | @Tests = grep {$_->[0] ne 'output-is-input.p'} @Tests; | |
4588 | +@Tests = grep {$_->[0] ne 'output-is-input-mb.p'} @Tests; | |
4589 | ||
4590 | my $save_temps = $ENV{DEBUG}; | |
4591 | my $verbose = $ENV{VERBOSE}; | |
4592 | diff -Naurp coreutils-8.25-orig/tests/misc/unexpand.pl coreutils-8.25/tests/misc/unexpand.pl | |
4593 | --- coreutils-8.25-orig/tests/misc/unexpand.pl 2016-01-16 12:18:14.000000000 -0600 | |
4594 | +++ coreutils-8.25/tests/misc/unexpand.pl 2016-02-08 19:07:10.317944671 -0600 | |
4595 | @@ -27,6 +27,14 @@ my $limits = getlimits (); | |
4596 | ||
4597 | my $prog = 'unexpand'; | |
4598 | ||
4599 | +# comment out next line to disable multibyte tests | |
4600 | +my $mb_locale = $ENV{LOCALE_FR_UTF8}; | |
4601 | +! defined $mb_locale || $mb_locale eq 'none' | |
4602 | + and $mb_locale = 'C'; | |
4603 | + | |
4604 | +my $try = "Try \`$prog --help' for more information.\n"; | |
4605 | +my $inval = "$prog: invalid byte, character or field list\n$try"; | |
4606 | + | |
4607 | my @Tests = | |
4608 | ( | |
4609 | ['a1', {IN=> ' 'x 1 ."y\n"}, {OUT=> ' 'x 1 ."y\n"}], | |
4610 | @@ -92,6 +100,37 @@ my @Tests = | |
4611 | {EXIT => 1}, {ERR => "$prog: tab stop value is too large\n"}], | |
4612 | ); | |
4613 | ||
4614 | +if ($mb_locale ne 'C') | |
4615 | + { | |
4616 | + # Duplicate each test vector, appending "-mb" to the test name and | |
4617 | + # inserting {ENV => "LC_ALL=$mb_locale"} in the copy, so that we | |
4618 | + # provide coverage for the distro-added multi-byte code paths. | |
4619 | + my @new; | |
4620 | + foreach my $t (@Tests) | |
4621 | + { | |
4622 | + my @new_t = @$t; | |
4623 | + my $test_name = shift @new_t; | |
4624 | + | |
4625 | + # Depending on whether unexpand is multi-byte-patched, | |
4626 | + # it emits different diagnostics: | |
4627 | + # non-MB: invalid byte or field list | |
4628 | + # MB: invalid byte, character or field list | |
4629 | + # Adjust the expected error output accordingly. | |
4630 | + if (grep {ref $_ eq 'HASH' && exists $_->{ERR} && $_->{ERR} eq $inval} | |
4631 | + (@new_t)) | |
4632 | + { | |
4633 | + my $sub = {ERR_SUBST => 's/, character//'}; | |
4634 | + push @new_t, $sub; | |
4635 | + push @$t, $sub; | |
4636 | + } | |
4637 | + next if ($test_name =~ 'b-1'); | |
4638 | + push @new, ["$test_name-mb", @new_t, {ENV => "LC_ALL=$mb_locale"}]; | |
4639 | + } | |
4640 | + push @Tests, @new; | |
4641 | + } | |
4642 | + | |
4643 | +@Tests = triple_test \@Tests; | |
4644 | + | |
4645 | my $save_temps = $ENV{DEBUG}; | |
4646 | my $verbose = $ENV{VERBOSE}; | |
4647 | ||
4648 | diff -Naurp coreutils-8.25-orig/tests/misc/uniq.pl coreutils-8.25/tests/misc/uniq.pl | |
4649 | --- coreutils-8.25-orig/tests/misc/uniq.pl 2016-01-16 12:18:14.000000000 -0600 | |
4650 | +++ coreutils-8.25/tests/misc/uniq.pl 2016-02-08 19:07:10.317944671 -0600 | |
4651 | @@ -23,9 +23,17 @@ my $limits = getlimits (); | |
4652 | my $prog = 'uniq'; | |
4653 | my $try = "Try '$prog --help' for more information.\n"; | |
4654 | ||
