Page MenuHomePhabricator

LocalisationUpdate V2
Closed, DeclinedPublic

Description

There is the LocalisationUpdate extension [1] but only few people use it. It is slow and needs a special configuration with cron. If we could integrate it into core, make it fast enough so that cron would not be needed it would allow a lot of third parties to enjoy the blazingly fast localisation updates (under 36 hours) that Wikimedia projects currently have. To make it fast enough, it is likely that a separate service needs to be implemented. It could be standalone or part of some MediaWiki instance. It should be secure and allow querying only needed data. Ideally, wikis would no longer need git to update their l10n.

http://www.mediawiki.org/wiki/Extension:LocalisationUpdate

Proposed by the Language team at https://www.mediawiki.org/w/index.php?title=Outreach_programs/Possible_projects&oldid=902079#Generic.2C_efficient_Localisation_Update_service , see also https://www.mediawiki.org/wiki/LUv2

Since MediaWiki core is affected, we need the buy-in from the maintainers before proposing it as a GSOC / OPW project.

See also:

Details

Reference
bz46653

Event Timeline

bzimport raised the priority of this task from to Low.Nov 22 2014, 1:25 AM
bzimport set Reference to bz46653.

vladjohn2013 wrote:

Hi, this project is still listed at https://www.mediawiki.org/wiki/Mentorship_programs/Possible_projects#MediaWiki_LocalisationUpdate_for_all

Should this project be still listed in that page? If not, please remove it. If it still makes sense, then it could be moved to the "Featured projects" section if it has community support and mentors.

According to GSoC 2014 results:

A project was created: https://www.mediawiki.org/wiki/Extension:LocalisationUpdate/LUv2

Code was published at https://github.com/konarak/LUv2/tree/epicmess

The mentors considered the project PASSED, but it is unclear from the updates in which status it is now:

https://www.mediawiki.org/wiki/Extension:LocalisationUpdate/LUv2/Updates

We are also missing the required blog post summarizing all this.

Please wrap up properly your project.

konarak.11 wrote:

Hi Quim, sorry for the delays.

The work on the LUv2 service is done. I am yet to start working on writing code for the extension (which can be integrated with core) due to some meatspace work. I'll make sure to publish the final report by the end of this week (SF time).

Nikerabbit, please revert T48653#1146478 if it doesn't fit your suggestion of yesterday on #wikimedia-releng.

Nikerabbit, do you prefer to work on this alone when you have time or should we look for a GSoC application?

I do not want to propose many topics because I cannot mentor them all.

Finishing this task means bringing it to "production" which means setting up secure and stable service, which can be problematic for GSoC task.

Krinkle renamed this task from Integrate LocalisationUpdate to MediaWiki core to LocalisationUpdate V2.Jul 8 2017, 2:20 AM
Krinkle subscribed.

Rephrasing summary. I don't think this functionality should be in core. However, once stable and proven, I think it would make a great candidate for bundling with core (by default) – as was originally requested and how this task came to be (T45815).

See also: T158360: RFC: Reevaluate LocalisationUpdate extension for WMF

Re the new summary, note there was already a project called LocalisationUpdate v2 or LUv2 (never finished/put into use).

In case anyone is interested, I am working on translation backport support, which would let third parties to get new translations with point releases or on a regular schedule if they are pulling from release branches in git.

Jdforrester-WMF subscribed.

The LocalisationUpdate extensions has been archived. See T309694 for details.