Annotation of src/usr.bin/cvs/diff.c, Revision 1.17
1.17 ! jfb 1: /* $OpenBSD: diff.c,v 1.16 2004/12/28 22:07:31 xsa Exp $ */
1.1 jfb 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"
1.4 jfb 149: #include "proto.h"
1.1 jfb 150:
151:
152: #define CVS_DIFF_DEFCTX 3 /* default context length */
153:
154:
155: /*
156: * Output format options
157: */
158: #define D_NORMAL 0 /* Normal output */
159: #define D_CONTEXT 1 /* Diff with context */
160: #define D_UNIFIED 2 /* Unified context diff */
161: #define D_IFDEF 3 /* Diff with merged #ifdef's */
162: #define D_BRIEF 4 /* Say if the files differ */
163:
164: /*
165: * Status values for print_status() and diffreg() return values
166: */
167: #define D_SAME 0 /* Files are the same */
168: #define D_DIFFER 1 /* Files are different */
169: #define D_BINARY 2 /* Binary files are different */
170: #define D_COMMON 3 /* Subdirectory common to both dirs */
171: #define D_ONLY 4 /* Only exists in one directory */
172: #define D_MISMATCH1 5 /* path1 was a dir, path2 a file */
173: #define D_MISMATCH2 6 /* path1 was a file, path2 a dir */
174: #define D_ERROR 7 /* An error occurred */
175: #define D_SKIPPED1 8 /* path1 was a special file */
176: #define D_SKIPPED2 9 /* path2 was a special file */
177:
178: struct cand {
179: int x;
180: int y;
181: int pred;
182: } cand;
183:
184: struct line {
185: int serial;
186: int value;
187: } *file[2];
188:
189: /*
190: * The following struct is used to record change information when
191: * doing a "context" or "unified" diff. (see routine "change" to
192: * understand the highly mnemonic field names)
193: */
194: struct context_vec {
195: int a; /* start line in old file */
196: int b; /* end line in old file */
197: int c; /* start line in new file */
198: int d; /* end line in new file */
199: };
200:
1.4 jfb 201: struct diff_arg {
202: char *rev1;
203: char *rev2;
204: char *date1;
205: char *date2;
206: };
207:
1.1 jfb 208:
209: struct excludes {
210: char *pattern;
211: struct excludes *next;
212: };
213:
214:
215: char *splice(char *, char *);
216: int cvs_diffreg(const char *, const char *);
1.9 jfb 217: int cvs_diff_file(struct cvs_file *, void *);
218: int cvs_diff_sendflags(struct cvsroot *, struct diff_arg *);
1.1 jfb 219: static void output(const char *, FILE *, const char *, FILE *);
220: static void check(FILE *, FILE *);
221: static void range(int, int, char *);
222: static void uni_range(int, int);
223: static void dump_context_vec(FILE *, FILE *);
224: static void dump_unified_vec(FILE *, FILE *);
225: static void prepare(int, FILE *, off_t);
226: static void prune(void);
227: static void equiv(struct line *, int, struct line *, int, int *);
228: static void unravel(int);
229: static void unsort(struct line *, int, int *);
230: static void change(const char *, FILE *, const char *, FILE *, int, int, int, int);
231: static void sort(struct line *, int);
232: static int ignoreline(char *);
233: static int asciifile(FILE *);
234: static int fetch(long *, int, int, FILE *, int, int);
235: static int newcand(int, int, int);
236: static int search(int *, int, int);
237: static int skipline(FILE *);
238: static int isqrt(int);
239: static int stone(int *, int, int *, int *);
240: static int readhash(FILE *);
241: static int files_differ(FILE *, FILE *);
1.17 ! jfb 242: static char *match_function(const long *, int, FILE *);
1.1 jfb 243: static char *preadline(int, size_t, off_t);
244:
245:
246: extern int cvs_client;
247:
1.17 ! jfb 248: static int aflag, bflag, dflag, iflag, Nflag, pflag, tflag, Tflag, wflag;
1.1 jfb 249: static int context, status;
250: static int format = D_NORMAL;
251: static struct stat stb1, stb2;
1.2 jfb 252: static char *ifdefname, *ignore_pats, diffargs[128];
253: static const char *diff_file;
1.1 jfb 254: regex_t ignore_re;
255:
256: static int *J; /* will be overlaid on class */
257: static int *class; /* will be overlaid on file[0] */
258: static int *klist; /* will be overlaid on file[0] after class */
259: static int *member; /* will be overlaid on file[1] */
260: static int clen;
261: static int inifdef; /* whether or not we are in a #ifdef block */
262: static int len[2];
263: static int pref, suff; /* length of prefix and suffix */
264: static int slen[2];
265: static int anychange;
266: static long *ixnew; /* will be overlaid on file[1] */
267: static long *ixold; /* will be overlaid on klist */
268: static struct cand *clist; /* merely a free storage pot for candidates */
269: static int clistlen; /* the length of clist */
270: static struct line *sfile[2]; /* shortened by pruning common prefix/suffix */
271: static u_char *chrtran; /* translation table for case-folding */
272: static struct context_vec *context_vec_start;
273: static struct context_vec *context_vec_end;
274: static struct context_vec *context_vec_ptr;
275:
276: #define FUNCTION_CONTEXT_SIZE 41
1.17 ! jfb 277: static char lastbuf[FUNCTION_CONTEXT_SIZE];
! 278: static int lastline;
! 279: static int lastmatchline;
1.1 jfb 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: * cvs_diff()
343: *
344: * Handler for the `cvs diff' command.
