]>
Commit | Line | Data |
---|---|---|
1 | /* | |
2 | * "$Id$" | |
3 | * | |
4 | * Search routines for CUPS. | |
5 | * | |
6 | * Copyright 2007-2014 by Apple Inc. | |
7 | * Copyright 1997-2006 by Easy Software Products. | |
8 | * | |
9 | * These coded instructions, statements, and computer programs are the | |
10 | * property of Apple Inc. and are protected by Federal copyright | |
11 | * law. Distribution and use rights are outlined in the file "LICENSE.txt" | |
12 | * which should have been included with this file. If this file is | |
13 | * file is missing or damaged, see the license at "http://www.cups.org/". | |
14 | */ | |
15 | ||
16 | /* | |
17 | * Include necessary headers... | |
18 | */ | |
19 | ||
20 | #include "cgi-private.h" | |
21 | #include <regex.h> | |
22 | ||
23 | ||
24 | /* | |
25 | * 'cgiCompileSearch()' - Compile a search string. | |
26 | */ | |
27 | ||
28 | void * /* O - Search context */ | |
29 | cgiCompileSearch(const char *query) /* I - Query string */ | |
30 | { | |
31 | regex_t *re; /* Regular expression */ | |
32 | char *s, /* Regular expression string */ | |
33 | *sptr, /* Pointer into RE string */ | |
34 | *sword; /* Pointer to start of word */ | |
35 | size_t slen; /* Allocated size of RE string */ | |
36 | const char *qptr, /* Pointer into query string */ | |
37 | *qend; /* End of current word */ | |
38 | const char *prefix; /* Prefix to add to next word */ | |
39 | int quoted; /* Word is quoted */ | |
40 | size_t wlen; /* Word length */ | |
41 | char *lword; /* Last word in query */ | |
42 | ||
43 | ||
44 | DEBUG_printf(("cgiCompileSearch(query=\"%s\")\n", query)); | |
45 | ||
46 | /* | |
47 | * Range check input... | |
48 | */ | |
49 | ||
50 | if (!query) | |
51 | return (NULL); | |
52 | ||
53 | /* | |
54 | * Allocate a regular expression storage structure... | |
55 | */ | |
56 | ||
57 | if ((re = (regex_t *)calloc(1, sizeof(regex_t))) == NULL) | |
58 | return (NULL); | |
59 | ||
60 | /* | |
61 | * Allocate a buffer to hold the regular expression string, starting | |
62 | * at 1024 bytes or 3 times the length of the query string, whichever | |
63 | * is greater. We'll expand the string as needed... | |
64 | */ | |
65 | ||
66 | slen = strlen(query) * 3; | |
67 | if (slen < 1024) | |
68 | slen = 1024; | |
69 | ||
70 | if ((s = (char *)malloc(slen)) == NULL) | |
71 | { | |
72 | free(re); | |
73 | return (NULL); | |
74 | } | |
75 | ||
76 | /* | |
77 | * Copy the query string to the regular expression, handling basic | |
78 | * AND and OR logic... | |
79 | */ | |
80 | ||
81 | prefix = ".*"; | |
82 | qptr = query; | |
83 | sptr = s; | |
84 | lword = NULL; | |
85 | ||
86 | while (*qptr) | |
87 | { | |
88 | /* | |
89 | * Skip leading whitespace... | |
90 | */ | |
91 | ||
92 | while (isspace(*qptr & 255)) | |
93 | qptr ++; | |
94 | ||
95 | if (!*qptr) | |
96 | break; | |
97 | ||
98 | /* | |
99 | * Find the end of the current word... | |
100 | */ | |
101 | ||
102 | if (*qptr == '\"' || *qptr == '\'') | |
