From 3163f7d94d498b022b28744bf8f5edd69cc1c069 Mon Sep 17 00:00:00 2001
From: hjp <hjp>
Date: Sat, 24 Apr 2010 15:25:15 +0000
Subject: [PATCH] create histogram of file sizes

---
 filesize/filesize | 29 +++++++++++++++++++++++++++++
 1 file changed, 29 insertions(+)
 create mode 100755 filesize/filesize

diff --git a/filesize/filesize b/filesize/filesize
new file mode 100755
index 0000000..114882c
--- /dev/null
+++ b/filesize/filesize
@@ -0,0 +1,29 @@
+#!/usr/bin/perl
+use warnings;
+use strict;
+
+use File::Find;
+
+my %seen;
+my %hist;
+
+sub collect {
+    my ($dev,$ino,$mode,$nlink,$uid,$gid,$rdev,$size,
+	$atime,$mtime,$ctime,$blksize,$blocks) = lstat($_);
+    return if $seen{"$dev:$ino"};
+    $hist{$size}++;
+}
+
+find(\&collect, @ARGV ? @ARGV : ("."));
+
+my $total_count;
+for my $c (values %hist) {
+    $total_count += $c;
+}
+
+my $ac = 0;
+for my $s (sort {$a <=> $b } keys %hist) {
+    $ac += $hist{$s};
+    printf "%g %g %6.2f\n", $s, $ac, 100 * $ac/$total_count;
+}
+