345: *
346: * SYNOPSIS: cvs [args] diff [-clipu] [-D date] [-r rev]
347: */
348: int
349: cvs_diff(int argc, char **argv)
350: {
1.4 jfb 351: int ch, recurse, flags;
352: struct diff_arg darg;
353: struct cvsroot *root;
1.1 jfb 354:
355: context = CVS_DIFF_DEFCTX;
1.4 jfb 356: flags = CF_RECURSE|CF_IGNORE|CF_SORT|CF_KNOWN;
357: recurse = 1;
358:
359: memset(&darg, 0, sizeof(darg));
1.2 jfb 360: strlcpy(diffargs, argv[0], sizeof(diffargs));
1.1 jfb 361:
1.17 ! jfb 362: while ((ch = getopt(argc, argv, "cD:liNpr:u")) != -1) {
1.1 jfb 363: switch (ch) {
364: case 'c':
1.2 jfb 365: strlcat(diffargs, " -c", sizeof(diffargs));
1.1 jfb 366: format = D_CONTEXT;
367: break;
368: case 'D':
1.4 jfb 369: if (darg.date1 == NULL && darg.rev1 == NULL)
370: darg.date1 = optarg;
371: else if (darg.date2 == NULL && darg.rev2 == NULL)
372: darg.date2 = optarg;
1.1 jfb 373: else {
374: cvs_log(LP_ERR,
375: "no more than two revisions/dates can "
376: "be specified");
377: }
378: break;
379: case 'l':
1.2 jfb 380: strlcat(diffargs, " -l", sizeof(diffargs));
1.1 jfb 381: recurse = 0;
1.4 jfb 382: flags &= ~CF_RECURSE;
1.1 jfb 383: break;
384: case 'i':
1.2 jfb 385: strlcat(diffargs, " -i", sizeof(diffargs));
1.1 jfb 386: iflag = 1;
387: break;
1.8 jfb 388: case 'N':
389: strlcat(diffargs, " -N", sizeof(diffargs));
390: Nflag = 1;
391: break;
1.17 ! jfb 392: case 'p':
! 393: strlcat(diffargs, " -p", sizeof(diffargs));
! 394: pflag = 1;
! 395: break;
1.1 jfb 396: case 'r':
1.4 jfb 397: if ((darg.rev1 == NULL) && (darg.date1 == NULL))
398: darg.rev1 = optarg;
399: else if ((darg.rev2 == NULL) && (darg.date2 == NULL))
400: darg.rev2 = optarg;
1.1 jfb 401: else {
402: cvs_log(LP_ERR,
403: "no more than two revisions/dates can "
404: "be specified");
405: return (EX_USAGE);
406: }
407: break;
408: case 'u':
1.2 jfb 409: strlcat(diffargs, " -u", sizeof(diffargs));
1.1 jfb 410: format = D_UNIFIED;
411: break;
412: default:
413: return (EX_USAGE);
414: }
415: }
416:
417: argc -= optind;
418: argv += optind;
419:
420: if (argc == 0) {
1.6 jfb 421: cvs_files = cvs_file_get(".", flags);
1.10 deraadt 422: } else
1.6 jfb 423: cvs_files = cvs_file_getspec(argv, argc, 0);
1.7 jfb 424: if (cvs_files == NULL)
425: return (EX_DATAERR);
1.1 jfb 426:
1.6 jfb 427: cvs_file_examine(cvs_files, cvs_diff_file, &darg);
1.1 jfb 428:
1.6 jfb 429: root = cvs_files->cf_ddat->cd_root;
1.5 jfb 430: if (root->cr_method != CVS_METHOD_LOCAL) {
1.6 jfb 431: cvs_senddir(root, cvs_files);
1.4 jfb 432: cvs_sendreq(root, CVS_REQ_DIFF, NULL);
1.5 jfb 433: }
1.1 jfb 434:
1.4 jfb 435: return (0);
436: }
1.1 jfb 437:
438:
1.4 jfb 439: /*
440: * cvs_diff_sendflags()
441: *
442: */
443: int
444: cvs_diff_sendflags(struct cvsroot *root, struct diff_arg *dap)
445: {
446: /* send the flags */
1.17 ! jfb 447: if (Nflag && (cvs_sendarg(root, "-N", 0) < 0))
! 448: return (-1);
! 449: if (pflag && (cvs_sendarg(root, "-p", 0) < 0))
! 450: return (-1);
! 451:
1.4 jfb 452: if (format == D_CONTEXT)
453: cvs_sendarg(root, "-c", 0);
454: else if (format == D_UNIFIED)
455: cvs_sendarg(root, "-u", 0);
456:
457: if (dap->rev1 != NULL) {
458: cvs_sendarg(root, "-r", 0);
1.14 jfb 459: cvs_sendarg(root, dap->rev1, 0);
1.10 deraadt 460: } else if (dap->date1 != NULL) {
1.4 jfb 461: cvs_sendarg(root, "-D", 0);
1.14 jfb 462: cvs_sendarg(root, dap->date1, 0);
1.4 jfb 463: }
464: if (dap->rev2 != NULL) {
465: cvs_sendarg(root, "-r", 0);
1.14 jfb 466: cvs_sendarg(root, dap->rev2, 0);
1.10 deraadt 467: } else if (dap->date2 != NULL) {
1.4 jfb 468: cvs_sendarg(root, "-D", 0);
1.14 jfb 469: cvs_sendarg(root, dap->date2, 0);
1.1 jfb 470: }
471:
472: return (0);
473: }
474:
475:
476: /*
477: * cvs_diff_file()
478: *
479: * Diff a single file.
480: */
481: int
1.4 jfb 482: cvs_diff_file(struct cvs_file *cfp, void *arg)
1.1 jfb 483: {
1.8 jfb 484: char *dir, *repo, buf[64];
485: char fpath[MAXPATHLEN], dfpath[MAXPATHLEN], rcspath[MAXPATHLEN];
1.12 djm 486: char path_tmp1[MAXPATHLEN], path_tmp2[MAXPATHLEN];
1.1 jfb 487: BUF *b1, *b2;
488: RCSNUM *r1, *r2;
489: RCSFILE *rf;
1.4 jfb 490: struct diff_arg *dap;
1.1 jfb 491: struct cvs_ent *entp;
1.4 jfb 492: struct cvsroot *root;
1.1 jfb 493:
1.4 jfb 494: dap = (struct diff_arg *)arg;
1.1 jfb 495:
1.4 jfb 496: if (cfp->cf_type == DT_DIR) {
1.7 jfb 497: if (cfp->cf_cvstat == CVS_FST_UNKNOWN) {
498: root = cfp->cf_parent->cf_ddat->cd_root;
1.8 jfb 499: cvs_sendreq(root, CVS_REQ_QUESTIONABLE,
500: CVS_FILE_NAME(cfp));
1.10 deraadt 501: } else {
1.7 jfb 502: root = cfp->cf_ddat->cd_root;
503: if ((cfp->cf_parent == NULL) ||
504: (root != cfp->cf_parent->cf_ddat->cd_root)) {
505: cvs_connect(root);
506: cvs_diff_sendflags(root, dap);
507: }
508:
509: cvs_senddir(root, cfp);
1.4 jfb 510: }
1.1 jfb 511:
1.13 jfb 512: return (0);
513: }
514:
515: if (cfp->cf_cvstat == CVS_FST_LOST) {
516: cvs_log(LP_WARN, "cannot find file %s", CVS_FILE_NAME(cfp));
1.4 jfb 517: return (0);
1.1 jfb 518: }
519:
1.4 jfb 520: rf = NULL;
1.8 jfb 521: diff_file = cvs_file_getpath(cfp, fpath, sizeof(fpath));
1.7 jfb 522:
1.4 jfb 523: if (cfp->cf_parent != NULL) {
1.8 jfb 524: dir = cvs_file_getpath(cfp->cf_parent, dfpath, sizeof(dfpath));
1.7 jfb 525: root = cfp->cf_parent->cf_ddat->cd_root;
1.4 jfb 526: repo = cfp->cf_parent->cf_ddat->cd_repo;
1.10 deraadt 527: } else {
1.4 jfb 528: dir = ".";
1.7 jfb 529: root = NULL;
1.4 jfb 530: repo = NULL;
1.1 jfb 531: }
532:
1.4 jfb 533: if (cfp->cf_cvstat == CVS_FST_UNKNOWN) {
534: if (root->cr_method == CVS_METHOD_LOCAL)
535: cvs_log(LP_WARN, "I know nothing about %s", diff_file);
536: else
1.8 jfb 537: cvs_sendreq(root, CVS_REQ_QUESTIONABLE,
538: CVS_FILE_NAME(cfp));
1.4 jfb 539: return (0);
1.1 jfb 540: }
541:
1.4 jfb 542: entp = cvs_ent_getent(diff_file);
543: if (entp == NULL)
544: return (-1);
1.1 jfb 545:
1.4 jfb 546: if (root->cr_method != CVS_METHOD_LOCAL) {
547: if (cvs_sendentry(root, entp) < 0) {
548: cvs_ent_free(entp);
549: return (-1);
550: }
1.1 jfb 551: }
552:
1.4 jfb 553: if (cfp->cf_cvstat == CVS_FST_UPTODATE) {
554: if (root->cr_method != CVS_METHOD_LOCAL)
1.8 jfb 555: cvs_sendreq(root, CVS_REQ_UNCHANGED,
556: CVS_FILE_NAME(cfp));
1.4 jfb 557: cvs_ent_free(entp);
1.1 jfb 558: return (0);
559: }
560:
561: /* at this point, the file is modified */
1.4 jfb 562: if (root->cr_method != CVS_METHOD_LOCAL) {
1.8 jfb 563: cvs_sendreq(root, CVS_REQ_MODIFIED, CVS_FILE_NAME(cfp));
1.4 jfb 564: cvs_sendfile(root, diff_file);
1.10 deraadt 565: } else {
1.1 jfb 566: snprintf(rcspath, sizeof(rcspath), "%s/%s/%s%s",
1.4 jfb 567: root->cr_dir, repo, diff_file, RCS_FILE_EXT);
1.1 jfb 568:
569: rf = rcs_open(rcspath, RCS_MODE_READ);
1.4 jfb 570: if (rf == NULL) {
571: cvs_ent_free(entp);
1.1 jfb 572: return (-1);
1.4 jfb 573: }
1.1 jfb 574:
1.4 jfb 575: cvs_printf("Index: %s\n%s\nRCS file: %s\n", diff_file,
1.1 jfb 576: RCS_DIFF_DIV, rcspath);
577:
1.4 jfb 578: if (dap->rev1 == NULL)
1.1 jfb 579: r1 = entp->ce_rev;
580: else {
581: r1 = rcsnum_alloc();
1.4 jfb 582: rcsnum_aton(dap->rev1, NULL, r1);
1.1 jfb 583: }
584:
1.4 jfb 585: cvs_printf("retrieving revision %s\n",
1.1 jfb 586: rcsnum_tostr(r1, buf, sizeof(buf)));
587: b1 = rcs_getrev(rf, r1);
588:
1.4 jfb 589: if (dap->rev2 != NULL) {
590: cvs_printf("retrieving revision %s\n", dap->rev2);
1.1 jfb 591: r2 = rcsnum_alloc();
1.4 jfb 592: rcsnum_aton(dap->rev2, NULL, r2);
1.1 jfb 593: b2 = rcs_getrev(rf, r2);
1.10 deraadt 594: } else {
1.4 jfb 595: b2 = cvs_buf_load(diff_file, BUF_AUTOEXT);
1.1 jfb 596: }
597:
1.4 jfb 598: rcs_close(rf);
599:
1.2 jfb 600: printf("%s", diffargs);
601: printf(" -r%s", buf);
1.4 jfb 602: if (dap->rev2 != NULL)
603: printf(" -r%s", dap->rev2);
604: printf(" %s\n", diff_file);
1.12 djm 605: strlcpy(path_tmp1, "/tmp/diff1.XXXXXXXXXX", sizeof(path_tmp1));
606: if (cvs_buf_write_stmp(b1, path_tmp1, 0600) == -1)
607: return (-1);
608: strlcpy(path_tmp2, "/tmp/diff2.XXXXXXXXXX", sizeof(path_tmp1));
609: if (cvs_buf_write_stmp(b2, path_tmp2, 0600) == -1) {
610: (void)unlink(path_tmp1);
611: return (-1);
612: }
613: cvs_diffreg(path_tmp1, path_tmp2);
614: (void)unlink(path_tmp1);
615: (void)unlink(path_tmp2);
1.1 jfb 616: }
617:
1.4 jfb 618: cvs_ent_free(entp);
1.1 jfb 619: return (0);
620: }
621:
622:
623: int
624: cvs_diffreg(const char *file1, const char *file2)
625: {
626: FILE *f1, *f2;
627: int i, rval;
628:
629: f1 = f2 = NULL;
630: rval = D_SAME;
631: anychange = 0;
632: lastline = 0;
633: lastmatchline = 0;
634: context_vec_ptr = context_vec_start - 1;
635: chrtran = (iflag ? cup2low : clow2low);
636:
637: f1 = fopen(file1, "r");
638: if (f1 == NULL) {
639: cvs_log(LP_ERRNO, "%s", file1);
640: status |= 2;
641: goto closem;
642: }
643:
644: f2 = fopen(file2, "r");
645: if (f2 == NULL) {
646: cvs_log(LP_ERRNO, "%s", file2);
647: status |= 2;
648: goto closem;
649: }
650:
651: switch (files_differ(f1, f2)) {
652: case 0:
653: goto closem;
654: case 1:
655: break;
656: default:
657: /* error */
658: status |= 2;
659: goto closem;
660: }
661:
662: if (!asciifile(f1) || !asciifile(f2)) {
663: rval = D_BINARY;
664: status |= 1;
665: goto closem;
666: }
667: prepare(0, f1, stb1.st_size);
668: prepare(1, f2, stb2.st_size);
669: prune();
670: sort(sfile[0], slen[0]);
671: sort(sfile[1], slen[1]);
672:
673: member = (int *)file[1];
674: equiv(sfile[0], slen[0], sfile[1], slen[1], member);
675: member = realloc(member, (slen[1] + 2) * sizeof(int));
676:
677: class = (int *)file[0];
678: unsort(sfile[0], slen[0], class);
679: class = realloc(class, (slen[0] + 2) * sizeof(int));
680:
681: klist = malloc((slen[0] + 2) * sizeof(int));
682: clen = 0;
683: clistlen = 100;
684: clist = malloc(clistlen * sizeof(cand));
685: i = stone(class, slen[0], member, klist);
686: free(member);
687: free(class);
688:
689: J = realloc(J, (len[0] + 2) * sizeof(int));
690: unravel(klist[i]);
691: free(clist);
692: free(klist);
693:
694: ixold = realloc(ixold, (len[0] + 2) * sizeof(long));
695: ixnew = realloc(ixnew, (len[1] + 2) * sizeof(long));
696: check(f1, f2);
697: output(file1, f1, file2, f2);
698:
699: closem:
700: if (anychange) {
701: status |= 1;
702: if (rval == D_SAME)
703: rval = D_DIFFER;
704: }
705: if (f1 != NULL)
706: fclose(f1);
707: if (f2 != NULL)
708: fclose(f2);
709: return (rval);
710: }
711:
712: /*
713: * Check to see if the given files differ.
