4.3reno diff. This is free because of the Caldera license. Nasty, but a

place to start.

6 files changed, 2190 insertions, 0 deletions

diff --git a/usr.bin/diff/diff.1 b/usr.bin/diff/diff.1
new file mode 100644
index 00000000000..e570bcf9c03
--- /dev/null
+++ b/usr.bin/diff/diff.1
@@ -0,0 +1,402 @@
+.\" Copyright (c) 1980, 1990 The Regents of the University of California.
+.\" All rights reserved.
+.\"
+.\" Redistribution and use in source and binary forms are permitted provided
+.\" that: (1) source distributions retain this entire copyright notice and
+.\" comment, and (2) distributions including binaries display the following
+.\" acknowledgement:  ``This product includes software developed by the
+.\" University of California, Berkeley and its contributors'' in the
+.\" documentation or other materials provided with the distribution and in
+.\" all advertising materials mentioning features or use of this software.
+.\" Neither the name of the University nor the names of its contributors may
+.\" be used to endorse or promote products derived from this software without
+.\" specific prior written permission.
+.\" THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR IMPLIED
+.\" WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF
+.\" MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
+.\"
+.\"     @(#)diff.1	6.6 (Berkeley) 7/24/90
+.\"
+.Dd July 24, 1990
+.Dt DIFF 1
+.Os BSD 4
+.Sh NAME
+.Nm diff
+.Nd differential file and directory comparator
+.Sh SYNOPSIS
+.Nm diff
+.Op Fl cefhn
+.Op Fl biwt
+.Ar file1 file2
+.Nm diff
+.Op Fl D Ar string
+.Op Fl biw
+.Ar file1 file2
+.Nm diff
+.Op Fl l
+.Op Fl r
+.Op Fl s
+.Op Fl cefhn
+.Op Fl biwt
+.Ar dir1 dir2
+.Sh DESCRIPTION
+The
+.Nm diff
+utility compares the contents of
+.Ar file1
+and
+.Ar file2
+and writes to the standard output the list of changes necessary to
+convert one file into the other.
+No output is produced if the files are identical.
+.Pp
+Output options (mutually exclusive):
+.Tw Fl
+.Tp Fl c
+produces a diff with lines of context.
+The default is to present 3 lines of context and may be changed, e.g to 10, by
+.Fl c10 .
+With
+.Fl c
+the output format is modified slightly:
+the output beginning with identification of the files involved and
+their creation dates and then each change is separated
+by a line with a dozen *'s.
+The lines removed from
+.Ar file1
+are marked with `\(mi '; those added to
+.Ar file2
+are marked `+ '.  Lines which are changed from one
+file to the other are marked in both files with with `! '.
+Changes which lie within <context> lines of each other are grouped
+together on output.  (This is a change from the previous ``diff -c''
+but the resulting output is usually much easier to interpret.)
+.Tp Fl e
+produces output in a form suitable as input for the editor utility,
+.Xr ed 1 ,
+which can then be used to convert file1 into file2.
+.Pp
+Extra commands are added to the output when comparing directories with
+.Fl e ,
+so that the result is a
+.Xr sh  1
+script for converting text files which are common to the two directories
+from their state in
+.Ar dir1
+to their state in
+.Ar dir2 .
+.Tp Fl f
+identical output of the
+.Fl e
+flag, but in reverse order.  It cannot
+be digested by
+.Xr ed 1 .
+.Tp Fl h
+Invokes an alternate algorithm which can handle files of very long lengths.
+There is a trade off. The algorithm can only deal with changes which are
+clearly delimited and brief. Long sections of changes and overlaps will
+confuse it.
+.Tp Fl n
+produces a script similar to that of
+.Fl e ,
+but in the opposite order and with a count of changed lines on each
+insert or delete command.  This is the form used by
+.Xr rcsdiff  1  .
+.Tc Fl D
+.Ar string
+.Cx
+creates a merged version of
+.Ar file1
+and
+.Ar file2
+on the standard output, with C preprocessor controls included so that
+a compilation of the result without defining
+.Ar string
+is equivalent
+to compiling
+.Ar file1 ,
+while defining
+.Ar string
+will yield
+.Ar file2 .
+.Tp
+.Pp
+Comparison options:
+.Tp Fl b
+causes trailing blanks (spaces and tabs) to be ignored, and other
+strings of blanks to compare equal.
+.Tp Fl i
+ignores the case of letters.  E.g., ``A'' will compare equal to ``a''.
+.Tp Fl t
+will expand tabs in output lines.  Normal or
+.Fl c
+output adds character(s) to the front of each line which may screw up
+the indentation of the original source lines and make the output listing
+difficult to interpret.  This option will preserve the original source's
+indentation.
+.Tp Fl w
+is similar to
+.Fl b
+but causes whitespace (blanks and tabs) to be totally ignored.  E.g.,
+``if\ (\ a\ ==\ b\ )'' will compare equal to ``if(a==b)''.
+.Tp
+.Pp
+Directory comparison options:
+.Tw Fl
+.Tp Fl l
+long output format; each text file
+.Sf Nm diff \'d
+is piped through
+.Xr pr  1
+to paginate it,
+other differences are remembered and summarized
+after all text file differences are reported.
+.Tp Fl r
+causes application of
+.Nm diff
+recursively to common subdirectories encountered.
+.Tp Fl s
+causes
+.Nm diff
+to report files which are the same, which are otherwise not mentioned.
+.Tc Fl S
+.Ar name
+.Cx
+re-starts a directory
+.Nm diff
+in the middle beginning with file
+.Ar name .
+.Tp
+.Pp
+If both arguments are directories,
+.Nm diff
+sorts the contents of the directories by name, and then runs the
+regular file
+.Nm diff
+algorithm, producing a change list,
+on text files which are different.
+Binary files which differ,
+common subdirectories, and files which appear in only one directory
+are described as such.
+.Pp
+If only one of
+.Ar file1
+and
+.Ar file2
+is a directory,
+.Nm diff
+is applied to the non-directory file and the file contained in
+the directory file with a filename that is the same as the
+last component of the non-directory file.
+.Pp
+If either the
+.Ar file1
+or
+.Ar file2
+is
+.Fl ,
+the standard input is
+used in its place.
+.Ss Output Style
+The default (without
+.Fl e ,
+.Fl c ,
+or
+.Fl n
+.\"  -C
+options)
+output contains lines of these forms, where
+.Va XX , YY , ZZ , QQ
+are line numbers respective of file order.
+.Pp
+.Dp Cx Li XX
+.Ic a
+.Li YY
+.Cx
+At (the end of) line
+.Va XX
+of
+.Ar file1 ,
+append the contents
+of line
+.Va YY
+of
+.Ar file2
+to make them equal.
+.Dp Cx Li XX
+.Ic a
+.Li YY ,
+.Li ZZ
+.Cx
+Same as above, but append the range of lines,
+.Va YY
+through
+.Va ZZ
+of
+.Ar file2
+to line
+.Va XX
+of file1.
+.Dp Cx Li XX
+.Ic d
+.Li YY
+.Cx
+At line
+.Va XX
+delete
+the line. The value
+.Va YY
+tells to which line the change
+would bring
+.Ar file1
+in line with
+.Ar file1 .
+.Dp Cx Li XX ,
+.Li YY
+.Ic d
+.Li ZZ
+.Cx
+Delete the range of lines
+.Va XX
+through
+.Va YY
+in
+.Ar file1 .
+.Dp Cx Li XX
+.Ic c
+.Li YY
+.Cx
+Change the line
+.Va XX
+in
+.Ar file1
+to the line
+.Va YY
+in
+.Ar file2.
+.Dp Cx Li XX ,
+.Li YY
+.Ic c
+.Li ZZ
+.Cx
+Replace the range of specified lines with the line
+.Va ZZ .
+.Dp Cx Li XX ,
+.Li YY
+.Ic c
+.Li ZZ ,
+.Li QQ
+.Cx
+Replace the range
+.Cx Va XX ,
+.Va YY
+.Cx
+from
+.Ar file1
+with the range
+.Cx Va ZZ ,
+.Va QQ
+.Cx
+from
+.Ar file2 .
+.Dp
+.Pp
+These lines resemble
+.Xr ed 1
+subcommands to convert
+.Ar file1
+into
+.Ar file2 .
+The line numbers before the action letters pertain to
+.Ar file1 ;
+those after pertain to
+.Ar file2 .
+Thus, by exchanging
+.Ic a
+for
+.Ic d
+and reading the line in reverse order, one can also
+determine how to convert
+.Ar file2
+into
+.Ar file1 .
+As in
+.Xr ed 1 ,
+identical
+pairs (where num1 = num2) are abbreviated as a single
+number.
