Page MenuHomePhabricator

Cannot select certain combined/modifier characters in enwiki's [[:Category:Redirects from Unicode characters]] list; 0px width of link
Open, LowestPublic

Description

https://en.wikipedia.org/wiki/Category:Redirects_from_Unicode_characters

When I view Category:Redirects from Unicode characters on the second page, starting 10 entries down on the third column, the entries become unclickable/selectable. This continues onto the third page. These appear to be the ones that run from wiki/%CC%80 to /wiki/%CD%AC which appear to be the diacritics. I'm using Windows 7 SP1 and Chrome 38.0.2125.111.

Originally reported at VP:Technical at https://en.wikipedia.org/wiki/Wikipedia:Village_pump_(technical)#Inability_to_select_on_Diacritics.28.3F.29_in_Category:Redirects_from_Unicode_characters

User on different system/browser verified.

Event Timeline

Naraht raised the priority of this task from to Needs Triage.
Naraht updated the task description. (Show Details)
Naraht changed Security from none to None.
Naraht subscribed.

I can confirm that: span.redirect-in-category is 17px wide for ˿ but then 0px wide for ` (and if that's modifier input than it's applied on the letter before on the left which... yeah, isn't here in this special case.)

Funny enough, I cannot select that symbol in the Developer Tools DOM view of Firefox 33 either where it's a line on its own between the <li></li> tags.

Aklapper renamed this task from Inability to select on Diacritics(?) in [[:Category:Redirects from Unicode characters]] on enwiki to Cannot select certain combined/modifier characters in enwiki's [[:Category:Redirects from Unicode characters]] list; 0px width of link.Dec 3 2014, 11:48 PM
Aklapper lowered the priority of this task from Low to Lowest.

Well there's really two approaches we could take here.

If the title starts with a combining character, we could either add a non breaking space so it has something to combine with. This would look like " ́". Or we could use a dotted circle - "◌́", to show how it would combine.

I'm leaning towards the non breaking space.

I think we can detect combining characters with preg_match( '/^[\p{Mn}\p{Me}]/u', $foo );. Not 100% sure if that gets all cases, but should catch most.

This is still an issue; updated link: https://en.wikipedia.org/w/index.php?title=Category:Redirects_from_Unicode_characters&pagefrom=000300#mw-pages

The first item is:

̀

and that is not clickable.

Removing good first task, anyone feel free to add more pointers where exactly in the code base to fix this.