714: * Returns 0 if they are the same, 1 if different, and -1 on error.
715: * XXX - could use code from cmp(1) [faster]
716: */
717: static int
718: files_differ(FILE *f1, FILE *f2)
719: {
720: char buf1[BUFSIZ], buf2[BUFSIZ];
721: size_t i, j;
722:
723: if (stb1.st_size != stb2.st_size)
724: return (1);
725: for (;;) {
726: i = fread(buf1, 1, sizeof(buf1), f1);
727: j = fread(buf2, 1, sizeof(buf2), f2);
728: if (i != j)
729: return (1);
730: if (i == 0 && j == 0) {
731: if (ferror(f1) || ferror(f2))
732: return (1);
733: return (0);
734: }
735: if (memcmp(buf1, buf2, i) != 0)
736: return (1);
737: }
738: }
739:
740:
741: char *
1.15 jfb 742: splice(char *dir, char *filename)
1.1 jfb 743: {
744: char *tail, *buf;
745:
1.15 jfb 746: if ((tail = strrchr(filename, '/')) == NULL)
747: tail = filename;
1.1 jfb 748: else
749: tail++;
750: asprintf(&buf, "%s/%s", dir, tail);
751: return (buf);
752: }
753:
754: static void
755: prepare(int i, FILE *fd, off_t filesize)
756: {
757: struct line *p;
758: int j, h;
759: size_t sz;
760:
761: rewind(fd);
762:
1.16 xsa 763: sz = ((size_t)filesize <= SIZE_MAX ? (size_t)filesize : SIZE_MAX) / 25;
1.1 jfb 764: if (sz < 100)
765: sz = 100;
766:
767: p = malloc((sz + 3) * sizeof(struct line));
768: for (j = 0; (h = readhash(fd));) {
769: if (j == (int)sz) {
770: sz = sz * 3 / 2;
771: p = realloc(p, (sz + 3) * sizeof(struct line));
772: }
773: p[++j].value = h;
774: }
775: len[i] = j;
776: file[i] = p;
777: }
778:
779: static void
780: prune(void)
781: {
782: int i, j;
783:
784: for (pref = 0; pref < len[0] && pref < len[1] &&
785: file[0][pref + 1].value == file[1][pref + 1].value;
786: pref++)
787: ;
788: for (suff = 0; suff < len[0] - pref && suff < len[1] - pref &&
789: file[0][len[0] - suff].value == file[1][len[1] - suff].value;
790: suff++)
791: ;
792: for (j = 0; j < 2; j++) {
793: sfile[j] = file[j] + pref;
794: slen[j] = len[j] - pref - suff;
795: for (i = 0; i <= slen[j]; i++)
796: sfile[j][i].serial = i;
797: }
798: }
799:
800: static void
801: equiv(struct line *a, int n, struct line *b, int m, int *c)
802: {
803: int i, j;
804:
805: i = j = 1;
806: while (i <= n && j <= m) {
807: if (a[i].value < b[j].value)
808: a[i++].value = 0;
809: else if (a[i].value == b[j].value)
810: a[i++].value = j;
811: else
812: j++;
813: }
814: while (i <= n)
815: a[i++].value = 0;
816: b[m + 1].value = 0;
817: j = 0;
818: while (++j <= m) {
819: c[j] = -b[j].serial;
820: while (b[j + 1].value == b[j].value) {
821: j++;
822: c[j] = b[j].serial;
823: }
824: }
825: c[j] = -1;
826: }
827:
828: /* Code taken from ping.c */
829: static int
830: isqrt(int n)
831: {
832: int y, x = 1;
833:
834: if (n == 0)
835: return(0);
836:
837: do { /* newton was a stinker */
838: y = x;
839: x = n / x;
840: x += y;
841: x /= 2;
842: } while ((x - y) > 1 || (x - y) < -1);
843:
844: return (x);
845: }
846:
847: static int
848: stone(int *a, int n, int *b, int *c)
849: {
850: int i, k, y, j, l;
851: int oldc, tc, oldl;
852: u_int numtries;
853:
1.4 jfb 854: const u_int bound = dflag ? UINT_MAX : MAX(256, isqrt(n));
1.1 jfb 855:
856: k = 0;
857: c[0] = newcand(0, 0, 0);
858: for (i = 1; i <= n; i++) {
859: j = a[i];
860: if (j == 0)
861: continue;
862: y = -b[j];
863: oldl = 0;
864: oldc = c[0];
865: numtries = 0;
866: do {
867: if (y <= clist[oldc].y)
868: continue;
869: l = search(c, k, y);
870: if (l != oldl + 1)
871: oldc = c[l - 1];
872: if (l <= k) {
873: if (clist[c[l]].y <= y)
874: continue;
875: tc = c[l];
876: c[l] = newcand(i, y, oldc);
877: oldc = tc;
878: oldl = l;
879: numtries++;
880: } else {
881: c[l] = newcand(i, y, oldc);
882: k++;
883: break;
884: }
885: } while ((y = b[++j]) > 0 && numtries < bound);
886: }
887: return (k);
888: }
889:
890: static int
891: newcand(int x, int y, int pred)
892: {
893: struct cand *q;
894:
895: if (clen == clistlen) {
896: clistlen = clistlen * 11 / 10;
897: clist = realloc(clist, clistlen * sizeof(cand));
898: }
899: q = clist + clen;
900: q->x = x;
901: q->y = y;
902: q->pred = pred;
903: return (clen++);
904: }
905:
906: static int
907: search(int *c, int k, int y)
908: {
909: int i, j, l, t;
910:
911: if (clist[c[k]].y < y) /* quick look for typical case */
912: return (k + 1);
913: i = 0;
914: j = k + 1;
