How to maintain this repository?

[ddeboer]

I'm torn, I'm not sure we actually want more things maintained here.

This repo is a collection of partially broken things, the promise for users is fairly low unfortunately. (because it's impossible to maintain all these extensions)

I'm wondering if we should just encourage people to have repos for the extensions they are actually maintaining instead of giving a false promise to users.

Anyway, that's a sidebar…

Originally posted by @mnapoli in #515 (comment)

I'm wondering if we should just encourage people to have repos for the extensions they are actually maintaining instead of giving a false promise to users.

Makes sense. Should the Bref docs then be updated to describe how to build your own extension layer?

Originally posted by @ddeboer in #515 (comment)

This repo is a collection of partially broken things, the promise for users is fairly low unfortunately. (because it's impossible to maintain all these extensions)

True, but most extensions have some basic test to validate if it’s been installed correctly.

Overall, this repository seems to have two problems:

1. its maintenance burden, including merging PRs for each new PHP/extension version and publishing releases
2. its unreliability, as stated by @mnapoli.

Would it make sense to phase out this repository? (Even though I’m using some layers from it.) If so, it would be nice if the Bref docs could describe to users how to build (and maintain) their own layers.

Or perhaps there’s some other solution?

[mnapoli]

Ideally, my preference would be this scenario:

• the community proposes layers on their own GitHub repositories/accounts
  • ideally it shouldn't be ONE company/person, to avoid making the same mistake: I'd love for companies to propose only the layers that maintain (because they use it)
• we update the Bref documentation to reference community-maintained extensions
• this repository becomes a fallback for some extensions until the major ones are maintained outside
• eventually this repository is phased out

This is long-term obviously, which is fine by me.

Hopefully by writing this down this will encourage people to start their own layers.