+.Sh ENVIRONMENT
+.Tw Fl
+.Tp Ev TMPDIR
+If the environment variable
+.Ev TMPDIR
+exists,
+.Nm Diff
+will use the name specified by
+.Ev TMPDIR
+as an alternate temporary directory.
+.Tp
+.Sh FILES
+.Dw /usr/bin/diffh
+.Di L
+.Dp Pa /tmp/d?????
+.br
+.Dp Pa /usr/bin/diffh
+Alternate algorithm version (used by option
+.Fl h ) .
+.Dp Pa /usr/bin/diff
+for directory diffs
+.Dp Pa /usr/bin/pr
+used by the
+.Fl l
+option.
+.Dp
+.Sh SEE ALSO
+.Xr cmp 1 ,
+.Xr cc 1 ,
+.Xr comm 1 ,
+.Xr ed 1 ,
+.Xr diff3 1
+.Sh HISTORY
+.Nm Diff
+appeared in Version 6 AT&T Unix.
+.Sh DIAGNOSTICS
+The
+.Nm diff
+utility exits with one of the following values:
+.Dw Ds
+.Dp \&0
+No differences were found.
+.Dp \&1
+Differences were found.
+.Dp \&>\&1
+An error occurred.
+.Dp
+.Sh BUGS
+The
+.Fl f
+and
+.Fl e
+options
+do not provide special handling for lines on which the
+first and only character is
+.Dq Li \&. .
+This can cause problems for
+.Xr ed 1 .
+.Pp
+When comparing directories with the
+.Fl b ,
+.Fl w
+or
+.Fl i
+options specified,
+.Nm diff
+first compares the files ala
+.Ar cmp ,
+and then decides to run the
+.Nm diff
+algorithm if they are not equal.
+This may cause a small amount of spurious output if the files
+then turn out to be identical because the only differences are
+insignificant white space or case differences.
+.\" .Sh STANDARDS
diff --git a/usr.bin/diff/diff.c b/usr.bin/diff/diff.c
new file mode 100644
index 00000000000..ddab61fac1f
--- /dev/null
+++ b/usr.bin/diff/diff.c
@@ -0,0 +1,212 @@
+static	char sccsid[] = "@(#)diff.c 4.7 5/11/89";
+
+#include "diff.h"
+#include "pathnames.h"
+
+/*
+ * diff - driver and subroutines
+ */
+
+char	diff[] = _PATH_DIFF;
+char	diffh[] = _PATH_DIFFH;
+char	pr[] = _PATH_PR;
+
+main(argc, argv)
+	int argc;
+	char **argv;
+{
+	register char *argp;
+
+	ifdef1 = "FILE1"; ifdef2 = "FILE2";
+	status = 2;
+	diffargv = argv;
+	argc--, argv++;
+	while (argc > 2 && argv[0][0] == '-') {
+		argp = &argv[0][1];
+		argv++, argc--;
+		while (*argp) switch(*argp++) {
+
+#ifdef notdef
+		case 'I':
+			opt = D_IFDEF;
+			wantelses = 0;
+			continue;
+		case 'E':
+			opt = D_IFDEF;
+			wantelses = 1;
+			continue;
+		case '1':
+			opt = D_IFDEF;
+			ifdef1 = argp;
+			*--argp = 0;
+			continue;
+#endif
+		case 'D':
+			/* -Dfoo = -E -1 -2foo */
+			wantelses = 1;
+			ifdef1 = "";
+			/* fall through */
+#ifdef notdef
+		case '2':
+#endif
+			opt = D_IFDEF;
+			ifdef2 = argp;
+			*--argp = 0;
+			continue;
+		case 'e':
+			opt = D_EDIT;
+			continue;
+		case 'f':
+			opt = D_REVERSE;
+			continue;
+		case 'n':
+			opt = D_NREVERSE;
+			continue;
+		case 'b':
+			bflag = 1;
+			continue;
+		case 'w':
+			wflag = 1;
+			continue;
+		case 'i':
+			iflag = 1;
+			continue;
+		case 't':
+			tflag = 1;
+			continue;
+		case 'c':
+			opt = D_CONTEXT;
+			if (isdigit(*argp)) {
+				context = atoi(argp);
+				while (isdigit(*argp))
+					argp++;
+				if (*argp) {
+					fprintf(stderr,
+					    "diff: -c: bad count\n");
+					done();
+				}
+				argp = "";
+			} else
+				context = 3;
+			continue;
+		case 'h':
+			hflag++;
+			continue;
+		case 'S':
+			if (*argp == 0) {
+				fprintf(stderr, "diff: use -Sstart\n");
+				done();
+			}
+			start = argp;
+			*--argp = 0;		/* don't pass it on */
+			continue;
+		case 'r':
+			rflag++;
+			continue;
+		case 's':
+			sflag++;
+			continue;
+		case 'l':
+			lflag++;
+			continue;
+		default:
+			fprintf(stderr, "diff: -%s: unknown option\n",
+			    --argp);
+			done();
+		}
+	}
+	if (argc != 2) {
+		fprintf(stderr, "diff: two filename arguments required\n");
+		done();
+	}
+	file1 = argv[0];
+	file2 = argv[1];
+	if (hflag && opt) {
+		fprintf(stderr,
+		    "diff: -h doesn't support -e, -f, -n, -c, or -I\n");
+		done();
+	}
+	if (!strcmp(file1, "-"))
+		stb1.st_mode = S_IFREG;
+	else if (stat(file1, &stb1) < 0) {
+		fprintf(stderr, "diff: ");
+		perror(file1);
+		done();
+	}
+	if (!strcmp(file2, "-"))
+		stb2.st_mode = S_IFREG;
+	else if (stat(file2, &stb2) < 0) {
+		fprintf(stderr, "diff: ");
+		perror(file2);
+		done();
+	}
+	if ((stb1.st_mode & S_IFMT) == S_IFDIR &&
+	    (stb2.st_mode & S_IFMT) == S_IFDIR) {
+		diffdir(argv);
+	} else
+		diffreg();
+	done();
+}
+
+char *
+savestr(cp)
+	register char *cp;
+{
+	register char *dp = malloc(strlen(cp)+1);
+
+	if (dp == 0) {
+		fprintf(stderr, "diff: ran out of memory\n");
+		done();
+	}
+	strcpy(dp, cp);
+	return (dp);
+}
+
+min(a,b)
+	int a,b;
+{
+
+	return (a < b ? a : b);
+}
+
+max(a,b)
+	int a,b;
+{
+
+	return (a > b ? a : b);
+}
+
+done()
+{
+	if (tempfile)
+		unlink(tempfile);
+	exit(status);
+}
+
+char *
+talloc(n)
+{
+	register char *p;
+
+	if ((p = malloc((unsigned)n)) != NULL)
+		return(p);
+	noroom();
+}
+
+char *
+ralloc(p,n)
+char *p;
+{
+	register char *q;
+	char *realloc();
+
+	if ((q = realloc(p, (unsigned)n)) == NULL)
+		noroom();
+	return(q);
+}
+
+noroom()
+{
+	fprintf(stderr, "diff: files too big, try -h\n");
+	done();
+}
diff --git a/usr.bin/diff/diff.h b/usr.bin/diff/diff.h
new file mode 100644
index 00000000000..7c11a5fb50d
--- /dev/null
+++ b/usr.bin/diff/diff.h
@@ -0,0 +1,85 @@
+/*	diff.h	4.7	85/08/16	*/
+
+/*
+ * diff - common declarations
+ */
+
+#include <stdio.h>
+#include <ctype.h>
+#include <sys/param.h>
+#include <sys/stat.h>
+#include <sys/dir.h>
+#include <signal.h>
+
+/*
+ * Output format options
+ */
+int	opt;
+
+#define	D_NORMAL	0	/* Normal output */
+#define	D_EDIT		-1	/* Editor script out */
+#define	D_REVERSE	1	/* Reverse editor script */
+#define	D_CONTEXT	2	/* Diff with context */
+#define	D_IFDEF		3	/* Diff with merged #ifdef's */
+#define	D_NREVERSE	4	/* Reverse ed script with numbered
+				   lines and no trailing . */
+
+int	tflag;			/* expand tabs on output */
+
+/*
+ * Algorithm related options
+ */
+int	hflag;			/* -h, use halfhearted DIFFH */
+int	bflag;			/* ignore blanks in comparisons */
+int	wflag;			/* totally ignore blanks in comparisons */
+int	iflag;			/* ignore case in comparisons */
+
+/*
+ * Options on hierarchical diffs.
+ */
+int	lflag;			/* long output format with header */
+int	rflag;			/* recursively trace directories */
+int	sflag;			/* announce files which are same */
+char	*start;			/* do file only if name >= this */
+
+/*
+ * Variables for -I D_IFDEF option.