4655 | +my $inval = "$prog: invalid byte, character or field list\n$try"; | |
4656 | + | |
4657 | # Turn off localization of executable's output. | |
4658 | @ENV{qw(LANGUAGE LANG LC_ALL)} = ('C') x 3; | |
4659 | ||
4660 | +my $mb_locale; | |
4661 | +#Comment out next line to disable multibyte tests | |
4662 | +$mb_locale = $ENV{LOCALE_FR_UTF8}; | |
4663 | +! defined $mb_locale || $mb_locale eq 'none' | |
4664 | + and $mb_locale = 'C'; | |
4665 | + | |
4666 | # When possible, create a "-z"-testing variant of each test. | |
4667 | sub add_z_variants($) | |
4668 | { | |
4669 | @@ -262,6 +270,53 @@ foreach my $t (@Tests) | |
4670 | and push @$t, {ENV=>'_POSIX2_VERSION=199209'}; | |
4671 | } | |
4672 | ||
4673 | +if ($mb_locale ne 'C') | |
4674 | + { | |
4675 | + # Duplicate each test vector, appending "-mb" to the test name and | |
4676 | + # inserting {ENV => "LC_ALL=$mb_locale"} in the copy, so that we | |
4677 | + # provide coverage for the distro-added multi-byte code paths. | |
4678 | + my @new; | |
4679 | + foreach my $t (@Tests) | |
4680 | + { | |
4681 | + my @new_t = @$t; | |
4682 | + my $test_name = shift @new_t; | |
4683 | + | |
4684 | + # Depending on whether uniq is multi-byte-patched, | |
4685 | + # it emits different diagnostics: | |
4686 | + # non-MB: invalid byte or field list | |
4687 | + # MB: invalid byte, character or field list | |
4688 | + # Adjust the expected error output accordingly. | |
4689 | + if (grep {ref $_ eq 'HASH' && exists $_->{ERR} && $_->{ERR} eq $inval} | |
4690 | + (@new_t)) | |
4691 | + { | |
4692 | + my $sub = {ERR_SUBST => 's/, character//'}; | |
4693 | + push @new_t, $sub; | |
4694 | + push @$t, $sub; | |
4695 | + } | |
4696 | + # In test #145, replace the each ‘...’ by '...'. | |
4697 | + if ($test_name =~ "145") | |
4698 | + { | |
4699 | + my $sub = { ERR_SUBST => "s/‘([^’]+)’/'\$1'/g"}; | |
4700 | + push @new_t, $sub; | |
4701 | + push @$t, $sub; | |
4702 | + } | |
4703 | + next if ( $test_name =~ "schar" | |
4704 | + or $test_name =~ "^obs-plus" | |
4705 | + or $test_name =~ "119"); | |
4706 | + push @new, ["$test_name-mb", @new_t, {ENV => "LC_ALL=$mb_locale"}]; | |
4707 | + } | |
4708 | + push @Tests, @new; | |
4709 | + } | |
4710 | + | |
4711 | +# Remember that triple_test creates from each test with exactly one "IN" | |
4712 | +# file two more tests (.p and .r suffix on name) corresponding to reading | |
4713 | +# input from a file and from a pipe. The pipe-reading test would fail | |
4714 | +# due to a race condition about 1 in 20 times. | |
4715 | +# Remove the IN_PIPE version of the "output-is-input" test above. | |
4716 | +# The others aren't susceptible because they have three inputs each. | |
4717 | + | |
4718 | +@Tests = grep {$_->[0] ne 'output-is-input.p'} @Tests; | |
4719 | + | |
4720 | @Tests = add_z_variants \@Tests; | |
4721 | @Tests = triple_test \@Tests; | |
4722 | ||
4723 | diff -Naurp coreutils-8.25-orig/tests/pr/pr-tests.pl coreutils-8.25/tests/pr/pr-tests.pl | |
