X-Git-Url: http://git.hungrycats.org/cgi-bin/gitweb.cgi?p=dupemerge;a=blobdiff_plain;f=faster-dupemerge;h=4cdef9716fd5e5de600b9810a198936ca9d46301;hp=9920a26bf2bb531bd2a22f2339b83c601aee2cb3;hb=de7efbca76dfc8ca745d8dabc2f9a99d9af0e4b4;hpb=af4afacf460ad2f6c2eafe7632bb55f2060823c6 diff --git a/faster-dupemerge b/faster-dupemerge index 9920a26..4cdef97 100755 --- a/faster-dupemerge +++ b/faster-dupemerge @@ -4,7 +4,7 @@ use Fcntl qw(:DEFAULT :flock); use File::Compare; use File::Temp; -# Copyright (C) 2003-2010 Zygo Blaxell +# Copyright (C) 2002-2010 Zygo Blaxell # 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 @@ -86,7 +86,9 @@ my $collapse_access = 0; my $collapse_timestamp = 0; my $collapse_zero = 0; my $skip_compares = 0; +my $skip_compare_preference = 0; my $skip_hashes = 0; +my $skip_hashes_threshold = 0; my $progress = 0; my $verbose = 0; my $debug = 0; @@ -153,6 +155,7 @@ hard links). S, s - lstat(2) (see source for details) U - unlink(2) . - all inodes with similar attributes done + (123456) - current file size in bytes --sort pass next options (up to --) to sort command @@ -160,7 +163,10 @@ hard links). --skip-compare skip byte-by-byte file comparisons - --skip-hash skip calculation of hash function on files + --skip-hash[=N] skip calculation of hash function on files + larger than N bytes (default 1M). + Scalars KMGT specify KiB, MiB, GiB, and TiB. + Scalars kmgt specify KB, MB, GB, and TB. --trust old name for --skip-compare (trust the hash function) @@ -180,9 +186,24 @@ while ($#ARGV >= 0) { } elsif ($arg eq '--zeros') { $collapse_zero = 1; } elsif ($arg eq '--trust' || $arg eq '--skip-compare') { - $skip_compares = 1; - } elsif ($arg eq '--skip-hash') { - $skip_hashes = 1; + $skip_compares = $skip_compare_preference = 1; + } elsif ($arg =~ /^--skip-hash(?:=(\d+)([KkMmGgTt]?))?$/os) { + my ($quantity, $unit) = ($1, $2); + $unit ||= '_'; + $quantity ||= 1048576; + my %scale = ( + _ => 1, + k => 1000, + K => 1024, + m => 1000*1000, + M => 1024*1024, + g => 1000*1000*1000, + G => 1024*1024*1024, + t => 1000*1000*1000*1000, + T => 1024*1024*1024*1024, + ); + $skip_hashes = 0; + $skip_hashes_threshold = $quantity * $scale{$unit}; } elsif ($arg eq '--progress') { $progress = 1; } elsif ($arg eq '--verbose') { @@ -221,11 +242,7 @@ while ($#ARGV >= 0) { } } -if ($skip_hashes && $skip_compares) { - die "Cannot skip both hashes and compares.\n"; -} - -@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: $!"; @@ -339,11 +356,13 @@ hash_file: foreach my $filename (sort keys(%{$inode_to_file_name{$candidate}})) { print STDERR "\t\tDigesting file $filename\n" if $debug; if ((-l $filename) || ! -f _) { + print STDERR "\n" if $progress; warn "Bogon file " . tick_quote($filename); - $surprises++; + $input_bogons++; + delete $inode_to_file_name{$candidate}->{$filename}; next; } - eval { + eval { $digest = digest($filename); }; if ($@) { @@ -379,14 +398,14 @@ incumbent_file: my ($incumbent_dev,$incumbent_ino,$incumbent_mode,$incumbent_nlink,$incumbent_uid,$incumbent_gid,$incumbent_rdev,$incumbent_size,$incumbent_atime,$incumbent_mtime,$incumbent_ctime,$incumbent_blksize,$incumbent_blocks) = lstat($incumbent_file); print STDERR "\t\tINCUMBENT