Hi folks! It's all set; you can check out the umbrella JIRA for further information: https://issues.apache.org/jira/browse/KNOX-1759
Additionally, you might want to read the enhanced documentation on how to contribute using GitHub PRs here: https://cwiki.apache.org/confluence/display/KNOX/Contribution+Process#ContributionProcess-GithubWorkflow Regards, Sandor On Fri, Feb 8, 2019 at 4:10 PM Sandor Molnar <[email protected]> wrote: > "We can't do this unless we only accept PRs. It is entirely possible to > commit to the repository without opening a PR so we want tests to run. The > Knox-master-daily job also runs full integration release tests and not just > the unit tests that are run on PRs currently." > > Got it; thanks for the clarification! > > On Fri, Feb 8, 2019 at 3:22 PM Kevin Risden <[email protected]> wrote: > >> I created https://issues.apache.org/jira/browse/KNOX-1759 with some >> subtasks for concrete action items to take to move forward. >> >> "If that happened we might get rid of the >> 'Knox-master-daily' job since it being executed after a commit is merged >> into master (AFAIK) which makes no sense if we only allow a commit to be >> merged if all tests were successfully passed already" >> >> We can't do this unless we only accept PRs. It is entirely possible to >> commit to the repository without opening a PR so we want tests to run. The >> Knox-master-daily job also runs full integration release tests and not >> just >> the unit tests that are run on PRs currently. >> Kevin Risden >> >> >> On Fri, Feb 8, 2019 at 4:34 AM Sandor Molnar <[email protected] >> > >> wrote: >> >> > +1 for PRs. >> > >> > My two cents on Kevin's list: >> > - PR template is a good idea; Ambari also has one here: >> > >> > >> https://github.com/apache/ambari/blob/trunk/.github/PULL_REQUEST_TEMPLATE.md >> > It would also be great if test steps are described in a detailed >> manner >> > (it helped me many times in case I had to reproduce something months >> after >> > the PR was merged) >> > >> > - comments on the PR: in case of Ambari they go to the 'Worklog' tab in >> the >> > corresponding JIRA, which - IMO - was better than put all of these stuff >> > within the comments; it gave us a clear separation and did not spam the >> > comments in the JIRA where other useful information may be found (i.e. a >> > design history, open point clarification, etc...). Not to mention that >> the >> > worklogs contain many information >> > >> > - link the PRs to the JIRA automatically is essential IMO; thanks for >> > pointing that out Kevin! >> > >> > - I'm not sure if it is feasible (currently does not seem to be the >> case) >> > but it would be great if contributors could invite others for review >> (i.e. >> > not only committers) >> > >> > - Apache has a Jenkins instance to run CI checks on its projects (Ambari >> > sample: >> https://builds.apache.org/job/Ambari-Github-PullRequest-Builder/). >> > Since Knox already has some jobs here (for instance >> > https://builds.apache.org/job/Knox-master-daily/) we might consider >> > creating Knox's PR Builder here too (so that all of our CI related jobs >> > would be in one place). If that happened we might get rid of the >> > 'Knox-master-daily' job since it being executed after a commit is merged >> > into master (AFAIK) which makes no sense if we only allow a commit to be >> > merged if all tests were successfully passed already >> > >> > Cheers, >> > Sandor >> > >> > On Fri, Feb 8, 2019 at 5:53 AM Jeffrey Rodriguez <[email protected]> >> > wrote: >> > >> > > +1 It is great that we are considering Pull request that would help to >> > > increase community collaboration. >> > > Jeffrey E Rodriguez >> > > >> > > On Thu, Feb 7, 2019 at 3:43 PM Robert Levas >> <[email protected] >> > > >> > > wrote: >> > > >> > > > +1. I think this is a great idea. >> > > > >> > > > On Thu, Feb 7, 2019 at 5:29 PM larry mccay <[email protected]> >> wrote: >> > > > >> > > > > Great list of ideas/practices there, Kevin! >> > > > > >> > > > > I for one would want comments added as comments to JIRA. >> > > > > I hate coming across a JIRA that would address something that I am >> > > > looking >> > > > > for and then find no meaningful comments. >> > > > > >> > > > > >> > > > > On Thu, Feb 7, 2019 at 4:20 PM Phil Zampino <[email protected]> >> > > wrote: >> > > > > >> > > > > > +1, let's follow good models from the community, and save >> ourselves >> > > > those >> > > > > > headaches which can be avoided. >> > > > > > >> > > > > > On Thu, Feb 7, 2019 at 4:17 PM >> > Kevin Risden >> > <[email protected]> >> > > > wrote: >> > > > > > >> > > > > > > I think PRs are a good improvement since we get Travis CI >> checks >> > by >> > > > > > default >> > > > > > > currently. Something that we currently don't get with patches >> > > > > > > >> > > > > > > If we go this route we should make sure we have the following >> in >> > > > place: >> > > > > > > >> > > > > > > - PR Github Template with useful info >> > > > > > > - >> > > > > > > >> > > > > > > >> > > > > > >> > > > > >> > > > >> > > >> > >> https://help.github.com/articles/creating-a-pull-request-template-for-your-repository/ >> > > > > > > - Livy has an example of this >> > > > > > > - Ensure that PRs are automatically linked to JIRA >> > > > > > > - Not currently done today and a pain since it should >> > happen >> > > > > > > automatically. >> > > > > > > - Calcite has this. Might be a simple INFRA ticket >> > > > > > > - Documentation for contributors/committers >> > > > > > > - Committers - linked github/asf accounts, how to merge >> a >> > PR >> > > > > > > - Contributors what to expect/updated docs to move from >> > patch >> > > > -> >> > > > > PR >> > > > > > > - Ensure that only squash/rebase/merge commits are allowed >> > > > > > > - Lot of nuance here and Calcite recently had INFRA >> disable >> > > the >> > > > > > > buttons for types that didn't fit their model >> > > > > > > - Decided what to do with PR comments >> > > > > > > - Some projects have PR comments go directly to JIRA >> > > comments. >> > > > > > > - Others have them go to worklog in JIRA. >> > > > > > > - Others don't capture PR comments in JIRA >> > > > > > > >> > > > > > > So all in all in favor just need to make sure we have the >> > plumbing >> > > in >> > > > > > > place. >> > > > > > > Kevin Risden >> > > > > > > >> > > > > > > On Thu, Feb 7, 2019 at 4:13 PM Sandeep Moré < >> > [email protected] >> > > > >> > > > > > wrote: >> > > > > > > >> > > > > > > > I second Phil. Personally, I am more comfortable with the >> > patches >> > > > > > mostly >> > > > > > > > because of their simplistic nature but do like PRs as they >> are >> > > more >> > > > > > > > community friendly (helps people review, comment, critique) >> and >> > > > looks >> > > > > > > like >> > > > > > > > they have become OSS standard as Phil pointed out. >> > > > > > > > So +1 from me. >> > > > > > > > >> > > > > > > > Best, >> > > > > > > > Sandeep >> > > > > > > > >> > > > > > > > On Thu, Feb 7, 2019 at 4:06 PM Phil Zampino < >> > [email protected] >> > > > >> > > > > > wrote: >> > > > > > > > >> > > > > > > > > I like the PR model, and it is familiar to many who >> > contribute >> > > to >> > > > > OSS >> > > > > > > > > projects. I suppose we could continue to support the patch >> > > > attached >> > > > > > to >> > > > > > > a >> > > > > > > > > Jira model, but we should encourage the PR model, IMHO. >> > > > > > > > > >> > > > > > > > > On Thu, Feb 7, 2019 at 4:03 PM larry mccay < >> > [email protected]> >> > > > > > wrote: >> > > > > > > > > >> > > > > > > > > > All - >> > > > > > > > > > >> > > > > > > > > > There has been interest from the Knox community in >> support >> > of >> > > > > Pull >> > > > > > > > > Requests >> > > > > > > > > > from github. >> > > > > > > > > > Our move to gitbox recently makes this easier to do. >> > > > > > > > > > >> > > > > > > > > > What are your thoughts on enabling PRs in general? >> > > > > > > > > > Should we support both patches in JIRA as well as github >> > > based >> > > > > PRs? >> > > > > > > > > > >> > > > > > > > > > thanks, >> > > > > > > > > > >> > > > > > > > > > --larry >> > > > > > > > > > >> > > > > > > > > >> > > > > > > > >> > > > > > > >> > > > > > >> > > > > >> > > > >> > > >> > >> >
