Annotation of src/usr.bin/diff3/diff3prog.c, Revision 1.6
1.6 ! millert 1: /* $OpenBSD: diff3prog.c,v 1.5 2004/01/07 18:16:42 canacar Exp $ */
1.1 millert 2:
3: /*
4: * Copyright (C) Caldera International Inc. 2001-2002.
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 and documentation must retain the above
11: * copyright 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: * 3. All advertising materials mentioning features or use of this software
16: * must display the following acknowledgement:
17: * This product includes software developed or owned by Caldera
18: * International, Inc.
19: * 4. Neither the name of Caldera International, Inc. nor the names of other
20: * contributors may be used to endorse or promote products derived from
21: * this software without specific prior written permission.
22: *
23: * USE OF THE SOFTWARE PROVIDED FOR UNDER THIS LICENSE BY CALDERA
24: * INTERNATIONAL, INC. AND CONTRIBUTORS ``AS IS'' AND ANY EXPRESS OR
25: * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
26: * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
27: * IN NO EVENT SHALL CALDERA INTERNATIONAL, INC. BE LIABLE FOR ANY DIRECT,
28: * INDIRECT INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
29: * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
30: * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
31: * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
32: * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
33: * IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
34: * POSSIBILITY OF SUCH DAMAGE.
35: */
36: /*-
37: * Copyright (c) 1991, 1993
38: * The Regents of the University of California. All rights reserved.
39: *
40: * Redistribution and use in source and binary forms, with or without
41: * modification, are permitted provided that the following conditions
42: * are met:
43: * 1. Redistributions of source code must retain the above copyright
44: * notice, this list of conditions and the following disclaimer.
45: * 2. Redistributions in binary form must reproduce the above copyright
46: * notice, this list of conditions and the following disclaimer in the
47: * documentation and/or other materials provided with the distribution.
48: * 3. Neither the name of the University nor the names of its contributors
49: * may be used to endorse or promote products derived from this software
50: * without specific prior written permission.
51: *
52: * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
53: * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
54: * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
55: * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
56: * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
57: * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
58: * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
59: * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
60: * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
61: * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
62: * SUCH DAMAGE.
63: *
64: * @(#)diff3.c 8.1 (Berkeley) 6/6/93
65: */
66:
67: #ifndef lint
68: static const char copyright[] =
69: "@(#) Copyright (c) 1991, 1993\n\
70: The Regents of the University of California. All rights reserved.\n";
71: #endif /* not lint */
72:
73: #ifndef lint
1.6 ! millert 74: static const char rcsid[] = "$OpenBSD: diff3prog.c,v 1.5 2004/01/07 18:16:42 canacar Exp $";
1.1 millert 75: #endif /* not lint */
76:
77: #include <stdio.h>
78: #include <stdlib.h>
1.6 ! millert 79: #include <string.h>
1.1 millert 80: #include <ctype.h>
81: #include <err.h>
82:
83: /* diff3 - 3-way differential file comparison */
84:
85: /* diff3 [-ex3EX] d13 d23 f1 f2 f3 [m1 m3]
86: *
87: * d13 = diff report on f1 vs f3
88: * d23 = diff report on f2 vs f3
89: * f1, f2, f3 the 3 files
90: * if changes in f1 overlap with changes in f3, m1 and m3 are used
91: * to mark the overlaps; otherwise, the file names f1 and f3 are used
92: * (only for options E and X).
93: */
94:
95: /*
96: * "from" is first in range of changed lines; "to" is last+1
97: * from=to=line after point of insertion for added lines.
98: */
99: struct range {
100: int from;
101: int to;
102: };
103: struct diff {
104: struct range old;
105: struct range new;
106: };
107:
108: #define NC 200
109: struct diff d13[NC];
110: struct diff d23[NC];
111: /*
112: * "de" is used to gather editing scripts. These are later spewed out in
113: * reverse order. Its first element must be all zero, the "new" component
114: * of "de" contains line positions or byte positions depending on when you
115: * look (!?). Array overlap indicates which sections in "de" correspond to
116: * lines that are different in all three files.
