Annotation of src/usr.bin/cvs/diff.c, Revision 1.2
1.1 jfb 1: /* $OpenBSD$ */
2: /*
3: * Copyright (C) Caldera International Inc. 2001-2002.
4: * All rights reserved.
5: *
6: * Redistribution and use in source and binary forms, with or without
7: * modification, are permitted provided that the following conditions
8: * are met:
9: * 1. Redistributions of source code and documentation must retain the above
10: * copyright notice, this list of conditions and the following disclaimer.
11: * 2. Redistributions in binary form must reproduce the above copyright
12: * notice, this list of conditions and the following disclaimer in the
13: * documentation and/or other materials provided with the distribution.
14: * 3. All advertising materials mentioning features or use of this software
15: * must display the following acknowledgement:
16: * This product includes software developed or owned by Caldera
17: * International, Inc.
18: * 4. Neither the name of Caldera International, Inc. nor the names of other
19: * contributors may be used to endorse or promote products derived from
20: * this software without specific prior written permission.
21: *
22: * USE OF THE SOFTWARE PROVIDED FOR UNDER THIS LICENSE BY CALDERA
23: * INTERNATIONAL, INC. AND CONTRIBUTORS ``AS IS'' AND ANY EXPRESS OR
24: * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
25: * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
26: * IN NO EVENT SHALL CALDERA INTERNATIONAL, INC. BE LIABLE FOR ANY DIRECT,
27: * INDIRECT INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
28: * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
29: * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
30: * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
31: * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
32: * IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
33: * POSSIBILITY OF SUCH DAMAGE.
34: */
35: /*-
36: * Copyright (c) 1991, 1993
37: * The Regents of the University of California. All rights reserved.
38: * Copyright (c) 2004 Jean-Francois Brousseau. 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: * @(#)diffreg.c 8.1 (Berkeley) 6/6/93
65: */
66: /*
67: * Uses an algorithm due to Harold Stone, which finds
68: * a pair of longest identical subsequences in the two
69: * files.
70: *
71: * The major goal is to generate the match vector J.
72: * J[i] is the index of the line in file1 corresponding
73: * to line i file0. J[i] = 0 if there is no
74: * such line in file1.
75: *
76: * Lines are hashed so as to work in core. All potential
77: * matches are located by sorting the lines of each file
78: * on the hash (called ``value''). In particular, this
79: * collects the equivalence classes in file1 together.
80: * Subroutine equiv replaces the value of each line in
81: * file0 by the index of the first element of its
82: * matching equivalence in (the reordered) file1.
83: * To save space equiv squeezes file1 into a single
84: * array member in which the equivalence classes
85: * are simply concatenated, except that their first
86: * members are flagged by changing sign.
87: *
88: * Next the indices that point into member are unsorted into
89: * array class according to the original order of file0.
90: *
91: * The cleverness lies in routine stone. This marches
92: * through the lines of file0, developing a vector klist
93: * of "k-candidates". At step i a k-candidate is a matched
94: * pair of lines x,y (x in file0 y in file1) such that
95: * there is a common subsequence of length k
96: * between the first i lines of file0 and the first y
97: * lines of file1, but there is no such subsequence for
98: * any smaller y. x is the earliest possible mate to y
99: * that occurs in such a subsequence.
100: *
101: * Whenever any of the members of the equivalence class of
102: * lines in file1 matable to a line in file0 has serial number
103: * less than the y of some k-candidate, that k-candidate
104: * with the smallest such y is replaced. The new
105: * k-candidate is chained (via pred) to the current
106: * k-1 candidate so that the actual subsequence can
107: * be recovered. When a member has serial number greater
108: * that the y of all k-candidates, the klist is extended.
109: * At the end, the longest subsequence is pulled out
110: * and placed in the array J by unravel
111: *
112: * With J in hand, the matches there recorded are
113: * check'ed against reality to assure that no spurious
114: * matches have crept in due to hashing. If they have,
115: * they are broken, and "jackpot" is recorded--a harmless
116: * matter except that a true match for a spuriously
117: * mated line may now be unnecessarily reported as a change.
118: *
119: * Much of the complexity of the program comes simply
120: * from trying to minimize core utilization and
121: * maximize the range of doable problems by dynamically
122: * allocating what is needed and reusing what is not.
123: * The core requirements for problems larger than somewhat
124: * are (in words) 2*length(file0) + length(file1) +
125: * 3*(number of k-candidates installed), typically about
126: * 6n words for files of length n.
127: */
128:
129: #include <sys/param.h>
130: #include <sys/stat.h>
131: #include <sys/wait.h>
132:
133: #include <errno.h>
134: #include <ctype.h>
135: #include <stdio.h>
136: #include <fcntl.h>
137: #include <paths.h>
138: #include <regex.h>
139: #include <dirent.h>
140: #include <stdlib.h>
141: #include <stddef.h>
142: #include <unistd.h>
143: #include <string.h>
144: #include <sysexits.h>
145:
146: #include "cvs.h"
147: #include "log.h"
148: #include "buf.h"
149:
150:
151: #define CVS_DIFF_DEFCTX 3 /* default context length */
152:
153:
154: /*
155: * Output format options
156: */
157: #define D_NORMAL 0 /* Normal output */
158: #define D_CONTEXT 1 /* Diff with context */
159: #define D_UNIFIED 2 /* Unified context diff */
160: #define D_IFDEF 3 /* Diff with merged #ifdef's */
161: #define D_BRIEF 4 /* Say if the files differ */
162:
163: /*
164: * Status values for print_status() and diffreg() return values
165: */
166: #define D_SAME 0 /* Files are the same */
167: #define D_DIFFER 1 /* Files are different */
168: #define D_BINARY 2 /* Binary files are different */
169: #define D_COMMON 3 /* Subdirectory common to both dirs */
170: #define D_ONLY 4 /* Only exists in one directory */
171: #define D_MISMATCH1 5 /* path1 was a dir, path2 a file */
172: #define D_MISMATCH2 6 /* path1 was a file, path2 a dir */
173: #define D_ERROR 7 /* An error occurred */
174: #define D_SKIPPED1 8 /* path1 was a special file */
175: #define D_SKIPPED2 9 /* path2 was a special file */
176:
177: struct cand {
178: int x;
179: int y;
180: int pred;
181: } cand;
182:
183: struct line {
184: int serial;
185: int value;
186: } *file[2];
187:
188: /*
189: * The following struct is used to record change information when
190: * doing a "context" or "unified" diff. (see routine "change" to
191: * understand the highly mnemonic field names)
192: */
193: struct context_vec {
194: int a; /* start line in old file */
195: int b; /* end line in old file */
196: int c; /* start line in new file */
197: int d; /* end line in new file */
198: };
199:
200:
201: struct excludes {
202: char *pattern;
203: struct excludes *next;
204: };
205:
206:
207: char *splice(char *, char *);
208: int cvs_diffreg(const char *, const char *);
209: int cvs_diff_file (const char *, const char *, const char *);
210: int cvs_diff_dir (const char *, int);
211: static void output(const char *, FILE *, const char *, FILE *);
212: static void check(FILE *, FILE *);
213: static void range(int, int, char *);
214: static void uni_range(int, int);
215: static void dump_context_vec(FILE *, FILE *);
216: static void dump_unified_vec(FILE *, FILE *);
217: static void prepare(int, FILE *, off_t);
218: static void prune(void);
219: static void equiv(struct line *, int, struct line *, int, int *);
220: static void unravel(int);
221: static void unsort(struct line *, int, int *);
222: static void change(const char *, FILE *, const char *, FILE *, int, int, int, int);
223: static void sort(struct line *, int);
224: static int ignoreline(char *);
225: static int asciifile(FILE *);
226: static int fetch(long *, int, int, FILE *, int, int);
227: static int newcand(int, int, int);
228: static int search(int *, int, int);
229: static int skipline(FILE *);
230: static int isqrt(int);
231: static int stone(int *, int, int *, int *);
232: static int readhash(FILE *);
233: static int files_differ(FILE *, FILE *);
234: static __inline int min(int, int);
235: static __inline int max(int, int);
236: static char *match_function(const long *, int, FILE *);
237: static char *preadline(int, size_t, off_t);
238:
239:
240:
241: extern int cvs_client;
242: extern struct cvsroot *cvs_root;
243:
244:
245:
246: static int aflag, bflag, dflag, iflag, tflag, Tflag, wflag;
247: static int context, status;
248: static int format = D_NORMAL;
249: static struct stat stb1, stb2;
1.2 ! jfb 250: static char *ifdefname, *ignore_pats, diffargs[128];
! 251: static const char *diff_file;
1.1 jfb 252: regex_t ignore_re;
253:
254: static int *J; /* will be overlaid on class */
255: static int *class; /* will be overlaid on file[0] */
256: static int *klist; /* will be overlaid on file[0] after class */
257: static int *member; /* will be overlaid on file[1] */
258: static int clen;
259: static int inifdef; /* whether or not we are in a #ifdef block */
260: static int len[2];
261: static int pref, suff; /* length of prefix and suffix */
262: static int slen[2];
263: static int anychange;
264: static long *ixnew; /* will be overlaid on file[1] */
265: static long *ixold; /* will be overlaid on klist */
266: static struct cand *clist; /* merely a free storage pot for candidates */
267: static int clistlen; /* the length of clist */
268: static struct line *sfile[2]; /* shortened by pruning common prefix/suffix */
269: static u_char *chrtran; /* translation table for case-folding */
270: static struct context_vec *context_vec_start;
271: static struct context_vec *context_vec_end;
272: static struct context_vec *context_vec_ptr;
273:
274: #define FUNCTION_CONTEXT_SIZE 41
275: static char lastbuf[FUNCTION_CONTEXT_SIZE];
276: static int lastline;
277: static int lastmatchline;
278:
279:
280:
281:
282: /*
283: * chrtran points to one of 2 translation tables: cup2low if folding upper to
284: * lower case clow2low if not folding case
285: */
286: u_char clow2low[256] = {
287: 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a,
288: 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, 0x10, 0x11, 0x12, 0x13, 0x14, 0x15,
289: 0x16, 0x17, 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, 0x20,
290: 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, 0x28, 0x29, 0x2a, 0x2b,
291: 0x2c, 0x2d, 0x2e, 0x2f, 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36,
292: 0x37, 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0x3e, 0x3f, 0x40, 0x41,
293: 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, 0x48, 0x49, 0x4a, 0x4b, 0x4c,
294: 0x4d, 0x4e, 0x4f, 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57,
295: 0x58, 0x59, 0x5a, 0x5b, 0x5c, 0x5d, 0x5e, 0x5f, 0x60, 0x61, 0x62,
296: 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d,
297: 0x6e, 0x6f, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78,
298: 0x79, 0x7a, 0x7b, 0x7c, 0x7d, 0x7e, 0x7f, 0x80, 0x81, 0x82, 0x83,
299: 0x84, 0x85, 0x86, 0x87, 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x8d, 0x8e,
300: 0x8f, 0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, 0x98, 0x99,
301: 0x9a, 0x9b, 0x9c, 0x9d, 0x9e, 0x9f, 0xa0, 0xa1, 0xa2, 0xa3, 0xa4,
302: 0xa5, 0xa6, 0xa7, 0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf,
303: 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7, 0xb8, 0xb9, 0xba,
304: 0xbb, 0xbc, 0xbd, 0xbe, 0xbf, 0xc0, 0xc1, 0xc2, 0xc3, 0xc4, 0xc5,
305: 0xc6, 0xc7, 0xc8, 0xc9, 0xca, 0xcb, 0xcc, 0xcd, 0xce, 0xcf, 0xd0,
306: 0xd1, 0xd2, 0xd3, 0xd4, 0xd5, 0xd6, 0xd7, 0xd8, 0xd9, 0xda, 0xdb,
307: 0xdc, 0xdd, 0xde, 0xdf, 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6,
308: 0xe7, 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, 0xf0, 0xf1,
309: 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, 0xf8, 0xf9, 0xfa, 0xfb, 0xfc,
310: 0xfd, 0xfe, 0xff
311: };
312:
313: u_char cup2low[256] = {
314: 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a,
315: 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, 0x10, 0x11, 0x12, 0x13, 0x14, 0x15,
316: 0x16, 0x17, 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, 0x20,
317: 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, 0x28, 0x29, 0x2a, 0x2b,
318: 0x2c, 0x2d, 0x2e, 0x2f, 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36,
319: 0x37, 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0x3e, 0x3f, 0x60, 0x61,
320: 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6a, 0x6b, 0x6c,
321: 0x6d, 0x6e, 0x6f, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77,
322: 0x78, 0x79, 0x7a, 0x7b, 0x7c, 0x7d, 0x7e, 0x7f, 0x60, 0x61, 0x62,
323: 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d,
324: 0x6e, 0x6f, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78,
325: 0x79, 0x7a, 0x7b, 0x7c, 0x7d, 0x7e, 0x7f, 0x80, 0x81, 0x82, 0x83,
326: 0x84, 0x85, 0x86, 0x87, 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x8d, 0x8e,
327: 0x8f, 0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, 0x98, 0x99,
328: 0x9a, 0x9b, 0x9c, 0x9d, 0x9e, 0x9f, 0xa0, 0xa1, 0xa2, 0xa3, 0xa4,
329: 0xa5, 0xa6, 0xa7, 0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf,
330: 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7, 0xb8, 0xb9, 0xba,
331: 0xbb, 0xbc, 0xbd, 0xbe, 0xbf, 0xc0, 0xc1, 0xc2, 0xc3, 0xc4, 0xc5,
332: 0xc6, 0xc7, 0xc8, 0xc9, 0xca, 0xcb, 0xcc, 0xcd, 0xce, 0xcf, 0xd0,
333: 0xd1, 0xd2, 0xd3, 0xd4, 0xd5, 0xd6, 0xd7, 0xd8, 0xd9, 0xda, 0xdb,
334: 0xdc, 0xdd, 0xde, 0xdf, 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6,
335: 0xe7, 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, 0xf0, 0xf1,
336: 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, 0xf8, 0xf9, 0xfa, 0xfb, 0xfc,
337: 0xfd, 0xfe, 0xff
338: };
339:
340:
341:
342: /*
343: * cvs_diff()
344: *
345: * Handler for the `cvs diff' command.
346: *
347: * SYNOPSIS: cvs [args] diff [-clipu] [-D date] [-r rev]
348: */
349:
350: int
351: cvs_diff(int argc, char **argv)
352: {
353: int i, ch, recurse;
354: size_t dalen;
355: char dir[MAXPATHLEN], file[MAXPATHLEN], *d1, *d2, *r1, *r2;
356:
357: context = CVS_DIFF_DEFCTX;
1.2 ! jfb 358: strlcpy(diffargs, argv[0], sizeof(diffargs));
1.1 jfb 359:
360: d1 = d2 = NULL;
361: r1 = r2 = NULL;
362: recurse = 1;
363:
364: while ((ch = getopt(argc, argv, "cD:lir:u")) != -1) {
365: switch (ch) {
366: case 'c':
1.2 ! jfb 367: strlcat(diffargs, " -c", sizeof(diffargs));
1.1 jfb 368: format = D_CONTEXT;
369: break;
370: case 'D':
371: if (d1 == NULL && r1 == NULL)
372: d1 = optarg;
373: else if (d2 == NULL && r2 == NULL)
374: d2 = optarg;
375: else {
376: cvs_log(LP_ERR,
377: "no more than two revisions/dates can "
378: "be specified");
379: }
380: break;
381: case 'l':
1.2 ! jfb 382: strlcat(diffargs, " -l", sizeof(diffargs));
1.1 jfb 383: recurse = 0;
384: break;
385: case 'i':
1.2 ! jfb 386: strlcat(diffargs, " -i", sizeof(diffargs));
1.1 jfb 387: iflag = 1;
388: break;
389: case 'r':
390: if ((r1 == NULL) && (d1 == NULL))
391: r1 = optarg;
392: else if ((r2 == NULL) && (d2 == NULL))
393: r2 = optarg;
394: else {
395: cvs_log(LP_ERR,
396: "no more than two revisions/dates can "
397: "be specified");
398: return (EX_USAGE);
399: }
400: break;
401: case 'u':
1.2 ! jfb 402: strlcat(diffargs, " -u", sizeof(diffargs));
1.1 jfb 403: format = D_UNIFIED;
404: break;
405: default:
406: return (EX_USAGE);
407: }
408: }
409:
410: argc -= optind;
411: argv += optind;
412:
413: if (argc == 0) {
414: /* get the CVSROOT from current dir */
415: strlcpy(dir, ".", sizeof(dir));
416:
417: cvs_root = cvsroot_get(dir);
418: if (cvs_root == NULL)
419: return (EX_USAGE);
420:
421: if (cvs_root->cr_method != CVS_METHOD_LOCAL) {
422: cvs_client_connect();
423:
424: /* send the flags */
425: if (format == D_CONTEXT)
426: cvs_client_sendarg("-c", 0);
427: else if (format == D_UNIFIED)
428: cvs_client_sendarg("-u", 0);
429:
430: if (r1 != NULL) {
431: cvs_client_sendarg("-r", 0);
432: cvs_client_sendarg(r1, 1);
433: }
434: else if (d1 != NULL) {
435: cvs_client_sendarg("-D", 0);
436: cvs_client_sendarg(d1, 1);
437: }
438: if (r2 != NULL) {
439: cvs_client_sendarg("-r", 0);
440: cvs_client_sendarg(r2, 1);
441: }
442: else if (d2 != NULL) {
443: cvs_client_sendarg("-D", 0);
444: cvs_client_sendarg(d2, 1);
445: }
446: }
447:
448: cvs_diff_dir(dir, recurse);
449: }
450: else {
451: for (i = 0; i < argc; i++) {
452: cvs_splitpath(argv[i], dir, sizeof(dir),
453: file, sizeof(file));
454: cvs_root = cvsroot_get(dir);
455: if (cvs_root == NULL)
456: return (EX_USAGE);
457:
458: if (cvs_root->cr_method != CVS_METHOD_LOCAL) {
459: cvs_client_connect();
460:
461: if (i == 0) {
462: /* send the flags */
463: if (format == D_CONTEXT)
464: cvs_client_sendarg("-c", 0);
465: else if (format == D_UNIFIED)
466: cvs_client_sendarg("-u", 0);
467: }
468: }
469:
470: cvs_diff_file(argv[i], r1, r2);
471: }
472: }
473:
474: if (cvs_root->cr_method != CVS_METHOD_LOCAL)
475: cvs_client_sendreq(CVS_REQ_DIFF, NULL, 1);
476:
477: return (0);
478: }
479:
480:
481: /*
482: * cvs_diff_file()
483: *
484: * Diff a single file.
485: */
486:
487: int
488: cvs_diff_file(const char *path, const char *rev1, const char *rev2)
489: {
490: int modif;
491: char dir[MAXPATHLEN], file[MAXPATHLEN], rcspath[MAXPATHLEN];
492: char repo[MAXPATHLEN], buf[64];
493: time_t tsec;
494: BUF *b1, *b2;
495: RCSNUM *r1, *r2;
496: RCSFILE *rf;
497: CVSENTRIES *entf;
498: struct tm tmstamp;
499: struct stat fst;
500: struct cvs_ent *entp;
501:
502: rf = NULL;
503: diff_file = path;
504:
505: if (stat(path, &fst) == -1) {
506: cvs_log(LP_ERRNO, "cannot find %s", path);
507: return (-1);
508: }
509:
510: cvs_splitpath(path, dir, sizeof(dir), file, sizeof(file));
511: cvs_readrepo(dir, repo, sizeof(repo));
512:
513: entf = cvs_ent_open(dir);
514: if (entf == NULL) {
515: cvs_log(LP_ERR, "no CVS/Entries file in `%s'", dir);
516: return (-1);
517: }
518:
519: entp = cvs_ent_get(entf, file);
520: if ((entp == NULL) && (cvs_root->cr_method == CVS_METHOD_LOCAL)) {
521: cvs_log(LP_WARN, "I know nothing about %s", path);
522: return (-1);
523: }
524:
525: if (cvs_root->cr_method != CVS_METHOD_LOCAL) {
526: if (cvs_client_senddir(dir) < 0)
527: return (-1);
528: }
529:
530: tsec = (time_t)fst.st_mtimespec.tv_sec;
531:
532: if ((gmtime_r(&tsec, &tmstamp) == NULL) ||
533: (asctime_r(&tmstamp, buf) == NULL)) {
534: cvs_log(LP_ERR, "failed to generate file timestamp");
535: return (-1);
536: }
537: modif = (strcmp(buf, entp->ce_timestamp) == 0) ? 0 : 1;
538:
539: if (cvs_root->cr_method != CVS_METHOD_LOCAL)
540: cvs_client_sendentry(entp);
541:
542: if (!modif) {
543: if (cvs_root->cr_method != CVS_METHOD_LOCAL)
544: cvs_client_sendreq(CVS_REQ_UNCHANGED, file, 0);
545: cvs_ent_close(entf);
546: return (0);
547: }
548:
549: /* at this point, the file is modified */
550: if (cvs_root->cr_method != CVS_METHOD_LOCAL) {
551: cvs_client_sendreq(CVS_REQ_MODIFIED, file, 0);
552: cvs_sendfile(path);
553: }
554: else {
555: snprintf(rcspath, sizeof(rcspath), "%s/%s/%s%s",
556: cvs_root->cr_dir, repo, path, RCS_FILE_EXT);
557:
558: rf = rcs_open(rcspath, RCS_MODE_READ);
559: if (rf == NULL)
560: return (-1);
561:
562: printf("Index: %s\n%s\nRCS file: %s\n", path,
563: RCS_DIFF_DIV, rcspath);
564:
565: if (rev1 == NULL)
566: r1 = entp->ce_rev;
567: else {
568: r1 = rcsnum_alloc();
569: rcsnum_aton(rev1, NULL, r1);
570: }
571:
572: printf("retrieving revision %s\n",
573: rcsnum_tostr(r1, buf, sizeof(buf)));
574: b1 = rcs_getrev(rf, r1);
575:
576: if (rev2 != NULL) {
577: printf("retrieving revision %s\n", rev2);
578: r2 = rcsnum_alloc();
579: rcsnum_aton(rev2, NULL, r2);
580: b2 = rcs_getrev(rf, r2);
581: }
582: else {
583: b2 = cvs_buf_load(path, BUF_AUTOEXT);
584: }
585:
1.2 ! jfb 586: printf("%s", diffargs);
! 587: printf(" -r%s", buf);
! 588: if (rev2 != NULL)
! 589: printf(" -r%s", rev2);
! 590: printf(" %s\n", path);
1.1 jfb 591: cvs_buf_write(b1, "/tmp/diff1", 0600);
592: cvs_buf_write(b2, "/tmp/diff2", 0600);
593: cvs_diffreg("/tmp/diff1", "/tmp/diff2");
594: }
595:
596: cvs_ent_close(entf);
597:
598: return (0);
599: }
600:
601:
602: /*
603: * cvs_diff_dir()
604: *
605: */
606:
607: int
608: cvs_diff_dir(const char *dir, int recurse)
609: {
610: char path[MAXPATHLEN];
611: DIR *dirp;
612: CVSENTRIES *entf;
613: struct dirent *dentp;
614: struct cvs_ent *entp;
615:
616: printf("cvs_diff_dir(%s)\n", dir);
617:
618: dirp = opendir(dir);
619: if (dirp == NULL) {
620: cvs_log(LP_ERRNO, "failed to open directory `%s'", dir);
621: return (-1);
622: }
623:
624: entf = cvs_ent_open(dir);
625: if (entf == NULL) {
626: cvs_log(LP_ERR, "no CVS/Entries file in `%s'", dir);
627: (void)closedir(dirp);
628: return (-1);
629: }
630:
631: while ((dentp = readdir(dirp)) != NULL) {
632: if ((strcmp(dentp->d_name, "CVS") == 0) ||
633: (dentp->d_name[0] == '.'))
634: continue;
635:
636: if (strcmp(dir, ".") != 0) {
637: strlcpy(path, dir, sizeof(path));
638: strlcat(path, "/", sizeof(path));
639: }
640: else
641: path[0] = '\0';
642: strlcat(path, dentp->d_name, sizeof(path));
643: if (dentp->d_type == DT_DIR) {
644: if (!recurse)
645: continue;
646: cvs_diff_dir(path, recurse);
647: }
648: else {
649: entp = cvs_ent_get(entf, dentp->d_name);
650: if (entp == NULL) {
651: cvs_client_sendreq(CVS_REQ_QUESTIONABLE, path,
652: 0);
653: }
654: else {
655: #if 0
656: cvs_diff_file(path);
657: #endif
658: }
659: }
660: }
661:
662: return (0);
663: }
664:
665:
666:
667:
668: int
669: cvs_diffreg(const char *file1, const char *file2)
670: {
671: FILE *f1, *f2;
672: int i, rval;
673:
674: f1 = f2 = NULL;
675: rval = D_SAME;
676: anychange = 0;
677: lastline = 0;
678: lastmatchline = 0;
679: context_vec_ptr = context_vec_start - 1;
680: chrtran = (iflag ? cup2low : clow2low);
681:
682: f1 = fopen(file1, "r");
683: if (f1 == NULL) {
684: cvs_log(LP_ERRNO, "%s", file1);
685: status |= 2;
686: goto closem;
687: }
688:
689: f2 = fopen(file2, "r");
690: if (f2 == NULL) {
691: cvs_log(LP_ERRNO, "%s", file2);
692: status |= 2;
693: goto closem;
694: }
695:
696: switch (files_differ(f1, f2)) {
697: case 0:
698: goto closem;
699: case 1:
700: break;
701: default:
702: /* error */
703: status |= 2;
704: goto closem;
705: }
706:
707: if (!asciifile(f1) || !asciifile(f2)) {
708: rval = D_BINARY;
709: status |= 1;
710: goto closem;
711: }
712: prepare(0, f1, stb1.st_size);
713: prepare(1, f2, stb2.st_size);
714: prune();
715: sort(sfile[0], slen[0]);
716: sort(sfile[1], slen[1]);
717:
718: member = (int *)file[1];
719: equiv(sfile[0], slen[0], sfile[1], slen[1], member);
720: member = realloc(member, (slen[1] + 2) * sizeof(int));
721:
722: class = (int *)file[0];
723: unsort(sfile[0], slen[0], class);
724: class = realloc(class, (slen[0] + 2) * sizeof(int));
725:
726: klist = malloc((slen[0] + 2) * sizeof(int));
727: clen = 0;
728: clistlen = 100;
729: clist = malloc(clistlen * sizeof(cand));
730: i = stone(class, slen[0], member, klist);
731: free(member);
732: free(class);
733:
734: J = realloc(J, (len[0] + 2) * sizeof(int));
735: unravel(klist[i]);
736: free(clist);
737: free(klist);
738:
739: ixold = realloc(ixold, (len[0] + 2) * sizeof(long));
740: ixnew = realloc(ixnew, (len[1] + 2) * sizeof(long));
741: check(f1, f2);
742: output(file1, f1, file2, f2);
743:
744: closem:
745: if (anychange) {
746: status |= 1;
747: if (rval == D_SAME)
748: rval = D_DIFFER;
749: }
750: if (f1 != NULL)
751: fclose(f1);
752: if (f2 != NULL)
753: fclose(f2);
754: return (rval);
755: }
756:
757: /*
758: * Check to see if the given files differ.
759: * Returns 0 if they are the same, 1 if different, and -1 on error.
760: * XXX - could use code from cmp(1) [faster]
761: */
762: static int
763: files_differ(FILE *f1, FILE *f2)
764: {
765: char buf1[BUFSIZ], buf2[BUFSIZ];
766: size_t i, j;
767:
768: if (stb1.st_size != stb2.st_size)
769: return (1);
770: for (;;) {
771: i = fread(buf1, 1, sizeof(buf1), f1);
772: j = fread(buf2, 1, sizeof(buf2), f2);
773: if (i != j)
774: return (1);
775: if (i == 0 && j == 0) {
776: if (ferror(f1) || ferror(f2))
777: return (1);
778: return (0);
779: }
780: if (memcmp(buf1, buf2, i) != 0)
781: return (1);
782: }
783: }
784:
785:
786: char *
787: splice(char *dir, char *file)
788: {
789: char *tail, *buf;
790:
791: if ((tail = strrchr(file, '/')) == NULL)
792: tail = file;
793: else
794: tail++;
795: asprintf(&buf, "%s/%s", dir, tail);
796: return (buf);
797: }
798:
799: static void
800: prepare(int i, FILE *fd, off_t filesize)
801: {
802: struct line *p;
803: int j, h;
804: size_t sz;
805:
806: rewind(fd);
807:
808: sz = (filesize <= SIZE_MAX ? filesize : SIZE_MAX) / 25;
809: if (sz < 100)
810: sz = 100;
811:
812: p = malloc((sz + 3) * sizeof(struct line));
813: for (j = 0; (h = readhash(fd));) {
814: if (j == (int)sz) {
815: sz = sz * 3 / 2;
816: p = realloc(p, (sz + 3) * sizeof(struct line));
817: }
818: p[++j].value = h;
819: }
820: len[i] = j;
821: file[i] = p;
822: }
823:
824: static void
825: prune(void)
826: {
827: int i, j;
828:
829: for (pref = 0; pref < len[0] && pref < len[1] &&
830: file[0][pref + 1].value == file[1][pref + 1].value;
831: pref++)
832: ;
833: for (suff = 0; suff < len[0] - pref && suff < len[1] - pref &&
834: file[0][len[0] - suff].value == file[1][len[1] - suff].value;
835: suff++)
836: ;
837: for (j = 0; j < 2; j++) {
838: sfile[j] = file[j] + pref;
839: slen[j] = len[j] - pref - suff;
840: for (i = 0; i <= slen[j]; i++)
841: sfile[j][i].serial = i;
842: }
843: }
844:
845: static void
846: equiv(struct line *a, int n, struct line *b, int m, int *c)
847: {
848: int i, j;
849:
850: i = j = 1;
851: while (i <= n && j <= m) {
852: if (a[i].value < b[j].value)
853: a[i++].value = 0;
854: else if (a[i].value == b[j].value)
855: a[i++].value = j;
856: else
857: j++;
858: }
859: while (i <= n)
860: a[i++].value = 0;
861: b[m + 1].value = 0;
862: j = 0;
863: while (++j <= m) {
864: c[j] = -b[j].serial;
865: while (b[j + 1].value == b[j].value) {
866: j++;
867: c[j] = b[j].serial;
868: }
869: }
870: c[j] = -1;
871: }
872:
873: /* Code taken from ping.c */
874: static int
875: isqrt(int n)
876: {
877: int y, x = 1;
878:
879: if (n == 0)
880: return(0);
881:
882: do { /* newton was a stinker */
883: y = x;
884: x = n / x;
885: x += y;
886: x /= 2;
887: } while ((x - y) > 1 || (x - y) < -1);
888:
889: return (x);
890: }
891:
892: static int
893: stone(int *a, int n, int *b, int *c)
894: {
895: int i, k, y, j, l;
896: int oldc, tc, oldl;
897: u_int numtries;
898:
899: const u_int bound = dflag ? UINT_MAX : max(256, isqrt(n));
900:
901: k = 0;
902: c[0] = newcand(0, 0, 0);
903: for (i = 1; i <= n; i++) {
904: j = a[i];
905: if (j == 0)
906: continue;
907: y = -b[j];
908: oldl = 0;
909: oldc = c[0];
910: numtries = 0;
911: do {
912: if (y <= clist[oldc].y)
913: continue;
914: l = search(c, k, y);
915: if (l != oldl + 1)
916: oldc = c[l - 1];
917: if (l <= k) {
918: if (clist[c[l]].y <= y)
919: continue;
920: tc = c[l];
921: c[l] = newcand(i, y, oldc);
922: oldc = tc;
923: oldl = l;
924: numtries++;
925: } else {
926: c[l] = newcand(i, y, oldc);
927: k++;
928: break;
929: }
930: } while ((y = b[++j]) > 0 && numtries < bound);
931: }
932: return (k);
933: }
934:
935: static int
936: newcand(int x, int y, int pred)
937: {
938: struct cand *q;
939:
940: if (clen == clistlen) {
941: clistlen = clistlen * 11 / 10;
942: clist = realloc(clist, clistlen * sizeof(cand));
943: }
944: q = clist + clen;
945: q->x = x;
946: q->y = y;
947: q->pred = pred;
948: return (clen++);
949: }
950:
951: static int
952: search(int *c, int k, int y)
953: {
954: int i, j, l, t;
955:
956: if (clist[c[k]].y < y) /* quick look for typical case */
957: return (k + 1);
958: i = 0;
959: j = k + 1;
960: while (1) {
961: l = i + j;
962: if ((l >>= 1) <= i)
963: break;
964: t = clist[c[l]].y;
965: if (t > y)
966: j = l;
967: else if (t < y)
968: i = l;
969: else
970: return (l);
971: }
972: return (l + 1);
973: }
974:
975: static void
976: unravel(int p)
977: {
978: struct cand *q;
979: int i;
980:
981: for (i = 0; i <= len[0]; i++)
982: J[i] = i <= pref ? i :
983: i > len[0] - suff ? i + len[1] - len[0] : 0;
984: for (q = clist + p; q->y != 0; q = clist + q->pred)
985: J[q->x + pref] = q->y + pref;
986: }
987:
988: /*
989: * Check does double duty:
990: * 1. ferret out any fortuitous correspondences due
991: * to confounding by hashing (which result in "jackpot")
992: * 2. collect random access indexes to the two files
993: */
994: static void
995: check(FILE *f1, FILE *f2)
996: {
997: int i, j, jackpot, c, d;
998: long ctold, ctnew;
999:
1000: rewind(f1);
1001: rewind(f2);
1002: j = 1;
1003: ixold[0] = ixnew[0] = 0;
1004: jackpot = 0;
1005: ctold = ctnew = 0;
1006: for (i = 1; i <= len[0]; i++) {
1007: if (J[i] == 0) {
1008: ixold[i] = ctold += skipline(f1);
1009: continue;
1010: }
1011: while (j < J[i]) {
1012: ixnew[j] = ctnew += skipline(f2);
1013: j++;
1014: }
1015: if (bflag || wflag || iflag) {
1016: for (;;) {
1017: c = getc(f1);
1018: d = getc(f2);
1019: /*
1020: * GNU diff ignores a missing newline
1021: * in one file if bflag || wflag.
1022: */
1023: if ((bflag || wflag) &&
1024: ((c == EOF && d == '\n') ||
1025: (c == '\n' && d == EOF))) {
1026: break;
1027: }
1028: ctold++;
1029: ctnew++;
1030: if (bflag && isspace(c) && isspace(d)) {
1031: do {
1032: if (c == '\n')
1033: break;
1034: ctold++;
1035: } while (isspace(c = getc(f1)));
1036: do {
1037: if (d == '\n')
1038: break;
1039: ctnew++;
1040: } while (isspace(d = getc(f2)));
1041: } else if (wflag) {
1042: while (isspace(c) && c != '\n') {
1043: c = getc(f1);
1044: ctold++;
1045: }
1046: while (isspace(d) && d != '\n') {
1047: d = getc(f2);
1048: ctnew++;
1049: }
1050: }
1051: if (chrtran[c] != chrtran[d]) {
1052: jackpot++;
1053: J[i] = 0;
1054: if (c != '\n' && c != EOF)
1055: ctold += skipline(f1);
1056: if (d != '\n' && c != EOF)
1057: ctnew += skipline(f2);
1058: break;
1059: }
1060: if (c == '\n' || c == EOF)
1061: break;
1062: }
1063: } else {
1064: for (;;) {
1065: ctold++;
1066: ctnew++;
1067: if ((c = getc(f1)) != (d = getc(f2))) {
1068: /* jackpot++; */
1069: J[i] = 0;
1070: if (c != '\n' && c != EOF)
1071: ctold += skipline(f1);
1072: if (d != '\n' && c != EOF)
1073: ctnew += skipline(f2);
1074: break;
1075: }
1076: if (c == '\n' || c == EOF)
1077: break;
1078: }
1079: }
1080: ixold[i] = ctold;
1081: ixnew[j] = ctnew;
1082: j++;
1083: }
1084: for (; j <= len[1]; j++)
1085: ixnew[j] = ctnew += skipline(f2);
1086: /*
1087: * if (jackpot)
1088: * fprintf(stderr, "jackpot\n");
1089: */
1090: }
1091:
1092: /* shellsort CACM #201 */
1093: static void
1094: sort(struct line *a, int n)
1095: {
1096: struct line *ai, *aim, w;
1097: int j, m = 0, k;
1098:
1099: if (n == 0)
1100: return;
1101: for (j = 1; j <= n; j *= 2)
1102: m = 2 * j - 1;
1103: for (m /= 2; m != 0; m /= 2) {
1104: k = n - m;
1105: for (j = 1; j <= k; j++) {
1106: for (ai = &a[j]; ai > a; ai -= m) {
1107: aim = &ai[m];
1108: if (aim < ai)
1109: break; /* wraparound */
1110: if (aim->value > ai[0].value ||
1111: (aim->value == ai[0].value &&
1112: aim->serial > ai[0].serial))
1113: break;
1114: w.value = ai[0].value;
1115: ai[0].value = aim->value;
1116: aim->value = w.value;
1117: w.serial = ai[0].serial;
1118: ai[0].serial = aim->serial;
1119: aim->serial = w.serial;
1120: }
1121: }
1122: }
1123: }
1124:
1125: static void
1126: unsort(struct line *f, int l, int *b)
1127: {
1128: int *a, i;
1129:
1130: a = malloc((l + 1) * sizeof(int));
1131: for (i = 1; i <= l; i++)
1132: a[f[i].serial] = f[i].value;
1133: for (i = 1; i <= l; i++)
1134: b[i] = a[i];
1135: free(a);
1136: }
1137:
1138: static int
1139: skipline(FILE *f)
1140: {
1141: int i, c;
1142:
1143: for (i = 1; (c = getc(f)) != '\n' && c != EOF; i++)
1144: continue;
1145: return (i);
1146: }
1147:
1148: static void
1149: output(const char *file1, FILE *f1, const char *file2, FILE *f2)
1150: {
1151: int m, i0, i1, j0, j1;
1152:
1153: rewind(f1);
1154: rewind(f2);
1155: m = len[0];
1156: J[0] = 0;
1157: J[m + 1] = len[1] + 1;
1158: for (i0 = 1; i0 <= m; i0 = i1 + 1) {
1159: while (i0 <= m && J[i0] == J[i0 - 1] + 1)
1160: i0++;
1161: j0 = J[i0 - 1] + 1;
1162: i1 = i0 - 1;
1163: while (i1 < m && J[i1 + 1] == 0)
1164: i1++;
1165: j1 = J[i1 + 1] - 1;
1166: J[i1] = j1;
1167: change(file1, f1, file2, f2, i0, i1, j0, j1);
1168: }
1169: if (m == 0)
1170: change(file1, f1, file2, f2, 1, 0, 1, len[1]);
1171: if (format == D_IFDEF) {
1172: for (;;) {
1173: #define c i0
1174: if ((c = getc(f1)) == EOF)
1175: return;
1176: putchar(c);
1177: }
1178: #undef c
1179: }
1180: if (anychange != 0) {
1181: if (format == D_CONTEXT)
1182: dump_context_vec(f1, f2);
1183: else if (format == D_UNIFIED)
1184: dump_unified_vec(f1, f2);
1185: }
1186: }
1187:
1188: static __inline void
1189: range(int a, int b, char *separator)
1190: {
1191: printf("%d", a > b ? b : a);
1192: if (a < b)
1193: printf("%s%d", separator, b);
1194: }
1195:
1196: static __inline void
1197: uni_range(int a, int b)
1198: {
1199: if (a < b)
1200: printf("%d,%d", a, b - a + 1);
1201: else if (a == b)
1202: printf("%d", b);
1203: else
1204: printf("%d,0", b);
1205: }
1206:
1207: static char *
1208: preadline(int fd, size_t len, off_t off)
1209: {
1210: char *line;
1211: ssize_t nr;
1212:
1213: line = malloc(len + 1);
1214: if (line == NULL) {
1215: cvs_log(LP_ERRNO, "failed to allocate line");
1216: return (NULL);
1217: }
1218: if ((nr = pread(fd, line, len, off)) < 0) {
1219: cvs_log(LP_ERRNO, "preadline failed");
1220: return (NULL);
1221: }
1222: line[nr] = '\0';
1223: return (line);
1224: }
1225:
1226: static int
1227: ignoreline(char *line)
1228: {
1229: int ret;
1230:
1231: ret = regexec(&ignore_re, line, 0, NULL, 0);
1232: free(line);
1233: return (ret == 0); /* if it matched, it should be ignored. */
1234: }
1235:
1236: /*
1237: * Indicate that there is a difference between lines a and b of the from file
1238: * to get to lines c to d of the to file. If a is greater then b then there
1239: * are no lines in the from file involved and this means that there were
1240: * lines appended (beginning at b). If c is greater than d then there are
1241: * lines missing from the to file.
1242: */
1243: static void
1244: change(const char *file1, FILE *f1, const char *file2, FILE *f2,
1245: int a, int b, int c, int d)
1246: {
1247: static size_t max_context = 64;
1248: int i;
1249:
1250: if (format != D_IFDEF && a > b && c > d)
1251: return;
1252: if (ignore_pats != NULL) {
1253: char *line;
1254: /*
1255: * All lines in the change, insert, or delete must
1256: * match an ignore pattern for the change to be
1257: * ignored.
