From: Dave Rolsky <autarch@urth.org>
Date: Mon, 14 Sep 2009 21:29:03 +0000 (-0500)
Subject: So apparently the case I thought could never happen _does_ happen.
X-Git-Tag: 0.93~3
X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=commitdiff_plain;h=018cba1d0623503e40de57ccf23fad368935f63f;p=gitmo%2FClass-MOP.git

So apparently the case I thought could never happen _does_ happen.

This fixes it, but I don't really understand why.
---

diff --git a/lib/Class/MOP/Package.pm b/lib/Class/MOP/Package.pm
index ec1aedd..441f8c9 100644
--- a/lib/Class/MOP/Package.pm
+++ b/lib/Class/MOP/Package.pm
@@ -379,12 +379,11 @@ sub get_method {
         }
     );
 
-    # we should never have a blessed map entry but no $code in the package
-    if ( blessed( $map_entry ) && !$code ) {
-        my $method = sprintf '%s::%s', $self->name, $method_name;
-        confess "Found a meta method object in the method map but no"
-            . " corresponding code entry in the symbol table for $method";
-    }
+    # This seems to happen in some weird cases where methods modifiers are
+    # added via roles or some other such bizareness. Honestly, I don't totally
+    # understand this, but returning the entry works, and keeps various MX
+    # modules from blowing up. - DR
+    return $map_entry if blessed $map_entry && !$code;
 
     return $map_entry if blessed $map_entry && $map_entry->body == $code;