117: */
118: struct diff de[NC];
119: char overlap[NC];
120: int overlapcnt;
121: FILE *fp[3];
122: int cline[3]; /* # of the last-read line in each file (0-2) */
123: /*
124: * the latest known correspondence between line numbers of the 3 files
125: * is stored in last[1-3];
126: */
127: int last[4];
128: int eflag;
1.2 deraadt 129: int oflag; /* indicates whether to mark overlaps (-E or -X)*/
1.1 millert 130: int debug = 0;
131: char f1mark[40], f3mark[40]; /* markers for -E and -X */
132:
133: int duplicate(struct range *, struct range *);
134: int edit(struct diff *, int, int);
1.6 ! millert 135: char *getchange(FILE *);
! 136: char *getline(FILE *, size_t *);
1.1 millert 137: int number(char **);
138: int readin(char *, struct diff *);
139: int skip(int, int, char *);
140: void change(int, struct range *, int);
141: void keep(int, struct range *);
142: void merge(int, int);
143: void prange(struct range *);
144: void repos(int);
145: void separate(const char *);
146: __dead void edscript(int);
147: __dead void trouble(void);
148: __dead void usage(void);
149:
150: int
151: main(int argc, char **argv)
152: {
153: int ch, i, m, n;
154:
1.2 deraadt 155: eflag = 0;
1.1 millert 156: oflag = 0;
157: while ((ch = getopt(argc, argv, "EeXx3")) != -1) {
158: switch (ch) {
1.2 deraadt 159: case 'E':
160: eflag = 3;
161: oflag = 1;
162: break;
1.1 millert 163: case 'e':
164: eflag = 3;
165: break;
1.2 deraadt 166: case 'X':
167: oflag = eflag = 1;
168: break;
1.1 millert 169: case 'x':
170: eflag = 1;
1.2 deraadt 171: break;
1.1 millert 172: case '3':
173: eflag = 2;
174: break;
175: }
176: }
177: argc -= optind;
178: argv += optind;
179: /* XXX - argc usage seems wrong here */
180: if (argc < 5)
181: usage();
182:
1.2 deraadt 183: if (oflag) {
184: (void)snprintf(f1mark, sizeof(f1mark), "<<<<<<< %s",
1.1 millert 185: argc >= 6 ? argv[5] : argv[2]);
1.2 deraadt 186: (void)snprintf(f3mark, sizeof(f3mark), ">>>>>>> %s",
1.1 millert 187: argc >= 7 ? argv[6] : argv[4]);
1.2 deraadt 188: }
1.1 millert 189:
190: m = readin(argv[0], d13);
191: n = readin(argv[1], d23);
192: for (i = 0; i <= 2; i++) {
193: if ((fp[i] = fopen(argv[i + 2], "r")) == NULL) {
194: printf("diff3: can't open %s\n", argv[i + 2]);
195: exit(EXIT_FAILURE);
196: }
197: }
198: merge(m, n);
199: exit(EXIT_SUCCESS);
200: }
201:
202: /*
1.4 otto 203: * Pick up the line numbers of all changes from one change file.
1.1 millert 204: * (This puts the numbers in a vector, which is not strictly necessary,
205: * since the vector is processed in one sequential pass.
206: * The vector could be optimized out of existence)
207: */
208: int
209: readin(char *name, struct diff *dd)
210: {
1.6 ! millert 211: int a, b, c, d, i;
! 212: char kind, *p;
! 213:
1.1 millert 214: fp[0] = fopen(name, "r");
1.6 ! millert 215: for (i=0; (p = getchange(fp[0])); i++) {
1.1 millert 216: if (i >= NC)
217: err(EXIT_FAILURE, "too many changes");
218: a = b = number(&p);
219: if (*p == ',') {
220: p++;
221: b = number(&p);
222: }
223: kind = *p++;
224: c = d = number(&p);
225: if (*p==',') {
226: p++;
227: d = number(&p);
228: }
229: if (kind == 'a')
230: a++;
231: if (kind == 'd')
232: c++;
233: b++;
234: d++;
235: dd[i].old.from = a;
236: dd[i].old.to = b;
237: dd[i].new.from = c;
238: dd[i].new.to = d;
239: }
240: dd[i].old.from = dd[i-1].old.to;
241: dd[i].new.from = dd[i-1].new.to;
242: (void)fclose(fp[0]);
243: return (i);
244: }
245:
246: int
247: number(char **lc)
248: {
249: int nn;
250: nn = 0;
251: while (isdigit((unsigned char)(**lc)))
252: nn = nn*10 + *(*lc)++ - '0';
253: return (nn);
254: }
255:
1.6 ! millert 256: char *
1.1 millert 257: getchange(FILE *b)
258: {
1.6 ! millert 259: char *line;
! 260:
! 261: while ((line = getline(b, NULL))) {
1.1 millert 262: if (isdigit((unsigned char)line[0]))
1.6 ! millert 263: return (line);
1.1 millert 264: }
1.6 ! millert 265: return (NULL);
1.1 millert 266: }
267:
1.6 ! millert 268: char *
! 269: getline(FILE *b, size_t *n)
1.1 millert 270: {
1.6 ! millert 271: char *cp;
! 272: size_t len;
! 273: static char *buf;
! 274: static size_t bufsize;
! 275:
! 276: if ((cp = fgetln(b, &len)) == NULL)
! 277: return (NULL);
! 278:
! 279: if (cp[len - 1] != '\n')
! 280: len++;
! 281: if (len + 1 > bufsize) {
