User Details
- User Since
- Apr 8 2016, 2:47 PM (505 w, 5 d)
- Availability
- Available
- LDAP User
- Sebastian Berlin (WMSE)
- MediaWiki User
- Sebastian Berlin (WMSE) [ Global Accounts ]
Yesterday
Mon, Dec 15
This works now for both http and https. The former redirects to the latter, which in turn redirects to http://wikilovesearth.se/. The certificate lives on the sites server.
Now it's disabled for all users.
Removed the rows for map.wikilovesearth.se. It's not being maintained, see T381763.
Fri, Dec 12
An important aspect of this is the actual content of the popup. What information should go there and what should be saved fot the help page? It may be worth looking at other well established extensions, like VisualEditor, and other functions.
Thu, Dec 11
It's tricky to know when you actually click on a piece of text. Using getSelection() gives a selection even if you click outside a node.
Wed, Dec 10
Tue, Dec 2
I realised that we don't need to load the utterance. The segments have the information about nodes which should be enough to see if where you click is something that can be played.
Mon, Dec 1
I tried using getRedirectTarget(), but that doesn't seem to work.
Yes, it would require making sure that all utterances are synthesised, which could take a while depending on if the page has been visited before and how long it is. It would also mean that you download the data for all the utterances and the audio could be quite a bit if it's a long page.
Yes, it works now.
Fri, Nov 28
The introduction of asynchronous logic (to make sure token information is loaded) caused some headache when I was writing tests. I ended up with something that I think is the third way of doing among our tests. May be worth having a look at using the same method for all of them.
Thu, Nov 27
@Ainali, any thoughts on the question above?
Wed, Nov 26
I did a test with 1,000 random files that were used in the 100,000 first declarations. For 73 of them no license could be determined. I went through and checked them manually and for most there was no license in the SDC.
There are some licenses on WD that don't have the trailing slash at the end of the URL, e.g. Creative Commons Attribution-ShareAlike 3.0 Unported. To cover those the check for allowed licenses accepts both with and without.
Tue, Nov 25
I think that would work. There may be something in the documentation about doing something like this. It might be worth having a look if you haven't already.
It looks like you locked down a bit too much. Now I can't edit or create pages as an admin 🙂 I think you may have to add the permissions you removed to the admin group.
What does the code look like where you use the config variable? What does getConfig()->get() return?
Mon, Nov 24
! In T182289#11394901, @Viktoria_Hillerud_WMSE wrote:
[...] Either way, I think the "button" that appears should clearly be signaling that: "Do you want to start playing from here?"
I also believe suspending means that you can't log into the account or use it in any other way.
I disagree.
Fri, Nov 21
@Jopparn, any thoughts on this?
There are default values for group permissions: mw:Manual:User rights. You can also look at Special:ListGroupRights to see settings for that wiki.
That said, I’m starting to think that a better user experience could be achieved by having a player button appear when the user clicks anywhere in the text.
Thu, Nov 20
I've done some hacking just to see how this could work and have gotten some of the way there:
Yes, for editing. We still want to keep reading and listening permissions for everyone.
I don't think there's anything that needs to be added to any documentation for this task. It's really just the start. T402526 is where there's a noticeable change for the end user so we probably can wait until that's done with documentation.
It may make more sense to add a column to Wikispeech-Text-to-Speech for the segmenter. This will be one of those ongoing things that we'll realistically not finish (i.e. create a perfect segmenter).
Wed, Nov 19
Yes, that makes sense.
No, but we probably should start making note of them. I think there were one or two during the user test.
After looking around for a bit, I couldn't find any segmenter that we could easily add to the extension. The ones in the description could be added as services, but then we need to take into consideration how much that would complicate development, maintenance and setup.
I asked the providers of our Wordpress installation on wikimedia.se about a warning about outdated PHP. They have extended support for PHP so that's not something we need to worry about. I made a note in the checklist about this.
Tue, Nov 18
I've gotten it to work locally. It required creating the page first and adding the table to it. Before that I received non-specific error: "Pronunciation error report could not be saved.". I think it would be easiest to just add some instructions for this for now. I may be wort to automate to some extent in the future.
Nov 17 2025
Thanks for letting me know. That task was more general, but pointed me in the right direction.
Adding the class skin-invert works here too:
Done in be78fa1.
Nov 14 2025
I ended up not writing any tests for this. I tried, but there was too much external things that were hard to mock. There isn't that much logic anyway.
Yes. When you specify the error page it could be either a local one or one on another wiki. In case of the former you'd use the logic you have in the patch. In the latter you'd use the other wikis API.
It looks like this only works on the wiki where Wikispeech is installed. That's good as the default, but the way we have it set up now we'd want to use a page on an external wiki. Is that something you plan for this task or will that be a follow up?
Nov 13 2025
This could be useful for changing settings: https://www.mediawiki.org/wiki/API:Options.
I'm not sure we need to, but I was wondering if you've investigated it. For instance how long are the longest messages? I'd guess most of them are short (a few words to a sentence), but there may be longer ones.
When you use the API to retrieve utterances the response can include multiple segments. For content you only ever have one. Have you considered how this may impact network usage? If you have a multi sentence message it'll fetch the whole thing before it can start to play.
Nov 12 2025
I was specifically wondering about items on WD. I forgot that SDC uses it too.

