projects
/
dupemerge
/ blobdiff
commit
grep
author
committer
pickaxe
?
search:
re
summary
|
shortlog
|
log
|
commit
|
commitdiff
|
tree
raw
|
inline
| side by side
dupemerge: inodes are now non-numeric
[dupemerge]
/
faster-dupemerge
diff --git
a/faster-dupemerge
b/faster-dupemerge
index 5069297630ee8b5d15b930f1a1eedc29debe07b9..7e77776d59474a93e560d97989d39f2601c9c899 100755
(executable)
--- a/
faster-dupemerge
+++ b/
faster-dupemerge
@@
-238,7
+238,7
@@
my @find_command = ('find', @directories, @extra_find_opts, '-type', 'f');
my $printf_string = '%s ' .
($collapse_access ? '0 0 0 ' : '%U %G %m ') .
($collapse_timestamp ? '0 ' : '%T@ ') .
my $printf_string = '%s ' .
($collapse_access ? '0 0 0 ' : '%U %G %m ') .
($collapse_timestamp ? '0 ' : '%T@ ') .
- '%D
:
%i %p\0';
+ '%D
%i %p\0';
push(@find_command, '!', '-empty') unless $collapse_zero;
push(@find_command, '-printf', $printf_string);
push(@find_command, '!', '-empty') unless $collapse_zero;
push(@find_command, '-printf', $printf_string);
@@
-288,6
+288,12
@@
sub link_files {
}
}
}
}
+# Convert $dev,$ino into a single string where lexical and numeric orderings are equivalent
+sub format_inode ($$) {
+ my ($dev, $ino) = @_;
+ return sprintf('%016x:%016x', $dev, $ino);
+}
+
# Process all known files so far.
sub merge_files {
$merges_attempted++;
# Process all known files so far.
sub merge_files {
$merges_attempted++;
@@
-335,7
+341,7
@@
hash_file:
my $incumbent_list = ($hash_to_inode{$digest} ||= []);
my $incumbent_matched = 0;
my $incumbent_list = ($hash_to_inode{$digest} ||= []);
my $incumbent_matched = 0;
- for my $incumbent (@$incumbent_list) {
+ for my $incumbent (
sort
@$incumbent_list) {
print STDERR "\t\tInodes $incumbent and $candidate have same hash\n" if $debug;
my $finished = 0;
print STDERR "\t\tInodes $incumbent and $candidate have same hash\n" if $debug;
my $finished = 0;
@@
-360,7
+366,7
@@
incumbent_file:
next incumbent_file;
}
next incumbent_file;
}
- if (
$incumbent_ino !=
$incumbent) {
+ if (
format_inode($incumbent_dev, $incumbent_ino) ne
$incumbent) {
warn "$incumbent_file: expected inode $incumbent, found $incumbent_ino";
$surprises++;
next incumbent_file;
warn "$incumbent_file: expected inode $incumbent, found $incumbent_ino";
$surprises++;
next incumbent_file;
@@
-380,7
+386,7
@@
candidate_file:
next candidate_file;
}
next candidate_file;
}
- if (
$candidate_ino !=
$candidate) {
+ if (
format_inode($candidate_dev, $candidate_ino) ne
$candidate) {
warn "$candidate_file: expected inode $candidate, found $candidate_ino";
$surprises++;
next candidate_file;
warn "$candidate_file: expected inode $candidate, found $candidate_ino";
$surprises++;
next candidate_file;
@@
-543,9
+549,12
@@
end_merge:
}
while (<FIND>) {
}
while (<FIND>) {
- my ($weak_key, $
inode, $name) = m/^(\d+ \d+ \d+ \d+ -?[\d.]+) (\d+:
\d+) (.+)\0$/so;
+ my ($weak_key, $
dev, $ino, $name) = m/^(\d+ \d+ \d+ \d+ -?[\d.]+) (\d+) (
\d+) (.+)\0$/so;
die "read error: $!\nLast input line was '$_'" unless defined($name);
die "read error: $!\nLast input line was '$_'" unless defined($name);
+ # 64 bits out to be enough for everybody!
+ my $inode = format_inode($dev, $ino);
+
print STDERR "weak_key=$weak_key inode=$inode name=$name\n" if $debug;
unless (! (-l $name) && (-f _)) {
print STDERR "weak_key=$weak_key inode=$inode name=$name\n" if $debug;
unless (! (-l $name) && (-f _)) {