Addshore added a comment.
So the main thing that we are seeing here (badly cached pages in the parser cache) seems like it could have been some temporary issue (although I have no idea what the issue was).
All of the pages I linked to in T205556#4654519 are now fine after re parses, and hitting
Billinghurst added a comment.
@Addshore As I mentioned above, whilst I a having some issues with WD directly, I am having greater issues with WEF framework tool's lookups. As I create items and do the lookup for matching terms it is finding items successfully, though the rendering of the name into
Addshore added a comment.
Here are a collection of example items that have the issue and the associated parse times.
https://www.wikidata.org/wiki/Q16985906 20180924223337
https://www.wikidata.org/wiki/Q53450252 20180925194242
https://www.wikidata.org/wiki/Q4807677 20180927074158
Addshore added a comment.
In T205556#4651725, @Magnus wrote:
To reproduce, I just click "random item" a few times, one will show up very quickly.
Example right now: The "published in" statement on https://www.wikidata.org/wiki/Q40220984 . The statement target (Q27713832) has an en label.
Magnus added a comment.
To reproduce, I just click "random item" a few times, one will show up very quickly.
Example right now: The "published in" statement on https://www.wikidata.org/wiki/Q40220984 . The statement target (Q27713832) has an en label.TASK
Addshore added a comment.
Does this occur for multiple people on single items?
Could you write some steps to try to reproduce? Or example links if it is happening right now?TASK DETAILhttps://phabricator.wikimedia.org/T205556EMAIL
Billinghurst added a comment.
I regularly use the WEF Framework gadget, and over the last several days I am seeing this problem come and go.
There is a variation in the problems the range I am seeing is
no labels render at all through the gadget; (least often)
the labels in the form render,
WMDE-leszek added a comment.
Could this be related to #wikidata-ugly-cat-trailblaze_wb_terms_trail_blazing
hmm, looking at the screenshot I'd tend to say "probably not", as the new formatter code is only use to format items Q1..Q100,000 in statements, and all Q-IDs in the picture are way beyond
