[GitHub] [beam] tvalentyn commented on pull request #12196: [release-2.23.0][BEAM-10308] Make component ID assignments consistent across Pipeline…

2020-07-14 Thread GitBox


tvalentyn commented on pull request #12196:
URL: https://github.com/apache/beam/pull/12196#issuecomment-658441886


   > To be honest I think this situation is somewhat of a gap in the guidance 
there. There may be cases where its worth delaying a release for a severe 
bugfix even if it's a longstanding issue and not a regression, and that's not 
addressed in our release guide. We could draw a line somewhere: e.g. maybe data 
loss/incorrect result bugs can delay, crashes can not?
   
   In my opinion for these situations releasing a patch version (2.23.1) 
following a release (2.23.0), would be most preferable to users: new 
features/bugfixes already in RC1 will become available earlier, and cherry-pick 
worthy long-standing issues that were recently fixed and easy to back-port 
won't have to wait 6 weeks to be released. If anything goes wrong (cherry-pick 
conflict, new bug), 2.23.0 will still stay around.
   For example, for this particular change I think we would only need to update 
Python SDK artifacts. We can propose to make 2.23.1 with this fix right after 
2.23.0, without waiting for 2.24.0, if you think it's urgent. Thoughts 
@TheNeuralBit @robertwb ?



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:
us...@infra.apache.org




[GitHub] [beam] tvalentyn commented on pull request #12196: [release-2.23.0][BEAM-10308] Make component ID assignments consistent across Pipeline…

2020-07-09 Thread GitBox


tvalentyn commented on pull request #12196:
URL: https://github.com/apache/beam/pull/12196#issuecomment-656278107


   This change is currently is not in RC1, and we can consider if for RC2 if 
there has to be an RC2. The release guildeline clearly say that regressions in 
new functionality should not be blocking the release. I have applied the same 
criteria to at least 3 or 4 other CP requests, so it would not be fair to have 
to select which new features are worth delaying the release and which are not. 
Also, are we confident that the fix would not introduce new issues given that 
it was merged recently?
   



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:
us...@infra.apache.org