Hi, Team,
An update: I’ve got the portConjugationVisuals_bug498282 test working properly, now.
For the other test, I took a guess at what the problem might have been, but that turned out to be wrong. This test imports the same RSA model as another test, opens the same composite structure diagram, and makes the same assertion about being able to find the diagram frame as that other test. The other test passes, this one fails to find the structure frame. And, of course, it only fails on the build server. I can’t imagine why.
So, I’m now setting up a Hudson build on a Linux virtual machine to emulate the Eclipse.org build environment. Hopefully I shall be able to reproduce the test failure, debug it, and learn how to avoid the same problem in the future. Wish me luck.
Christian
On 27 July, 2016 at 09:08:46, Christian Damus (give.a.damus@xxxxxxxxx) wrote: Hi, Quentin,
Yes, that’s the one I’m working on. It passes when the Papyrus-RT bundles are installed but not when they are absent, because the test accidentally relies on the Papyrus-RT transformation extension to fill in the Port::isConjugated attribute defined by UML 2.5 (the source model is a UML 2.2 model, which doesn’t have this Port attribute). On 27 July, 2016 at 09:06:40, LE MENEZ Quentin (quentin.lemenez@xxxxxx) wrote:
Hi Everyone,
I just launched a test on my computer (with the latests Neon version and on the Master branch- same issue anyway) and
the only one that failed was portConjugationVisuals_bug498282. I’ll try to make a new fresh install for Oxygen and test it there.
I’ll have a look at these. One I know how to deal with, but the other seems to be a failure only on the build server. I don’t suppose you could confirm whether
org.eclipse.papyrus.migration.rsa.tests.regression.StructureDiagramTest.dividerBetweenNameAndStructureCompartment_bug496653
fails on your local Windows system? (it doesn’t fail on my Mac) This one depends on opening a diagram in the editor being able to convert the old-style numeric visual IDs
to the Neon view types, but I think other tests also rely on this, so I have some difficult investigation to do.
On 27 July, 2016 at 02:42:20, MAGGI Benoit (benoit.maggi@xxxxxx) wrote:
_______________________________________________
mdt-papyrus.dev mailing list
mdt-papyrus.dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/mdt-papyrus.dev
|