+ */
+int	wantelses;		/* -E */
+char	*ifdef1;		/* String for -1 */
+char	*ifdef2;		/* String for -2 */
+char	*endifname;		/* What we will print on next #endif */
+int	inifdef;
+
+/*
+ * Variables for -c context option.
+ */
+int	context;		/* lines of context to be printed */
+
+/*
+ * State for exit status.
+ */
+int	status;
+int	anychange;
+char	*tempfile;		/* used when comparing against std input */
+
+/*
+ * Variables for diffdir.
+ */
+char	**diffargv;		/* option list to pass to recursive diffs */
+
+/*
+ * Input file names.
+ * With diffdir, file1 and file2 are allocated BUFSIZ space,
+ * and padded with a '/', and then efile0 and efile1 point after
+ * the '/'.
+ */
+char	*file1, *file2, *efile1, *efile2;
+struct	stat stb1, stb2;
+
+char	*malloc(), *talloc(), *ralloc();
+char	*savestr(), *splice(), *splicen();
+char	*mktemp(), *copytemp(), *rindex();
+int	done();
+
+extern	char diffh[], diff[], pr[];
diff --git a/usr.bin/diff/diffdir.c b/usr.bin/diff/diffdir.c
new file mode 100644
index 00000000000..7be6578da6f
--- /dev/null
+++ b/usr.bin/diff/diffdir.c
@@ -0,0 +1,419 @@
+static	char *sccsid = "@(#)diffdir.c	4.12 (Berkeley) 4/30/89";
+
+#include "diff.h"
+/*
+ * diff - directory comparison
+ */
+#define	d_flags	d_ino
+
+#define	ONLY	1		/* Only in this directory */
+#define	SAME	2		/* Both places and same */
+#define	DIFFER	4		/* Both places and different */
+#define	DIRECT	8		/* Directory */
+
+struct dir {
+	u_long	d_ino;
+	short	d_reclen;
+	short	d_namlen;
+	char	*d_entry;
+};
+
+struct	dir *setupdir();
+int	header;
+static int	dirstatus;		/* exit status from diffdir */
+extern int	status;
+char	title[2*BUFSIZ], *etitle;
+
+diffdir(argv)
+	char **argv;
+{
+	register struct dir *d1, *d2;
+	struct dir *dir1, *dir2;
+	register int i;
+	int cmp;
+
+	if (opt == D_IFDEF) {
+		fprintf(stderr, "diff: can't specify -I with directories\n");
+		done();
+	}
+	if (opt == D_EDIT && (sflag || lflag))
+		fprintf(stderr,
+		    "diff: warning: shouldn't give -s or -l with -e\n");
+	title[0] = 0;
+	strcpy(title, "diff ");
+	for (i = 1; diffargv[i+2]; i++) {
+		if (!strcmp(diffargv[i], "-"))
+			continue;	/* was -S, dont look silly */
+		strcat(title, diffargv[i]);
+		strcat(title, " ");
+	}
+	for (etitle = title; *etitle; etitle++)
+		;
+	setfile(&file1, &efile1, file1);
+	setfile(&file2, &efile2, file2);
+	argv[0] = file1;
+	argv[1] = file2;
+	dir1 = setupdir(file1);
+	dir2 = setupdir(file2);
+	d1 = dir1; d2 = dir2;
+	while (d1->d_entry != 0 || d2->d_entry != 0) {
+		if (d1->d_entry && useless(d1->d_entry)) {
+			d1++;
+			continue;
+		}
+		if (d2->d_entry && useless(d2->d_entry)) {
+			d2++;
+			continue;
+		}
+		if (d1->d_entry == 0)
+			cmp = 1;
+		else if (d2->d_entry == 0)
+			cmp = -1;
+		else
+			cmp = strcmp(d1->d_entry, d2->d_entry);
+		if (cmp < 0) {
+			if (lflag)
+				d1->d_flags |= ONLY;
+			else if (opt == 0 || opt == 2)
+				only(d1, 1);
+			d1++;
+			dirstatus |= 1;
+		} else if (cmp == 0) {
+			compare(d1);
+			d1++;
+			d2++;
+		} else {
+			if (lflag)
+				d2->d_flags |= ONLY;
+			else if (opt == 0 || opt == 2)
+				only(d2, 2);
+			d2++;
+			dirstatus |= 1;
+		}
+	}
+	if (lflag) {
+		scanpr(dir1, ONLY, "Only in %.*s", file1, efile1, 0, 0);
+		scanpr(dir2, ONLY, "Only in %.*s", file2, efile2, 0, 0);
+		scanpr(dir1, SAME, "Common identical files in %.*s and %.*s",
+		    file1, efile1, file2, efile2);
+		scanpr(dir1, DIFFER, "Binary files which differ in %.*s and %.*s",
+		    file1, efile1, file2, efile2);
+		scanpr(dir1, DIRECT, "Common subdirectories of %.*s and %.*s",
+		    file1, efile1, file2, efile2);
+	}
+	if (rflag) {
+		if (header && lflag)
+			printf("\f");
+		for (d1 = dir1; d1->d_entry; d1++)  {
+			if ((d1->d_flags & DIRECT) == 0)
+				continue;
+			strcpy(efile1, d1->d_entry);
+			strcpy(efile2, d1->d_entry);
+			calldiff(0);
+		}
+	}
+	status = dirstatus;
+}
+
+setfile(fpp, epp, file)
+	char **fpp, **epp;
+	char *file;
+{
+	register char *cp;
+
+	*fpp = malloc(BUFSIZ);
+	if (*fpp == 0) {
+		fprintf(stderr, "diff: ran out of memory\n");
+		exit(1);
+	}
+	strcpy(*fpp, file);
+	for (cp = *fpp; *cp; cp++)
+		continue;
+	*cp++ = '/';
+	*epp = cp;
+}
+
+scanpr(dp, test, title, file1, efile1, file2, efile2)
+	register struct dir *dp;
+	int test;
+	char *title, *file1, *efile1, *file2, *efile2;
+{
+	int titled = 0;
+
+	for (; dp->d_entry; dp++) {
+		if ((dp->d_flags & test) == 0)
+			continue;
+		if (titled == 0) {
+			if (header == 0)
+				header = 1;
+			else
+				printf("\n");
+			printf(title,
+			    efile1 - file1 - 1, file1,
+			    efile2 - file2 - 1, file2);
+			printf(":\n");
+			titled = 1;
+		}
+		printf("\t%s\n", dp->d_entry);
+	}
+}
+
+only(dp, which)
+	struct dir *dp;
+	int which;
+{
+	char *file = which == 1 ? file1 : file2;
+	char *efile = which == 1 ? efile1 : efile2;
+
+	printf("Only in %.*s: %s\n", efile - file - 1, file, dp->d_entry);
+	
+}
+
+int	entcmp();
+
+struct dir *
+setupdir(cp)
+	char *cp;
+{
+	register struct dir *dp, *ep;
+	register struct direct *rp;
+	register int nitems, n;
+	DIR *dirp;
+
+	dirp = opendir(cp);
+	if (dirp == NULL) {
+		fprintf(stderr, "diff: ");
+		perror(cp);
+		done();
+	}
+	nitems = 0;
+	dp = (struct dir *)malloc(sizeof (struct dir));
+	if (dp == 0) {
+		fprintf(stderr, "diff: ran out of memory\n");
+		done();
+	}
+	while (rp = readdir(dirp)) {
+		ep = &dp[nitems++];
+		ep->d_reclen = rp->d_reclen;
+		ep->d_namlen = rp->d_namlen;
+		ep->d_entry = 0;
+		ep->d_flags = 0;
+		if (ep->d_namlen > 0) {
+			ep->d_entry = malloc(ep->d_namlen + 1);
+			if (ep->d_entry == 0) {
+				fprintf(stderr, "diff: out of memory\n");
+				done();
+			}
+			strcpy(ep->d_entry, rp->d_name);
+		}
+		dp = (struct dir *)realloc((char *)dp,
+			(nitems + 1) * sizeof (struct dir));
+		if (dp == 0) {
+			fprintf(stderr, "diff: ran out of memory\n");
+			done();
+		}
+	}
+	dp[nitems].d_entry = 0;		/* delimiter */
+	closedir(dirp);
+	qsort(dp, nitems, sizeof (struct dir), entcmp);
+	return (dp);
+}
+
+entcmp(d1, d2)
+	struct dir *d1, *d2;
+{
+	return (strcmp(d1->d_entry, d2->d_entry));
+}
+
+compare(dp)
+	register struct dir *dp;
+{
+	register int i, j;
+	int f1, f2, fmt1, fmt2;
+	struct stat stb1, stb2;
+	int flag = 0;
+	char buf1[BUFSIZ], buf2[BUFSIZ];
+
+	strcpy(efile1, dp->d_entry);
+	strcpy(efile2, dp->d_entry);
+	f1 = open(file1, 0);
+	if (f1 < 0) {
+		perror(file1);
+		return;
+	}
+	f2 = open(file2, 0);
+	if (f2 < 0) {
+		perror(file2);
