potiuk edited a comment on pull request #9961:
URL: https://github.com/apache/airflow/pull/9961#issuecomment-663444449


   No worries QP: It's not an easy one to deal with the mutlttude of ways PR 
can be run :). And GitHub does not make it easy either. And the bad thing is 
that it's not easy to test how it will behave after you merge it to master I 
thin @zikun nicely shown all the different cases you might have when you try to 
interact/pull/push another branch:
   
   ```
   PR: user:nonmaster -> apache:master
   PR: user:master -> apache:master
   PR: user:nonmaster -> user:master
   PR: apache:nonmaster -> apache:master
   CI build: apache:master
   ```
   
   I'd also say that we have another one: scheduled nightly build (but this is 
a special cases for the last one "CI build". 
   
   And this is not even touching the case where we have v1-10-test and 
v1-10-stable branches involved - then it gets really messy as we can also have 
PRs towards v1-10-stable and direct pushes to v1-10-test. And sometimes we have 
PRs for v1-10-test :face_with_head_bandage:  
   
   Luckily witth to openapi we are only concerned about master :).
   
   One thing that I found useful though is that I have my own fork, and we have 
"PolideaInternal" fork for our organization so I try to reproduce the cases 
sometimes by pushing my branch as master to my own fork: `git push -f potiuk 
MYBRANCH:master` to simulate merge/scheduled push. That helps a bit (and I 
discussed yesterday similar strategy with Apache Beam team - I am consulting CI 
builds for them and they have very similar kinds of problems) 


----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
[email protected]


Reply via email to