Re: [qvtd-dev] RC1

[Horacio Hoyos Rodriguez <hhr502@xxxxxxxxxx>]

Hi,

I think the current model compare method is not fit for purpose.  I think
the discrepancies will come from collections that contain objects of
subtypes of the collection's type. In UmlToRdbms, a Package has a
collection of elements of PackageElement, which can be Classes, DataTypes,
or Associations. The schedule will impose an order in which the subtypes
are transformed, e.g. DataTypes before Classes before Associations, thus,
the output order will always depend on the schedule. Although the
scheduler should be deterministic, the correct order cannot me known a
priori. Hence, the Oracle/comparison model might not put the elements of
the package in the same order that the transformation puts them.

Yes, one could "cheat" and use the output to "fine tune" the oracle
models, but this seems not the correct approach.

Regards,


Horacio Hoyos Rodríguez
EngD Student
University of York

http://www.york.ac.uk/docs/disclaimer/email.htm

-----Original Message-----
From: qvtd-dev-bounces@xxxxxxxxxxx [mailto:qvtd-dev-bounces@xxxxxxxxxxx]
On Behalf Of Ed Willink
Sent: 18 May 2015 14:52
To: QVTD developers mailing list
Subject: [qvtd-dev] RC1

Hi

Potential OCL builds from OCL/QVTd master are available as N-builds.

I've managed to revise Bugzilla 464714 so that graphs are now available.

I've managed to rebased the two non-duiplicate commits of Bugzilla 465913,
but the result has genuine differences. It would be good to get this out
of the way.

     Regards

         Ed Willink
_______________________________________________
qvtd-dev mailing list
qvtd-dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe
from this list, visit https://dev.eclipse.org/mailman/listinfo/qvtd-dev