+		close(f1);
+		return;
+	}
+	fstat(f1, &stb1); fstat(f2, &stb2);
+	fmt1 = stb1.st_mode & S_IFMT;
+	fmt2 = stb2.st_mode & S_IFMT;
+	if (fmt1 != S_IFREG || fmt2 != S_IFREG) {
+		if (fmt1 == fmt2) {
+			if (fmt1 != S_IFDIR && stb1.st_rdev == stb2.st_rdev)
+				goto same;
+			if (fmt1 == S_IFDIR) {
+				dp->d_flags = DIRECT;
+				if (lflag || opt == D_EDIT)
+					goto closem;
+				printf("Common subdirectories: %s and %s\n",
+				    file1, file2);
+				goto closem;
+			}
+		}
+		goto notsame;
+	}
+	if (stb1.st_size != stb2.st_size)
+		goto notsame;
+	for (;;) {
+		i = read(f1, buf1, BUFSIZ);
+		j = read(f2, buf2, BUFSIZ);
+		if (i < 0 || j < 0 || i != j)
+			goto notsame;
+		if (i == 0 && j == 0)
+			goto same;
+		for (j = 0; j < i; j++)
+			if (buf1[j] != buf2[j])
+				goto notsame;
+	}
+same:
+	if (sflag == 0)
+		goto closem;
+	if (lflag)
+		dp->d_flags = SAME;
+	else
+		printf("Files %s and %s are identical\n", file1, file2);
+	goto closem;
+notsame:
+	dirstatus |= 1;
+	if (!ascii(f1) || !ascii(f2)) {
+		if (lflag)
+			dp->d_flags |= DIFFER;
+		else if (opt == D_NORMAL || opt == D_CONTEXT)
+			printf("Binary files %s and %s differ\n",
+			    file1, file2);
+		goto closem;
+	}
+	close(f1); close(f2);
+	anychange = 1;
+	if (lflag)
+		calldiff(title);
+	else {
+		if (opt == D_EDIT) {
+			printf("ed - %s << '-*-END-*-'\n", dp->d_entry);
+			calldiff(0);
+		} else {
+			printf("%s%s %s\n", title, file1, file2);
+			calldiff(0);
+		}
+		if (opt == D_EDIT)
+			printf("w\nq\n-*-END-*-\n");
+	}
+	return;
+closem:
+	close(f1); close(f2);
+}
+
+char	*prargs[] = { "pr", "-h", 0, "-f", 0, 0 };
+
+calldiff(wantpr)
+	char *wantpr;
+{
+	int pid, lstatus, lstatus2, pv[2];
+
+	prargs[2] = wantpr;
+	fflush(stdout);
+	if (wantpr) {
+		(void)sprintf(etitle, "%s %s", file1, file2);
+		pipe(pv);
+		pid = fork();
+		if (pid == -1) {
+			fprintf(stderr, "No more processes");
+			done();
+		}
+		if (pid == 0) {
+			close(0);
+			dup(pv[0]);
+			close(pv[0]);
+			close(pv[1]);
+			execv(pr+4, prargs);
+			execv(pr, prargs);
+			perror(pr);
+			done();
+		}
+	}
+	pid = fork();
+	if (pid == -1) {
+		fprintf(stderr, "diff: No more processes\n");
+		done();
+	}
+	if (pid == 0) {
+		if (wantpr) {
+			close(1);
+			dup(pv[1]);
+			close(pv[0]);
+			close(pv[1]);
+		}
+		execv(diff+4, diffargv);
+		execv(diff, diffargv);
+		perror(diff);
+		done();
+	}
+	if (wantpr) {
+		close(pv[0]);
+		close(pv[1]);
+	}
+	while (wait(&lstatus) != pid)
+		continue;
+	while (wait(&lstatus2) != -1)
+		continue;
+/*
+	if ((lstatus >> 8) >= 2)
+		done();
+*/
+	dirstatus |= lstatus >> 8;
+}
+
+#include <a.out.h>
+
+ascii(f)
+	int f;
+{
+	char buf[BUFSIZ];
+	register int cnt;
+	register char *cp;
+
+	lseek(f, (long)0, 0);
+	cnt = read(f, buf, BUFSIZ);
+	if (cnt >= sizeof (struct exec)) {
+		struct exec hdr;
+		hdr = *(struct exec *)buf;
+		if (!N_BADMAG(hdr))
+			return (0);
+	}
+	cp = buf;
+	while (--cnt >= 0)
+		if (*cp++ & 0200)
+			return (0);
+	return (1);
+}
+
+/*
+ * THIS IS CRUDE.
+ */
+useless(cp)
+register char *cp;
+{
+
+	if (cp[0] == '.') {
+		if (cp[1] == '\0')
+			return (1);	/* directory "." */
+		if (cp[1] == '.' && cp[2] == '\0')
+			return (1);	/* directory ".." */
+	}
+	if (start && strcmp(start, cp) > 0)
+		return (1);
+	return (0);
+}
diff --git a/usr.bin/diff/diffreg.c b/usr.bin/diff/diffreg.c
new file mode 100644
index 00000000000..fc49f6e4a29
--- /dev/null
+++ b/usr.bin/diff/diffreg.c
@@ -0,0 +1,1048 @@
+static	char sccsid[] = "@(#)diffreg.c 4.21 4/6/90";
+
+#include "diff.h"
+#include "pathnames.h"
+/*
+ * diff - compare two files.
+ */
+
+/*
+ *	Uses an algorithm due to Harold Stone, which finds
+ *	a pair of longest identical subsequences in the two
+ *	files.
+ *
+ *	The major goal is to generate the match vector J.
+ *	J[i] is the index of the line in file1 corresponding
+ *	to line i file0. J[i] = 0 if there is no
+ *	such line in file1.
+ *
+ *	Lines are hashed so as to work in core. All potential
+ *	matches are located by sorting the lines of each file
+ *	on the hash (called ``value''). In particular, this
+ *	collects the equivalence classes in file1 together.
+ *	Subroutine equiv replaces the value of each line in
+ *	file0 by the index of the first element of its 
+ *	matching equivalence in (the reordered) file1.
+ *	To save space equiv squeezes file1 into a single
+ *	array member in which the equivalence classes
+ *	are simply concatenated, except that their first
+ *	members are flagged by changing sign.
+ *
+ *	Next the indices that point into member are unsorted into
+ *	array class according to the original order of file0.
+ *
+ *	The cleverness lies in routine stone. This marches
+ *	through the lines of file0, developing a vector klist
+ *	of "k-candidates". At step i a k-candidate is a matched
+ *	pair of lines x,y (x in file0 y in file1) such that
+ *	there is a common subsequence of length k
+ *	between the first i lines of file0 and the first y 
+ *	lines of file1, but there is no such subsequence for
+ *	any smaller y. x is the earliest possible mate to y
+ *	that occurs in such a subsequence.
+ *
+ *	Whenever any of the members of the equivalence class of
+ *	lines in file1 matable to a line in file0 has serial number 
+ *	less than the y of some k-candidate, that k-candidate 
+ *	with the smallest such y is replaced. The new 
+ *	k-candidate is chained (via pred) to the current
+ *	k-1 candidate so that the actual subsequence can
+ *	be recovered. When a member has serial number greater
+ *	that the y of all k-candidates, the klist is extended.
+ *	At the end, the longest subsequence is pulled out
+ *	and placed in the array J by unravel
+ *
+ *	With J in hand, the matches there recorded are
+ *	check'ed against reality to assure that no spurious
+ *	matches have crept in due to hashing. If they have,
+ *	they are broken, and "jackpot" is recorded--a harmless
+ *	matter except that a true match for a spuriously
+ *	mated line may now be unnecessarily reported as a change.
+ *
+ *	Much of the complexity of the program comes simply
+ *	from trying to minimize core utilization and
+ *	maximize the range of doable problems by dynamically
+ *	allocating what is needed and reusing what is not.
+ *	The core requirements for problems larger than somewhat
+ *	are (in words) 2*length(file0) + length(file1) +
+ *	3*(number of k-candidates installed),  typically about
+ *	6n words for files of length n. 