! 282: do {
! 283: bufsize += 1024;
! 284: } while (len + 1 > bufsize);
! 285: if ((buf = realloc(buf, bufsize)) == NULL)
! 286: err(EXIT_FAILURE, NULL);
1.1 millert 287: }
1.6 ! millert 288: memcpy(buf, cp, len - 1);
! 289: buf[len - 1] = '\n';
! 290: buf[len] = '\0';
! 291: if (n != NULL)
! 292: *n = len;
! 293: return (buf);
1.1 millert 294: }
295:
296: void
297: merge(int m1, int m2)
298: {
299: struct diff *d1, *d2, *d3;
300: int dup, j, t1, t2;
301:
302: d1 = d13;
303: d2 = d23;
304: j = 0;
305: while ((t1 = d1 < d13 + m1) | (t2 = d2 < d23 + m2)) {
306: if (debug) {
307: printf("%d,%d=%d,%d %d,%d=%d,%d\n",
308: d1->old.from,d1->old.to,
309: d1->new.from,d1->new.to,
310: d2->old.from,d2->old.to,
311: d2->new.from,d2->new.to);
312: }
313: /* first file is different from others */
1.5 canacar 314: if (!t2 || (t1 && d1->new.to < d2->new.from)) {
1.1 millert 315: /* stuff peculiar to 1st file */
316: if (eflag==0) {
317: separate("1");
318: change(1, &d1->old, 0);
319: keep(2, &d1->new);
320: change(3, &d1->new, 0);
321: }
322: d1++;
323: continue;
324: }
325: /* second file is different from others */
1.5 canacar 326: if (!t1 || (t2 && d2->new.to < d1->new.from)) {
1.1 millert 327: if (eflag==0) {
328: separate("2");
329: keep(1, &d2->new);
330: change(2, &d2->old, 0);
331: change(3, &d2->new, 0);
332: }
333: d2++;
334: continue;
335: }
336: /*
337: * Merge overlapping changes in first file
338: * this happens after extension (see below).
339: */
340: if (d1 + 1 < d13 + m1 && d1->new.to >= d1[1].new.from) {
341: d1[1].old.from = d1->old.from;
342: d1[1].new.from = d1->new.from;
343: d1++;
344: continue;
345: }
346:
347: /* merge overlapping changes in second */
348: if (d2 + 1 < d23 + m2 && d2->new.to >= d2[1].new.from) {
349: d2[1].old.from = d2->old.from;
350: d2[1].new.from = d2->new.from;
351: d2++;
352: continue;
353: }
354: /* stuff peculiar to third file or different in all */
355: if (d1->new.from == d2->new.from && d1->new.to == d2->new.to) {
356: dup = duplicate(&d1->old,&d2->old);
357: /*
358: * dup = 0 means all files differ
1.4 otto 359: * dup = 1 means files 1 and 2 identical
1.1 millert 360: */
361: if (eflag==0) {
362: separate(dup ? "3" : "");
363: change(1, &d1->old, dup);
364: change(2, &d2->old, 0);
365: d3 = d1->old.to > d1->old.from ? d1 : d2;
366: change(3, &d3->new, 0);
367: } else
368: j = edit(d1, dup, j);
369: d1++;
370: d2++;
371: continue;
372: }
373: /*
374: * Overlapping changes from file 1 and 2; extend changes
375: * appropriately to make them coincide.
376: */
1.2 deraadt 377: if (d1->new.from < d2->new.from) {
1.1 millert 378: d2->old.from -= d2->new.from-d1->new.from;
379: d2->new.from = d1->new.from;
380: } else if (d2->new.from < d1->new.from) {
381: d1->old.from -= d1->new.from-d2->new.from;
382: d1->new.from = d2->new.from;
383: }
384: if (d1->new.to > d2->new.to) {
385: d2->old.to += d1->new.to - d2->new.to;
386: d2->new.to = d1->new.to;
387: } else if (d2->new.to > d1->new.to) {
388: d1->old.to += d2->new.to - d1->new.to;
389: d1->new.to = d2->new.to;
390: }
391: }
392: if (eflag)
393: edscript(j);
394: }
395:
396: void
397: separate(const char *s)
398: {
399: printf("====%s\n", s);
400: }
401:
402: /*
403: * The range of lines rold.from thru rold.to in file i is to be changed.
404: * It is to be printed only if it does not duplicate something to be
405: * printed later.
406: */
407: void
408: change(int i, struct range *rold, int dup)
409: {
410: printf("%d:", i);
411: last[i] = rold->to;
412: prange(rold);
413: if (dup || debug)
414: return;
415: i--;
416: (void)skip(i, rold->from, NULL);
417: (void)skip(i, rold->to, " ");
418: }
419:
420: /*
421: * print the range of line numbers, rold.from thru rold.to, as n1,n2 or n1
422: */
423: void
424: prange(struct range *rold)
425: {
426: if (rold->to <= rold->from)
427: printf("%da\n", rold->from - 1);
428: else {
429: printf("%d", rold->from);
430: if (rold->to > rold->from+1)
431: printf(",%d", rold->to - 1);
432: printf("c\n");
433: }
434: }
435:
436: /*
437: * No difference was reported by diff between file 1 (or 2) and file 3,
438: * and an artificial dummy difference (trange) must be ginned up to
439: * correspond to the change reported in the other file.
