Closed (fixed)
Project:
Link checker
Version:
5.x-2.1
Component:
Code
Priority:
Normal
Category:
Bug report
Assigned:
Unassigned
Reporter:
Created:
31 Mar 2009 at 15:43 UTC
Updated:
19 Apr 2009 at 13:50 UTC
Jump to comment: Most recent file
Comments
Comment #1
rsvelko commentedsame with the 6.x. branch
Comment #2
hass commentedPlease enable "Check full qualified domain names only" and provide a patch. The problem is that mailto and others pass http://api.drupal.org/api/function/valid_url/6 as valid URL... :-(
Comment #3
jromine commentedYes, valid_url() only checks ftp and http[s] schemes. Patch attached.
Comment #4
jromine commentedSorry, I reversed the logic in the fix. Here's a corrected fix which skips mailto:, javascript:, etc. links.
Comment #5
rsvelko commentedI patched and after several cron runs I cannot make mailto: links disappear from the reports page and the links table.
What would you suggest? How does this module work as a workflow?
Scan nodes - find links -> log each to the linkchecker_links and linkchecker_nodes table
But what is missing in the above workflow (and I am sure linkchecker does it somehow) is how are manually fixed links removed from the table? Or they re just made status 200 which is ignored in the report?
So, will the table get smaller bec of that patch?
I still have those "http://example.commailto:office@example.com" links in the url field of the links table...
Comment #6
hass commentedWith your patch applied you need to re-save the node where this broken links are in... that's all. The cleanup task remove the entry later from the table. Otherwise you can clear all links and rescan all nodes... there is a button on the link checker config page. Not sure how many nodes have emails inside...
Comment #7
hass commentedFixed in CVS. I hope this works well as it haven't been tested very much by me...
Comment #8
hass commentedOnly to note for a .test we may need to write, URLs like http://www.example.com:8080/foo are extracted correctly.
Comment #9
hass commented