prawnpdf / prawn-table

Provides support for tables in Prawn
Other
205 stars 98 forks source link

Fix build setup and pivot to GitHub Actions #131

Closed leoarnold closed 2 years ago

leoarnold commented 2 years ago

This fixes unit tests which failes because they tried to use an image files from another gem and did not find it in the expected location. We simply add the file as a fixture to the project at hand.

Also, this project was still using Travis CI dot org which was deprecated in favor of dot com, and then again Travis CI introduced restrictions to protect their infrastructure which are just detremental to OSS development:

https://www.jeffgeerling.com/blog/2020/travis-cis-new-pricing-plan-threw-wrench-my-open-source-works

Hence we move all tests over to GitHub Actions.

leoarnold commented 2 years ago

Before merge: Make sure GitHub Actions is enabled for this repo and reconfigure branch protection for master: Drop Travis CI, but require GitHub Actions to pass.

petergoldstein commented 2 years ago

@pointlessone I would prioritize merging this.

pointlessone commented 2 years ago

@petergoldstein Can we have a copy of the prawnpdf/prawn setup here? Maybe apart from RuboCop part. I don't believe this repos even has RuboCop set up in it.

petergoldstein commented 2 years ago

@pointlessone Sure. I'll set up Rubocop as well with the same config and just add a .rubocop_todo.yml for all existing lints.

One quick question - can I change the pull_request_target to pull_request, as that's the more standard configuration?

I'll put up a PR with that in the next 24 hours.

pointlessone commented 2 years ago

@petergoldstein Yes, we probably should change that on other repos to.

I suggest splitting RuboCop into a separate PR or even skipping it altogether. I imagine it will take some effort to fix all the issues there and it's not immediately valuable.

petergoldstein commented 2 years ago

@pointlessone Ok. I'll split Rubocop out separately.

Any objection to me removing the custom bundle cache logic? There's a built in bundle-cache in the ruby/setup-ruby action that does this all automatically. Any objection to me using that? It works fine with multiple rubies. That will simplify the GitHub Actions file significantly.

pointlessone commented 2 years ago

Please keep the custom bundle cache logic. It has been discussed extensively in prawnpdf/prawn#1190 and ruby/setup-ruby#136. It's been a while since those discussions but I don't think much has changed since.