A commit that was "co-authored-by" 6+ people and has three thousand lines of code: this is a total wreck of a development workflow. This feature should have been implemented with a series of about 20 patches. Awful.
You are being downvoted, but you are entirely correct. This is also explicitly not allowed in FFmpeg, but this was pushed after many months, with no heads up on the list, no final review sign off, and with some developers expressing (and continuing to express) reservations about its quality on the list and IRC.
That's really unfortunate to hear. I'm a huge fan of Webrtc and Pion, and was very excited to get some ffmpeg integration -- hopefully some of the quality issues will be ironed out before the next ffmpeg release
I mean, it probably was a branch that several people contributed commits to that was squashed prior to merge into mainline. Folks sometimes have thoughts about whether there's value in squashing or not, but it's a pretty common and sensible workflow.
Perhaps "common and technically works" would be a better way to put that (similarly for rebase). I suspect people would stop squashing if git gained the ability to tag groups of commits with topics in either a nested or overlapping manner.
Daemon404|9 months ago
You are being downvoted, but you are entirely correct. This is also explicitly not allowed in FFmpeg, but this was pushed after many months, with no heads up on the list, no final review sign off, and with some developers expressing (and continuing to express) reservations about its quality on the list and IRC.
bigfishrunning|9 months ago
jpk|9 months ago
fc417fc802|9 months ago
Perhaps "common and technically works" would be a better way to put that (similarly for rebase). I suspect people would stop squashing if git gained the ability to tag groups of commits with topics in either a nested or overlapping manner.