1258: */
1259: if (a <= b) { /* Changes and deletes. */
1260: for (i = a; i <= b; i++) {
1261: line = preadline(fileno(f1),
1262: ixold[i] - ixold[i - 1], ixold[i - 1]);
1263: if (!ignoreline(line))
1264: goto proceed;
1265: }
1266: }
1267: if (a > b || c <= d) { /* Changes and inserts. */
1268: for (i = c; i <= d; i++) {
1269: line = preadline(fileno(f2),
1270: ixnew[i] - ixnew[i - 1], ixnew[i - 1]);
1271: if (!ignoreline(line))
1272: goto proceed;
1273: }
1274: }
1275: return;
1276: }
1277: proceed:
1278: if (format == D_CONTEXT || format == D_UNIFIED) {
1279: /*
1280: * Allocate change records as needed.
1281: */
1282: if (context_vec_ptr == context_vec_end - 1) {
1283: ptrdiff_t offset = context_vec_ptr - context_vec_start;
1284: max_context <<= 1;
1285: context_vec_start = realloc(context_vec_start,
1286: max_context * sizeof(struct context_vec));
1287: context_vec_end = context_vec_start + max_context;
1288: context_vec_ptr = context_vec_start + offset;
1289: }
1290: if (anychange == 0) {
1291: /*
1292: * Print the context/unidiff header first time through.
1293: */
1294: printf("%s %s %s",
1295: format == D_CONTEXT ? "***" : "---", diff_file,
1296: ctime(&stb1.st_mtime));
1297: printf("%s %s %s",
1298: format == D_CONTEXT ? "---" : "+++", diff_file,
1299: ctime(&stb2.st_mtime));
1300: anychange = 1;
1301: } else if (a > context_vec_ptr->b + (2 * context) + 1 &&
1302: c > context_vec_ptr->d + (2 * context) + 1) {
1303: /*
1304: * If this change is more than 'context' lines from the
1305: * previous change, dump the record and reset it.
1306: */
1307: if (format == D_CONTEXT)
1308: dump_context_vec(f1, f2);
1309: else
1310: dump_unified_vec(f1, f2);
1311: }
1312: context_vec_ptr++;
1313: context_vec_ptr->a = a;
1314: context_vec_ptr->b = b;
1315: context_vec_ptr->c = c;
1316: context_vec_ptr->d = d;
1317: return;
1318: }
1319: if (anychange == 0)
1320: anychange = 1;
1321: switch (format) {
1322: case D_BRIEF:
1323: return;
1324: case D_NORMAL:
1325: range(a, b, ",");
1326: putchar(a > b ? 'a' : c > d ? 'd' : 'c');
1327: if (format == D_NORMAL)
1328: range(c, d, ",");
1329: putchar('\n');
1330: break;
1331: }
1332: if (format == D_NORMAL || format == D_IFDEF) {
1333: fetch(ixold, a, b, f1, '<', 1);
1334: if (a <= b && c <= d && format == D_NORMAL)
1335: puts("---");
1336: }
1337: i = fetch(ixnew, c, d, f2, format == D_NORMAL ? '>' : '\0', 0);
1338: if (inifdef) {
1339: printf("#endif /* %s */\n", ifdefname);
1340: inifdef = 0;
1341: }
1342: }
1343:
1344: static int
1345: fetch(long *f, int a, int b, FILE *lb, int ch, int oldfile)
1346: {
1347: int i, j, c, lastc, col, nc;
1348:
1349: /*
1350: * When doing #ifdef's, copy down to current line
1351: * if this is the first file, so that stuff makes it to output.
1352: */
1353: if (format == D_IFDEF && oldfile) {
1354: long curpos = ftell(lb);
1355: /* print through if append (a>b), else to (nb: 0 vs 1 orig) */
1356: nc = f[a > b ? b : a - 1] - curpos;
1357: for (i = 0; i < nc; i++)
1358: putchar(getc(lb));
1359: }
1360: if (a > b)
1361: return (0);
1362: if (format == D_IFDEF) {
1363: if (inifdef) {
1364: printf("#else /* %s%s */\n",
1365: oldfile == 1 ? "!" : "", ifdefname);
1366: } else {
1367: if (oldfile)
1368: printf("#ifndef %s\n", ifdefname);
1369: else
1370: printf("#ifdef %s\n", ifdefname);
1371: }
1372: inifdef = 1 + oldfile;
1373: }
1374: for (i = a; i <= b; i++) {
1375: fseek(lb, f[i - 1], SEEK_SET);
1376: nc = f[i] - f[i - 1];
1377: if (format != D_IFDEF && ch != '\0') {
1378: putchar(ch);
1379: if (Tflag && (format == D_NORMAL || format == D_CONTEXT
1380: || format == D_UNIFIED))
1381: putchar('\t');
1382: else if (format != D_UNIFIED)
1383: putchar(' ');
1384: }
1385: col = 0;
1386: for (j = 0, lastc = '\0'; j < nc; j++, lastc = c) {
1387: if ((c = getc(lb)) == EOF) {
1388: puts("\n\\ No newline at end of file");
1389: return (0);
1390: }
1391: if (c == '\t' && tflag) {
1392: do {
1393: putchar(' ');
1394: } while (++col & 7);
1395: } else {
1396: putchar(c);
1397: col++;
1398: }
1399: }
1400: }
1401: return (0);
1402: }
1403:
1404: /*
1405: * Hash function taken from Robert Sedgewick, Algorithms in C, 3d ed., p 578.
1406: */
1407: static int
1408: readhash(FILE *f)
1409: {
1410: int i, t, space;
1411: int sum;
1412:
1413: sum = 1;
1414: space = 0;
1415: if (!bflag && !wflag) {
1416: if (iflag)
1417: for (i = 0; (t = getc(f)) != '\n'; i++) {
1418: if (t == EOF) {
1419: if (i == 0)
1420: return (0);
1421: break;
1422: }
1423: sum = sum * 127 + chrtran[t];
1424: }
1425: else
1426: for (i = 0; (t = getc(f)) != '\n'; i++) {
1427: if (t == EOF) {
1428: if (i == 0)
1429: return (0);
1430: break;
1431: }
1432: sum = sum * 127 + t;
1433: }
1434: } else {
1435: for (i = 0;;) {
1436: switch (t = getc(f)) {
1437: case '\t':
1438: case ' ':
1439: space++;
1440: continue;
1441: default:
1442: if (space && !wflag) {
1443: i++;
1444: space = 0;
1445: }
1446: sum = sum * 127 + chrtran[t];
1447: i++;
1448: continue;
1449: case EOF:
1450: if (i == 0)
1451: return (0);
1452: /* FALLTHROUGH */
1453: case '\n':
1454: break;
1455: }
1456: break;
1457: }
1458: }
1459: /*
1460: * There is a remote possibility that we end up with a zero sum.
1461: * Zero is used as an EOF marker, so return 1 instead.
