Thank you for your feedback, Lars. Knox does follow a CTR policy for committers in Knox which is why you see much of what you mention. KNOX-640 was communicated on the dev@ list [1] to solicit feedback and opinions.
We watch the commits to the Apache git repository via the dev@ list and easily review those changes post commit. I don't recall ever using review board for contributions in Knox. I will apply the patch, review the changes and provide comments back to the contributor via JIRA comments. Perhaps others have used review board for Knox. Generally speaking, changes that are in line with current design, are bug fixes, etc do not get reviewed. Larger, architectural and security related changes should have patches attached and get reviewed. Knox certainly has a number of external contributions from folks that consist of bug fixes to larger features. I think that it would be inaccurate to characterize such contributions as blocked. However, we would love to encourage more. I am aware of the Sentry thread and found it to be concerning and unfortunate for that community, to be frank. The incubator can certainly define what is required/desired for graduation criteria though. Our committer workflow process should probably be reworded to better reflect the CTR policy of the Knox project. As for a 24 hr cool down period, I personally don't see the benefit to it in a project that follows CTR. The patches are readily available via the git notification and available for anyone to review. I do believe that as the community gets larger that we will need to revisit the CTR/RTC policy decision. All of that said, I have been consciously trying to communicate more via dev@ and within the JIRAs - as you can see by [1] and KNOX-640 - which reflect the content of that email thread. I think that we can discuss a cool off period and the possible need for attaching patches for every commit even though they are available through git. I'll start a separate DISCUSS thread for those topics. I would urge you to provide your insights - especially as to how the lack of either in a CTR process blocks external contribution. I will also look for well documented CTR policies within Apache and see how well we align with those. Again, thank you for your feedback. I hope that we can make you feel comfortable enough to make more contributions to Knox. Please feel free to express interest in any existing JIRAs, proposing features or whatever you like. [1] http://mail-archives.apache.org/mod_mbox/knox-dev/201511.mbox/%3CCACRbFygWLh2xVXrEWAD5ae%2BdoLm7iKf3a3e8kOkWSnYKPMkj4g%40mail.gmail.com%3E On Fri, Dec 11, 2015 at 4:27 AM, Lars Francke <lars.fran...@gmail.com> wrote: > Hi, > > I've been digging into Knox for the first time yesterday. I'm a contributor > to various other Apache projects and a Hive committer. > > I looked at a couple of JIRAs and then a few more random ones and I see a > few concerning things: > > * Some/lots of tickets don't have the final patch attached and only point > to a git or svn commit (KNOX-615, KNOX-601, KNOX-640) > * I haven't seen a single review being done. Knox Reviewboard is empty[1]. > Reviews are hard to do when there's no patch to review anywhere :) > * I found a couple of instances where patches were modified before > committing without attaching the patch that's really been committed > * I see issues being created, committed and closed within minutes ( > https://issues.apache.org/jira/browse/KNOX-640) which does not really > invite contributions and does not allow anyone to do any reviews or raise > objections to issues. > > This last issue has recently been discussed at length in the Incubator > mailing list in a thread called "Concerning Sentry"[2] and I urge you to > read that. The suggestion in that thread was to wait at least 24h before > committing a change and I think that's reasonable. > > Your contribution process[3] explicitly mentions that patches have to be > attached to JIRA and that reviews are necessary. > > Just FYI I'm seeing the same practice in the Ambari project and raised the > topic there as well[4,5] > > I understand that you're doing this to make fast progress and you're not > doing this to intentionally block external contributions but that's what > the net effect may be. So I urge you to look at your development practices > and consider changing them. > > Thanks! > > Cheers, > Lars > > [1] <https://reviews.apache.org/groups/Knox/> > [2] < > > http://thread.gmane.org/gmane.comp.apache.incubator.general/52126/focus=52351 > > > [3] <https://cwiki.apache.org/confluence/display/KNOX/Contribution+Process > > > [4]< > > http://mail-archives.apache.org/mod_mbox/ambari-dev/201512.mbox/%3CCAD-Ua_gt9TgKTxr12vrKO2kCEad3r97-GUwyB_LmUtFviFHt7A%40mail.gmail.com%3E > > > [5]< > > http://mail-archives.apache.org/mod_mbox/ambari-dev/201512.mbox/%3CCAD-Ua_hpjfnYGF6HNuOQqTaUrqXqb0fc97BpqJ%3Dy%2BBw59csR4g%40mail.gmail.com%3E > > >
