diff options
author | nobody <nobody@localhost> | 2004-10-13 09:42:10 +0000 |
---|---|---|
committer | nobody <nobody@localhost> | 2004-10-13 09:42:10 +0000 |
commit | 8c59a0bf0e9e2d87b0ff273ea3f0bf05bbbf6373 (patch) | |
tree | 1826706cd4fd009fcd14f4f8021005ec8ec0fa59 /busybox/coreutils/cut.c | |
download | busybox-w32-8c59a0bf0e9e2d87b0ff273ea3f0bf05bbbf6373.tar.gz busybox-w32-8c59a0bf0e9e2d87b0ff273ea3f0bf05bbbf6373.tar.bz2 busybox-w32-8c59a0bf0e9e2d87b0ff273ea3f0bf05bbbf6373.zip |
This commit was manufactured by cvs2svn to create tag 'busybox_1_00'.
Diffstat (limited to 'busybox/coreutils/cut.c')
-rw-r--r-- | busybox/coreutils/cut.c | 344 |
1 files changed, 344 insertions, 0 deletions
diff --git a/busybox/coreutils/cut.c b/busybox/coreutils/cut.c new file mode 100644 index 000000000..d26e80eee --- /dev/null +++ b/busybox/coreutils/cut.c | |||
@@ -0,0 +1,344 @@ | |||
1 | /* vi: set sw=8 ts=8: */ | ||
2 | /* | ||
3 | * cut.c - minimalist version of cut | ||
4 | * | ||
5 | * Copyright (C) 1999,2000,2001 by Lineo, inc. | ||
6 | * Written by Mark Whitley <markw@codepoet.org> | ||
7 | * | ||
8 | * This program is free software; you can redistribute it and/or modify | ||
9 | * it under the terms of the GNU General Public License as published by | ||
10 | * the Free Software Foundation; either version 2 of the License, or | ||
11 | * (at your option) any later version. | ||
12 | * | ||
13 | * This program is distributed in the hope that it will be useful, | ||
14 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | ||
15 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | ||
16 | * General Public License for more details. | ||
17 | * | ||
18 | * You should have received a copy of the GNU General Public License | ||
19 | * along with this program; if not, write to the Free Software | ||
20 | * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA | ||
21 | * | ||
22 | */ | ||
23 | |||
24 | #include <stdio.h> | ||
25 | #include <stdlib.h> | ||
26 | #include <getopt.h> | ||
27 | #include <unistd.h> | ||
28 | #include <string.h> | ||
29 | #include <limits.h> | ||
30 | #include "busybox.h" | ||
31 | |||
32 | |||
33 | /* option vars */ | ||
34 | static const char optstring[] = "b:c:f:d:sn"; | ||
35 | #define OPT_BYTE_FLGS 1 | ||
36 | #define OPT_CHAR_FLGS 2 | ||
37 | #define OPT_FIELDS_FLGS 4 | ||
38 | #define OPT_DELIM_FLGS 8 | ||
39 | #define OPT_SUPRESS_FLGS 16 | ||
40 | static char part; /* (b)yte, (c)har, (f)ields */ | ||
41 | static unsigned int supress_non_delimited_lines; | ||
42 | static char delim = '\t'; /* delimiter, default is tab */ | ||
43 | |||
44 | struct cut_list { | ||
45 | int startpos; | ||
46 | int endpos; | ||
47 | }; | ||
48 | |||
49 | static const int BOL = 0; | ||
50 | static const int EOL = INT_MAX; | ||
51 | static const int NON_RANGE = -1; | ||
52 | |||
53 | static struct cut_list *cut_lists = NULL; /* growable array holding a series of lists */ | ||
54 | static unsigned int nlists = 0; /* number of elements in above list */ | ||
55 | |||
56 | |||
57 | static int cmpfunc(const void *a, const void *b) | ||
58 | { | ||
59 | struct cut_list *la = (struct cut_list *)a; | ||
60 | struct cut_list *lb = (struct cut_list *)b; | ||
61 | |||
62 | if (la->startpos > lb->startpos) | ||
63 | return 1; | ||
64 | if (la->startpos < lb->startpos) | ||