4724 | --- coreutils-8.25-orig/tests/pr/pr-tests.pl 2016-01-16 12:18:14.000000000 -0600 | |
4725 | +++ coreutils-8.25/tests/pr/pr-tests.pl 2016-02-08 19:07:10.318944674 -0600 | |
4726 | @@ -24,6 +24,15 @@ use strict; | |
4727 | my $prog = 'pr'; | |
4728 | my $normalize_strerror = "s/': .*/'/"; | |
4729 | ||
4730 | +my $mb_locale; | |
4731 | +#Uncomment the following line to enable multibyte tests | |
4732 | +$mb_locale = $ENV{LOCALE_FR_UTF8}; | |
4733 | +! defined $mb_locale || $mb_locale eq 'none' | |
4734 | + and $mb_locale = 'C'; | |
4735 | + | |
4736 | +my $try = "Try \`$prog --help' for more information.\n"; | |
4737 | +my $inval = "$prog: invalid byte, character or field list\n$try"; | |
4738 | + | |
4739 | my @tv = ( | |
4740 | ||
4741 | # -b option is no longer an official option. But it's still working to | |
4742 | @@ -467,8 +476,48 @@ push @Tests, | |
4743 | {IN=>{3=>"x\ty\tz\n"}}, | |
4744 | {OUT=>join("\t", qw(a b c m n o x y z)) . "\n"} ]; | |
4745 | ||
4746 | +# Add _POSIX2_VERSION=199209 to the environment of each test | |
4747 | +# that uses an old-style option like +1. | |
4748 | +if ($mb_locale ne 'C') | |
4749 | + { | |
4750 | + # Duplicate each test vector, appending "-mb" to the test name and | |
4751 | + # inserting {ENV => "LC_ALL=$mb_locale"} in the copy, so that we | |
4752 | + # provide coverage for the distro-added multi-byte code paths. | |
4753 | + my @new; | |
4754 | + foreach my $t (@Tests) | |
4755 | + { | |
4756 | + my @new_t = @$t; | |
4757 | + my $test_name = shift @new_t; | |
4758 | + | |
4759 | + # Depending on whether pr is multi-byte-patched, | |
4760 | + # it emits different diagnostics: | |
4761 | + # non-MB: invalid byte or field list | |
4762 | + # MB: invalid byte, character or field list | |
4763 | + # Adjust the expected error output accordingly. | |
4764 | + if (grep {ref $_ eq 'HASH' && exists $_->{ERR} && $_->{ERR} eq $inval} | |
4765 | + (@new_t)) | |
4766 | + { | |
4767 | + my $sub = {ERR_SUBST => 's/, character//'}; | |
4768 | + push @new_t, $sub; | |
4769 | + push @$t, $sub; | |
4770 | + } | |
4771 | + #temporarily skip some failing tests | |
4772 | + next if ($test_name =~ "col-0" or $test_name =~ "col-inval"); | |
4773 | + push @new, ["$test_name-mb", @new_t, {ENV => "LC_ALL=$mb_locale"}]; | |
4774 | + } | |
4775 | + push @Tests, @new; | |
4776 | + } | |
4777 | + | |
4778 | @Tests = triple_test \@Tests; | |
4779 | ||
4780 | +# Remember that triple_test creates from each test with exactly one "IN" | |
4781 | +# file two more tests (.p and .r suffix on name) corresponding to reading | |
4782 | +# input from a file and from a pipe. The pipe-reading test would fail | |
4783 | +# due to a race condition about 1 in 20 times. | |
4784 | +# Remove the IN_PIPE version of the "output-is-input" test above. | |
4785 | +# The others aren't susceptible because they have three inputs each. | |
4786 | +@Tests = grep {$_->[0] ne 'output-is-input.p'} @Tests; | |
4787 | + | |
4788 | my $save_temps = $ENV{DEBUG}; | |
4789 | my $verbose = $ENV{VERBOSE}; | |
4790 |