Fix Github-101 documentation page following GitHub redesign #58
Comments
I can take this one.... I need it for Shanghai anyway :/ |
It might be good, if you have time, to add just a few words about what I'm guessing that you need to check the web interface for the words: [[ You're all set—the r12a:r12a/html5-character-encoding branch can be I'm assuming that that means It would be nice (if i'm right) to have a paragraph that confirmed those Cheers, On 19/07/2013 21:56, Larry McLister wrote:
Richard Ishida, W3C |
re: deleting the branch after a PR is accepted, I think it probably depends upon the workflow. I'd like to get this into the git doc but want to make sure we agree, so take a look at the cases below and lmk what you think. Also, what cases are missing? Say I'm writing flexbox tests,... I create the branch lmclister/flexboxtests. I complete some tests and submit a PR. Case 1 - Nobody has looked at my PR but I want to create more flexbox tests. This is just the normal update PR workflow. I reuse the branch / update the PR. Case 2 - Somebody has provided me feedback on my PR. Now I really should go address that feedback on the PR but I really just want to write a few more tests. In this case, make a new branch, agreed? if so, any preferences on branch naming? Case 3 - Somebody has accepted my PR. I reuse the branch for my next set of flexbox tests. Case 4 - I want to create some mulitcol tests (or some other feature) and there are no outstanding PRs on lmclister/flexboxtests. Nevertheless, I shouldn't reuse lmclister/flexboxtests, instead I should create a lmclister/multicoltests branch. If the branch name was more general it wouldn't matter, so I suppose it is a matter of if we care about branch names. |
We take the risk of bumping into hard merge problems when we tell people to reuse branches. 90% of my time helping folks with Git is to fix pull requests which branching issues. We make it a lot easier on ourselves if we have one unique consistent story to tell: start from master every time, make small topic branches, name them precisely, delete them once they've been pulled.
You want to choose a more precise name for reasons you outline later yourself.
In general, I would advise against this. The smaller the PR the faster it'll get reviewed and the bigger the sense of accomplishment for everyone involved.
Absolutely.
Pick a name that describes precisely what it is you're working on. Note that you might want to do something like:
No, you really want to start a new branch from master. Branches are cheap in Git. More precise naming would allow you to avoid this temptation.
We care about branch names to avoid name collisions and to scope the tests to something palatable. That's pretty much all there is to it. In general, the Git flow is about small commits, and small topic branches. This makes it a much easier to get early feedback, maintain momentum, etc. |
Very much yes about small branches/patches. I don't want to review a PR to fix all outstanding issues with all XHR tests (like web-platform-tests/wpt#103), though I'd probably have been happy to review any of those commits on their own. (The workflow I use is:
.) |
Great, thank you for the feedback. |
Sounds like #65 closes this one. |
/cc @r12a
The text was updated successfully, but these errors were encountered: