Jenkins and Windows

[Rod Widdowson]

Apologies for the delay in getting back to the list, I have been out on vacation these past three days.

> So, thanks for reading, and for your feedback. I’ll head further down the
> “platform axis” path and see what comes up, but I’m open to suggestions.

I saw this mail earlier and thought about it offline.  I think that I conclusion that (IMO) we actually want to test two things: integration and regression.

For the latter I would agree that adding Windoze as an axis to the multi-JDK is the correct thing to do.  I actually think that moving the build triggering to be "nightly no matter what" rather than SCM driven (which is what I understand you to mean) is an advantage - I fixed a regression in an OpenSAML test over the break which was triggered by a JDK upgrade and not by an SCM change - so the regression had been around for a long time before I made the checkin which triggered the build which failed.

For the former I think we should add windows as a <something> to the integration tests.  I have an JIRA case which I want to address this quarter to expand these to Jetty9.3 (and to test our installation technology at the same time - the container part of which being something I want to at remove from the IdP) and I'd like to expand this to test against Tomcat (longer term).  At this stage having windows as an axis for idp-integration-tests makes sense, but it may not be sensible to run this one automatically.  Currently, integration tests are pretty light weight, but if I add another two containers (together with the time taken to start/stop them) it may begin to add up.  Precisely how this should be done is less obvious to me perhaps it can be a platform (or add integration tests to multi-jdk?), but at least in this case there is no downstream project.

/Rod