dev=$incumbent_dev ino=$incumbent_ino mode=$incumbent_mode nlink=$incumbent_nlink uid=$incumbent_uid gid=$incumbent_gid rdev=$incumbent_rdev size=$incumbent_size atime=$incumbent_atime mtime=$incumbent_mtime ctime=$incumbent_ctime blksize=$incumbent_blksize blocks=$incumbent_blocks _=$incumbent_file\n" if $debug; - if (!defined($incumbent_blocks)) { + if (!defined($incumbent_blocks) || ! (-f _)) { warn "lstat: $incumbent_file: $!"; $surprises++; next incumbent_file; } if (format_inode($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); $surprises++; next incumbent_file; } @@ -400,14 +419,14 @@ candidate_file: my ($candidate_dev,$candidate_ino,$candidate_mode,$candidate_nlink,$candidate_uid,$candidate_gid,$candidate_rdev,$candidate_size,$candidate_atime,$candidate_mtime,$candidate_ctime,$candidate_blksize,$candidate_blocks) = lstat($candidate_file); print STDERR "\t\t\tCANDIDATE dev=$candidate_dev ino=$candidate_ino mode=$candidate_mode nlink=$candidate_nlink uid=$candidate_uid gid=$candidate_gid rdev=$candidate_rdev size=$candidate_size atime=$candidate_atime mtime=$candidate_mtime ctime=$candidate_ctime blksize=$candidate_blksize blocks=$candidate_blocks _=$candidate_file\n" if $debug; - if (!defined($candidate_blocks)) { + if (!defined($candidate_blocks) || ! (-f _)) { warn "lstat: $candidate_file: $!"; $surprises++; next candidate_file; } if (format_inode($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); $surprises++; next candidate_file; } @@ -523,7 +542,7 @@ candidate_file: if ($link_done) { delete $inode_to_file_name{$to_inode}->{$to_file}; - $inode_to_file_name{$from_inode}->{$to_file} = undef; + $inode_to_file_name{$from_inode}->{$to_file} = undef unless ($dry_run); $hash_to_inode{$digest} = [ $from_inode ]; $hard_links++; @@ -562,26 +581,40 @@ candidate_file: delete $inode_to_file_name{$candidate}; } } + print STDERR '.' if $progress; end_merge: - print STDERR '.' if $progress; print STDERR "Merge done.\n" if $debug; undef %inode_to_file_name; } +my $last_time = 0; +my $last_size = 0; + while () { - my ($weak_key, $dev, $ino, $name) = m/^(\d+ \d+ \d+ \d+ -?[\d.]+) (\d+) (\d+) (.+)\0$/so; + my ($weak_key, $size, $dev, $ino, $name) = m/^((\d+) \d+ \d+ \d+ -?[\d.]+) (\d+) (\d+) (.+)\0$/so; die "read error: $!\nLast input line was '$_'" unless defined($name); my $inode = format_inode($dev, $ino); print STDERR "weak_key=$weak_key inode=$inode name=$name\n" if $debug; - unless (! (-l $name) && (-f _)) { - warn "Bogon file " . tick_quote($name); - $input_bogons++; - next; + if ($skip_hashes_threshold && $size >= $skip_hashes_threshold) { + $skip_hashes = 1; + $skip_compares = 0; + } else { + $skip_hashes = 0; + $skip_compares = $skip_compare_preference; + } + + if ($progress) { + my $this_time = time(); + if ($this_time != $last_time && $size != $last_size) { + $last_time = $this_time; + $last_size = $size; + print STDERR "($size)"; + } } $input_links++; @@ -632,7 +665,7 @@ if ($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 $num; } @@ -642,6 +675,7 @@ if ($humane) { $stats_blob =~ s/([^\n]*\n[^\n]*? )(\s+)( [^\n]*\n)/$1 . ('.' x length($2)) . $3/oemg; +print STDERR "\n" if $progress; print STDERR $stats_blob; exit(0);