Ladsgroup added a comment.
In T255882#6240846 <https://phabricator.wikimedia.org/T255882#6240846>, @Lucas_Werkmeister_WMDE wrote: > If the API behavior should stay the same, then I imagine we could resolve the direct Client→Repo dependency by registering some kind of “term buffer + entity ID lookup” stash in Lib, and Repo fills it with its services upon initialization, and Client fills it with its services only if it wasn’t already initialized by Repo. So the two extensions would coordinate via Lib instead of talk directly to one another, so to speak. Something similar could probably also be done if Product says the API should first return the terms of the current entity and otherwise the terms of the entity linked to the current page, or something like that. I would be very much in favor of having simpler logic in our codebase, this seems complicated (meaning harder to maintain and more likely to have bugs and harder onboarding). Sometimes complicated logic is inevitable, like high concurrency, large scale caching, etc. But this case can be resolved by having other API modules (all boils down to what product decide of course) but I would say complicated logic is a negative point for having everything in one place. TASK DETAIL https://phabricator.wikimedia.org/T255882 EMAIL PREFERENCES https://phabricator.wikimedia.org/settings/panel/emailpreferences/ To: Ladsgroup Cc: Ladsgroup, Aklapper, Lucas_Werkmeister_WMDE, darthmon_wmde, Nandana, Lahi, Gq86, GoranSMilovanovic, QZanden, LawExplorer, _jensen, rosalieper, Scott_WUaS, Wikidata-bugs, aude, Lydia_Pintscher, Mbch331
_______________________________________________ Wikidata-bugs mailing list Wikidata-bugs@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-bugs
