The debug versions of Link Harvest have yielded a mess of inconsistent data. It appears that the link harvest action can fail for any number of reasons.
Some of them:
- PHP running out of memory.
- Hitting the script execution time limit.
- Downloading a media file (.mp3, .mov, etc.).
- Overzealous firewalls and plugins like Bad Behavior.
- A myriad of unknown reasons.
I’m going to address the media file issue and explicitly set a long script timeout in a 1.1 release and see if that makes much of a difference. I’m not sure when that will be available, but I wanted to post the results of the debug efforts and thank those that sent in their debug logs.
This post is part of the thread: Content Presentation – an ongoing story on this site. View the thread timeline for more context on this post.
This post is part of the project: Link Harvest. View the project timeline for more context on this post.
You could add to that plugins like Ultimate Tag Warrior. I had to de-activate Link Harvest as it doesn’t appear to work with UTW. For example I could not add any tags when I had LH turned on. Once I turned it off, I could add tags to my posts without a problem. It currently remains deactivated.
Thanks for all your efforts Alex. You write some of the best plugins available.
Glad to read you can reach some point with debugs, even if not all the problems will be solved.
Thanks for this good plugin.
I was able to get mine to work by disabling other plugins just while I ran the harvest and also told it to ignore a few links.
Any progress on this?
At some point I’ll have time to finish coding in the exceptions and release a new version. I think I checked some stuff in to the SVN repo, but I don’t recall if it’s in a good state or not.
Hi Alex, and thanks for your efforts. I’d suggest you that you find out where in the code the error is triggered. In other words, which of all the calls to harvest_error() is the culprit. In my tests, I found out that it’s in
if (progress == “”)
(that’s line 797)
The other errors are never triggered.
I also found out that the sum of posts+pages in my site is 14 (=6+8), but LH reports 9 when doing the harvest.
I have nearly a dozen different types of errors reported, some are due to broken server configs, some are due to timeouts, some are due to running out of memory, some are due to script execution limits…
I really wish there was just one issue here.
Hi again. I discovered something which might be useful. I have my site at MochaHost, but also locally (with my own Apache+PHP server). When I try LH in my local site, it doesn’t run. When I try it in the remote site, it does run. Maybe it’s a different setting in PHP’s configuration?
Exactly.