You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Some of them (black, isort) can automatically fix issues they find, but in our current workflow we have to ask the user to run them and update the PR (although occasionally if the contributor is having trouble or it's really obvious to me where the flaw is, I'll either set up a "suggestion" or update the PR directly to fix issues).
On one hand, this little stepping stone usually helps new contributors get a development environment set up and helps them learn about the tools we use. On the other hand, it can mean a full day before code is merged because I have to leave a comment requesting changes. If the tools can auto-fix why don't we have a bot that does it so both the contributor and reviewer can spend more time on the parts of coding that actually need a human?
So... I'm opening this issue for people to discuss two things:
Is this a thing we want?
I'm leaning towards yes, but I'd appreciate some feedback from folk who've done their first commits more recently)
If we do want to auto-sort and blacken code, how should it be implemented? (e.g. what tools, should it suggest and comment, update the PR directly, etc?)
The text was updated successfully, but these errors were encountered:
terriko
added
the
CI
Related to our continuous integration service (GitHub Actions)
label
Apr 19, 2023
If by "work on it" you mean "investigate some options and post a couple of viable solutions here so people can discuss them" then sure.
I've got hackathon folk working on this project through April 28 and I don't want to change the CI system in the middle of the event, so I'm not really interested in actual PRs related to this until after that ends. But the good news is that leaves us at least a week to leave this issue open for discussion!
We currently run a whole series of linters on every PR:
https://github.com/intel/cve-bin-tool/blob/main/CONTRIBUTING.md#running-linters
Some of them (black, isort) can automatically fix issues they find, but in our current workflow we have to ask the user to run them and update the PR (although occasionally if the contributor is having trouble or it's really obvious to me where the flaw is, I'll either set up a "suggestion" or update the PR directly to fix issues).
On one hand, this little stepping stone usually helps new contributors get a development environment set up and helps them learn about the tools we use. On the other hand, it can mean a full day before code is merged because I have to leave a comment requesting changes. If the tools can auto-fix why don't we have a bot that does it so both the contributor and reviewer can spend more time on the parts of coding that actually need a human?
So... I'm opening this issue for people to discuss two things:
The text was updated successfully, but these errors were encountered: