aurora-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "John Sirois" <john.sir...@gmail.com>
Subject Re: Review Request 41659: Support in-flight review chains via `depends-on`.
Date Sun, 27 Dec 2015 20:53:57 GMT


> On Dec. 23, 2015, 9:48 a.m., Joshua Cohen wrote:
> > build-support/jenkins/review_feedback.py, lines 142-143
> > <https://reviews.apache.org/r/41659/diff/2/?file=1174743#file1174743line142>
> >
> >     Not sure there's anything we can do about it, but if the review has been committed
but the RB was not marked as submitted we'll likely fail to apply the patch causing ReviewBot
to ask if it needs to be rebased.
> >     
> >     Is there potentially a better way to message that edge case? If not, it's just
something we need to be aware of (stay on top of closing RB's, and keep an eye out for erroneous
"needs rebase" messages from ReviewBot).
> 
> John Sirois wrote:
>     If the concept of a patch chain were made explicit and passed around then the messaging
could fork based on "Are we in a patch chain or is this a simple patch with no dependents?".
>     
>     For the patch chain cases we could always mention the full patch chain in the error
messages, ie
>     ```
>     Error applying patch chain for RB#456 [master (sha) <- RB#123 <- RB#456].
>     Failed to patch RB#123 on master (sha):
>     ...[maybe include git apply stderr]...
>     
>     RB#123 may already have been submitted but the review is not marked as such.  If
so - please mark RB#123 submitted.
>     Otherwise you may need to rebase the patch chain.
>     ```
>     
>     Versus the simple cases:
>     ```
>     Error patching RB#789 on master (sha):
>     ...[maybe include git apply stderr]...
>     
>     Do you need to rebase?
>     ```
>     
>     Your call whether that should happen now or later.  I'd prefer later and will file
an issue against myself if you agree.
> 
> Joshua Cohen wrote:
>     Perfectly happy to address in a follow up as time allows. I'll commit this, feel
free to file an issue to track the follow up.

Thanks Josh - filed https://issues.apache.org/jira/browse/AURORA-1567 to track this improvement.


- John


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/41659/#review111790
-----------------------------------------------------------


On Dec. 22, 2015, 2:34 p.m., John Sirois wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/41659/
> -----------------------------------------------------------
> 
> (Updated Dec. 22, 2015, 2:34 p.m.)
> 
> 
> Review request for Aurora, Joshua Cohen and Bill Farner.
> 
> 
> Repository: aurora
> 
> 
> Description
> -------
> 
> This adds support for following `depends-on` chains of in-flight RBs to
> form patch sets ultimately based off master.
> 
> Request processing logic is factored up into a helper class that main
> drives in a loop over pending RBs.
> 
>  build-support/jenkins/review_feedback.py | 246 +++++++++++++++++++++++++++++---------------------
>  1 file changed, 143 insertions(+), 103 deletions(-)
> 
> 
> Diffs
> -----
> 
>   build-support/jenkins/review_feedback.py ee37742c78a7b28bc1ccc687afae17f711471fc4 
> 
> Diff: https://reviews.apache.org/r/41659/diff/
> 
> 
> Testing
> -------
> 
> Extensive testing against a local server.
> 
> 
> Thanks,
> 
> John Sirois
> 
>


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message