Bug 358709

Comment 1 Adrian Price 2011-09-23 07:53:23 EDT

Doh, premature submit. Here's the rest of the description:

If one transform extends a base transform, and the base transform calls resolveoneIn(mapping, type), where mapping has been overridden in the extension transform, resolveonIn does not yield the correct result.

I noticed this when experimenting with extension mappings for the Simpleuml_To_Rdb.qvto example. I have attached my extension transform, in which the override UML::Class::persistentClass2table() : RDB::Table mapping is identical to the base version. If you compare the results you will see that those produced by the extension lack the includedColumns attribute.

Comment 2 Adrian Price 2011-09-23 08:31:32 EDT

Created attachment 203893 [details]
Extension transformation

Comment 3 Adrian Price 2011-09-23 08:33:05 EDT

Created attachment 203894 [details]
Extension transformation launch configuration

Comment 4 Christopher Gerking 2012-10-22 05:37:21 EDT

Created attachment 222626 [details]
JUnit test patch

Comment 5 Christopher Gerking 2012-10-22 05:42:09 EDT

Created attachment 222627 [details]
Patch for resolveIn on overridden mappings

The problem was that on execution of resolve operations, the overriding mapping is simply not considered. Fixed by the proposed patch.

Comment 6 Christopher Gerking 2012-11-16 13:52:31 EST

Created attachment 223669 [details]
JUnit test patch

Updated for javaless mode

Comment 7 Christopher Gerking 2012-11-16 13:53:08 EST

Created attachment 223670 [details]
JUnit test patch

Updated for javaless mode

Comment 8 Adolfo Sanchez-Barbudo Herrera 2013-01-24 07:34:07 EST

+1

Trivial comment (no action required, just for the future):
- when doing (overridingOper instanceof MappingOperation) the previous overridingOpen != null is not required.

Comment 9 Adolfo Sanchez-Barbudo Herrera 2013-03-11 07:09:00 EDT

The following has been committed and pushed to master for M6:
- Bug fix patch with trivial change commented above.

Resolving as fixed.

Comment 10 Sergey Boyko 2013-03-15 07:03:51 EDT

See also Bug 403440.

Comment 11 Philipp W. Kutter 2013-03-19 01:41:39 EDT

What is missing to put this to verified?

Comment 12 Ed Willink 2013-03-19 05:26:24 EDT

See my response on https://bugs.eclipse.org/bugs/show_bug.cgi?id=309762#c20.

Please do not waste our time by spamming Bugzillas. We have a developers mailing list (https://dev.eclipse.org/mailman/listinfo/qvto-dev) that is much more suitable for philosophiocal and project discussions.