65 | return -1; | ||
66 | return 0; | ||
67 | } | ||
68 | |||
69 | |||
70 | /* | ||
71 | * parse_lists() - parses a list and puts values into startpos and endpos. | ||
72 | * valid list formats: N, N-, N-M, -M | ||
73 | * more than one list can be separated by commas | ||
74 | */ | ||
75 | static void parse_lists(char *lists) | ||
76 | { | ||
77 | char *ltok = NULL; | ||
78 | char *ntok = NULL; | ||
79 | char *junk; | ||
80 | int s = 0, e = 0; | ||
81 | |||
82 | /* take apart the lists, one by one (they are separated with commas */ | ||
83 | while ((ltok = strsep(&lists, ",")) != NULL) { | ||
84 | |||
85 | /* it's actually legal to pass an empty list */ | ||
86 | if (strlen(ltok) == 0) | ||
87 | continue; | ||
88 | |||
89 | /* get the start pos */ | ||
90 | ntok = strsep(<ok, "-"); | ||
91 | if (ntok == NULL) { | ||
92 | fprintf(stderr, "Help ntok is null for starting position! What do I do?\n"); | ||
93 | } else if (strlen(ntok) == 0) { | ||
94 | s = BOL; | ||
95 | } else { | ||
96 | s = strtoul(ntok, &junk, 10); | ||
97 | if(*junk != '\0' || s < 0) | ||
98 | bb_error_msg_and_die("invalid byte or field list"); | ||
99 | |||
100 | /* account for the fact that arrays are zero based, while the user | ||
101 | * expects the first char on the line to be char # 1 */ | ||
102 | if (s != 0) | ||
103 | s--; | ||
104 | } | ||
105 | |||
106 | /* get the end pos */ | ||
107 | ntok = strsep(<ok, "-"); | ||
108 | if (ntok == NULL) { | ||
109 | e = NON_RANGE; | ||
110 | } else if (strlen(ntok) == 0) { | ||
111 | e = EOL; | ||
112 | } else { | ||
113 | e = strtoul(ntok, &junk, 10); | ||
114 | if(*junk != '\0' || e < 0) | ||
115 | bb_error_msg_and_die("invalid byte or field list"); | ||
116 | /* if the user specified and end position of 0, that means "til the | ||
117 | * end of the line */ | ||
118 | if (e == 0) | ||
119 | e = INT_MAX; | ||
120 | e--; /* again, arrays are zero based, lines are 1 based */ | ||
121 | if (e == s) | ||
122 | e = NON_RANGE; | ||
123 | } | ||
124 | |||
125 | /* if there's something left to tokenize, the user past an invalid list */ | ||
126 | if (ltok) | ||
127 | bb_error_msg_and_die("invalid byte or field list"); | ||
128 | |||
129 | /* add the new list */ | ||
130 | cut_lists = xrealloc(cut_lists, sizeof(struct cut_list) * (++nlists)); | ||
131 | cut_lists[nlists-1].startpos = s; | ||
132 | cut_lists[nlists-1].endpos = e; | ||
133 | } | ||
134 | |||
135 | /* make sure we got some cut positions out of all that */ | ||
136 | if (nlists == 0) | ||
137 | bb_error_msg_and_die("missing list of positions"); | ||
138 | |||
139 | /* now that the lists are parsed, we need to sort them to make life easier | ||
140 | * on us when it comes time to print the chars / fields / lines */ | ||
141 | qsort(cut_lists, nlists, sizeof(struct cut_list), cmpfunc); | ||
142 | |||
143 | } | ||
144 | |||
145 | |||
146 | static void cut_line_by_chars(const char *line) | ||
147 | { | ||
148 | int c, l; | ||
149 | /* set up a list so we can keep track of what's been printed */ | ||
150 | char *printed = xcalloc(strlen(line), sizeof(char)); | ||
151 | |||
152 | /* print the chars specified in each cut list */ | ||
153 | for (c = 0; c < nlists; c++) { | ||
154 | l = cut_lists[c].startpos; | ||
155 | while (l < strlen(line)) { | ||