+ */
+
+#define	prints(s)	fputs(s,stdout)
+
+FILE	*input[2];
+FILE	*fopen();
+
+struct cand {
+	int	x;
+	int	y;
+	int	pred;
+} cand;
+struct line {
+	int	serial;
+	int	value;
+} *file[2], line;
+int	len[2];
+struct	line *sfile[2];	/* shortened by pruning common prefix and suffix */
+int	slen[2];
+int	pref, suff;	/* length of prefix and suffix */
+int	*class;		/* will be overlaid on file[0] */
+int	*member;	/* will be overlaid on file[1] */
+int	*klist;		/* will be overlaid on file[0] after class */
+struct	cand *clist;	/* merely a free storage pot for candidates */
+int	clen = 0;
+int	*J;		/* will be overlaid on class */
+long	*ixold;		/* will be overlaid on klist */
+long	*ixnew;		/* will be overlaid on file[1] */
+char	*chrtran;	/* translation table for case-folding */
+
+/* chrtran points to one of 2 translation tables:
+ *	cup2low if folding upper to lower case
+ *	clow2low if not folding case
+ */
+char	clow2low[256] = {
+0x00,0x01,0x02,0x03,0x04,0x05,0x06,0x07,0x08,0x09,0x0a,0x0b,0x0c,0x0d,0x0e,0x0f,
+0x10,0x11,0x12,0x13,0x14,0x15,0x16,0x17,0x18,0x19,0x1a,0x1b,0x1c,0x1d,0x1e,0x1f,
+0x20,0x21,0x22,0x23,0x24,0x25,0x26,0x27,0x28,0x29,0x2a,0x2b,0x2c,0x2d,0x2e,0x2f,
+0x30,0x31,0x32,0x33,0x34,0x35,0x36,0x37,0x38,0x39,0x3a,0x3b,0x3c,0x3d,0x3e,0x3f,
+0x40,0x41,0x42,0x43,0x44,0x45,0x46,0x47,0x48,0x49,0x4a,0x4b,0x4c,0x4d,0x4e,0x4f,
+0x50,0x51,0x52,0x53,0x54,0x55,0x56,0x57,0x58,0x59,0x5a,0x5b,0x5c,0x5d,0x5e,0x5f,
+0x60,0x61,0x62,0x63,0x64,0x65,0x66,0x67,0x68,0x69,0x6a,0x6b,0x6c,0x6d,0x6e,0x6f,
+0x70,0x71,0x72,0x73,0x74,0x75,0x76,0x77,0x78,0x79,0x7a,0x7b,0x7c,0x7d,0x7e,0x7f,
+0x80,0x81,0x82,0x83,0x84,0x85,0x86,0x87,0x88,0x89,0x8a,0x8b,0x8c,0x8d,0x8e,0x8f,
+0x90,0x91,0x92,0x93,0x94,0x95,0x96,0x97,0x98,0x99,0x9a,0x9b,0x9c,0x9d,0x9e,0x9f,
+0xa0,0xa1,0xa2,0xa3,0xa4,0xa5,0xa6,0xa7,0xa8,0xa9,0xaa,0xab,0xac,0xad,0xae,0xaf,
+0xb0,0xb1,0xb2,0xb3,0xb4,0xb5,0xb6,0xb7,0xb8,0xb9,0xba,0xbb,0xbc,0xbd,0xbe,0xbf,
+0xc0,0xc1,0xc2,0xc3,0xc4,0xc5,0xc6,0xc7,0xc8,0xc9,0xca,0xcb,0xcc,0xcd,0xce,0xcf,
+0xd0,0xd1,0xd2,0xd3,0xd4,0xd5,0xd6,0xd7,0xd8,0xd9,0xda,0xdb,0xdc,0xdd,0xde,0xdf,
+0xe0,0xe1,0xe2,0xe3,0xe4,0xe5,0xe6,0xe7,0xe8,0xe9,0xea,0xeb,0xec,0xed,0xee,0xef,
+0xf0,0xf1,0xf2,0xf3,0xf4,0xf5,0xf6,0xf7,0xf8,0xf9,0xfa,0xfb,0xfc,0xfd,0xfe,0xff
+};
+
+char	cup2low[256] = {
+0x00,0x01,0x02,0x03,0x04,0x05,0x06,0x07,0x08,0x09,0x0a,0x0b,0x0c,0x0d,0x0e,0x0f,
+0x10,0x11,0x12,0x13,0x14,0x15,0x16,0x17,0x18,0x19,0x1a,0x1b,0x1c,0x1d,0x1e,0x1f,
+0x20,0x21,0x22,0x23,0x24,0x25,0x26,0x27,0x28,0x29,0x2a,0x2b,0x2c,0x2d,0x2e,0x2f,
+0x30,0x31,0x32,0x33,0x34,0x35,0x36,0x37,0x38,0x39,0x3a,0x3b,0x3c,0x3d,0x3e,0x3f,
+0x60,0x61,0x62,0x63,0x64,0x65,0x66,0x67,0x68,0x69,0x6a,0x6b,0x6c,0x6d,0x6e,0x6f,
+0x70,0x71,0x72,0x73,0x74,0x75,0x76,0x77,0x78,0x79,0x7a,0x7b,0x7c,0x7d,0x7e,0x7f,
+0x60,0x61,0x62,0x63,0x64,0x65,0x66,0x67,0x68,0x69,0x6a,0x6b,0x6c,0x6d,0x6e,0x6f,
+0x70,0x71,0x72,0x73,0x74,0x75,0x76,0x77,0x78,0x79,0x7a,0x7b,0x7c,0x7d,0x7e,0x7f,
+0x80,0x81,0x82,0x83,0x84,0x85,0x86,0x87,0x88,0x89,0x8a,0x8b,0x8c,0x8d,0x8e,0x8f,
+0x90,0x91,0x92,0x93,0x94,0x95,0x96,0x97,0x98,0x99,0x9a,0x9b,0x9c,0x9d,0x9e,0x9f,
+0xa0,0xa1,0xa2,0xa3,0xa4,0xa5,0xa6,0xa7,0xa8,0xa9,0xaa,0xab,0xac,0xad,0xae,0xaf,
+0xb0,0xb1,0xb2,0xb3,0xb4,0xb5,0xb6,0xb7,0xb8,0xb9,0xba,0xbb,0xbc,0xbd,0xbe,0xbf,
+0xc0,0xc1,0xc2,0xc3,0xc4,0xc5,0xc6,0xc7,0xc8,0xc9,0xca,0xcb,0xcc,0xcd,0xce,0xcf,
+0xd0,0xd1,0xd2,0xd3,0xd4,0xd5,0xd6,0xd7,0xd8,0xd9,0xda,0xdb,0xdc,0xdd,0xde,0xdf,
+0xe0,0xe1,0xe2,0xe3,0xe4,0xe5,0xe6,0xe7,0xe8,0xe9,0xea,0xeb,0xec,0xed,0xee,0xef,
+0xf0,0xf1,0xf2,0xf3,0xf4,0xf5,0xf6,0xf7,0xf8,0xf9,0xfa,0xfb,0xfc,0xfd,0xfe,0xff
+};
+
+diffreg()
+{
+	register int i, j;
+	FILE *f1, *f2;
+	char buf1[BUFSIZ], buf2[BUFSIZ];
+
+	if (hflag) {
+		diffargv[0] = "diffh";
+		execv(diffh, diffargv);
+		fprintf(stderr, "diff: ");
+		perror(diffh);
+		done();
+	}
+	chrtran = (iflag? cup2low : clow2low);
+	if ((stb1.st_mode & S_IFMT) == S_IFDIR) {
+		file1 = splice(file1, file2);
+		if (stat(file1, &stb1) < 0) {
+			fprintf(stderr, "diff: ");
+			perror(file1);
+			done();
+		}
+	} else if ((stb2.st_mode & S_IFMT) == S_IFDIR) {
+		file2 = splice(file2, file1);
+		if (stat(file2, &stb2) < 0) {
+			fprintf(stderr, "diff: ");
+			perror(file2);
+			done();
+		}
+	} else if ((stb1.st_mode & S_IFMT) != S_IFREG || !strcmp(file1, "-")) {
+		if (!strcmp(file2, "-")) {
+			fprintf(stderr, "diff: can't specify - -\n");
+			done();
+		}
+		file1 = copytemp();
+		if (stat(file1, &stb1) < 0) {
+			fprintf(stderr, "diff: ");
+			perror(file1);
+			done();
+		}
+	} else if ((stb2.st_mode & S_IFMT) != S_IFREG || !strcmp(file2, "-")) {
+		file2 = copytemp();
+		if (stat(file2, &stb2) < 0) {
+			fprintf(stderr, "diff: ");
+			perror(file2);
+			done();
+		}
+	}
+	if ((f1 = fopen(file1, "r")) == NULL) {
+		fprintf(stderr, "diff: ");
+		perror(file1);
+		done();
+	}
+	if ((f2 = fopen(file2, "r")) == NULL) {
+		fprintf(stderr, "diff: ");
+		perror(file2);
+		fclose(f1);
+		done();
+	}
+	if (stb1.st_size != stb2.st_size)
+		goto notsame;
+	for (;;) {
+		i = fread(buf1, 1, BUFSIZ, f1);
+		j = fread(buf2, 1, BUFSIZ, f2);
+		if (i < 0 || j < 0 || i != j)
+			goto notsame;
+		if (i == 0 && j == 0) {
+			fclose(f1);
+			fclose(f2);
+			status = 0;		/* files don't differ */
+			goto same;
+		}
+		for (j = 0; j < i; j++)
+			if (buf1[j] != buf2[j])
+				goto notsame;
+	}
+notsame:
+	/*
+	 *	Files certainly differ at this point; set status accordingly
+	 */
+	status = 1;
+	if (!asciifile(f1) || !asciifile(f2)) {
+		printf("Binary files %s and %s differ\n", file1, file2);
+		fclose(f1);
+		fclose(f2);
+		done();
+	}
+	prepare(0, f1);
+	prepare(1, f2);
+	fclose(f1);
+	fclose(f2);
+	prune();
+	sort(sfile[0],slen[0]);
+	sort(sfile[1],slen[1]);
+
+	member = (int *)file[1];
+	equiv(sfile[0], slen[0], sfile[1], slen[1], member);
