uPortal IRC Logs-2014-01-10

[07:32:59 CST(-0600)] <tlev> yes, 11.2.0.3

[07:33:48 CST(-0600)] <tlev> to answer jwennmacher

[10:05:31 CST(-0600)] <apetro__> I think I want to take a shot at getting travis-ci.org really working for CI of uPortal builds. As in, full-fledged ant initportal's against a database and

[10:05:37 CST(-0600)] <apetro__> "deploying" to a temp directory

[10:05:45 CST(-0600)] <apetro__> anyone have cycles in on that / wisdom to share?

[10:06:51 CST(-0600)] <tlev> i like the idea, what would we gain on top of having the nightly build that we do?

[10:07:21 CST(-0600)] <apetro__> continuous integration build on roughly every commit and on the prospective merge of every pull request

[10:07:38 CST(-0600)] <apetro__> travis pulls the code, runs the build, reports on the results right back into the GitHub UI

[10:07:58 CST(-0600)] <apetro__> you're looking at a pull request. If the pull request breaks the build, there's a big fat warning around the Merge button

[10:08:26 CST(-0600)] <apetro__> not that we routinely have pull requests that break the build, of course. but still. one more layer of automated safety net.

[10:10:18 CST(-0600)] <apetro__> basically this is CI / testing integration with GitHub, serving a slightly different purpose than a nightly build that actually deploys to a live Tomcat

[10:18:05 CST(-0600)] <tlev> I think it would be a good thing to do.

[10:37:33 CST(-0600)] <jwennmacher> If someone submits a pull request that breaks the build, but we don't get to it for a while (say weeks) does it keep building/warning for weeks, or does it do the build only once when the pull request was submitted?

[10:42:48 CST(-0600)] <jwennmacher> Does anyone have the ability to log onto the Amazon elastic instances to look into the periodic 'waiting for tomcat to start' issue? I tried to, but it won't accept my SSH (which is odd because I tried the private key I found in the bamboo configuration). See https://developer.jasig.org/bamboo/browse/UP-LIVEBUILD-822. The hanging issue happens ever week or so. I stopped the build so we could release the instance.

[10:47:03 CST(-0600)] <tlev> i think that it would build only when changes happen

[10:49:34 CST(-0600)] <jwennmacher> That's what I would think as well, but I wanted to verify. If so, it is a great enhancement. We don't want anything to be a nuisance (smile)

[12:50:45 CST(-0600)] <jwennmacher> tlev, petro_, does your uportal-war's persistence.xml have hibernate.properties set within it? And what version of hibernate-core are you using?

[13:40:38 CST(-0600)] <apetro__> tlev, jwennmacher, in answer to the travis-ci question I overlooked: travis-ci builds are triggered by commits (or manually). So it will build once-ish. It can re-build pull requests on commit to the branch to which they will be merged, since that changes the resulting merged code that needs tested. But perhaps more importantly, travis-ci can be configured about whether and how to email on brokenness. So, a typic

[13:40:38 CST(-0600)] <apetro__> al configuration is to alert on state changes, for a given branch/PR. clean to broken? Email. Broken to clean? Email.

[13:41:33 CST(-0600)] <jwennmacher> excellent.

[13:47:03 CST(-0600)] <apetro__> here's our persistence.xml , https://gist.github.com/8361302 , jwennmacher

[13:48:10 CST(-0600)] <jwennmacher> Thanks. Looks like the standard one. Not sure why I'm having issues with this client needing to have hibernate.dialect defined, but I'm adding it via UP-3909