915: while (1) {
916: l = i + j;
917: if ((l >>= 1) <= i)
918: break;
919: t = clist[c[l]].y;
920: if (t > y)
921: j = l;
922: else if (t < y)
923: i = l;
924: else
925: return (l);
926: }
927: return (l + 1);
928: }
929:
930: static void
931: unravel(int p)
932: {
933: struct cand *q;
934: int i;
935:
936: for (i = 0; i <= len[0]; i++)
937: J[i] = i <= pref ? i :
938: i > len[0] - suff ? i + len[1] - len[0] : 0;
939: for (q = clist + p; q->y != 0; q = clist + q->pred)
940: J[q->x + pref] = q->y + pref;
941: }
942:
943: /*
944: * Check does double duty:
945: * 1. ferret out any fortuitous correspondences due
946: * to confounding by hashing (which result in "jackpot")
947: * 2. collect random access indexes to the two files
948: */
949: static void
950: check(FILE *f1, FILE *f2)
951: {
952: int i, j, jackpot, c, d;
953: long ctold, ctnew;
954:
955: rewind(f1);
956: rewind(f2);
957: j = 1;
958: ixold[0] = ixnew[0] = 0;
959: jackpot = 0;
960: ctold = ctnew = 0;
961: for (i = 1; i <= len[0]; i++) {
962: if (J[i] == 0) {
963: ixold[i] = ctold += skipline(f1);
964: continue;
965: }
966: while (j < J[i]) {
967: ixnew[j] = ctnew += skipline(f2);
968: j++;
969: }
970: if (bflag || wflag || iflag) {
971: for (;;) {
972: c = getc(f1);
973: d = getc(f2);
974: /*
975: * GNU diff ignores a missing newline
976: * in one file if bflag || wflag.
977: */
978: if ((bflag || wflag) &&
979: ((c == EOF && d == '\n') ||
980: (c == '\n' && d == EOF))) {
981: break;
982: }
983: ctold++;
984: ctnew++;
985: if (bflag && isspace(c) && isspace(d)) {
986: do {
987: if (c == '\n')
988: break;
989: ctold++;
990: } while (isspace(c = getc(f1)));
991: do {
992: if (d == '\n')
993: break;
994: ctnew++;
995: } while (isspace(d = getc(f2)));
996: } else if (wflag) {
997: while (isspace(c) && c != '\n') {
998: c = getc(f1);
999: ctold++;
1000: }
1001: while (isspace(d) && d != '\n') {
1002: d = getc(f2);
1003: ctnew++;
1004: }
1005: }
1006: if (chrtran[c] != chrtran[d]) {
1007: jackpot++;
1008: J[i] = 0;
1009: if (c != '\n' && c != EOF)
1010: ctold += skipline(f1);
1011: if (d != '\n' && c != EOF)
1012: ctnew += skipline(f2);
1013: break;
1014: }
1015: if (c == '\n' || c == EOF)
1016: break;
1017: }
1018: } else {
1019: for (;;) {
1020: ctold++;
1021: ctnew++;
1022: if ((c = getc(f1)) != (d = getc(f2))) {
1023: /* jackpot++; */
1024: J[i] = 0;
1025: if (c != '\n' && c != EOF)
1026: ctold += skipline(f1);
1027: if (d != '\n' && c != EOF)
1028: ctnew += skipline(f2);
1029: break;
1030: }
1031: if (c == '\n' || c == EOF)
1032: break;
1033: }
1034: }
1035: ixold[i] = ctold;
1036: ixnew[j] = ctnew;
1037: j++;
1038: }
1039: for (; j <= len[1]; j++)
1040: ixnew[j] = ctnew += skipline(f2);
1041: /*
1042: * if (jackpot)
1043: * fprintf(stderr, "jackpot\n");
1044: */
1045: }
1046:
1047: /* shellsort CACM #201 */
1048: static void
1049: sort(struct line *a, int n)
1050: {
1051: struct line *ai, *aim, w;
1052: int j, m = 0, k;
1053:
1054: if (n == 0)
1055: return;
1056: for (j = 1; j <= n; j *= 2)
1057: m = 2 * j - 1;
1058: for (m /= 2; m != 0; m /= 2) {
1059: k = n - m;
1060: for (j = 1; j <= k; j++) {
1061: for (ai = &a[j]; ai > a; ai -= m) {
1062: aim = &ai[m];
1063: if (aim < ai)
1064: break; /* wraparound */
1065: if (aim->value > ai[0].value ||
1066: (aim->value == ai[0].value &&
1067: aim->serial > ai[0].serial))
1068: break;
1069: w.value = ai[0].value;
1070: ai[0].value = aim->value;
1071: aim->value = w.value;
1072: w.serial = ai[0].serial;
1073: ai[0].serial = aim->serial;
1074: aim->serial = w.serial;
1075: }
1076: }
1077: }
1078: }
1079:
1080: static void
1081: unsort(struct line *f, int l, int *b)
1082: {
1083: int *a, i;
1084:
1085: a = malloc((l + 1) * sizeof(int));
1086: for (i = 1; i <= l; i++)
1087: a[f[i].serial] = f[i].value;
1088: for (i = 1; i <= l; i++)
1089: b[i] = a[i];
1090: free(a);
1091: }
1092:
1093: static int
1094: skipline(FILE *f)
1095: {
1096: int i, c;
1097:
1098: for (i = 1; (c = getc(f)) != '\n' && c != EOF; i++)
1099: continue;
1100: return (i);
1101: }
1102:
1103: static void
1104: output(const char *file1, FILE *f1, const char *file2, FILE *f2)
1105: {
1106: int m, i0, i1, j0, j1;
1107:
1108: rewind(f1);
1109: rewind(f2);
1110: m = len[0];
1111: J[0] = 0;
