luckoverthere <[email protected]> wrote: > On Friday, 19 October 2018 at 03:53:12 UTC, Walter Bright wrote: >> On 10/15/2018 2:23 PM, Walter Bright wrote: >>> I'm giving a presentation at: >>> >>> http://nwcpp.org/ >>> >>> See you there! >> >> Had a nice crowd there last night. Apparently lots of people >> were interested in this topic! >> >> Video: >> https://www.youtube.com/watch?v=lbp6vwdnE0k&feature=youtu.be >> >> Slides: http://nwcpp.org/talks/2018/code_smells.pdf > > https://www.youtube.com/watch?v=lbp6vwdnE0k#t=42m01s > > Or it's a pain in the ass to try to get more than one pull > request merged in due to lack of overseers. It's hard enough > trying to get one pull request merged in, but trying to split one > pull request into 4 and then try to micro manage it so that they > are merged in order as the changes depend on each other (eg > spelling error for a function). That becomes a nightmare when you > don't have merging permission for a repo, especially for > something like DMD where it's hard enough to get even a simple > fix through the door.
I unfortunately ran into this with my make-ddoc-support-markdown PR. The original PR was considered too big, and I was asked to split it into multiple PRs against a feature branch, which would then be merged into master. I was fine with this, and started the process. It was a bit slow, but seemed like it would work. Oh, what a fool I was! When my second PR to the feature branch was approved, it couldn’t be merged because it caused unrelated errors in the tests. My best theory is that the feature branch had drifted too far away from master, and since dmd, druntime and phobos have dependencies on each other it was somehow causing the test servers to fail. There’s a bit more to the story, but it boils down to the fact that since I don’t have permissions to update the feature branch, I’m basically dead in the water. So the work is all done and ready, but I have no path forward to get it in. And it’s a bit disheartening to be foiled by the machines :)
