Replies: 7 comments
-
I agree. I believe a potential setup would look like:
I don't necessarily think we need to do PRs during the competition due to the time limitations and availability of people to review but between competitions we absolutely should keep up our PR practices. |
Beta Was this translation helpful? Give feedback.
-
I absolutely agree with this idea, but one thing I will say is that unless we explicitly give everyone a quick 5 minute training/practice on how this development model works (maybe the meeting before the comp, or in the pits before stuff actually gets going), next comp will probably descend back into madness. Just something to make sure we do |
Beta Was this translation helpful? Give feedback.
-
Agreed. Before this closes out, I would like someone to write-up a markdown document in our repo here to clearly display how our competition processes will go and submit a PR. This can be a part of #40 |
Beta Was this translation helpful? Give feedback.
-
Probably not PRs during comp, I meant PRs after the comp for each of the feature branches. If Github forces us to use PRs, we can just set the required number of reviewers for that rule to be 0. Also, could we not name feature branches event//? Github rules are based on the name of a branch, so having just event in the name would be difficult to work around. Could we instead do event-feature//? |
Beta Was this translation helpful? Give feedback.
-
I don't think we'll be forced to use PRs. The branch will just be locked to the dev laptop so the code that should be merged will be pulled in from that laptop directly to the event branch. We can definitely do a different naming scheme. Outside of competitions, we should likely be labeling branches |
Beta Was this translation helpful? Give feedback.
-
If anyone is curious, I've kept this issue open because even during last competition, our version control was not great(though definitely better). Part of it was due to the fact that we had so many things to test but not a lot of time. Accordingly, we need to come up with a better procedure for actually testing code and get through our backlog of testing during comps. |
Beta Was this translation helpful? Give feedback.
-
Definitely still needs to stay open until we at least have a document showing how we handle competition changes, which will include how to handle the git side, unit testing, etc. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
Especially with the last competition, we have a large amount of different features all jumbled together in one branch, and the issue is that we made so much changes during comp to each of those features that we would be losing a huge amount of progress if we just decided to go back to the original branches those features came from.
In an ideal scenario, each of the features we would want for a comp would already be in a complete state and merged to main, and we would just add the smaller changes during a comp to main afterwards. However, when features need to be devved during comp, things become messy, as we often just merge to the comp branch and work on that branch rather than the feature branch. This is because it allows us to work with new features without having to worry about merging even more than we are currently.
I propose that we enforce some sort of branch rule on event branches. Perhaps we limit who can push to these branches to only an account that is only on the dev drive laptops? Everyone else has to code on feature branches, and when they finish with their changes, they can merge to the comp branch. This would actually reduce the amount of merge conflicts we deal with, as we only merge when every change to a feature is done, allowing us to still push our changes safely. This would also allow us to make prs for the feature branches first before the event branch, meaning that code review would be a lot more meaningful and actually allow the reviewers to see if any parts of the code need changes.
Beta Was this translation helpful? Give feedback.
All reactions