1112: J[m + 1] = len[1] + 1;
1113: for (i0 = 1; i0 <= m; i0 = i1 + 1) {
1114: while (i0 <= m && J[i0] == J[i0 - 1] + 1)
1115: i0++;
1116: j0 = J[i0 - 1] + 1;
1117: i1 = i0 - 1;
1118: while (i1 < m && J[i1 + 1] == 0)
1119: i1++;
1120: j1 = J[i1 + 1] - 1;
1121: J[i1] = j1;
1122: change(file1, f1, file2, f2, i0, i1, j0, j1);
1123: }
1124: if (m == 0)
1125: change(file1, f1, file2, f2, 1, 0, 1, len[1]);
1126: if (format == D_IFDEF) {
1127: for (;;) {
1128: #define c i0
1129: if ((c = getc(f1)) == EOF)
1130: return;
1131: putchar(c);
1132: }
1133: #undef c
1134: }
1135: if (anychange != 0) {
1136: if (format == D_CONTEXT)
1137: dump_context_vec(f1, f2);
1138: else if (format == D_UNIFIED)
1139: dump_unified_vec(f1, f2);
1140: }
1141: }
1142:
1143: static __inline void
1144: range(int a, int b, char *separator)
1145: {
1146: printf("%d", a > b ? b : a);
1147: if (a < b)
1148: printf("%s%d", separator, b);
1149: }
1150:
1151: static __inline void
1152: uni_range(int a, int b)
1153: {
1154: if (a < b)
1155: printf("%d,%d", a, b - a + 1);
1156: else if (a == b)
1157: printf("%d", b);
1158: else
1159: printf("%d,0", b);
1160: }
1161:
1162: static char *
1.15 jfb 1163: preadline(int fd, size_t rlen, off_t off)
1.1 jfb 1164: {
1165: char *line;
1166: ssize_t nr;
1167:
1.15 jfb 1168: line = malloc(rlen + 1);
1.1 jfb 1169: if (line == NULL) {
1170: cvs_log(LP_ERRNO, "failed to allocate line");
1171: return (NULL);
1172: }
1.15 jfb 1173: if ((nr = pread(fd, line, rlen, off)) < 0) {
1.1 jfb 1174: cvs_log(LP_ERRNO, "preadline failed");
1175: return (NULL);
1176: }
1177: line[nr] = '\0';
1178: return (line);
1179: }
1180:
1181: static int
1182: ignoreline(char *line)
1183: {
1184: int ret;
1185:
1186: ret = regexec(&ignore_re, line, 0, NULL, 0);
1187: free(line);
1188: return (ret == 0); /* if it matched, it should be ignored. */
1189: }
1190:
1191: /*
1192: * Indicate that there is a difference between lines a and b of the from file
1193: * to get to lines c to d of the to file. If a is greater then b then there
1194: * are no lines in the from file involved and this means that there were
1195: * lines appended (beginning at b). If c is greater than d then there are
1196: * lines missing from the to file.
1197: */
1198: static void
1199: change(const char *file1, FILE *f1, const char *file2, FILE *f2,
1200: int a, int b, int c, int d)
1201: {
1202: static size_t max_context = 64;
1203: int i;
1204:
1205: if (format != D_IFDEF && a > b && c > d)
1206: return;
1207: if (ignore_pats != NULL) {
1208: char *line;
1209: /*
1210: * All lines in the change, insert, or delete must
1211: * match an ignore pattern for the change to be
1212: * ignored.
1213: */
1214: if (a <= b) { /* Changes and deletes. */
1215: for (i = a; i <= b; i++) {
1216: line = preadline(fileno(f1),
1217: ixold[i] - ixold[i - 1], ixold[i - 1]);
1218: if (!ignoreline(line))
1219: goto proceed;
1220: }
1221: }
1222: if (a > b || c <= d) { /* Changes and inserts. */
1223: for (i = c; i <= d; i++) {
1224: line = preadline(fileno(f2),
1225: ixnew[i] - ixnew[i - 1], ixnew[i - 1]);
1226: if (!ignoreline(line))
1227: goto proceed;
1228: }
1229: }
1230: return;
1231: }
1232: proceed:
1233: if (format == D_CONTEXT || format == D_UNIFIED) {
1234: /*
1235: * Allocate change records as needed.
1236: */
1237: if (context_vec_ptr == context_vec_end - 1) {
1238: ptrdiff_t offset = context_vec_ptr - context_vec_start;
1239: max_context <<= 1;
1240: context_vec_start = realloc(context_vec_start,
1241: max_context * sizeof(struct context_vec));
1242: context_vec_end = context_vec_start + max_context;
1243: context_vec_ptr = context_vec_start + offset;
1244: }
1245: if (anychange == 0) {
1246: /*
1247: * Print the context/unidiff header first time through.
1248: */
1249: printf("%s %s %s",
1250: format == D_CONTEXT ? "***" : "---", diff_file,
1251: ctime(&stb1.st_mtime));
1252: printf("%s %s %s",
1253: format == D_CONTEXT ? "---" : "+++", diff_file,
1254: ctime(&stb2.st_mtime));
1255: anychange = 1;
1256: } else if (a > context_vec_ptr->b + (2 * context) + 1 &&
1257: c > context_vec_ptr->d + (2 * context) + 1) {
1258: /*
1259: * If this change is more than 'context' lines from the
1260: * previous change, dump the record and reset it.