440: */
441: void
442: keep(int i, struct range *rnew)
443: {
444: int delta;
445: struct range trange;
446:
447: delta = last[3] - last[i];
448: trange.from = rnew->from - delta;
449: trange.to = rnew->to - delta;
450: change(i, &trange, 1);
451: }
452:
453: /*
1.4 otto 454: * skip to just before line number from in file "i". If "pr" is non-NULL,
1.1 millert 455: * print all skipped stuff with string pr as a prefix.
456: */
457: int
458: skip(int i, int from, char *pr)
459: {
1.6 ! millert 460: size_t j, n;
! 461: char *line;
1.1 millert 462:
463: for (n = 0; cline[i] < from - 1; n += j) {
1.6 ! millert 464: if ((line = getline(fp[i], &j)) == NULL)
1.1 millert 465: trouble();
466: if (pr != NULL)
467: printf("%s%s", pr, line);
468: cline[i]++;
469: }
1.6 ! millert 470: return ((int) n);
1.1 millert 471: }
472:
473: /*
474: * Return 1 or 0 according as the old range (in file 1) contains exactly
475: * the same data as the new range (in file 2).
476: */
477: int
478: duplicate(struct range *r1, struct range *r2)
479: {
480: int c,d;
481: int nchar;
482: int nline;
483:
484: if (r1->to-r1->from != r2->to-r2->from)
485: return (0);
486: (void)skip(0, r1->from, NULL);
487: (void)skip(1, r2->from, NULL);
488: nchar = 0;
489: for (nline=0; nline < r1->to - r1->from; nline++) {
490: do {
491: c = getc(fp[0]);
492: d = getc(fp[1]);
493: if (c == -1 || d== -1)
494: trouble();
495: nchar++;
496: if (c != d) {
497: repos(nchar);
498: return (0);
499: }
500: } while (c != '\n');
501: }
502: repos(nchar);
503: return (1);
504: }
505:
506: void
507: repos(int nchar)
508: {
509: int i;
510:
1.2 deraadt 511: for (i = 0; i < 2; i++)
1.1 millert 512: (void)fseek(fp[i], (long)-nchar, 1);
513: }
514:
515: __dead void
516: trouble(void)
517: {
518: errx(EXIT_FAILURE, "logic error");
519: }
520:
521: /*
522: * collect an editing script for later regurgitation
523: */
524: int
525: edit(struct diff *diff, int dup, int j)
526: {
527: if (((dup + 1) & eflag) == 0)
528: return (j);
529: j++;
1.2 deraadt 530: overlap[j] = !dup;
531: if (!dup)
1.1 millert 532: overlapcnt++;
533: de[j].old.from = diff->old.from;
534: de[j].old.to = diff->old.to;
535: de[j].new.from = de[j-1].new.to + skip(2, diff->new.from, NULL);
536: de[j].new.to = de[j].new.from + skip(2, diff->new.to, NULL);
537: return (j);
538: }
539:
540: /* regurgitate */
541: __dead void
542: edscript(int n)
543: {
544: int j,k;
545: char block[BUFSIZ];
546:
547: for (n = n; n > 0; n--) {
1.2 deraadt 548: if (!oflag || !overlap[n])
549: prange(&de[n].old);
550: else
551: printf("%da\n=======\n", de[n].old.to -1);
1.1 millert 552: (void)fseek(fp[2], (long)de[n].new.from, 0);
553: for (k = de[n].new.to-de[n].new.from; k > 0; k-= j) {
554: j = k > BUFSIZ ? BUFSIZ : k;
555: if (fread(block, 1, j, fp[2]) != j)
556: trouble();
557: (void)fwrite(block, 1, j, stdout);
558: }
1.2 deraadt 559: if (!oflag || !overlap[n])
560: printf(".\n");
561: else {
562: printf("%s\n.\n", f3mark);
563: printf("%da\n%s\n.\n", de[n].old.from - 1, f1mark);
564: }
1.1 millert 565: }
1.2 deraadt 566: exit(overlapcnt);
1.1 millert 567: }
568:
569: __dead void
570: usage(void)
571: {
572: extern char *__progname;
573:
574: fprintf(stderr, "usage: %s [-exEX3] /tmp/d3a.?????????? "
575: "/tmp/d3b.?????????? file1 file2 file3\n", __progname);
576: exit(EXIT_FAILURE);
577: }