Annotation of src/usr.bin/grep/util.c, Revision 1.14
1.14 ! millert 1: /* $OpenBSD: util.c,v 1.13 2003/07/06 19:16:45 avsm Exp $ */
1.3 deraadt 2:
1.1 deraadt 3: /*-
4: * Copyright (c) 1999 James Howard and Dag-Erling Coïdan Smørgrav
5: * All rights reserved.
6: *
7: * Redistribution and use in source and binary forms, with or without
8: * modification, are permitted provided that the following conditions
9: * are met:
10: * 1. Redistributions of source code must retain the above copyright
11: * notice, this list of conditions and the following disclaimer.
12: * 2. Redistributions in binary form must reproduce the above copyright
13: * notice, this list of conditions and the following disclaimer in the
14: * documentation and/or other materials provided with the distribution.
15: *
16: * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
17: * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18: * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19: * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
20: * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21: * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22: * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23: * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24: * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25: * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26: * SUCH DAMAGE.
27: */
28:
29: #include <sys/types.h>
30: #include <sys/stat.h>
31:
32: #include <ctype.h>
33: #include <err.h>
34: #include <errno.h>
35: #include <fts.h>
36: #include <regex.h>
37: #include <stdio.h>
38: #include <stdlib.h>
39: #include <string.h>
40: #include <unistd.h>
41: #include <zlib.h>
42:
43: #include "grep.h"
44:
45: /*
46: * Process a file line by line...
47: */
48:
49: static int linesqueued;
1.4 tedu 50: static int procline(str_t *l, int);
1.9 millert 51: static int grep_search(fastgrep_t *, unsigned char *, int, regmatch_t *pmatch);
1.6 tedu 52: static int grep_cmp(const unsigned char *, const unsigned char *, size_t);
53: static void grep_revstr(unsigned char *, int);
1.1 deraadt 54:
1.2 deraadt 55: int
1.1 deraadt 56: grep_tree(char **argv)
57: {
1.10 deraadt 58: FTS *fts;
59: FTSENT *p;
60: int c, fts_flags;
1.1 deraadt 61:
62: c = fts_flags = 0;
63:
64: if (Hflag)
65: fts_flags = FTS_COMFOLLOW;
66: if (Pflag)
67: fts_flags = FTS_PHYSICAL;
68: if (Sflag)
69: fts_flags = FTS_LOGICAL;
70:
71: fts_flags |= FTS_NOSTAT | FTS_NOCHDIR;
72:
1.11 millert 73: if (!(fts = fts_open(argv, fts_flags, NULL)))
1.14 ! millert 74: err(2, NULL);
1.1 deraadt 75: while ((p = fts_read(fts)) != NULL) {
76: switch (p->fts_info) {
77: case FTS_DNR:
78: break;
79: case FTS_ERR:
1.14 ! millert 80: errx(2, "%s: %s", p->fts_path, strerror(p->fts_errno));
1.1 deraadt 81: break;
82: case FTS_DP:
83: break;
84: default:
85: c += procfile(p->fts_path);
86: break;
87: }
88: }
89:
90: return c;
91: }
92:
93: int
94: procfile(char *fn)
95: {
96: str_t ln;
97: file_t *f;
1.4 tedu 98: int c, t, z, nottext;
1.1 deraadt 99:
100: if (fn == NULL) {
101: fn = "(standard input)";
102: f = grep_fdopen(STDIN_FILENO, "r");
103: } else {
104: f = grep_open(fn, "r");
105: }
106: if (f == NULL) {
107: if (!sflag)
108: warn("%s", fn);
109: return 0;
110: }
1.4 tedu 111:
112: nottext = grep_bin_file(f);
113: if (nottext && binbehave == BIN_FILE_SKIP) {
1.1 deraadt 114: grep_close(f);
115: return 0;
116: }
117:
118: ln.file = fn;
119: ln.line_no = 0;
120: linesqueued = 0;
121: ln.off = -1;
122:
123: if (Bflag > 0)
124: initqueue();
125: for (c = 0; !(lflag && c);) {
126: ln.off += ln.len + 1;
127: if ((ln.dat = grep_fgetln(f, &ln.len)) == NULL)
128: break;
129: if (ln.len > 0 && ln.dat[ln.len - 1] == '\n')
130: --ln.len;
131: ln.line_no++;
132:
133: z = tail;
1.2 deraadt 134:
1.4 tedu 135: if ((t = procline(&ln, nottext)) == 0 && Bflag > 0 && z == 0) {
1.1 deraadt 136: enqueue(&ln);
137: linesqueued++;
138: }
139: c += t;
140: }
141: if (Bflag > 0)
142: clearqueue();
143: grep_close(f);
144:
145: if (cflag) {
146: if (!hflag)
147: printf("%s:", ln.file);
148: printf("%u\n", c);
149: }
150: if (lflag && c != 0)
151: printf("%s\n", fn);
152: if (Lflag && c == 0)
153: printf("%s\n", fn);
1.4 tedu 154: if (c && !cflag && !lflag && !Lflag &&
1.7 tedu 155: binbehave == BIN_FILE_BIN && nottext && !qflag)
1.4 tedu 156: printf("Binary file %s matches\n", fn);
157:
1.1 deraadt 158: return c;
159: }
160:
161:
162: /*
163: * Process an individual line in a file. Return non-zero if it matches.