+	member = (int *)ralloc((char *)member,(slen[1]+2)*sizeof(int));
+
+	class = (int *)file[0];
+	unsort(sfile[0], slen[0], class);
+	class = (int *)ralloc((char *)class,(slen[0]+2)*sizeof(int));
+
+	klist = (int *)talloc((slen[0]+2)*sizeof(int));
+	clist = (struct cand *)talloc(sizeof(cand));
+	i = stone(class, slen[0], member, klist);
+	free((char *)member);
+	free((char *)class);
+
+	J = (int *)talloc((len[0]+2)*sizeof(int));
+	unravel(klist[i]);
+	free((char *)clist);
+	free((char *)klist);
+
+	ixold = (long *)talloc((len[0]+2)*sizeof(long));
+	ixnew = (long *)talloc((len[1]+2)*sizeof(long));
+	check();
+	output();
+	status = anychange;
+same:
+	if (opt == D_CONTEXT && anychange == 0)
+		printf("No differences encountered\n");
+	done();
+}
+
+char *tempfile = _PATH_TMP;
+
+char *
+copytemp()
+{
+	char buf[BUFSIZ];
+	register int i, f;
+
+	signal(SIGHUP,done);
+	signal(SIGINT,done);
+	signal(SIGPIPE,done);
+	signal(SIGTERM,done);
+	mktemp(tempfile);
+	f = creat(tempfile,0600);
+	if (f < 0) {
+		fprintf(stderr, "diff: ");
+		perror(tempfile);
+		done();
+	}
+	while ((i = read(0,buf,BUFSIZ)) > 0)
+		if (write(f,buf,i) != i) {
+			fprintf(stderr, "diff: ");
+			perror(tempfile);
+			done();
+		}
+	close(f);
+	return (tempfile);
+}
+
+char *
+splice(dir, file)
+	char *dir, *file;
+{
+	char *tail;
+	char buf[BUFSIZ];
+
+	if (!strcmp(file, "-")) {
+		fprintf(stderr, "diff: can't specify - with other arg directory\n");
+		done();
+	}
+	tail = rindex(file, '/');
+	if (tail == 0)
+		tail = file;
+	else
+		tail++;
+	(void)sprintf(buf, "%s/%s", dir, tail);
+	return (savestr(buf));
+}
+
+prepare(i, fd)
+	int i;
+	FILE *fd;
+{
+	register struct line *p;
+	register j,h;
+
+	fseek(fd, (long)0, 0);
+	p = (struct line *)talloc(3*sizeof(line));
+	for(j=0; h=readhash(fd);) {
+		p = (struct line *)ralloc((char *)p,(++j+3)*sizeof(line));
+		p[j].value = h;
+	}
+	len[i] = j;
+	file[i] = p;
+}
+
+prune()
+{
+	register i,j;
+	for(pref=0;pref<len[0]&&pref<len[1]&&
+		file[0][pref+1].value==file[1][pref+1].value;
+		pref++ ) ;
+	for(suff=0;suff<len[0]-pref&&suff<len[1]-pref&&
+		file[0][len[0]-suff].value==file[1][len[1]-suff].value;
+		suff++) ;
+	for(j=0;j<2;j++) {
+		sfile[j] = file[j]+pref;
+		slen[j] = len[j]-pref-suff;
+		for(i=0;i<=slen[j];i++)
+			sfile[j][i].serial = i;
+	}
+}
+
+equiv(a,n,b,m,c)
+struct line *a, *b;
+int *c;
+{
+	register int i, j;
+	i = j = 1;
+	while(i<=n && j<=m) {
+		if(a[i].value <b[j].value)
+			a[i++].value = 0;
+		else if(a[i].value == b[j].value)
+			a[i++].value = j;
+		else
+			j++;
+	}
+	while(i <= n)
+		a[i++].value = 0;
+	b[m+1].value = 0;
+	j = 0;
+	while(++j <= m) {
+		c[j] = -b[j].serial;
+		while(b[j+1].value == b[j].value) {
+			j++;
+			c[j] = b[j].serial;
+		}
+	}
+	c[j] = -1;
+}
+
+stone(a,n,b,c)
+int *a;
+int *b;
+register int *c;
+{
+	register int i, k,y;
+	int j, l;
+	int oldc, tc;
+	int oldl;
+	k = 0;
+	c[0] = newcand(0,0,0);
+	for(i=1; i<=n; i++) {
+		j = a[i];
+		if(j==0)
+			continue;
+		y = -b[j];
+		oldl = 0;
+		oldc = c[0];
+		do {
+			if(y <= clist[oldc].y)
+				continue;
+			l = search(c, k, y);
+			if(l!=oldl+1)
+				oldc = c[l-1];
+			if(l<=k) {
+				if(clist[c[l]].y <= y)
+					continue;
+				tc = c[l];
+				c[l] = newcand(i,y,oldc);
+				oldc = tc;
+				oldl = l;
+			} else {
+				c[l] = newcand(i,y,oldc);
+				k++;
+				break;
+			}
+		} while((y=b[++j]) > 0);
+	}
+	return(k);
+}
+
+newcand(x,y,pred)
+{
+	register struct cand *q;
+	clist = (struct cand *)ralloc((char *)clist,++clen*sizeof(cand));
+	q = clist + clen -1;
+	q->x = x;
+	q->y = y;
+	q->pred = pred;
+	return(clen-1);
+}
+
+search(c, k, y)
+int *c;
+{
+	register int i, j, l;
+	int t;
+	if(clist[c[k]].y<y)	/*quick look for typical case*/
+		return(k+1);
+	i = 0;
+	j = k+1;
+	while (1) {
+		l = i + j;
+		if ((l >>= 1) <= i) 
+			break;
+		t = clist[c[l]].y;
+		if(t > y)
+			j = l;
+		else if(t < y)
+			i = l;
+		else
+			return(l);
+	}
+	return(l+1);
+}
+
+unravel(p)
+{
+	register int i;
+	register struct cand *q;
+	for(i=0; i<=len[0]; i++)
+		J[i] =	i<=pref ? i:
+			i>len[0]-suff ? i+len[1]-len[0]:
+			0;
+	for(q=clist+p;q->y!=0;q=clist+q->pred)
+		J[q->x+pref] = q->y+pref;
+}
+
+/* check does double duty:
+1.  ferret out any fortuitous correspondences due
+to confounding by hashing (which result in "jackpot")
+2.  collect random access indexes to the two files */
+
+check()
+{
+	register int i, j;
+	int jackpot;
+	long ctold, ctnew;
+	register int c,d;
+
+	if ((input[0] = fopen(file1,"r")) == NULL) {
+		perror(file1);
+		done();
+	}
+	if ((input[1] = fopen(file2,"r")) == NULL) {
+		perror(file2);
+		done();
+	}
+	j = 1;
+	ixold[0] = ixnew[0] = 0;
+	jackpot = 0;
+	ctold = ctnew = 0;
+	for(i=1;i<=len[0];i++) {
+		if(J[i]==0) {
+			ixold[i] = ctold += skipline(0);
+			continue;
+		}
+		while(j<J[i]) {
+			ixnew[j] = ctnew += skipline(1);
+			j++;
+		}
+		if(bflag || wflag || iflag) {
+			for(;;) {
+				c = getc(input[0]);
+				d = getc(input[1]);
+				ctold++;
+				ctnew++;
+				if(bflag && isspace(c) && isspace(d)) {
+					do {
+						if(c=='\n')
+							break;
+						ctold++;
+					} while(isspace(c=getc(input[0])));
+					do {
+						if(d=='\n')
+							break;
+						ctnew++;
+					} while(isspace(d=getc(input[1])));
+				} else if ( wflag ) {
+					while( isspace(c) && c!='\n' ) {
+						c=getc(input[0]);
+						ctold++;
+					}
+					while( isspace(d) && d!='\n' ) {
+						d=getc(input[1]);
+						ctnew++;
+					}
+				}
+				if(chrtran[c] != chrtran[d]) {
+					jackpot++;
+					J[i] = 0;
+					if(c!='\n')
+						ctold += skipline(0);
+					if(d!='\n')
+						ctnew += skipline(1);
+					break;
+				}
+				if(c=='\n')
+					break;
+			}
+		} else {
+			for(;;) {
+				ctold++;
+				ctnew++;
+				if((c=getc(input[0])) != (d=getc(input[1]))) {
+					/* jackpot++; */
+					J[i] = 0;
