[Rt-commit] r9269 - in rt/branches/3.7-RTIR-RELENG: lib

[ruz at bestpractical.com]

Author: ruz
Date: Wed Oct 10 21:41:49 2007
New Revision: 9269

Modified:
   rt/branches/3.7-RTIR-RELENG/html/autohandler
   rt/branches/3.7-RTIR-RELENG/lib/RT.pm.in

Log:
* get rid of the workaround as we have it in the proper place now
* move comment
* add RT::Principal to the list

Modified: rt/branches/3.7-RTIR-RELENG/html/autohandler
==============================================================================
--- rt/branches/3.7-RTIR-RELENG/html/autohandler	(original)
+++ rt/branches/3.7-RTIR-RELENG/html/autohandler	Wed Oct 10 21:41:49 2007
@@ -64,12 +64,6 @@
     $m->autoflush( $m->request_comp->attr('AutoFlush') );
 }
 
-# XXX: on a cold server (just after restart) people could have a principal object
-# in the session, as we deserialize it so we never call constructor of the principal
-# class, so the list of accessible fields is empty and we die with "Method xxx is
-# not implemented in RT::Principal"
-{ my $tmp = RT::Principal->new( $RT::SystemUser ) }
-
 %ARGS = map {
 
     # if they've passed multiple values, they'll be an array. if they've

Modified: rt/branches/3.7-RTIR-RELENG/lib/RT.pm.in
==============================================================================
--- rt/branches/3.7-RTIR-RELENG/lib/RT.pm.in	(original)
+++ rt/branches/3.7-RTIR-RELENG/lib/RT.pm.in	Wed Oct 10 21:41:49 2007
@@ -391,11 +391,16 @@
     require RT::ObjectCustomFieldValues;
     require RT::Attributes;
 
+    # on a cold server (just after restart) people could have an object
+    # in the session, as we deserialize it so we never call constructor
+    # of the class, so the list of accessible fields is empty and we die
+    # with "Method xxx is not implemented in RT::SomeClass"
     $_->_BuildTableAttributes foreach qw(
         RT::Ticket
         RT::Transaction
         RT::Attachment
         RT::User
+        RT::Principal
         RT::Template
         RT::Queue
         RT::ScripAction