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