To continue with a mono-repo handling various node packages, I think we're going to need to agree on some conventions and norms in order to avoid chaos.
The situation that precipitated this was @ahmedabdelkawy's #79 PR, which, given the very precise name, I was shocked to see was over 600 lines in 16 files! The cause of this was simple -- #79 is actually just built on top of @JNmpi's dev_joerg_from_main branch. My solution here was to (a) open a draft PR #80 of @JNmpi's branch into main, and then (b) Retarget #79 on top of #80. @ahmedabdelkawy, @JNmpi, I hope you're both amenable to that.
FYI from a technical perspective:
Branches with no associated PR have very low visibility -- other developers are not likely to notice that your changes exist. (@niklassiemer heads-up also re nsiemer_plot3d, which I now see exists but haven't looked at)
There is no barrier to stacking changes on top of existing PR's -- in this case, if @ahmedabdelkawy want's to use @JNmpi's changes, a new PR can be made into that branch instead of into main, keeping the new changes well isolated -- it's A-OK and GitHub lets us select targets either when the PR is created or by editing it post-facto if we forgot to choose the right target.
Questions from a social perspective:
Who will review PRs here?
What sort of time-frame can we expect from each other for getting those reviews and merging stuff in so it's available to everyone?
How strict and picky do we want to be before merging something to main? (e.g. must everything be PEP8 compliant?) Do we emphasize speed or stability/quality?
To continue with a mono-repo handling various node packages, I think we're going to need to agree on some conventions and norms in order to avoid chaos.
The situation that precipitated this was @ahmedabdelkawy's #79 PR, which, given the very precise name, I was shocked to see was over 600 lines in 16 files! The cause of this was simple -- #79 is actually just built on top of @JNmpi's
dev_joerg_from_mainbranch. My solution here was to (a) open a draft PR #80 of @JNmpi's branch into main, and then (b) Retarget #79 on top of #80. @ahmedabdelkawy, @JNmpi, I hope you're both amenable to that.FYI from a technical perspective:
Questions from a social perspective: