config.include works unpredictably when supplying more than one metadata to filter

[timdiggins]

config.include works unpredictably when supplying more than one metadata to filter

[timdiggins]

I've tested 4-0-dev against the inline code and it passes. 🎉

FYI using

  gem "rspec-support", github: "rspec/rspec-support", branch: "4-0-dev"
  gem "rspec-core", github: "rspec/rspec-core", branch: "4-0-dev"
  gem "rspec-expectations", github: "rspec/rspec-expectations", branch: "4-0-dev"
  gem "rspec-mocks", github: "rspec/rspec-mocks", branch: "4-0-dev"

So that's wonderful!

Also tested rspec 4 on a rails project with a decent number of tests and all ran fine (after adapting to rspec 4 changes).

@pirj Had no idea about RSpec 4.0 -- so much great work, and yet seemingly still so far away (looking at rspec/rspec-metagem#61). Yet looks like the main thing to complete on the 4.0 list is testing (there's also a post-4.0 list and a cleanup list - but are those optional?)

[pirj]

Awesome! Thanks for checking this.
I’ll close this as completed.

Almost there actually. We wanted to glue all repos together into one repo. It would allow to shuffle things between repos, simplify our builds, and simplify cross-gem fixes. but it’s tricky and time-consuming with all those tags and branches.

[JonRowe]

4.0 is mostly waiting on me to finish the monorepo setup and prepare the release from existing work, plus any preparatory work for 3.99 I hope to get back to this soon.

I'm going to close this as 4.0 seems to fix your issue and it would be a breaking change to fix it any earlier

[pirj]

This would be a breaking change, and we plan to release all this in RSpec 4.
Please see #2878 for trigger inclusion; #1821 / #2874 for unification if multi-condition filtering.

I kindly appreciate it if you could check if the code on 4.0-dev branch meets your expectations. The branch is behind our main branch and may miss some latest fixes. Please let me know if it doesn’t work properly, it will be a good motivation to forward-port fixes to it.