Front-end standards group/2016-03-09

Copied from https://etherpad.wikimedia.org/p/FrontEndStandardsGroup

Attending: Volker E., Timo T., RobLa, Bartosz D., Julien G.

Action & Code

• Front-end standards related ArchCom-RFCs: https://phabricator.wikimedia.org/project/board/52/query/X7D3dZrpZ9Ai/

Follow-up

UI changes: Consolidation of `progressive` and `constructive` flag got merged in OOjs UI and core (VE, BD) - https://phabricator.wikimedia.org/T126309

• TT: good direction to go, there are a few views, where the consolidation might have to be revisited.
• RobLa: would wait for a few weeks and then remove band-aid quickly

[RFC] Section headings should have a clickable anchor (VE) - https://phabricator.wikimedia.org/T18691

• VE Will be getting back to this in next couple of weeks. Current take: Collect the issues raised to all proposed solutions and question the task for our given users.
  TT: maybe as opt-in gadget instead of core given the raised questions by Volker about concerned user group last meeting.

[RFC] Standardize on how to access/register JavaScript interfaces (TT, RK) - https://phabricator.wikimedia.org/T108655

• Ongoing and on a good way

New quests

https://github.com/lesshint/lesshint Less style checker (VE)

• Syntax linter (for MW modules) already done as part of https://phabricator.wikimedia.org/T106780
  • VE: Current linter is checking for errors, but not there for implying guidelines. There is a current trend to move away from Less.
  • TT: Given the performance implications by easily running into overnesting selectors, jQuery team has been deciding to stick with plain CSS.
  • JG: In new Wikipedia Portal we are using PostCSS right now, but we also don't have big necessities to fulfill there. I share the concern of people "running away" from Less.
  • TT: We've been discussion how Less is performing in production.
    If we had to face the decision about including Less today, we would not deploy it in RL
    Lots of nesting as performance implicator.
    jQuery are sticking to CSS and PostCSS
  • JG: when sticking to plain CSS, you're more future-proof and avoid the selector nesting soup
  • VE: guideline syntax checker for avoiding nesting (with PostCSS again?)
    PostCSS RfC?
    Biggest concern when not using Less, is currently Less variables and mixins. They are widely used in f.e. OOjs UI and support needs to be provided for UI consistency.
  • TT: CSS: runs through min, janus, could first run through PostCSS

Topics for next meeting

Review, as promised in last meeting: 4 topics finished, 2 topics on the way