Taxonomy duplication
Eshilon Design Inc - May 21, 2008 - 22:56
| Project: | Google Client Geocoder |
| Version: | 5.x-0.1 |
| Component: | Code |
| Category: | bug report |
| Priority: | normal |
| Assigned: | Unassigned |
| Status: | needs review |
Description
The trouble is that if you change a node address, the module will (correctly) apply (or create) the correct taxonomy for that Country, Locality.. but the node also RETAINS the original taxonomy vocab/terms.
So if you had a node in Sydney, Australia
And then changed the address to Moscow, Russia
Your node terms would be something like:
NSW AUS, SYDNEY, AUSTRALIA, MOSCOW, RUSSIA
#1
I can confirm this on the 5.x-1.x-dev version as well
Ricardo
#2
If you tick the "delete address" tickbox on a already submitted node, the duplicate key entry in the taxonomy does not happen. But I also confirm this bug.
#3
Hi,
I've fixed this. The problem has to do with how node terms are cached when manipulating a node programatically. Luckily it's an easy fix.
Starting line 325 in gcg_taxonomy.module change the following:
if (!empty($terms)) {taxonomy_node_delete($node->nid);
$node->taxonomy[$vid] = $terms;
taxonomy_node_save($node->nid, $node->taxonomy);
}
to
if (!empty($terms)) {unset($node->taxonomy);
$node->taxonomy[$vid] = $terms;
taxonomy_node_save($node->nid, $node->taxonomy);
}
Attached is a patch for those that like to do it that way...
Michael
#4
Hi Michael,
Did you manage to get this module running on your site?
#5
Yes but not in production so I've yet to see whether it can stand up to the test of users.
#6
Did you have to make any more changes to the site besides this patch?
I am trying to figure out a way to set a default location and also to restrict the users to tag nodes only from a particular region...do you have any idea if that can be done?
#7
Just this patch.
I don't think those functionalities are built in. You will likely have to modify the module code to achieve that.
#8
I'm still confused as to the status of this project. Not sure it's being maintained. Either way, I've changed the status of this issue to a patch for review so that is might get incorporated into the code base.
Thanks
Mike