I think Justin was just giving rationale, not suggesting we shouldn't make the improvements. I think at one point we also had some ambiguity about whether or how Docker was going to run in Travis, but I believe some folks have found a path through that? I agree with you both - let's get this added to Travis. Good suggestion, Nick. +1.
On Wed, May 22, 2019 at 7:53 AM Nick Allen <n...@nickallen.org> wrote: > Yes, running up Full Dev is a manual verification that is required. And as > a manual verification sometimes that will get missed. And in this specific > case, tt does seem a bit silly that the addition of a parser should require > the contributor to run up Full Dev. > > That being said, anything we can do to reduce the amount of manual > verification that is required is a good thing. We should be pushing > ourselves to an end state where no manual verification is required for any > Metron PR. I think building the RPMs/DEBs as part of the Travis build is > at least a small step in the right direction. > > > On Wed, May 22, 2019 at 9:34 AM Justin Leet <justinjl...@gmail.com> wrote: > > > Theoretically, we didn't need to before there were both RPMs and DEBs > since > > running dev up (which necessitates building those) is part of the build > > process. Since they've been split apart, I agree we probably should be > > building them, because nobody is going to run both unless they > specifically > > done something they'd expect to affect both. > > > > On Wed, May 22, 2019 at 9:30 AM Nick Allen <n...@nickallen.org> wrote: > > > > > In light of issues like this > https://github.com/apache/metron/pull/1419, > > > has anyone looked into building our RPMs and DEBs in Travis? This is a > > > very common and easy mistake to make and our CI builds really should be > > > able to catch this. > > > > > >
