Thanks, Joel. Do you know why it doesn't seem to reflect the recent changes? For example, kafka-1928 still has jkreps listed as the reviewer.
Jun On Mon, May 18, 2015 at 3:10 PM, Joel Koshy <jjkosh...@gmail.com> wrote: > Here is (roughly) what I had in mind for this: > > https://cwiki.apache.org/confluence/display/KAFKA/Patch+submission+and+review#Patchsubmissionandreview-Reviewqueue > > Earlier versions of wiki/jira integration were much better in that the > inline table had pagination within the wiki page itself. Unfortunately > in newer versions if there are more than a certain number of issues in > the list it will give you a link to the jira search. We could split > the table up by reviewer which may be better for visibility and peer > pressure as well :) > > Thanks, > > Joel > > On Wed, Apr 08, 2015 at 10:43:54AM -0700, Neha Narkhede wrote: > > > > > > We can do both of course, but I think apache infra already has > > > what is needed for this. > > > > > > Yes, if Apache INFRA has this, great. I would pick something that is a > > *push* mechanism and leave the *pull* as an option for proactive folks to > > refer to. Honestly, today you can go and save the JIRA query that lists > you > > as a reviewer today, but doesn't look like that is something people are > > doing. I'll be honest, I don't remember to do that myself. > > > > I would really encourage people to make this an email that gets sent out > to > > the entire dev@ mailing list so everyone is on the loop about the > patches > > that need reviews, who is on-point and which ones have no reviewers. > > > > On Wed, Apr 8, 2015 at 10:22 AM, Joel Koshy <jjkosh...@gmail.com> wrote: > > > > > The weekly email sounds good, but I would much rather if that were > > > integrated into the wiki/jira system than run a custom script on > > > non-apache infra. We did something similar during 0.8 replication > > > development where we embedded JQL links in the wiki and had a nice > > > dashboard view of jiras on a wiki page. > > > > > > So I was thinking of something similar - i.e., a wiki page of > > > committers and contributors that have signed up to review patches and > > > have a jira table for each. We could allocate no more than five > > > minutes at the hangout we have been doing to go over the top > > > patches per reviewer and decide whether patch reviews need to move > > > around. I think having this view helps in the interest of > > > visibility. > > > > > > We can do both of course, but I think apache infra already has > > > what is needed for this. > > > > > > Thanks, > > > > > > Joel > > > > > > On Wed, Apr 08, 2015 at 09:07:00AM -0700, Neha Narkhede wrote: > > > > The other suggestion I'd like to make involves writing a simple JIRA > > > script > > > > (whaat?). Currently the weekly email we get from JIRA just tells you > what > > > > JIRAs have patches. But it doesn't tell you which > committers/reviewers > > > have > > > > signed up to review those patches. The 2nd piece of information can > be > > > > really helpful. Here's why- > > > > > > > > I'm sure everyone wants to know which patches are open and since our > > > > community is so active, that is often a long list even if each JIRA > is > > > > actively being reviewed. As a busy committer, I think if I knew the > list > > > of > > > > JIRAs I've signed up to review and commit, I would really appreciate > the > > > > weekly reminder. It allows me to either get to it or give it up to > > > someone > > > > else. > > > > > > > > If someone had some time, they can look at my patch-review script > that > > > uses > > > > some of JIRA's python APIs and write a little script that does this. > At > > > > Confluent, we are happy to host that script on some EC2 machine. > > > > > > > > Thanks, > > > > Neha > > > > > > > > On Wed, Apr 8, 2015 at 8:41 AM, Gwen Shapira <gshap...@cloudera.com> > > > wrote: > > > > > > > > > Hi Jun, > > > > > > > > > > Do reviewers have to be committers? > > > > > I'll be happy to review patches in areas that I'm familiar with, if > > > this > > > > > makes life easier for committers and helps us get improvements in > > > faster. > > > > > > > > > > Gwen > > > > > > > > > > On Tue, Apr 7, 2015 at 11:10 PM, Jun Rao <j...@confluent.io> wrote: > > > > > > > > > > > Hi, everyone, > > > > > > > > > > > > Currently, we have accumulated quite a few jira backlogs. We can > > > probably > > > > > > brainstorm a bit what we can do to improve this. A few thoughts. > > > > > > > > > > > > 1. There are quite a few jiras marked with "patch available", but > > > with > > > > > > unaddressed review comments. I made a pass and moved them back > to "in > > > > > > progress" (just click the "Resume Progress") button. Now, there > are > > > about > > > > > > 100 jiras with valid patches. As people review the patches, it > may be > > > > > > useful to keep the status updated as well (thanks Neha for the > > > > > suggestion). > > > > > > > > > > > > 2. Some of the jiras have trivial patches. Perhaps we can just > assign > > > > > them > > > > > > with "trivial" priority. This way, a committer with only limited > > > amount > > > > > of > > > > > > time may be able to help out. > > > > > > > > > > > > 3. Some of the jiras have been tagged with reviewers. It would be > > > good if > > > > > > those reviewers can spend time finishing the reviews. If they > can't > > > do > > > > > the > > > > > > review soon, they can unset the reviewer field so that someone > else > > > can > > > > > > pick up the review. > > > > > > > > > > > > 4. Perhaps we can also spend a bit of time to triage at least the > > > active > > > > > > jiras during our weekly hangout. > > > > > > > > > > > > Anything else that we can do? > > > > > > > > > > > > Thanks, > > > > > > > > > > > > Jun > > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > > > Thanks, > > > > Neha > > > > > > > > > > > > -- > > Thanks, > > Neha > >