164: */
165:
166: #define isword(x) (isalnum(x) || (x) == '_')
167:
168: static int
1.4 tedu 169: procline(str_t *l, int nottext)
1.1 deraadt 170: {
171: regmatch_t pmatch;
172: int c, i, r, t;
173:
174: if (matchall) {
175: c = !vflag;
176: goto print;
177: }
1.2 deraadt 178:
1.1 deraadt 179: t = vflag ? REG_NOMATCH : 0;
180: for (c = i = 0; i < patterns; i++) {
1.9 millert 181: pmatch.rm_so = 0;
182: pmatch.rm_eo = l->len;
1.6 tedu 183: if (fg_pattern[i].pattern)
184: r = grep_search(&fg_pattern[i], (unsigned char *)l->dat,
1.9 millert 185: l->len, &pmatch);
1.6 tedu 186: else
1.9 millert 187: r = regexec(&r_pattern[i], l->dat, 1, &pmatch, eflags);
1.1 deraadt 188: if (r == REG_NOMATCH && t == 0)
189: continue;
190: if (r == 0) {
191: if (wflag) {
1.5 deraadt 192: if ((pmatch.rm_so != 0 &&
193: isword(l->dat[pmatch.rm_so - 1])) ||
194: (pmatch.rm_eo != l->len &&
195: isword(l->dat[pmatch.rm_eo])))
1.1 deraadt 196: r = REG_NOMATCH;
197: }
198: if (xflag) {
199: if (pmatch.rm_so != 0 || pmatch.rm_eo != l->len)
200: r = REG_NOMATCH;
201: }
202: }
203: if (r == t) {
204: c++;
205: break;
206: }
207: }
1.2 deraadt 208:
1.1 deraadt 209: print:
1.4 tedu 210: if (c && binbehave == BIN_FILE_BIN && nottext)
211: return c; /* Binary file */
212:
1.1 deraadt 213: if ((tail > 0 || c) && !cflag && !qflag) {
214: if (c) {
1.5 deraadt 215: if (first > 0 && tail == 0 && (Bflag < linesqueued) &&
216: (Aflag || Bflag))
1.1 deraadt 217: printf("--\n");
218: first = 1;
219: tail = Aflag;
220: if (Bflag > 0)
221: printqueue();
222: linesqueued = 0;
223: printline(l, ':');
224: } else {
225: printline(l, '-');
226: tail--;
227: }
228: }
229: return c;
230: }
231:
1.6 tedu 232: /*
1.10 deraadt 233: * Returns: -1 on failure, 0 on success
1.6 tedu 234: */
235: int
236: fastcomp(fastgrep_t *fg, const char *pattern)
237: {
238: int i;
239: int bol = 0;
240: int eol = 0;
241: int origPatternLen;
242: int shiftPatternLen;
243: int hasDot = 0;
244: int firstHalfDot = -1;
245: int firstLastHalfDot = -1;
246: int lastHalfDot = 0;
247:
1.8 tedu 248: if (Fflag) {
249: fg->pattern = NULL;
250: return (-1);
251: }
252:
1.6 tedu 253: /* Initialize. */
254: origPatternLen = fg->patternLen = strlen(pattern);
255: fg->bol = 0;
256: fg->eol = 0;
257: fg->reversedSearch = 0;
258:
259: /* Remove end-of-line character ('$'). */
260: if (pattern[fg->patternLen - 1] == '$') {
261: eol++;
262: fg->eol = 1;
263: fg->patternLen--;
264: boleol = 1;
265: }
266:
267: /* Remove beginning-of-line character ('^'). */
268: if (pattern[0] == '^') {
269: bol++;
270: fg->bol = 1;
271: fg->patternLen--;
272: boleol = 1;
273: }
274:
275: /*
1.8 tedu 276: * Copy pattern minus '^' and '$' characters at the beginning and
277: * ending of the string respectively.