+					if(c!='\n')
+						ctold += skipline(0);
+					if(d!='\n')
+						ctnew += skipline(1);
+					break;
+				}
+				if(c=='\n')
+					break;
+			}
+		}
+		ixold[i] = ctold;
+		ixnew[j] = ctnew;
+		j++;
+	}
+	for(;j<=len[1];j++) {
+		ixnew[j] = ctnew += skipline(1);
+	}
+	fclose(input[0]);
+	fclose(input[1]);
+/*
+	if(jackpot)
+		fprintf(stderr, "jackpot\n");
+*/
+}
+
+sort(a,n)	/*shellsort CACM #201*/
+struct line *a;
+{
+	struct line w;
+	register int j,m;
+	struct line *ai;
+	register struct line *aim;
+	int k;
+
+	if (n == 0)
+		return;
+	for(j=1;j<=n;j*= 2)
+		m = 2*j - 1;
+	for(m/=2;m!=0;m/=2) {
+		k = n-m;
+		for(j=1;j<=k;j++) {
+			for(ai = &a[j]; ai > a; ai -= m) {
+				aim = &ai[m];
+				if(aim < ai)
+					break;	/*wraparound*/
+				if(aim->value > ai[0].value ||
+				   aim->value == ai[0].value &&
+				   aim->serial > ai[0].serial)
+					break;
+				w.value = ai[0].value;
+				ai[0].value = aim->value;
+				aim->value = w.value;
+				w.serial = ai[0].serial;
+				ai[0].serial = aim->serial;
+				aim->serial = w.serial;
+			}
+		}
+	}
+}
+
+unsort(f, l, b)
+struct line *f;
+int *b;
+{
+	register int *a;
+	register int i;
+	a = (int *)talloc((l+1)*sizeof(int));
+	for(i=1;i<=l;i++)
+		a[f[i].serial] = f[i].value;
+	for(i=1;i<=l;i++)
+		b[i] = a[i];
+	free((char *)a);
+}
+
+skipline(f)
+{
+	register i, c;
+
+	for(i=1;(c=getc(input[f]))!='\n';i++)
+		if (c < 0)
+			return(i);
+	return(i);
+}
+
+output()
+{
+	int m;
+	register int i0, i1, j1;
+	int j0;
+	input[0] = fopen(file1,"r");
+	input[1] = fopen(file2,"r");
+	m = len[0];
+	J[0] = 0;
+	J[m+1] = len[1]+1;
+	if(opt!=D_EDIT) for(i0=1;i0<=m;i0=i1+1) {
+		while(i0<=m&&J[i0]==J[i0-1]+1) i0++;
+		j0 = J[i0-1]+1;
+		i1 = i0-1;
+		while(i1<m&&J[i1+1]==0) i1++;
+		j1 = J[i1+1]-1;
+		J[i1] = j1;
+		change(i0,i1,j0,j1);
+	} else for(i0=m;i0>=1;i0=i1-1) {
+		while(i0>=1&&J[i0]==J[i0+1]-1&&J[i0]!=0) i0--;
+		j0 = J[i0+1]-1;
+		i1 = i0+1;
+		while(i1>1&&J[i1-1]==0) i1--;
+		j1 = J[i1-1]+1;
+		J[i1] = j1;
+		change(i1,i0,j1,j0);
+	}
+	if(m==0)
+		change(1,0,1,len[1]);
+	if (opt==D_IFDEF) {
+		for (;;) {
+#define	c i0
+			c = getc(input[0]);
+			if (c < 0)
+				return;
+			putchar(c);
+		}
+#undef c
+	}
+	if (anychange && opt == D_CONTEXT)
+		dump_context_vec();
+}
+
+/*
+ * The following struct is used to record change information when
+ * doing a "context" diff.  (see routine "change" to understand the
+ * highly mneumonic field names)
+ */
+struct context_vec {
+	int	a;	/* start line in old file */
+	int	b;	/* end line in old file */
+	int	c;	/* start line in new file */
+	int	d;	/* end line in new file */
+};
+
+struct	context_vec	*context_vec_start,
+			*context_vec_end,
+			*context_vec_ptr;
+
+#define	MAX_CONTEXT	128
+
+/* indicate that there is a difference between lines a and b of the from file
+   to get to lines c to d of the to file.
+   If a is greater then b then there are no lines in the from file involved
+   and this means that there were lines appended (beginning at b).
+   If c is greater than d then there are lines missing from the to file.
+*/
+change(a,b,c,d)
+{
+	int lowa,upb,lowc,upd;
+	struct stat stbuf;
+
+	if (opt != D_IFDEF && a>b && c>d)
+		return;
+	if (anychange == 0) {
+		anychange = 1;
+		if(opt == D_CONTEXT) {
+			printf("*** %s	", file1);
+			stat(file1, &stbuf);
+			printf("%s--- %s	",
+			    ctime(&stbuf.st_mtime), file2);
+			stat(file2, &stbuf);
+			printf("%s", ctime(&stbuf.st_mtime));
+
+			context_vec_start = (struct context_vec *) 
+						malloc(MAX_CONTEXT *
+						   sizeof(struct context_vec));
+			context_vec_end = context_vec_start + MAX_CONTEXT;
+			context_vec_ptr = context_vec_start - 1;
+		}
+	}
+	if(opt == D_CONTEXT) {
+		/*
+		 * if this new change is within 'context' lines of
+		 * the previous change, just add it to the change
+		 * record.  If the record is full or if this
+		 * change is more than 'context' lines from the previous
+		 * change, dump the record, reset it & add the new change.
+		 */
+		if ( context_vec_ptr >= context_vec_end ||
+		     ( context_vec_ptr >= context_vec_start &&
+		       a > (context_vec_ptr->b + 2*context) &&
+		       c > (context_vec_ptr->d + 2*context) ) )
+			dump_context_vec();
+
+		context_vec_ptr++;
+		context_vec_ptr->a = a;
+		context_vec_ptr->b = b;
+		context_vec_ptr->c = c;
+		context_vec_ptr->d = d;
+		return;
+	}
+	switch (opt) {
+
+	case D_NORMAL:
+	case D_EDIT:
+		range(a,b,",");
+		putchar(a>b?'a':c>d?'d':'c');
+		if(opt==D_NORMAL)
+			range(c,d,",");
+		putchar('\n');
+		break;
+	case D_REVERSE:
+		putchar(a>b?'a':c>d?'d':'c');
+		range(a,b," ");
+		putchar('\n');
+		break;
+        case D_NREVERSE:
+                if (a>b)
+                        printf("a%d %d\n",b,d-c+1);
+                else {
+                        printf("d%d %d\n",a,b-a+1);
+                        if (!(c>d))
+                           /* add changed lines */
+                           printf("a%d %d\n",b, d-c+1);
+                }
+                break;
+	}
+	if(opt == D_NORMAL || opt == D_IFDEF) {
+		fetch(ixold,a,b,input[0],"< ", 1);
+		if(a<=b&&c<=d && opt == D_NORMAL)
+			prints("---\n");
+	}
+	fetch(ixnew,c,d,input[1],opt==D_NORMAL?"> ":"", 0);
+	if ((opt ==D_EDIT || opt == D_REVERSE) && c<=d)
+		prints(".\n");
+	if (inifdef) {
+		fprintf(stdout, "#endif /* %s */\n", endifname);
+		inifdef = 0;
+	}
+}
+
+range(a,b,separator)
+char *separator;
+{
+	printf("%d", a>b?b:a);
+	if(a<b) {
+		printf("%s%d", separator, b);
+	}
+}
+
+fetch(f,a,b,lb,s,oldfile)
+long *f;
+FILE *lb;
+char *s;
+{
+	register int i, j;
+	register int c;
+	register int col;
+	register int nc;
+	int oneflag = (*ifdef1!='\0') != (*ifdef2!='\0');
+
+	/*
+	 * When doing #ifdef's, copy down to current line
+	 * if this is the first file, so that stuff makes it to output.
