Bug 317964 - Data loss when migrating UML 2.1 models into UML 2.2 if these models have instances of Template Parameter Substitution using multiple actuals
Summary: Data loss when migrating UML 2.1 models into UML 2.2 if these models have ins...
Status: NEW
Alias: None
Product: MDT.UML2
Classification: Modeling
Component: Core (show other bugs)
Version: 3.0.0   Edit
Hardware: PC Linux
: P3 major (vote)
Target Milestone: ---   Edit
Assignee: UML2 Inbox CLA
QA Contact:
URL:
Whiteboard:
Keywords: helpwanted
Depends on:
Blocks:
 
Reported: 2010-06-25 08:55 EDT by Anthony Hunter CLA
Modified: 2010-06-25 14:23 EDT (History)
3 users (show)

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Anthony Hunter CLA 2010-06-25 08:55:29 EDT
The "actual" and "ownedActual" fields of the TemplateParameterSubstitution metaclass have changed multiplicity from * to 1, essentially becoming scalars instead of collections in UML 2.2 spec.

Please refer to "http://www.eclipse.org/modeling/mdt/uml2/docs/guides/UML2_3.0_Migration_Guide/guide.html" and specifically the relevant change is:
·        TemplateParameterSubstitution::actual[1..*]->[1]   
·        TemplateParameterSubstitution::ownedActual[0..*]->[0..1]

If a UML 2.1 model with Template Parameter Substitution whose actual field has more than one element is migrated to UML 2.2 the value of the "actual" and "ownedActual" fields is not migrated correctly and there is consequently a data loss. Only one of the values from the collection (in 2.1) is picked (looks like the last one) and preserved in the now scalar field. That is reasonable as the multiplicity changed. However, in cases like this the migration process (the resource handler) applies the "UML" profile to the model and applies a specialized stereotype to the affected element. That stereotype then has a property which preserves data that would otherwise be lost. In case of Template Parameter Substitution migration there is no specialized stereotype and there is no way to recover the previous value.
Comment 1 Kenn Hussey CLA 2010-06-25 09:25:26 EDT
This looks like an oversight when the migration to UML 2.2 was done. :(
Comment 2 Anthony Hunter CLA 2010-06-25 11:20:58 EDT
Hi Kenn, what does the keyword 'helpwanted' indicate? This defect is a candidate for Helios SR1 right?

I imagine you are expecting a patch with the fix. Assuming we get one, are you or James available to review and create a SR1 over the summer?
Comment 3 Kenn Hussey CLA 2010-06-25 14:23:39 EDT
(In reply to comment #2)
> Hi Kenn, what does the keyword 'helpwanted' indicate? This defect is a
> candidate for Helios SR1 right?

It indicates that none of the committers currently has the bandwidth to implement a fix, so we're looking for help from the community (e.g., you).

> 
> I imagine you are expecting a patch with the fix. Assuming we get one, are you
> or James available to review and create a SR1 over the summer?

Yes.