103 | { | |
104 | /* | |
105 | * Scan quoted string... | |
106 | */ | |
107 | ||
108 | quoted = *qptr ++; | |
109 | for (qend = qptr; *qend && *qend != quoted; qend ++); | |
110 | ||
111 | if (!*qend) | |
112 | { | |
113 | /* | |
114 | * No closing quote, error out! | |
115 | */ | |
116 | ||
117 | free(s); | |
118 | free(re); | |
119 | ||
120 | if (lword) | |
121 | free(lword); | |
122 | ||
123 | return (NULL); | |
124 | } | |
125 | } | |
126 | else | |
127 | { | |
128 | /* | |
129 | * Scan whitespace-delimited string... | |
130 | */ | |
131 | ||
132 | quoted = 0; | |
133 | for (qend = qptr + 1; *qend && !isspace(*qend); qend ++); | |
134 | } | |
135 | ||
136 | wlen = (size_t)(qend - qptr); | |
137 | ||
138 | /* | |
139 | * Look for logic words: AND, OR | |
140 | */ | |
141 | ||
142 | if (wlen == 3 && !_cups_strncasecmp(qptr, "AND", 3)) | |
143 | { | |
144 | /* | |
145 | * Logical AND with the following text... | |
146 | */ | |
147 | ||
148 | if (sptr > s) | |
149 | prefix = ".*"; | |
150 | ||
151 | qptr = qend; | |
152 | } | |
153 | else if (wlen == 2 && !_cups_strncasecmp(qptr, "OR", 2)) | |
154 | { | |
155 | /* | |
156 | * Logical OR with the following text... | |
157 | */ | |
158 | ||
159 | if (sptr > s) | |
160 | prefix = ".*|.*"; | |
161 | ||
162 | qptr = qend; | |
163 | } | |
164 | else | |
165 | { | |
166 | /* | |
167 | * Add a search word, making sure we have enough room for the | |
168 | * string + RE overhead... | |
169 | */ | |
170 | ||
171 | wlen = (size_t)(sptr - s) + 2 * 4 * wlen + 2 * strlen(prefix) + 11; | |
172 | if (lword) | |
173 | wlen += strlen(lword); | |
174 | ||
175 | if (wlen > slen) | |
176 | { | |
177 | /* | |
178 | * Expand the RE string buffer... | |
179 | */ | |
180 | ||
181 | char *temp; /* Temporary string pointer */ | |
182 | ||
183 | ||
184 | slen = wlen + 128; | |
185 | temp = (char *)realloc(s, slen); | |
186 | if (!temp) | |
187 | { | |
188 | free(s); | |
189 | free(re); | |
190 | ||
191 | if (lword) | |
192 | free(lword); | |
193 | ||
194 | return (NULL); | |
195 | } | |
196 | ||
197 | sptr = temp + (sptr - s); | |
198 | s = temp; | |
199 | } | |
200 | ||
201 | /* | |
202 | * Add the prefix string... | |
203 | */ | |
204 | ||
205 | memcpy(sptr, prefix, strlen(prefix) + 1); | |
206 | sptr += strlen(sptr); | |
207 | ||
208 | /* | |
209 | * Then quote the remaining word characters as needed for the | |
210 | * RE... | |
211 | */ | |
212 | ||
213 | sword = sptr; | |
214 | ||
215 | while (qptr < qend) | |
216 | { | |
217 | /* | |
218 | * Quote: ^ . [ $ ( ) | * + ? { \ | |
219 | */ | |
220 | ||
221 | if (strchr("^.[$()|*+?{\\", *qptr)) | |
222 | *sptr++ = '\\'; | |
223 | ||
224 | *sptr++ = *qptr++; | |
225 | } | |
226 | ||
227 | *sptr = '\0'; | |
228 | ||
229 | /* | |
230 | * For "word1 AND word2", add reciprocal "word2 AND word1"... | |
231 | */ | |
232 | ||
233 | if (!strcmp(prefix, ".*") && lword) | |
234 | { | |
235 | char *lword2; /* New "last word" */ | |
236 | ||
237 | ||