+	 */
+	if (opt == D_IFDEF && oldfile){
+		long curpos = ftell(lb);
+		/* print through if append (a>b), else to (nb: 0 vs 1 orig) */
+		nc = f[a>b? b : a-1 ] - curpos;
+		for (i = 0; i < nc; i++)
+			putchar(getc(lb));
+	}
+	if (a > b)
+		return;
+	if (opt == D_IFDEF) {
+		if (inifdef)
+			fprintf(stdout, "#else /* %s%s */\n", oneflag && oldfile==1 ? "!" : "", ifdef2);
+		else {
+			if (oneflag) {
+				/* There was only one ifdef given */
+				endifname = ifdef2;
+				if (oldfile)
+					fprintf(stdout, "#ifndef %s\n", endifname);
+				else
+					fprintf(stdout, "#ifdef %s\n", endifname);
+			}
+			else {
+				endifname = oldfile ? ifdef1 : ifdef2;
+				fprintf(stdout, "#ifdef %s\n", endifname);
+			}
+		}
+		inifdef = 1+oldfile;
+	}
+
+	for(i=a;i<=b;i++) {
+		fseek(lb,f[i-1],0);
+		nc = f[i]-f[i-1];
+		if (opt != D_IFDEF)
+			prints(s);
+		col = 0;
+		for(j=0;j<nc;j++) {
+			c = getc(lb);
+			if (c == '\t' && tflag)
+				do
+					putchar(' ');
+				while (++col & 7);
+			else {
+				putchar(c);
+				col++;
+			}
+		}
+	}
+
+	if (inifdef && !wantelses) {
+		fprintf(stdout, "#endif /* %s */\n", endifname);
+		inifdef = 0;
+	}
+}
+
+#define POW2			/* define only if HALFLONG is 2**n */
+#define HALFLONG 16
+#define low(x)	(x&((1L<<HALFLONG)-1))
+#define high(x)	(x>>HALFLONG)
+
+/*
+ * hashing has the effect of
+ * arranging line in 7-bit bytes and then
+ * summing 1-s complement in 16-bit hunks 
+ */
+readhash(f)
+register FILE *f;
+{
+	register long sum;
+	register unsigned shift;
+	register t;
+	register space;
+
+	sum = 1;
+	space = 0;
+	if(!bflag && !wflag) {
+		if(iflag)
+			for(shift=0;(t=getc(f))!='\n';shift+=7) {
+				if(t==-1)
+					return(0);
+				sum += (long)chrtran[t] << (shift
+#ifdef POW2
+				    &= HALFLONG - 1);
+#else
+				    %= HALFLONG);
+#endif
+			}
+		else
+			for(shift=0;(t=getc(f))!='\n';shift+=7) {
+				if(t==-1)
+					return(0);
+				sum += (long)t << (shift
+#ifdef POW2
+				    &= HALFLONG - 1);
+#else
+				    %= HALFLONG);
+#endif
+			}
+	} else {
+		for(shift=0;;) {
+			switch(t=getc(f)) {
+			case -1:
+				return(0);
+			case '\t':
+			case ' ':
+				space++;
+				continue;
+			default:
+				if(space && !wflag) {
+					shift += 7;
+					space = 0;
+				}
+				sum += (long)chrtran[t] << (shift
+#ifdef POW2
+				    &= HALFLONG - 1);
+#else
+				    %= HALFLONG);
+#endif
+				shift += 7;
+				continue;
+			case '\n':
+				break;
+			}
+			break;
+		}
+	}
+	sum = low(sum) + high(sum);
+	return((short)low(sum) + (short)high(sum));
+}
+
+#include <a.out.h>
+
+asciifile(f)
+	FILE *f;
+{
+	char buf[BUFSIZ];
+	register int cnt;
+	register char *cp;
+
+	fseek(f, (long)0, 0);
+	cnt = fread(buf, 1, BUFSIZ, f);
+	if (cnt >= sizeof (struct exec)) {
+		struct exec hdr;
+		hdr = *(struct exec *)buf;
+		if (!N_BADMAG(hdr))
+			return (0);
+	}
+	cp = buf;
+	while (--cnt >= 0)
+		if (*cp++ & 0200)
+			return (0);
+	return (1);
+}
+
+
+/* dump accumulated "context" diff changes */
+dump_context_vec()
+{
+	register int	a, b, c, d;
+	register char	ch;
+	register struct	context_vec *cvp = context_vec_start;
+	register int	lowa, upb, lowc, upd;
+	register int	do_output;
+
+	if ( cvp > context_vec_ptr )
+		return;
+
+	lowa = max(1, cvp->a - context);
+	upb  = min(len[0], context_vec_ptr->b + context);
+	lowc = max(1, cvp->c - context);
+	upd  = min(len[1], context_vec_ptr->d + context);
+
+	printf("***************\n*** ");
+	range(lowa,upb,",");
+	printf(" ****\n");
+
+	/*
+	 * output changes to the "old" file.  The first loop suppresses
+	 * output if there were no changes to the "old" file (we'll see
+	 * the "old" lines as context in the "new" list).
+	 */
+	do_output = 0;
+	for ( ; cvp <= context_vec_ptr; cvp++)
+		if (cvp->a <= cvp->b) {
+			cvp = context_vec_start;
+			do_output++;
+			break;
+		}
+	
+	if ( do_output ) {
+		while (cvp <= context_vec_ptr) {
+			a = cvp->a; b = cvp->b; c = cvp->c; d = cvp->d;
+
+			if (a <= b && c <= d)
+				ch = 'c';
+			else
+				ch = (a <= b) ? 'd' : 'a';
+
+			if (ch == 'a')
+				fetch(ixold,lowa,b,input[0],"  ");
+			else {
+				fetch(ixold,lowa,a-1,input[0],"  ");
+				fetch(ixold,a,b,input[0],ch == 'c' ? "! " : "- ");
+			}
+			lowa = b + 1;
+			cvp++;
+		}
+		fetch(ixold, b+1, upb, input[0], "  ");
+	}
+
+	/* output changes to the "new" file */
+	printf("--- ");
+	range(lowc,upd,",");
+	printf(" ----\n");
+
+	do_output = 0;
+	for (cvp = context_vec_start; cvp <= context_vec_ptr; cvp++)
+		if (cvp->c <= cvp->d) {
+			cvp = context_vec_start;
+			do_output++;
+			break;
+		}
+	
+	if (do_output) {
+		while (cvp <= context_vec_ptr) {
+			a = cvp->a; b = cvp->b; c = cvp->c; d = cvp->d;
+
+			if (a <= b && c <= d)
+				ch = 'c';
+			else
+				ch = (a <= b) ? 'd' : 'a';
+
+			if (ch == 'd')
+				fetch(ixnew,lowc,d,input[1],"  ");
+			else {
+				fetch(ixnew,lowc,c-1,input[1],"  ");
+				fetch(ixnew,c,d,input[1],ch == 'c' ? "! " : "+ ");
+			}
+			lowc = d + 1;
+			cvp++;
+		}
+		fetch(ixnew, d+1, upd, input[1], "  ");
+	}
+
+	context_vec_ptr = context_vec_start - 1;
+}
diff --git a/usr.bin/diff/pathnames.h b/usr.bin/diff/pathnames.h
new file mode 100644
index 00000000000..97b9ae9717c
--- /dev/null
+++ b/usr.bin/diff/pathnames.h
@@ -0,0 +1,24 @@
+/*
+ * Copyright (c) 1989 The Regents of the University of California.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms are permitted
+ * provided that the above copyright notice and this paragraph are
+ * duplicated in all such forms and that any documentation,
+ * advertising materials, and other materials related to such
+ * distribution and use acknowledge that the software was developed
+ * by the University of California, Berkeley.  The name of the
+ * University may not be used to endorse or promote products derived
+ * from this software without specific prior written permission.
+ * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
+ *
+ *	@(#)pathnames.h	5.1 (Berkeley) 5/11/89
+ */
+
+#define	_PATH_DIFF	"/usr/bin/diff"
+#define	_PATH_DIFFH	"/usr/bin/diffh"
+#define	_PATH_PR	"/usr/bin/pr"
+#undef _PATH_TMP
+#define	_PATH_TMP	"/tmp/dXXXXX"