156 | if (!printed[l]) { | ||
157 | putchar(line[l]); | ||
158 | printed[l] = 'X'; | ||
159 | } | ||
160 | l++; | ||
161 | if (cut_lists[c].endpos == NON_RANGE || l > cut_lists[c].endpos) | ||
162 | break; | ||
163 | } | ||
164 | } | ||
165 | putchar('\n'); /* cuz we were handed a chomped line */ | ||
166 | free(printed); | ||
167 | } | ||
168 | |||
169 | |||
170 | static void cut_line_by_fields(char *line) | ||
171 | { | ||
172 | int c, f; | ||
173 | int ndelim = -1; /* zero-based / one-based problem */ | ||
174 | int nfields_printed = 0; | ||
175 | char *field = NULL; | ||
176 | char d[2] = { delim, 0 }; | ||
177 | char *printed; | ||
178 | |||
179 | /* test the easy case first: does this line contain any delimiters? */ | ||
180 | if (strchr(line, delim) == NULL) { | ||
181 | if (!supress_non_delimited_lines) | ||
182 | puts(line); | ||
183 | return; | ||
184 | } | ||
185 | |||
186 | /* set up a list so we can keep track of what's been printed */ | ||
187 | printed = xcalloc(strlen(line), sizeof(char)); | ||
188 | |||
189 | /* process each list on this line, for as long as we've got a line to process */ | ||
190 | for (c = 0; c < nlists && line; c++) { | ||
191 | f = cut_lists[c].startpos; | ||
192 | do { | ||
193 | |||
194 | /* find the field we're looking for */ | ||
195 | while (line && ndelim < f) { | ||
196 | field = strsep(&line, d); | ||
197 | ndelim++; | ||
198 | } | ||
199 | |||
200 | /* we found it, and it hasn't been printed yet */ | ||
201 | if (field && ndelim == f && !printed[ndelim]) { | ||
202 | /* if this isn't our first time through, we need to print the | ||
203 | * delimiter after the last field that was printed */ | ||
204 | if (nfields_printed > 0) | ||
205 | putchar(delim); | ||
206 | fputs(field, stdout); | ||
207 | printed[ndelim] = 'X'; | ||
208 | nfields_printed++; | ||
209 | } | ||
210 | |||
211 | f++; | ||
212 | |||
213 | /* keep going as long as we have a line to work with, this is a | ||
214 | * list, and we're not at the end of that list */ | ||
215 | } while (line && cut_lists[c].endpos != NON_RANGE && f <= cut_lists[c].endpos); | ||
216 | } | ||
217 | |||
218 | /* if we printed anything at all, we need to finish it with a newline cuz | ||
219 | * we were handed a chomped line */ | ||
220 | putchar('\n'); | ||
221 | |||
222 | free(printed); | ||
223 | } | ||
224 | |||
225 | |||
226 | static void cut_file_by_lines(const char *line, unsigned int linenum) | ||
227 | { | ||
228 | static int c = 0; | ||
229 | static int l = -1; | ||
230 | |||
231 | /* I can't initialize this above cuz the "initializer isn't | ||
232 | * constant" *sigh* */ | ||
233 | if (l == -1) | ||
234 | l = cut_lists[c].startpos; | ||
235 | |||
236 | /* get out if we have no more lists to process or if the lines are lower | ||
237 | * than what we're interested in */ | ||
238 | if (c >= nlists || linenum < l) | ||
239 | return; | ||
240 | |||
241 | /* if the line we're looking for is lower than the one we were passed, it | ||
242 | * means we displayed it already, so move on */ | ||
243 | while (l < linenum) { | ||
244 | l++; | ||
245 | /* move on to the next list if we're at the end of this one */ | ||
246 | if (cut_lists[c].endpos == NON_RANGE || l > cut_lists[c].endpos) { | ||
247 | c++; | ||
248 | /* get out if there's no more lists to process */ | ||
