Only inspects node->body which is empty for cck
moshe weitzman - September 29, 2006 - 04:00
| Project: | Related links |
| Version: | HEAD |
| Component: | Code |
| Category: | bug report |
| Priority: | normal |
| Assigned: | Unassigned |
| Status: | active |
Jump to:
Description
parsed link inspection only happens for $node->body, which is empty for some modules like CCK. i think we want to inspect the whole rendered text which is output by node_view(). perhaps move this feature to cron and emulate node indexing which works on the node_view() as well.
in 5.0, you could use nodeapi('alter') which will likely be a post_rendered string once eaton's patch gets in: http://drupal.org/node/82045
#1
i just saw the todo in the code to move parsed link examining to filter_process(). that might work. in the cck case, you would only see markup fields and not the whole node render but thats probably OK ... or use cron as i suggested.
#2
oops. i forgot that filters have no context for nid or anything else. they should. was discussed a long tikme ago but no commit. see http://www.acko.net/blog/drupal-filter-formats and http://drupal.org/node/8000
#3
It might be hard to fix this. Consider doing what search.module does and doing the link discovery during hook_cron(). Then you can build up a fully rendered node with node_view() and doing parsing on that.
#4
hmmm. is properly discovering links for cck 4.7 so this isn't such a serious issue. not sure about cck 5.0
it might still make sense to use the search.module technique so we can discover links for php nodes (among other advantages)
#5
I can't get the module (HEAD version) to discover links for node types created with the 4.7 version of CCK. Links are discovered for other content types just fine. I'd really like to be able to use the module with CCK node types; is there any sort of patch you applied to CCK to get it working?
#6
Sorry, I meant the module does not find parsed links for CCK node types (nothing to do with discovered links).