1261: */
1262: if (format == D_CONTEXT)
1263: dump_context_vec(f1, f2);
1264: else
1265: dump_unified_vec(f1, f2);
1266: }
1267: context_vec_ptr++;
1268: context_vec_ptr->a = a;
1269: context_vec_ptr->b = b;
1270: context_vec_ptr->c = c;
1271: context_vec_ptr->d = d;
1272: return;
1273: }
1274: if (anychange == 0)
1275: anychange = 1;
1276: switch (format) {
1277: case D_BRIEF:
1278: return;
1279: case D_NORMAL:
1280: range(a, b, ",");
1281: putchar(a > b ? 'a' : c > d ? 'd' : 'c');
1282: if (format == D_NORMAL)
1283: range(c, d, ",");
1284: putchar('\n');
1285: break;
1286: }
1287: if (format == D_NORMAL || format == D_IFDEF) {
1288: fetch(ixold, a, b, f1, '<', 1);
1289: if (a <= b && c <= d && format == D_NORMAL)
1290: puts("---");
1291: }
1292: i = fetch(ixnew, c, d, f2, format == D_NORMAL ? '>' : '\0', 0);
1293: if (inifdef) {
1294: printf("#endif /* %s */\n", ifdefname);
1295: inifdef = 0;
1296: }
1297: }
1298:
1299: static int
1300: fetch(long *f, int a, int b, FILE *lb, int ch, int oldfile)
1301: {
1302: int i, j, c, lastc, col, nc;
1303:
1304: /*
1305: * When doing #ifdef's, copy down to current line
1306: * if this is the first file, so that stuff makes it to output.
1307: */
1308: if (format == D_IFDEF && oldfile) {
1309: long curpos = ftell(lb);
1310: /* print through if append (a>b), else to (nb: 0 vs 1 orig) */
1311: nc = f[a > b ? b : a - 1] - curpos;
1312: for (i = 0; i < nc; i++)
1313: putchar(getc(lb));
1314: }
1315: if (a > b)
1316: return (0);
1317: if (format == D_IFDEF) {
1318: if (inifdef) {
1319: printf("#else /* %s%s */\n",
1320: oldfile == 1 ? "!" : "", ifdefname);
1321: } else {
1322: if (oldfile)
1323: printf("#ifndef %s\n", ifdefname);
1324: else
1325: printf("#ifdef %s\n", ifdefname);
1326: }
1327: inifdef = 1 + oldfile;
1328: }
1329: for (i = a; i <= b; i++) {
1330: fseek(lb, f[i - 1], SEEK_SET);
1331: nc = f[i] - f[i - 1];
1332: if (format != D_IFDEF && ch != '\0') {
1333: putchar(ch);
1334: if (Tflag && (format == D_NORMAL || format == D_CONTEXT
1335: || format == D_UNIFIED))
1336: putchar('\t');
1337: else if (format != D_UNIFIED)
1338: putchar(' ');
1339: }
1340: col = 0;
1341: for (j = 0, lastc = '\0'; j < nc; j++, lastc = c) {
1342: if ((c = getc(lb)) == EOF) {
1343: puts("\n\\ No newline at end of file");
1344: return (0);
1345: }
1346: if (c == '\t' && tflag) {
1347: do {
1348: putchar(' ');
1349: } while (++col & 7);
1350: } else {
1351: putchar(c);
1352: col++;
1353: }
1354: }
1355: }
1356: return (0);
1357: }
1358:
1359: /*
1360: * Hash function taken from Robert Sedgewick, Algorithms in C, 3d ed., p 578.
1361: */
1362: static int
1363: readhash(FILE *f)
1364: {
1365: int i, t, space;
1366: int sum;
1367:
1368: sum = 1;
1369: space = 0;
1370: if (!bflag && !wflag) {
1371: if (iflag)
1372: for (i = 0; (t = getc(f)) != '\n'; i++) {
1373: if (t == EOF) {
1374: if (i == 0)
1375: return (0);
1376: break;
1377: }
1378: sum = sum * 127 + chrtran[t];
1379: }
1380: else
1381: for (i = 0; (t = getc(f)) != '\n'; i++) {
1382: if (t == EOF) {
1383: if (i == 0)
1384: return (0);
1385: break;
1386: }
1387: sum = sum * 127 + t;
1388: }
1389: } else {
1390: for (i = 0;;) {
1391: switch (t = getc(f)) {
1392: case '\t':
1393: case ' ':
1394: space++;
1395: continue;
1396: default:
1397: if (space && !wflag) {
1398: i++;
1399: space = 0;
1400: }
1401: sum = sum * 127 + chrtran[t];
1402: i++;
1403: continue;
1404: case EOF:
1405: if (i == 0)
1406: return (0);
1407: /* FALLTHROUGH */
1408: case '\n':
1409: break;
1410: }
1411: break;
1412: }
1413: }
1414: /*
1415: * There is a remote possibility that we end up with a zero sum.
1416: * Zero is used as an EOF marker, so return 1 instead.
1417: */
1418: return (sum == 0 ? 1 : sum);
1419: }
1420:
1421: static int
1422: asciifile(FILE *f)
1423: {
1424: char buf[BUFSIZ];
1425: int i, cnt;
1426:
1427: if (aflag || f == NULL)
1428: return (1);
1429:
1430: rewind(f);
1431: cnt = fread(buf, 1, sizeof(buf), f);
1432: for (i = 0; i < cnt; i++)
1433: if (!isprint(buf[i]) && !isspace(buf[i]))
1434: return (0);
1435: return (1);
1436: }
1437:
1.17 ! jfb 1438: static char*
! 1439: match_function(const long *f, int pos, FILE *fp)
! 1440: {
! 1441: unsigned char buf[FUNCTION_CONTEXT_SIZE];
! 1442: size_t nc;
! 1443: int last = lastline;
! 1444: char *p;
! 1445:
! 1446: lastline = pos;
! 1447: while (pos > last) {
! 1448: fseek(fp, f[pos - 1], SEEK_SET);
! 1449: nc = f[pos] - f[pos - 1];
! 1450: if (nc >= sizeof(buf))
! 1451: nc = sizeof(buf) - 1;
! 1452: nc = fread(buf, 1, nc, fp);
! 1453: if (nc > 0) {
! 1454: buf[nc] = '\0';
! 1455: p = strchr(buf, '\n');
! 1456: if (p != NULL)
! 1457: *p = '\0';
! 1458: if (isalpha(buf[0]) || buf[0] == '_' || buf[0] == '$') {
! 1459: strlcpy(lastbuf, buf, sizeof lastbuf);
! 1460: lastmatchline = pos;
! 1461: return lastbuf;
! 1462: }
! 1463: }
! 1464: pos--;
! 1465: }
! 1466: return (lastmatchline > 0) ? lastbuf : NULL;
! 1467: }
! 1468:
1.1 jfb 1469:
1470: /* dump accumulated "context" diff changes */
1471: static void
1472: dump_context_vec(FILE *f1, FILE *f2)
1473: {
1474: struct context_vec *cvp = context_vec_start;
1475: int lowa, upb, lowc, upd, do_output;
1476: int a, b, c, d;
1.17 ! jfb 1477: char ch, *f;
1.1 jfb 1478:
1479: if (context_vec_start > context_vec_ptr)
1480: return;
1481:
1482: b = d = 0; /* gcc */
1.4 jfb 1483: lowa = MAX(1, cvp->a - context);
1484: upb = MIN(len[0], context_vec_ptr->b + context);
1485: lowc = MAX(1, cvp->c - context);
1486: upd = MIN(len[1], context_vec_ptr->d + context);
1.1 jfb 1487:
1488: printf("***************");
1.17 ! jfb 1489: if (pflag) {
! 1490: f = match_function(ixold, lowa - 1, f1);
! 1491: if (f != NULL) {
! 1492: putchar(' ');
! 1493: fputs(f, stdout);
! 1494: }
! 1495: }
1.1 jfb 1496: printf("\n*** ");
1497: range(lowa, upb, ",");
1498: printf(" ****\n");
1499:
1500: /*
1501: * Output changes to the "old" file. The first loop suppresses
1502: * output if there were no changes to the "old" file (we'll see
1503: * the "old" lines as context in the "new" list).