249 | if (c >= nlists) | ||
250 | return; | ||
251 | l = cut_lists[c].startpos; | ||
252 | /* get out if the current line is lower than the one we just became | ||
253 | * interested in */ | ||
254 | if (linenum < l) | ||
255 | return; | ||
256 | } | ||
257 | } | ||
258 | |||
259 | /* If we made it here, it means we've found the line we're looking for, so print it */ | ||
260 | puts(line); | ||
261 | } | ||
262 | |||
263 | |||
264 | /* | ||
265 | * snippy-snip | ||
266 | */ | ||
267 | static void cut_file(FILE *file) | ||
268 | { | ||
269 | char *line = NULL; | ||
270 | unsigned int linenum = 0; /* keep these zero-based to be consistent */ | ||
271 | |||
272 | /* go through every line in the file */ | ||
273 | while ((line = bb_get_chomped_line_from_file(file)) != NULL) { | ||
274 | |||
275 | /* cut based on chars/bytes XXX: only works when sizeof(char) == byte */ | ||
276 | if ((part & (OPT_CHAR_FLGS | OPT_BYTE_FLGS))) | ||
277 | cut_line_by_chars(line); | ||
278 | |||
279 | /* cut based on fields */ | ||
280 | else { | ||
281 | if (delim == '\n') | ||
282 | cut_file_by_lines(line, linenum); | ||
283 | else | ||
284 | cut_line_by_fields(line); | ||
285 | } | ||
286 | |||
287 | linenum++; | ||
288 | free(line); | ||
289 | } | ||
290 | } | ||
291 | |||
292 | |||
293 | extern int cut_main(int argc, char **argv) | ||
294 | { | ||
295 | unsigned long opt; | ||
296 | char *sopt, *sdopt; | ||
297 | |||
298 | bb_opt_complementaly = "b~bcf:c~bcf:f~bcf"; | ||
299 | opt = bb_getopt_ulflags(argc, argv, optstring, &sopt, &sopt, &sopt, &sdopt); | ||
300 | part = opt & (OPT_BYTE_FLGS|OPT_CHAR_FLGS|OPT_FIELDS_FLGS); | ||
301 | if(part == 0) | ||
302 | bb_error_msg_and_die("you must specify a list of bytes, characters, or fields"); | ||
303 | if(opt & 0x80000000UL) | ||
304 | bb_error_msg_and_die("only one type of list may be specified"); | ||
305 | parse_lists(sopt); | ||
306 | if((opt & (OPT_DELIM_FLGS))) { | ||
307 | if (strlen(sdopt) > 1) { | ||
308 | bb_error_msg_and_die("the delimiter must be a single character"); | ||
309 | } | ||
310 | delim = sdopt[0]; | ||
311 | } | ||
312 | supress_non_delimited_lines = opt & OPT_SUPRESS_FLGS; | ||
313 | |||
314 | /* non-field (char or byte) cutting has some special handling */ | ||
315 | if (part != OPT_FIELDS_FLGS) { | ||
316 | if (supress_non_delimited_lines) { | ||
317 | bb_error_msg_and_die("suppressing non-delimited lines makes sense" | ||
318 | " only when operating on fields"); | ||
319 | } | ||
320 | if (delim != '\t') { | ||
321 | bb_error_msg_and_die("a delimiter may be specified only when operating on fields"); | ||
322 | } | ||
323 | } | ||
324 | |||
325 | /* argv[(optind)..(argc-1)] should be names of file to process. If no | ||
326 | * files were specified or '-' was specified, take input from stdin. | ||
327 | * Otherwise, we process all the files specified. */ | ||
328 | if (argv[optind] == NULL || (strcmp(argv[optind], "-") == 0)) { | ||
329 | cut_file(stdin); | ||
330 | } | ||
331 | else { | ||
332 | int i; | ||
333 | FILE *file; | ||
334 | for (i = optind; i < argc; i++) { | ||
335 | file = bb_wfopen(argv[i], "r"); | ||
336 | if(file) { | ||
337 | cut_file(file); | ||
338 | fclose(file); | ||
339 | } | ||
340 | } | ||
341 | } | ||
342 | |||
343 | return EXIT_SUCCESS; | ||
344 | } | ||