[rt-users] RT errors on search functions / a.k.a. tipsontrackingdown the source of errors in RT...

Stephen Turner sturner at MIT.EDU
Fri Dec 30 10:29:33 EST 2005

At Thursday 12/29/2005 09:25 AM, Duncan Shannon wrote:
>I have now removed the Saved Search displays on the Home page (by
>removing the quips from index.html (and resorted back to the orig
>version) and my problems seem to have cleared up.
>What is it with the Saved Searches that is going awry (especially all of
>the sudden... nothing should have changed w/r/t my saved searches other
>than the following to chances: 1) I added the text based priorities
>(high/low etc) which I cant imagine affected the Saved Searches (and I
>subsequently rolled the changes back assuming it didn't touch the DB)
>and 2) I added a couple of saved searches that I think are pretty non
>exciting, other than I am now using "LastUpdated < 7 days ago". Could
>that be part of it?
>I saw the thread on rt-devel
>And it must be happening to a few people.
>I'd really like to help out how ever I can (that applies to all things
>Please I hope someone is reading this and can offer some advise on how
>to fix my problem AND/OR how to help out with RT.

Hello Duncan,

I don't believe the cause of this error (identified by the message 
"RT::Attribute::Name Unimplemented in RT::Attributes) has been definitely 
identified - it seems to be somewhere deep in the code that retrieves 
attributes for an object. Because Saved Searches are stored as user 
attributes, the error is often seen when working with them. We've seen the 
error with attribute settings other than saved searches that we've added to 
RT ourselves - I wouldn't spend time worrying about whether specific saved 
search criteria have caused a problem.

A fix that we tried (suggested by Jesse) was to upgrade to the latest 
DBIx::SearchBuilder (we are now at 1.36) and to remove one line of code 
from SB::Record->__Value:

   $field = $self->_Accessible($field, "column") if 
$self->_Accessible($field, "column");

This _seems_ to have gotten rid of the problem for us - however, I 
personally don't understand what the omission of the code changes and so I 
don't understand why this fix works. It's not clear that this will 
guarantee a solution to the problem, so if you try it I strongly recommend 
trying it on a test/devel setup first, and if you try it in production, 
make sure you can easily back the change out if you need to.

Good luck -

More information about the rt-users mailing list