There's long been discussion about us setting up a dedicated service (a la commons) that could serve as a centralized repository for code used on the wikis. This is broadly construed to include:
Some of the proposed benefits to doing this would be:
Scary transclusion was an attempt at this, but really is scary and not usable in its current form.
Duplicate of/related to T66475: Make crosswiki bits and pieces truly global (tracking)?
Adding #Engineering-Community because the lack of this feature has an impact in the way the communities of templates / modules / gadgets are (des)organized, duplicating plenty of work, and handling a lot of code that has a newer version somewhere.
Is #reading-Infrastructure-team considering to have this task as a quarterly goal? If not, how can we help pushing it to higher priority in our backlogs?
"Central Code Repository" sounds like git, and then why not use gerrit/diffusion for review? Then, the gadget, template, Lua module page could transclude a particular revision from the repository, e.g.
{{#git include: |ProveIt_Wikipedia.js |mediawiki/gadgets/proveit-js |wmf/group1}}
(file, project, branch). At that point T91626: Technology to transclude git content into wiki pages might be relevant.
The alternative mentioned in the Global scripts RFC is a shadow namespace using ForeignFileRepos. I don't know how hard it would be to support http access to a git repo as a ForeignAPIRepo.
Techniclly that is correct but note the social aspect here - people who know how to edit wiki pages might not all be keen on learning a system like git... Potential participation/contribution barriers, so to say.
Qgil moved this task to Proposed: July - September 2016 on the Developer-Relations workboard.
This would mean that a team includes this goal in their annual plan FY2016-17. I still wonder to which team would this goal belong, though.