| Re: [ecf-dev] JMS build |
On 10/01/2010 04:09 PM, Scott Lewis wrote:
<stuff deleted>
I removed the requires section of feature.xml...and that didn't help (i.e. still failed: https://ecf2.osuosl.org/hudson/job/C-HEAD-jms.feature/94/ The strange thing is the initial compile errors are for org.eclipse.ecf.provider.jms...which has always been in the JMS feature. It's not finding some ECF plugins for compile. What I meant was to try the last building feature.xml, not to remove requires all together. Does tree [0] build successfully?
I don't know....I'll fuss with trying that as soon as I can. I don't immediately see how the requires in feature.xml would have anything to do with the target platform contents though...and it appears to me that target platform contents is the problem here (although I admit I don't know why it was working previously).
I checked on the build machine, and I can see that the target platform path does *not* have all the needed ECF bundles...so I guess this is the problem. Missing (e.g.) org.eclipse.ecf.provider, org.eclipse.ecf.provider.remoteservice and some others. See below for the list of the bundles in /opt/hudson/jobs/C-HEAD-jms.feature/workspace/targetPlatformPath/plugins on ecf2. I guess the target platform has to be amended/added to. I can't quite understand where the target platform (bundles below) is actually being gotten from/created for this build.The TP is created by Buckminster during the build based on the feature.xmls, buckminster.cspex, *.mspec, *.cquery's and *.rmaps
Ok...so I guess I need to know...what are the right additions (to cspex, mspec, cquery and/or rmap) to include all of the ECF sdk (some successful recent build from HEAD I guess) in the JMS target platform? Don't we have some boilerplate for that (i.e. isn't it already like that in some other builds)? If so just point me in that direction.
Thanks, Scott