|Re: [mdt.dev] Can MDT/OCL make MDT/UML2-Tools useable on Helios?
In its actual state in the repository the U2T should not have any dependencies to outdated components in general and to emf.ocl in particular. It had been regenerated with the release version of the GMP-tooling and has remarkable number of fixes and improvements from the 0.9.0 release. It also had been locally tested against the Helios and _should be_ compiliable and installable over the Helios modeling package.The historical computed dependency probably arises through EMF Query whose dependence caused significant transitive anguish for the M4 buckybuild.
It's a pity, because I don't have basically any idea on how to make a build either at http://modeling.eclipse.org/modeling/mdt/uml2tools/build/ or from somewhere else. I have to admit that building always had been a real trouble for me, so that's very probable that I am (for a couple of months at least) struggling with something really simple.I sympathize; it's a strange unhelpful no fun area. I can build but have yet to promote M2M/QVTd.
Ed, do you think someone from OCL team (may be the one who was responsible for your builds) can take a look at the U2T build?I looked at the releng. Unfortunately it's the really old pre-Hudson build. Most projects moved to the Hudson Dash-Athena build. Most projects are now moving to Buckminster.
I would appreciate any help, and really hope that it should not take much time/efforts.If all auto-generated code is in CVS, the build could be a simple variation of an existing build, so it would make sense for an existing build team to respin.
Obeo were kindly offering to support MDT/OCL builds. Perhaps Obeo might help here. I'll send them a direct email to draw attention to this thread.
Regards Ed Willink
Back to the top