X-Git-Url: http://git.hungrycats.org/cgi-bin/gitweb.cgi?p=dupemerge;a=blobdiff_plain;f=faster-dupemerge;h=2c85685b7056ba3c55abb05df99d84faf7aa34bb;hp=fe23f8e70619443a8cc6fb3637dc6c07354b6ba5;hb=5d6afd72794a9be8dcf2040fd4cad7fb1ad75cd6;hpb=f8527d273c71e1dc1302cfe9e854d6a57cf9b48b

diff --git a/faster-dupemerge b/faster-dupemerge
index fe23f8e..2c85685 100755
--- a/faster-dupemerge
+++ b/faster-dupemerge
@@ -4,7 +4,7 @@ use Fcntl qw(:DEFAULT :flock);
 use File::Compare;
 use File::Temp;
 
-# Copyright (C) 2002-2010 Zygo Blaxell <faster-dupemerge@mailtoo.hungrycats.org>
+# Copyright (C) 2002-2012 Zygo Blaxell <faster-dupemerge@mailtoo.hungrycats.org>
 
 # This program is free software; you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
@@ -155,12 +155,14 @@ hard links).
                         S, s - lstat(2) (see source for details)
                         U - unlink(2)
                         . - all inodes with similar attributes done
+                        (123456) - current file size in bytes
 
         --sort          pass next options (up to --) to sort command
 
         --timestamps    mtime may be different for identical files
 
-        --skip-compare  skip byte-by-byte file comparisons
+        --skip-compare  skip byte-by-byte file comparisons,
+                        compare only file hashes
 
         --skip-hash[=N] skip calculation of hash function on files
                         larger than N bytes (default 1M).
@@ -173,6 +175,12 @@ hard links).
         --verbose       report files as they are considered
 
         --zeros         hard-link zero-length files too
+
+--skip-compare and --skip-hash can be combined, in which case a file is
+either hashed (if it is below the --skip-hash size threshold) or compared
+(if it is above), but never both.
+
+Version:  0.20120103
 USAGE
 }
 
@@ -185,7 +193,7 @@ while ($#ARGV >= 0) {
 	} elsif ($arg eq '--zeros') {
 		$collapse_zero = 1;
 	} elsif ($arg eq '--trust' || $arg eq '--skip-compare') {
-		$skip_compare_preference = 1;
+		$skip_compares = $skip_compare_preference = 1;
 	} elsif ($arg =~ /^--skip-hash(?:=(\d+)([KkMmGgTt]?))?$/os) {
 		my ($quantity, $unit) = ($1, $2);
 		$unit ||= '_';
@@ -241,7 +249,7 @@ while ($#ARGV >= 0) {
 	}
 }
 
-@directories or usage;
+@directories or usage($0);
 
 if (defined($lock_file) && !$dry_run) {
 	sysopen(LOCK_FILE, $lock_file, O_CREAT|O_RDONLY, 0666) or die "open: $lock_file: $!";
@@ -404,7 +412,7 @@ incumbent_file:
 						}
 
 						if (format_inode($incumbent_dev, $incumbent_ino) ne $incumbent) {
-							warn "$incumbent_file: expected inode $incumbent, found $incumbent_dev:$incumbent_ino";
+							warn "$incumbent_file: expected inode $incumbent, found ".format_inode($incumbent_dev, $incumbent_ino);
 							$surprises++;
 							next incumbent_file;
 						}
@@ -425,7 +433,7 @@ candidate_file:
 							}
 
 							if (format_inode($candidate_dev, $candidate_ino) ne $candidate) {
-								warn "$candidate_file: expected inode $candidate, found $candidate_dev:$candidate_ino";
+								warn "$candidate_file: expected inode $candidate, found ".format_inode($candidate_dev, $candidate_ino);
 								$surprises++;
 								next candidate_file;
 							}
@@ -589,6 +597,7 @@ end_merge:
 }
 
 my $last_time = 0;
+my $last_size = 0;
 
 while (<FIND>) {
 	my ($weak_key, $size, $dev, $ino, $name) = m/^((\d+) \d+ \d+ \d+ -?[\d.]+) (\d+) (\d+) (.+)\0$/so;
@@ -598,7 +607,7 @@ while (<FIND>) {
 
 	print STDERR "weak_key=$weak_key inode=$inode name=$name\n" if $debug;
 
-	if ($size >= $skip_hashes_threshold) {
+	if ($skip_hashes_threshold && $size >= $skip_hashes_threshold) {
 		$skip_hashes = 1;
 		$skip_compares = 0;
 	} else {
@@ -608,8 +617,9 @@ while (<FIND>) {
 
 	if ($progress) {
 		my $this_time = time();
-		if ($this_time != $last_time) {
+		if ($this_time != $last_time && $size != $last_size) {
 			$last_time = $this_time;
+			$last_size = $size;
 			print STDERR "($size)";
 		}
 	}
@@ -662,7 +672,7 @@ if ($humane) {
 
 	sub space_numbers {
 		my ($num) = @_;
-		1 while $num =~ s/(\d)(\d\d\d)( \d\d\d)*$/$1 $2$3/os;
+		1 while $num =~ s/(\d)(\d\d\d)((?: \d\d\d)*)$/$1 $2$3/os;
 		$num = ' ' x ($max_num_len - length($num)) . $num;
 		return $num;
 	}