On Tue, 16 Dec 2014, David Lang wrote:

I ask
because almost all pull requests are done against master branch, which
means I need to manually merge them to master-candidate and close the PR as
"unmerged".

It would probably much more efficient to have "master" be the experimental
branch, and when the testbench succeeds move it to something like
"master-ok" (or so).

the thing is that the testing branch gets recreated for each run. It doesn't have a long term history (or shouldn't), so you don't want people trying to develop against this.

after reading the rest of the thread, I'm not being clear here. And as a result, what I was trying to describe and what Rainer is probaly doing aren't likely to match.

(much of this is copied from the approch that git uses for development)

what I think that Rainer understood (and what I may have described) was

1. develop on branches

2. merge dev branches into test branch, test it

3. if all tests pass, merge test branch into master

4. otherwise, hack on dev or test branchs to fix problem, merge results into test, test it, goto 3


What I should have desribed is slightly different

1. develop on branches

2. create throwaway branch that merges all dev branches, test it

3. if all tests passes, merge those dev branches into master (since this is the same merge that was done in #2 it should be trivial)

4. otherwise, goto #2 but skip merging the branch that caused problems.


the reason to test against a throw-away branch instead of a long-lived branch is so that if you have a problem with a branch that's being merged in, you can just drop that branch rather than needing to revert it. This makes for a cleaner history as you won't have merge/revert commits in it.

David Lang


_______________________________________________
rsyslog mailing list
http://lists.adiscon.net/mailman/listinfo/rsyslog
http://www.rsyslog.com/professional-services/
What's up with rsyslog? Follow https://twitter.com/rgerhards
NOTE WELL: This is a PUBLIC mailing list, posts are ARCHIVED by a myriad of 
sites beyond our control. PLEASE UNSUBSCRIBE and DO NOT POST if you DON'T LIKE 
THAT.

Reply via email to