From ddbd3b5eb2b17434dc02af398372abd7684b5955 Mon Sep 17 00:00:00 2001
From: Ward Wouts <ward@wouts.nl>
Date: Thu, 30 Aug 2001 09:51:04 +0000
Subject: [PATCH] almost complete rewrite. much saner/faster now

---
 list_same/list_same | 61 ++++++++++++++++++++++++---------------------
 1 file changed, 32 insertions(+), 29 deletions(-)

diff --git a/list_same/list_same b/list_same/list_same
index 8aafbab..dacfb65 100755
--- a/list_same/list_same
+++ b/list_same/list_same
@@ -5,12 +5,15 @@
 # Handy for cleaning up pictures like this:
 # list_same|cut -f 1 -d " "|xargs rm
 
-# This still sucks big time!
-# Can it handle directories at all?
-# Why does it bother with calculating md5 checksums of everything,
-# instead of just the files that have the same size?
+# wishlist:
+#	- entering a list of file to check on the commandline
+
 
 # Changelog:
+# 2001-08-30:
+#	_much_ saner now. only calculate md5s if sizes of
+#	files are the same
+#	nearly complete rewrite
 # 2001-08-29:
 #	use MD5; instead of a shell call to md5
 #	some cleaning of code
@@ -19,42 +22,42 @@ use MD5;
 
 opendir(DIR, ".") or die "can't open . $!";
 while (defined($file = readdir(DIR))) {
-	&get_finfo;
+	push @filelist, $file;
 }
 closedir(DIR);
 
-$old_md5="";
-$old_size="";
-$old_name="";
+foreach (@filelist) {
+	if ( -f $_ ) {
+		push @{$sizes{&get_size($_)}}, $_;
+	}
+}
 
-foreach $line (sort(@file_lijstje)) {
-#  print $line;
-	$line .= "\n";
-	$line =~ /(.*?)\s(.*?)\s(.*)/;
-	$md5 = $1; $size = $2; $name = $3;
-	if (($old_md5 eq $md5) && ($old_size eq $size)) {
-		print "$old_name $name\n";
-	} else {
-		$old_md5=$md5;
-		$old_size=$size;
-		$old_name=$name;
+foreach (keys %sizes) {
+	if (@{$sizes{$_}} > 1) {
+		%md5s = ();
+		foreach (@{$sizes{$_}}) {
+			push @{$md5s{&calc_md5($_)}}, $_;
+		}
+		foreach (keys %md5s) {
+			if (@{$md5s{$_}} > 1) {
+				foreach (@{$md5s{$_}}) { print "$_ "; }
+				print "\n";
+			}
+		}
 	}
 }
 
 #################################
-# dain bread functions
+# functions
 
-sub get_finfo() {
-	my ($dev,$ino,$mode,$nlink,$uid,$gid,$rdev,$size, $atime,$mtime,$ctime,$blksize,$blocks);
-	my $line;
-	if (($file ne ".")&&($file ne "..")) {
-		($dev,$ino,$mode,$nlink,$uid,$gid,$rdev,$size,
-		$atime,$mtime,$ctime,$blksize,$blocks) = stat $file;
-		$line = &calc_md5($file);
-		push @file_lijstje, "$line $size $file";
-	}
+sub get_size {
+	my $file = shift;
+	my @stat;
+	@stat = stat $file;
+	return $stat[7];
 }
 
+# same md5 calculation i use in mv_wrap
 sub calc_md5($file) {
 	my ($file, $digest);
 	$file = shift;