[drupal-devel] [bug] Deleting feeds and categories leaves error-generating block

[drumm]

Issue status update for http://drupal.org/node/17160

 Project:      Drupal
 Version:      cvs
 Component:    aggregator.module
 Category:     bug reports
 Priority:     normal
 Assigned to:  drumm
 Reported by:  drumm
 Updated by:   drumm
 Status:       patch

These queries will go away as soon as someone saves the block position
settings. I don't think we want to go into block's tables and call a
"private" function from the aggregator module. This does not add
queries, although we could rewrite this to remove queries, but I do not
think we should for this edge case considering all the block module
innards it would require using.


drumm



Previous comments:
------------------------------------------------------------------------

February 10, 2005 - 18:32 : drumm

Attachment: http://drupal.org/files/issues/aggregator.module_1.diff (2.72 KB)

Make a feed with the block enabled. Delete the feed. The block remains
and spews a database error on every page view. This fixes that.
As for 4.5, the diff file at
http://trac.civicspacelabs.com/cgi-bin/trac.cgi/changeset/273?format=diff
might apply if the attached diff does not. The indentation is the only
thing which might have changed in the section of code from 4.5 to CVS
that I can see.


------------------------------------------------------------------------

February 11, 2005 - 13:56 : Dries

I'm OK with this patch but wouldn't it be better if we did not have to
execute those queries for deleted/non-existing blocks?  What do other
modules do to prevent this kind of behavior? (_block_rehash() appears
to be a private function.)


------------------------------------------------------------------------

February 11, 2005 - 14:19 : andremolnar

Block.module is required in Drupal - so _block_rehash is always
available - i've often called it from a different module when doing
tests and debugging (so as not to have to go to the block page to have
it fire and rehash the blocks). 
A little off topic, but could (or maybe should) block_rehash be moved
to common.inc???
andre