From: Peter Rabbitson <ribasushi@cpan.org>
Date: Tue, 19 May 2015 03:49:48 +0000 (+0200)
Subject: Make sure tests pass without a compiler present (another step to RT#74706)
X-Git-Tag: v0.082821~16
X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=commitdiff_plain;h=ab63014e90ff44053204ba629619838e5317720f;p=dbsrgits%2FDBIx-Class.git

Make sure tests pass without a compiler present (another step to RT#74706)
---

diff --git a/Changes b/Changes
index c887d97..fbf13dd 100644
--- a/Changes
+++ b/Changes
@@ -1,5 +1,8 @@
 Revision history for DBIx::Class
 
+    * Fixes
+        - Fix t/52leaks.t failures on compilerless systems (RT#104429)
+
 0.082820 2015-03-20 20:35 (UTC)
     * Fixes
         - Protect destructors from rare but possible double execution, and
diff --git a/t/52leaks.t b/t/52leaks.t
index af91f7d..b3b9b9c 100644
--- a/t/52leaks.t
+++ b/t/52leaks.t
@@ -447,6 +447,14 @@ for my $addr (keys %$weak_registry) {
     delete $weak_registry->{$addr}
       unless $cleared->{hash_merge_singleton}{$weak_registry->{$addr}{weakref}{behavior}}++;
   }
+  elsif ($names =~ /^B::Hooks::EndOfScope::PP::_TieHintHashFieldHash/m) {
+    # there is one tied lexical which stays alive until GC time
+    # https://metacpan.org/source/ETHER/B-Hooks-EndOfScope-0.15/lib/B/Hooks/EndOfScope/PP/FieldHash.pm#L24
+    # simply ignore it here, instead of teaching the leaktracer to examine ties
+    # the latter is possible yet terrible: https://github.com/dbsrgits/dbix-class/blob/v0.082820/t/lib/DBICTest/Util/LeakTracer.pm#L113-L117
+    delete $weak_registry->{$addr}
+      unless $cleared->{bheos_pptiehinthashfieldhash}++;
+  }
   elsif ($names =~ /^DateTime::TimeZone::UTC/m) {
     # DT is going through a refactor it seems - let it leak zones for now
     delete $weak_registry->{$addr};