1.6 tedu 278: */
279: fg->pattern = grep_strdup(pattern + bol);
280:
281: /* Look for ways to cheat...er...avoid the full regex engine. */
282: for (i = 0; i < fg->patternLen; i++)
283: {
284: /* Can still cheat? */
285: if ((isalnum(fg->pattern[i])) || isspace(fg->pattern[i]) ||
286: (fg->pattern[i] == '_') || (fg->pattern[i] == ',') ||
287: (fg->pattern[i] == '^') || (fg->pattern[i] == '$') ||
288: (fg->pattern[i] == '=') || (fg->pattern[i] == '-') ||
289: (fg->pattern[i] == ':') || (fg->pattern[i] == '/')) {
290: /* As long as it is good, upper case it for later. */
291: if (iflag)
292: fg->pattern[i] = toupper(fg->pattern[i]);
293: } else if (fg->pattern[i] == '.') {
294: hasDot = i;
295: if (i < fg->patternLen / 2) {
296: if (firstHalfDot < -1)
297: /* Closest dot to the beginning */
298: firstHalfDot = i;
299: } else {
300: /* Closest dot to the end of the pattern. */
301: lastHalfDot = i;
302: if (firstLastHalfDot < 0)
303: firstLastHalfDot = i;
304: }
305: } else {
306: /* Free memory and let others know this is empty. */
307: free(fg->pattern);
308: fg->pattern = NULL;
309: return (-1);
310: }
311: }
312:
313: /*
314: * Determine if a reverse search would be faster based on the placement
315: * of the dots.
316: */
317: if ((!(lflag || cflag)) && ((!(bol || eol)) &&
318: ((lastHalfDot) && ((firstHalfDot < 0) ||
319: ((fg->patternLen - (lastHalfDot + 1)) < firstHalfDot))))) {
320: fg->reversedSearch = 1;
321: hasDot = fg->patternLen - (firstHalfDot < 0 ?
322: firstLastHalfDot : firstHalfDot) - 1;
323: grep_revstr(fg->pattern, fg->patternLen);
324: }
325:
326: /*
327: * Normal Quick Search would require a shift based on the position the
328: * next character after the comparison is within the pattern. With
329: * wildcards, the position of the last dot effects the maximum shift
330: * distance.
331: * The closer to the end the wild card is the slower the search. A
1.10 deraadt 332: * reverse version of this algorithm would be useful for wildcards near
1.6 tedu 333: * the end of the string.
334: *
335: * Examples:
336: * Pattern Max shift
337: * ------- ---------
338: * this 5
339: * .his 4
340: * t.is 3
341: * th.s 2
342: * thi. 1
343: */
344:
345: /* Adjust the shift based on location of the last dot ('.'). */
346: shiftPatternLen = fg->patternLen - hasDot;
347:
348: /* Preprocess pattern. */
349: for (i = 0; i <= UCHAR_MAX; i++)
350: fg->qsBc[i] = shiftPatternLen;
351: for (i = hasDot + 1; i < fg->patternLen; i++) {
352: fg->qsBc[fg->pattern[i]] = fg->patternLen - i;
353: /*
354: * If case is ignored, make the jump apply to both upper and
355: * lower cased characters. As the pattern is stored in upper
356: * case, apply the same to the lower case equivalents.
357: */
358: if (iflag)
359: fg->qsBc[tolower(fg->pattern[i])] = fg->patternLen - i;
360: }
361:
362: /*
363: * Put pattern back to normal after pre-processing to allow for easy
364: * comparisons later.