1462: */
1463: return (sum == 0 ? 1 : sum);
1464: }
1465:
1466: static int
1467: asciifile(FILE *f)
1468: {
1469: char buf[BUFSIZ];
1470: int i, cnt;
1471:
1472: if (aflag || f == NULL)
1473: return (1);
1474:
1475: rewind(f);
1476: cnt = fread(buf, 1, sizeof(buf), f);
1477: for (i = 0; i < cnt; i++)
1478: if (!isprint(buf[i]) && !isspace(buf[i]))
1479: return (0);
1480: return (1);
1481: }
1482:
1483: static __inline int min(int a, int b)
1484: {
1485: return (a < b ? a : b);
1486: }
1487:
1488: static __inline int max(int a, int b)
1489: {
1490: return (a > b ? a : b);
1491: }
1492:
1493: static char *
1494: match_function(const long *f, int pos, FILE *file)
1495: {
1496: char buf[FUNCTION_CONTEXT_SIZE];
1497: size_t nc;
1498: int last = lastline;
1499: char *p;
1500:
1501: lastline = pos;
1502: while (pos > last) {
1503: fseek(file, f[pos - 1], SEEK_SET);
1504: nc = f[pos] - f[pos - 1];
1505: if (nc >= sizeof(buf))
1506: nc = sizeof(buf) - 1;
1507: nc = fread(buf, 1, nc, file);
1508: if (nc > 0) {
1509: buf[nc] = '\0';
1510: p = strchr(buf, '\n');
1511: if (p != NULL)
1512: *p = '\0';
1513: if (isalpha(buf[0]) || buf[0] == '_' || buf[0] == '$') {
1514: strlcpy(lastbuf, buf, sizeof lastbuf);
1515: lastmatchline = pos;
1516: return lastbuf;
1517: }
1518: }
1519: pos--;
1520: }
1521: return lastmatchline > 0 ? lastbuf : NULL;
1522: }
1523:
1524: /* dump accumulated "context" diff changes */
1525: static void
1526: dump_context_vec(FILE *f1, FILE *f2)
1527: {
1528: struct context_vec *cvp = context_vec_start;
1529: int lowa, upb, lowc, upd, do_output;
1530: int a, b, c, d;
1531: char ch, *f;
1532:
1533: if (context_vec_start > context_vec_ptr)
1534: return;
1535:
1536: b = d = 0; /* gcc */
1537: lowa = max(1, cvp->a - context);
1538: upb = min(len[0], context_vec_ptr->b + context);
1539: lowc = max(1, cvp->c - context);
1540: upd = min(len[1], context_vec_ptr->d + context);
1541:
1542: printf("***************");
1543: printf("\n*** ");
1544: range(lowa, upb, ",");
1545: printf(" ****\n");
1546:
1547: /*
1548: * Output changes to the "old" file. The first loop suppresses
1549: * output if there were no changes to the "old" file (we'll see
1550: * the "old" lines as context in the "new" list).
1551: */
1552: do_output = 0;
1553: for (; cvp <= context_vec_ptr; cvp++)
1554: if (cvp->a <= cvp->b) {
1555: cvp = context_vec_start;
1556: do_output++;
1557: break;
1558: }
1559: if (do_output) {
1560: while (cvp <= context_vec_ptr) {
1561: a = cvp->a;
1562: b = cvp->b;
1563: c = cvp->c;
1564: d = cvp->d;
1565:
1566: if (a <= b && c <= d)
1567: ch = 'c';
1568: else
1569: ch = (a <= b) ? 'd' : 'a';
1570:
1571: if (ch == 'a')
1572: fetch(ixold, lowa, b, f1, ' ', 0);
1573: else {
1574: fetch(ixold, lowa, a - 1, f1, ' ', 0);
1575: fetch(ixold, a, b, f1,
1576: ch == 'c' ? '!' : '-', 0);
1577: }
1578: lowa = b + 1;
1579: cvp++;
1580: }
1581: fetch(ixold, b + 1, upb, f1, ' ', 0);
1582: }
1583: /* output changes to the "new" file */
1584: printf("--- ");
1585: range(lowc, upd, ",");
1586: printf(" ----\n");
1587:
1588: do_output = 0;
1589: for (cvp = context_vec_start; cvp <= context_vec_ptr; cvp++)
1590: if (cvp->c <= cvp->d) {
1591: cvp = context_vec_start;
1592: do_output++;
1593: break;
1594: }
1595: if (do_output) {
1596: while (cvp <= context_vec_ptr) {
1597: a = cvp->a;
1598: b = cvp->b;
1599: c = cvp->c;
1600: d = cvp->d;
1601:
1602: if (a <= b && c <= d)
1603: ch = 'c';
1604: else
1605: ch = (a <= b) ? 'd' : 'a';
1606:
1607: if (ch == 'd')
1608: fetch(ixnew, lowc, d, f2, ' ', 0);
1609: else {
1610: fetch(ixnew, lowc, c - 1, f2, ' ', 0);
1611: fetch(ixnew, c, d, f2,
1612: ch == 'c' ? '!' : '+', 0);
1613: }
1614: lowc = d + 1;
1615: cvp++;
1616: }
1617: fetch(ixnew, d + 1, upd, f2, ' ', 0);
1618: }
1619: context_vec_ptr = context_vec_start - 1;
1620: }
1621:
1622: /* dump accumulated "unified" diff changes */
1623: static void
1624: dump_unified_vec(FILE *f1, FILE *f2)
1625: {
1626: struct context_vec *cvp = context_vec_start;
1627: int lowa, upb, lowc, upd;
1628: int a, b, c, d;
1629: char ch, *f;
1630:
1631: if (context_vec_start > context_vec_ptr)
1632: return;
1633:
1634: b = d = 0; /* gcc */
1635: lowa = max(1, cvp->a - context);
1636: upb = min(len[0], context_vec_ptr->b + context);
1637: lowc = max(1, cvp->c - context);
1638: upd = min(len[1], context_vec_ptr->d + context);
1639:
1640: fputs("@@ -", stdout);
1641: uni_range(lowa, upb);
1642: fputs(" +", stdout);
1643: uni_range(lowc, upd);
1644: fputs(" @@", stdout);
1645: putchar('\n');
1646:
1647: /*
1648: * Output changes in "unified" diff format--the old and new lines
1649: * are printed together.
1650: */
1651: for (; cvp <= context_vec_ptr; cvp++) {
1652: a = cvp->a;
1653: b = cvp->b;
1654: c = cvp->c;
1655: d = cvp->d;
1656:
1657: /*
1658: * c: both new and old changes
1659: * d: only changes in the old file
1660: * a: only changes in the new file
1661: */
1662: if (a <= b && c <= d)
1663: ch = 'c';
1664: else
1665: ch = (a <= b) ? 'd' : 'a';
1666:
1667: switch (ch) {
1668: case 'c':
1669: fetch(ixold, lowa, a - 1, f1, ' ', 0);
1670: fetch(ixold, a, b, f1, '-', 0);
1671: fetch(ixnew, c, d, f2, '+', 0);
1672: break;
1673: case 'd':
1674: fetch(ixold, lowa, a - 1, f1, ' ', 0);
1675: fetch(ixold, a, b, f1, '-', 0);
1676: break;
1677: case 'a':
1678: fetch(ixnew, lowc, c - 1, f2, ' ', 0);
1679: fetch(ixnew, c, d, f2, '+', 0);
1680: break;
1681: }
1682: lowa = b + 1;
1683: lowc = d + 1;
1684: }
1685: fetch(ixnew, d + 1, upd, f2, ' ', 0);
1686:
1687: context_vec_ptr = context_vec_start - 1;
1688: }