I still kind of like the build image historically, since not everybody who
interfaces with our project will _know_ that the build must always succeed
for a release, but I agree that this is a clean approach so I rolled back
the wiki changes.  Thanks,

Jon

On Tue, Oct 24, 2017 at 2:34 PM Matt Foley <ma...@apache.org> wrote:

> The release wouldn’t have been made if the build didn’t succeed.
> And the Release Manager doesn’t need one more fiddly manual edit to do.
> I recommend the Release Process instructions be put back, and instead we
> incorporate
> https://github.com/apache/metron/pull/815
> Rational in
> https://issues.apache.org/jira/browse/METRON-1278
>
> Thanks,
> --Matt
>
> On 10/24/17, 5:37 AM, "zeo...@gmail.com" <zeo...@gmail.com> wrote:
>
>     Hmm, I kind of like it as a historical validation/confirmation of build
>     success, but I can see where you are coming from.  Here
>     <
> https://cwiki.apache.org/confluence/pages/diffpagesbyversion.action?pageId=66854770&selectedPageVersions=34&selectedPageVersions=33
> >
>     is the wiki diff, feel free to critique/alter.
>
>     Jon
>
>     On Tue, Oct 24, 2017 at 7:07 AM Kyle Richardson <
> kylerichards...@gmail.com>
>     wrote:
>
>     > +1 I agree with Matt and Justin. The Travis build widget doesn't
> make sense
>     > in the published release documentation. No sense in fixing
> retrospectively.
>     >
>     > -Kyle
>     >
>     > On Mon, Oct 23, 2017 at 3:13 PM Matt Foley <mfo...@hortonworks.com>
> wrote:
>     >
>     > > I agree with Justin.  This micro-feature is intended as a github
> widget,
>     > > which causes the top-level README to give all viewers an immediate
> flag
>     > > whether the build is healthy or not.  It does not belong in a
> rendered
>     > > site-book.
>     > >
>     > > Removing the widget during site-book build, can be done with a
> one-line
>     > > addition to HREF_REWRITE_LIST in:
>     > >
>     > >
>     >
> https://github.com/apache/metron/blob/master/site-book/bin/generate-md.sh#L75
>     > >
>     > > Recommend not worrying about historical site-books (they naturally
>     > > obsolesce out of the “dist/release” area).
>     > >
>     > > Cheers,
>     > > --Matt
>     > >
>     > > On 10/23/17, 6:38 AM, "Justin Leet" <justinjl...@gmail.com> wrote:
>     > >
>     > >     I'd argue it shouldn't be in the site-book at all. Presumably
> we
>     > aren't
>     > >     releasing while Travis is broken, so it's not useful
> information for
>     > > anyone
>     > >     looking at docs. It just carries over from the main README.
> Seems
>     > > like we
>     > >     should just scrub it when we do the other fixes to the READMEs
> to
>     > make
>     > > them
>     > >     suitable for site-book.  At that point it's just gone
> entirely. from
>     > > the
>     > >     next release.
>     > >
>     > >     Doesn't solve the problem of prior releases (assuming we care
> enough
>     > > to do
>     > >     anything).
>     > >
>     > >     On Mon, Oct 23, 2017 at 9:32 AM, zeo...@gmail.com <
> zeo...@gmail.com>
>     > > wrote:
>     > >
>     > >     > Today I was poking around the Metron site and documentation,
> and I
>     > > noticed
>     > >     > that the site-book's travis build status image is pointing to
>     > master
>     > > for
>     > >     > all of our releases.  We should probably update the release
> process
>     > >     > <
>     > https://cwiki.apache.org/confluence/display/METRON/Release+Process>
>     > > to
>     > >     > pin
>     > >     > this to the specific release.
>     > >     >
>     > >     > I will happily handle the documentation update but I wanted
> to ask,
>     > > what
>     > >     > should it be pointing to?  Repointing is incredibly
> straightforward
>     > >     > <https://travis-ci.org/apache/metron/branches>, but I'm not
> sure
>     > > would be
>     > >     > more appropriate to use - the release branches (e.g.
> Metron_0.4.1),
>     > > or
>     > >     > the release tags (e.g. apache-metron-0.4.1-release)?  I'm
> not clear
>     > > on
>     > >     > their specific uses in our environment.  In reviewing our
> current
>     > > process,
>     > >     > it appears that we _could_ use either.
>     > >     >
>     > >     > I also wanted to ask, does anybody think that this should
> get fixed
>     > >     > historically?  I think that this might be an excessive
> amount of
>     > > hassle,
>     > >     > but I wanted to put it out there since I'm not intimately
> familiar
>     > > with
>     > >     > what we'd need to do in order to clean this up.
>     > >     >
>     > >     > Jon
>     > >     > --
>     > >     >
>     > >     > Jon
>     > >     >
>     > >
>     > >
>     > >
>     >
>     --
>
>     Jon
>
>
>
>
> --

Jon

Reply via email to