rundiff revision 338:6cf264d111b4
1#! /usr/bin/env perl
2
3# Copyright (c) 2003 The Regents of The University of Michigan
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 are
8# met: redistributions of source code must retain the above copyright
9# notice, this list of conditions and the following disclaimer;
10# redistributions in binary form must reproduce the above copyright
11# notice, this list of conditions and the following disclaimer in the
12# documentation and/or other materials provided with the distribution;
13# neither the name of the copyright holders nor the names of its
14# contributors may be used to endorse or promote products derived from
15# this software without specific prior written permission.
16#
17# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
18# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
19# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
20# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
21# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
22# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
23# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
24# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
25# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
26# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
27# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
28
29# Diff two streams.
30#
31# Unlike regular diff, this script does not read in the entire input
32# before doing a diff, so it can be used on lengthy outputs piped from
33# other programs (e.g., M5 traces).  The best way to do this is to
34# take advantage of the power of Perl's open function, which will
35# automatically fork a subprocess if the last character in the
36# "filename" is a pipe (|).  Thus to compare the instruction traces
37# from two versions of m5 (m5a and m5b), you can do this:
38#
39# rundiff 'm5a --trace:flags=InstExec |' 'm5b --trace:flags=InstExec |'
40#
41
42use strict;
43
44#
45# For the highest-quality (minimal) diffs, we can use the
46# Algorithm::Diff package.  If you don't have this installed, or want
47# the script to run faster (like 3-4x faster, based on informal
48# observation), set $use_complexdiff to 0; then a built-in, simple,
49# and generally quite adequate algorithm will be used instead.
50my $use_complexdiff = 0;
51
52#if ($use_complexdiff) {
53#    use Algorithm::Diff qw(traverse_sequences);
54#};
55
56my $lookahead_lines = 200;
57my $precontext_lines = 3;
58my $postcontext_lines = 3;
59
60my $file1 = $ARGV[0];
61my $file2 = $ARGV[1];
62
63die "Need two args." if (!(defined($file1) && defined($file2)));
64
65my ($fh1, $fh2);
66open($fh1, $file1) or die "Can't open $file1";
67open($fh2, $file2) or die "Can't open $file2";
68
69# buffer of matching lines for pre-diff context
70my @precontext = ();
71# number of post-diff matching lines remaining to print
72my $postcontext = 0;
73
74# lookahead buffers for $file1 and $file2 respectively
75my @lines1 = ();
76my @lines2 = ();
77
78# Next line number available to print from each file.  Generally this
79# corresponds to the oldest line in @precontext, or the oldest line in
80# @lines1 and @lines2 if @precontext is empty.
81my $lineno1 = 1;
82my $lineno2 = 1;
83
84# Fill a lookahead buffer to $lookahead_lines lines (or until EOF).
85sub fill
86{
87    my ($fh, $array) = @_;
88
89    while (@$array < $lookahead_lines) {
90	my $line = <$fh>;
91	last if (!defined($line));
92	push @$array, $line;
93    }
94}
95
96# Print and delete n lines from front of given array with given prefix.
97sub printlines
98{
99    my ($array, $n, $prefix) = @_;
100
101    while ($n--) {
102	my $line = shift @$array;
103	last if (!defined($line));
104	print $prefix, $line;
105    }
106}
107
108# Print a difference region where n1 lines of file1 were replaced by
109# n2 lines of file2 (where either n1 or n2 could be zero).
110sub printdiff
111{
112    my ($n1, $n2)= @_;
113
114    # If the precontext buffer is full or we're at the beginning of a
115    # file, then this is a new diff region, so we should print a
116    # header indicating the current line numbers.  If we're past the
117    # beginning and the precontext buffer isn't full, then whatever
118    # we're about to print is contiguous with the end of the last
119    # region we printed, so we just concatenate them on the output.
120    if (@precontext == $precontext_lines || ($lineno1 == 0 && $lineno2 == 0)) {
121	print "@@ -$lineno1 +$lineno2 @@\n";
122    }
123
124    # Print and clear the precontext buffer.
125    if (@precontext) {
126	print ' ', join(' ', @precontext);
127	$lineno1 += scalar(@precontext);
128	$lineno2 += scalar(@precontext);
129	@precontext = ();
130    }
131
132    # Print the differing lines.
133    printlines(\@lines1, $n1, '-');
134    printlines(\@lines2, $n2, '+');
135    $lineno1 += $n1;
136    $lineno2 += $n2;
137
138    # Set $postcontext to print the next $postcontext_lines matching lines.
139    $postcontext = $postcontext_lines;
140}
141
142
143########################
144#
145# Complex diff algorithm
146#
147########################
148
149{
150    my $match_found;
151    my $discard_lines1;
152    my $discard_lines2;
153
154    sub match { $match_found = 1; }
155    sub discard1 { $discard_lines1++ unless $match_found; }
156    sub discard2 { $discard_lines2++ unless $match_found; }
157
158    sub complex_diff
159    {
160	$match_found = 0;
161	$discard_lines1 = 0;
162	$discard_lines2 = 0;
163
164	# See Diff.pm.  Note that even though this call generates a
165	# complete diff of both lookahead buffers, all we use it for
166	# is to figure out how many lines to discard off the front of
167	# each buffer to resync the streams.
168	traverse_sequences( \@lines1, \@lines2,
169			    { MATCH => \&match,
170			      DISCARD_A => \&discard1,
171			      DISCARD_B => \&discard2 });
172
173	die "Lost sync!" if (!$match_found);
174
175	# Since we shouldn't get here unless the first lines of the
176	# buffers are different, then we must discard some lines off
177	# at least one of the buffers.
178	die if ($discard_lines1 == 0 && $discard_lines2 == 0);
179
180	printdiff($discard_lines1, $discard_lines2);
181    }
182}
183
184#######################
185#
186# Simple diff algorithm
187#
188#######################
189
190# Check for a pair of matching lines; if found, generate appropriate
191# diff output.
192sub checkmatch
193{
194    my ($n1, $n2) = @_;
195
196    # Check if two adjacent lines match, to reduce false resyncs
197    # (particularly on unrelated blank lines).  This generates
198    # larger-than-necessary diffs when a single line really should be
199    # treated as common; if that bugs you, use Algorithm::Diff.
200    if ($lines1[$n1] eq $lines2[$n2] && $lines1[$n1+1] eq $lines2[$n2+1]) {
201	printdiff($n1, $n2);
202    }
203}
204
205sub simple_diff
206{
207    # Look for differences of $cnt lines to resync,
208    # increasing $cnt from 1 to $lookahead_lines until we find
209    # something.
210    for (my $cnt = 1; $cnt < $lookahead_lines-1; ++$cnt) {
211	# Check for n lines in one file being replaced by
212	# n lines in the other.
213	return if checkmatch($cnt, $cnt);
214	# Find differences where n lines in one file were
215	# replaced by m lines in the other.  We let m = $cnt
216	# and iterate for n = 0 to $cnt-1.
217	for (my $n = 0; $n < $cnt; ++$n) {
218	    return if checkmatch($n, $cnt);
219	    return if checkmatch($cnt, $n);
220	}
221    }
222    die "Lost sync!";
223}
224
225# Set the pointer to the appropriate diff function.
226#
227# Note that in either case the function determines how many lines to
228# discard from the front of each lookahead buffer to resync the
229# streams, then prints the appropriate diff output and discards them.
230# After the function returns, it should always be the case that
231# $lines1[0] eq $lines2[0].
232my $find_diff = $use_complexdiff ? \&complex_diff : \&simple_diff;
233
234# The main loop.
235while (1) {
236    # keep lookahead buffers topped up
237    fill($fh1, \@lines1);
238    fill($fh2, \@lines2);
239
240    # peek at first line in each buffer
241    my $l1 = $lines1[0];
242    my $l2 = $lines2[0];
243
244    if (!defined($l1) && !defined($l2)) {
245	# reached EOF on both streams: exit
246	exit(1);
247    }
248
249    if ($l1 eq $l2) {
250	# matching lines: delete from lookahead buffer
251	shift @lines1;
252	shift @lines2;
253	# figure out what to do with this line
254	if ($postcontext > 0) {
255	    # we're in the post-context of a diff: print it
256	    $postcontext--;
257	    print ' ', $l1;
258	    $lineno1++;
259	    $lineno2++;
260	}
261	else {
262	    # we're in the middle of a matching region... save this
263	    # line for precontext in case we run into a difference.
264	    push @precontext, $l1;
265	    # don't let precontext buffer get bigger than needed
266	    while (@precontext > $precontext_lines) {
267		shift @precontext;
268		$lineno1++;
269		$lineno2++;
270	    }
271	}
272    }
273    else {
274	# Mismatch.  Deal with it.
275	&$find_diff();
276    }
277}
278