238 | if ((lword2 = strdup(sword)) == NULL) | |
239 | { | |
240 | free(lword); | |
241 | free(s); | |
242 | free(re); | |
243 | return (NULL); | |
244 | } | |
245 | ||
246 | memcpy(sptr, ".*|.*", 6); | |
247 | sptr += 5; | |
248 | ||
249 | memcpy(sptr, lword2, strlen(lword2) + 1); | |
250 | sptr += strlen(sptr); | |
251 | ||
252 | memcpy(sptr, ".*", 3); | |
253 | sptr += 2; | |
254 | ||
255 | memcpy(sptr, lword, strlen(lword) + 1); | |
256 | sptr += strlen(sptr); | |
257 | ||
258 | free(lword); | |
259 | lword = lword2; | |
260 | } | |
261 | else | |
262 | { | |
263 | if (lword) | |
264 | free(lword); | |
265 | ||
266 | lword = strdup(sword); | |
267 | } | |
268 | ||
269 | prefix = ".*|.*"; | |
270 | } | |
271 | ||
272 | /* | |
273 | * Advance to the next string... | |
274 | */ | |
275 | ||
276 | if (quoted) | |
277 | qptr ++; | |
278 | } | |
279 | ||
280 | if (lword) | |
281 | free(lword); | |
282 | ||
283 | if (sptr > s) | |
284 | memcpy(sptr, ".*", 3); | |
285 | else | |
286 | { | |
287 | /* | |
288 | * No query data, return NULL... | |
289 | */ | |
290 | ||
291 | free(s); | |
292 | free(re); | |
293 | ||
294 | return (NULL); | |
295 | } | |
296 | ||
297 | /* | |
298 | * Compile the regular expression... | |
299 | */ | |
300 | ||
301 | DEBUG_printf((" s=\"%s\"\n", s)); | |
302 | ||
303 | if (regcomp(re, s, REG_EXTENDED | REG_ICASE)) | |
304 | { | |
305 | free(re); | |
306 | free(s); | |
307 | ||
308 | return (NULL); | |
309 | } | |
310 | ||
311 | /* | |
312 | * Free the RE string and return the new regular expression we compiled... | |
313 | */ | |
314 | ||
315 | free(s); | |
316 | ||
317 | return ((void *)re); | |
318 | } | |
319 | ||
320 | ||
321 | /* | |
322 | * 'cgiDoSearch()' - Do a search of some text. | |
323 | */ | |
324 | ||
325 | int /* O - Number of matches */ | |
326 | cgiDoSearch(void *search, /* I - Search context */ | |
327 | const char *text) /* I - Text to search */ | |
328 | { | |
329 | int i; /* Looping var */ | |
330 | regmatch_t matches[100]; /* RE matches */ | |
331 | ||
332 | ||
333 | /* | |
334 | * Range check... | |
335 | */ | |
336 | ||
337 | if (!search || !text) | |
338 | return (0); | |
339 | ||
340 | /* | |
341 | * Do a lookup... | |
342 | */ | |
343 | ||
344 | if (!regexec((regex_t *)search, text, sizeof(matches) / sizeof(matches[0]), | |
345 | matches, 0)) | |
346 | { | |
347 | /* | |
348 | * Figure out the number of matches in the string... | |
349 | */ | |
350 | ||
351 | for (i = 0; i < (int)(sizeof(matches) / sizeof(matches[0])); i ++) | |
352 | if (matches[i].rm_so < 0) | |
353 | break; | |
354 | ||
355 | return (i); | |
356 | } | |
357 | else | |
358 | return (0); | |
359 | } | |
360 | ||
361 | ||
362 | /* | |
363 | * 'cgiFreeSearch()' - Free a compiled search context. | |
364 | */ | |
365 | ||
366 | void | |
367 | cgiFreeSearch(void *search) /* I - Search context */ | |
368 | { | |
369 | regfree((regex_t *)search); | |
370 | } | |
371 | ||
372 | ||
373 | /* | |
374 | * End of "$Id$". | |
375 | */ |