Re: [glassfish-dev] PRs and reviews

[arjan tijms <arjan.tijms@xxxxxxxxx>]

Hi,

Indeed, last few weeks it's a larger amount of commits with typically a tagged reviewer from the same company. However if you look a little bit further back it's typically reviewers from 4 organisations (Payara, Fujitsu, Oracle and OmniFaces) and an individual. Reviews have indeed come in from all those. We recently voted in Piotr Żygieło as an individual, who's been doing really well in reviewing, even when not always explicitly tagged for review.

That all being said, I do believe in the openness of open source and the repo, and the ability of anyone to raise concerns and do PRs to revert things if needed. Mark Thomas (CC'ed) started a discussion about that a couple of times, and most repos he contributes to now work like that. It has advantages as that the pace of development is not unnecessarily constrained for many trivial things such as a version update of say the compiler plug-in. OpenMQ for example does a large number of commits as well, and would be severely slowed down if every commit needed a review.

There's also a minimum window to merge things in the GlassFish project, as we run a great number of tests on every commit, which need to pass first.

At any length, I'll make sure to always tag a representative number of people for PRs, and after the GlassFish 7 release we can allow for some more time.

Kind regards,

Arjan Tijms

On Thu, May 19, 2022 at 1:13 PM Steve Millidge (Payara) <steve.millidge@xxxxxxxxxxx> wrote:

Hi,

 

I know people are working rapidly to achieve the GlassFish 7.0.0 release for Jakarta EE 10. However post the 7 release we need to ensure that PRs when raised have a broad set of reviewers. Currently many PRs are being raised, solely requested  for review by a committer from the same company, then merged and closed all within a very short time. This does not enable a plurality of views to be taken into account. I think this is fine as new EE 10 apis are merged in. However beyond the EE 10 release I believe issues should be raised and reviews requested by members of the team from a different organisation. Happy to have some rules on lazy consensus to not slow everybody up. This should encompass all repositories within the GlassFish project.

 

Thanks

 

Steve Millidge

 

 

_______________________________________________
glassfish-dev mailing list
glassfish-dev@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/glassfish-dev