Closed Bug 1066176 Opened 11 years ago Closed 11 years ago

migrate loop webclient standalone Makefile infrastructure to Grunt

Categories

(Hello (Loop) :: Client, defect)

x86
macOS
defect
Not set
normal

Tracking

(Not tracked)

RESOLVED INVALID
backlog backlog+

People

(Reporter: dmosedale, Unassigned)

References

Details

(Whiteboard: [tech-debt])

In bug 1044411, grunt and bower were introduced to build localization stuff in the . We get significant network effect benefits (being able to easily use and update packages and tasks in npm and bower.io). For now, we still have the Makefile infrastructure around, which leaves us with more complexity and maintainability than we'd like. We should port the stuff currently done by the Makefile to Grunt and eventually kill the Makefile.
Depends on: 1044411
Shouldn't we first move the standalone client to live outside of m-c to open a whole range of possibilities? Else we should wait on having NodeJS available on the build infra.
I should have mentioned here what I added as a comment to the Makefile itself during review: this Makefile doesn't get used by the Mozilla invoked by the build infrastructure at all. It's only used for development and deployment at the moment. You're quite right that the full benefits don't accrue until later, but I think that's somewhat orthogonal to the maintainability improvements that this cleanup offers.
Whiteboard: [tech-debt]
backlog: --- → Fx36+
backlog: Fx36+ → Fx37+
backlog: Fx37+ → backlog
Enough experience with Grunt suggests that debugging it is so painful that it offsets the benefits. We discussed this at retrospective a while back, and there's no plan to invest further in Grunt. We should probably remove it once we stop depending it on for the ToS stuff.
Status: NEW → RESOLVED
Closed: 11 years ago
Resolution: --- → INVALID
You need to log in before you can comment on or make changes to this bug.