1504: */
1505: do_output = 0;
1506: for (; cvp <= context_vec_ptr; cvp++)
1507: if (cvp->a <= cvp->b) {
1508: cvp = context_vec_start;
1509: do_output++;
1510: break;
1511: }
1512: if (do_output) {
1513: while (cvp <= context_vec_ptr) {
1514: a = cvp->a;
1515: b = cvp->b;
1516: c = cvp->c;
1517: d = cvp->d;
1518:
1519: if (a <= b && c <= d)
1520: ch = 'c';
1521: else
1522: ch = (a <= b) ? 'd' : 'a';
1523:
1524: if (ch == 'a')
1525: fetch(ixold, lowa, b, f1, ' ', 0);
1526: else {
1527: fetch(ixold, lowa, a - 1, f1, ' ', 0);
1528: fetch(ixold, a, b, f1,
1529: ch == 'c' ? '!' : '-', 0);
1530: }
1531: lowa = b + 1;
1532: cvp++;
1533: }
1534: fetch(ixold, b + 1, upb, f1, ' ', 0);
1535: }
1536: /* output changes to the "new" file */
1537: printf("--- ");
1538: range(lowc, upd, ",");
1539: printf(" ----\n");
1540:
1541: do_output = 0;
1542: for (cvp = context_vec_start; cvp <= context_vec_ptr; cvp++)
1543: if (cvp->c <= cvp->d) {
1544: cvp = context_vec_start;
1545: do_output++;
1546: break;
1547: }
1548: if (do_output) {
1549: while (cvp <= context_vec_ptr) {
1550: a = cvp->a;
1551: b = cvp->b;
1552: c = cvp->c;
1553: d = cvp->d;
1554:
1555: if (a <= b && c <= d)
1556: ch = 'c';
1557: else
1558: ch = (a <= b) ? 'd' : 'a';
1559:
1560: if (ch == 'd')
1561: fetch(ixnew, lowc, d, f2, ' ', 0);
1562: else {
1563: fetch(ixnew, lowc, c - 1, f2, ' ', 0);
1564: fetch(ixnew, c, d, f2,
1565: ch == 'c' ? '!' : '+', 0);
1566: }
1567: lowc = d + 1;
1568: cvp++;
1569: }
1570: fetch(ixnew, d + 1, upd, f2, ' ', 0);
1571: }
1572: context_vec_ptr = context_vec_start - 1;
1573: }
1574:
1575: /* dump accumulated "unified" diff changes */
1576: static void
1577: dump_unified_vec(FILE *f1, FILE *f2)
1578: {
1579: struct context_vec *cvp = context_vec_start;
1580: int lowa, upb, lowc, upd;
1581: int a, b, c, d;
1.17 ! jfb 1582: char ch, *f;
1.1 jfb 1583:
1584: if (context_vec_start > context_vec_ptr)
1585: return;
1586:
1587: b = d = 0; /* gcc */
1.4 jfb 1588: lowa = MAX(1, cvp->a - context);
1589: upb = MIN(len[0], context_vec_ptr->b + context);
1590: lowc = MAX(1, cvp->c - context);
1591: upd = MIN(len[1], context_vec_ptr->d + context);
1.1 jfb 1592:
1593: fputs("@@ -", stdout);
1594: uni_range(lowa, upb);
1595: fputs(" +", stdout);
1596: uni_range(lowc, upd);
1597: fputs(" @@", stdout);
1.17 ! jfb 1598: if (pflag) {
! 1599: f = match_function(ixold, lowa - 1, f1);
! 1600: if (f != NULL) {
! 1601: putchar(' ');
! 1602: fputs(f, stdout);
! 1603: }
! 1604: }
1.1 jfb 1605: putchar('\n');
1606:
1607: /*
1608: * Output changes in "unified" diff format--the old and new lines
1609: * are printed together.
1610: */
1611: for (; cvp <= context_vec_ptr; cvp++) {
1612: a = cvp->a;
1613: b = cvp->b;
1614: c = cvp->c;
1615: d = cvp->d;
1616:
1617: /*
1618: * c: both new and old changes
1619: * d: only changes in the old file
1620: * a: only changes in the new file
1621: */
1622: if (a <= b && c <= d)
1623: ch = 'c';
1624: else
1625: ch = (a <= b) ? 'd' : 'a';
1626:
1627: switch (ch) {
1628: case 'c':
1629: fetch(ixold, lowa, a - 1, f1, ' ', 0);
1630: fetch(ixold, a, b, f1, '-', 0);
1631: fetch(ixnew, c, d, f2, '+', 0);
1632: break;
1633: case 'd':
1634: fetch(ixold, lowa, a - 1, f1, ' ', 0);
1635: fetch(ixold, a, b, f1, '-', 0);
1636: break;
1637: case 'a':
1638: fetch(ixnew, lowc, c - 1, f2, ' ', 0);
1639: fetch(ixnew, c, d, f2, '+', 0);
1640: break;
1641: }
1642: lowa = b + 1;
1643: lowc = d + 1;
1644: }
1645: fetch(ixnew, d + 1, upd, f2, ' ', 0);
1646:
1647: context_vec_ptr = context_vec_start - 1;
1648: }