From 35cd477fea7b31a721e257bebdadc75245706435 Mon Sep 17 00:00:00 2001
From: Johannes Schauer Marin Rodrigues <josch@mister-muffin.de>
Date: Thu, 1 Feb 2024 05:53:51 +0100
Subject: [PATCH] Take hard links into account when computing disk usage based
 on dpkg-gencontrol.pl

Thanks: Guillem Jover <guillem@debian.org>, Sven Joachim <svenjoac@gmx.de>
---
 mmdebstrap | 20 ++++++++++++++------
 1 file changed, 14 insertions(+), 6 deletions(-)

diff --git a/mmdebstrap b/mmdebstrap
index a862a51..2da5c3a 100755
--- a/mmdebstrap
+++ b/mmdebstrap
@@ -4393,19 +4393,27 @@ sub approx_disk_usage {
     # We ignore /dev because depending on the mode, the directory might be
     # populated or not and we want consistent disk usage results independent
     # of the mode.
-    my $installed_size      = 0;
+    my $installed_size = 0;
+    my %hardlink;
     my $scan_installed_size = sub {
         if ($File::Find::name eq "$directory/dev") {
             # add all entries of @devfiles once
             $installed_size += scalar @devfiles;
+            return;
         } elsif ($File::Find::name =~ /^$directory\/dev\//) {
             # ignore everything below /dev
-        } elsif (-l $File::Find::name) {
-            # -f follows symlinks, so we first check if we have a symlink
-            $installed_size += 1;
-        } elsif (-f $File::Find::name) {
+            return;
+        }
+
+        lstat or error "cannot stat $File::Find::name";
+
+        if (-f _ or -l _) {
+            my ($dev, $ino, $nlink) = (lstat _)[0, 1, 3];
+            return if exists $hardlink{"$dev:$ino"};
+            # Track hardlinks to avoid repeated additions.
+            $hardlink{"$dev:$ino"} = 1 if $nlink > 1;
             # add file size in 1024 byte blocks, rounded up
-            $installed_size += int(((-s $File::Find::name) + 1024) / 1024);
+            $installed_size += int(((-s _) + 1024) / 1024);
         } else {
             # all other entries are assumed to only take up one block
             $installed_size += 1;