365: */
366: if (fg->reversedSearch)
367: grep_revstr(fg->pattern, fg->patternLen);
368:
369: return (0);
370: }
371:
1.9 millert 372: static int
373: grep_search(fastgrep_t *fg, unsigned char *data, int dataLen, regmatch_t *pmatch)
1.6 tedu 374: {
375: int j;
376: int rtrnVal = REG_NOMATCH;
377:
1.9 millert 378: pmatch->rm_so = -1;
379: pmatch->rm_eo = -1;
380:
1.6 tedu 381: /* No point in going farther if we do not have enough data. */
382: if (dataLen < fg->patternLen)
383: return (rtrnVal);
384:
385: /* Only try once at the beginning or ending of the line. */
386: if (fg->bol || fg->eol) {
387: /* Simple text comparison. */
388: /* Verify data is >= pattern length before searching on it. */
389: if (dataLen >= fg->patternLen) {
390: /* Determine where in data to start search at. */
391: if (fg->eol)
392: j = dataLen - fg->patternLen;
393: else
394: j = 0;
395: if (!((fg->bol && fg->eol) && (dataLen != fg->patternLen)))
1.9 millert 396: if (grep_cmp(fg->pattern, data + j, fg->patternLen) == -1) {
1.6 tedu 397: rtrnVal = 0;
1.9 millert 398: pmatch->rm_so = j;
399: pmatch->rm_eo = j + fg->patternLen;
400: }
1.6 tedu 401: }
402: } else if (fg->reversedSearch) {
403: /* Quick Search algorithm. */
404: j = dataLen;
405: do {
406: if (grep_cmp(fg->pattern, data + j - fg->patternLen,
407: fg->patternLen) == -1) {
408: rtrnVal = 0;
1.9 millert 409: pmatch->rm_so = j - fg->patternLen;
410: pmatch->rm_eo = j;
1.6 tedu 411: break;
412: }
413:
414: /* Shift if within bounds, otherwise, we are done. */
1.12 millert 415: if (j - fg->patternLen - 1 < 0)
1.6 tedu 416: break;
1.12 millert 417: j -= fg->qsBc[data[j - fg->patternLen - 1]];
1.6 tedu 418: } while (j >= 0);
419: } else {
420: /* Quick Search algorithm. */
421: j = 0;
422: do {
423: if (grep_cmp(fg->pattern, data + j, fg->patternLen) == -1) {
424: rtrnVal = 0;
1.9 millert 425: pmatch->rm_so = j;
426: pmatch->rm_eo = j + fg->patternLen;
1.6 tedu 427: break;
428: }
429:
430: /* Shift if within bounds, otherwise, we are done. */
431: if (j + fg->patternLen == dataLen)
432: break;
433: else
434: j += fg->qsBc[data[j + fg->patternLen]];
435: } while (j <= (dataLen - fg->patternLen));
436: }
437:
438: return (rtrnVal);
439: }
440:
441:
1.1 deraadt 442: void *
443: grep_malloc(size_t size)
444: {
1.10 deraadt 445: void *ptr;
1.1 deraadt 446:
447: if ((ptr = malloc(size)) == NULL)
1.14 ! millert 448: err(2, "malloc");
1.1 deraadt 449: return ptr;
450: }
451:
452: void *
453: grep_realloc(void *ptr, size_t size)
454: {
455: if ((ptr = realloc(ptr, size)) == NULL)
1.14 ! millert 456: err(2, "realloc");
1.1 deraadt 457: return ptr;
1.6 tedu 458: }
459:
460: unsigned char *
461: grep_strdup(const char *str)
462: {
463: unsigned char *ptr;
464:
465: if ((ptr = (unsigned char *)strdup(str)) == NULL)
1.14 ! millert 466: err(2, "strdup");
1.6 tedu 467: return ptr;
468: }
469:
470: /*
471: * Returns: i >= 0 on failure (position that it failed)
472: * -1 on success
473: */
474: int
1.9 millert 475: grep_cmp(const unsigned char *pattern, const unsigned char *data, size_t len)
1.6 tedu 476: {
477: int i;
478:
479: for (i = 0; i < len; i++) {
480: if (((pattern[i] == data[i]) || (pattern[i] == '.')) ||
481: (iflag && pattern[i] == toupper(data[i])))
482: continue;
483: return (i);
484: }
485:
486: return (-1);
487: }
488:
489: static void
490: grep_revstr(unsigned char *str, int len)
491: {
492: int i;
493: char c;
494:
495: for (i = 0; i < len / 2; i++) {
496: c = str[i];
497: str[i] = str[len - i - 1];
498: str[len - i - 1] = c;
499: }
1.1 deraadt 500: }
501:
502: void
503: printline(str_t *line, int sep)
504: {
505: int n;
1.2 deraadt 506:
1.1 deraadt 507: n = 0;
508: if (!hflag) {
509: fputs(line->file, stdout);
510: ++n;
511: }
512: if (nflag) {
513: if (n)
514: putchar(sep);
515: printf("%d", line->line_no);
516: ++n;
517: }
518: if (bflag) {
519: if (n)
520: putchar(sep);
521: printf("%lu", (unsigned long)line->off);
522: }
523: if (n)
524: putchar(sep);
525: fwrite(line->dat, line->len, 1, stdout);
526: putchar('\n');
527: }