|
Re: [QVTO] Applying stereotypes to out uml model [message #100251 is a reply to message #100172] |
Wed, 18 February 2009 14:25 |
Radomil Dvorak Messages: 249 Registered: July 2009 |
Senior Member |
|
|
Hi Matt,
I have scheduled that bug for M6.
The cause of the problem is that the contents of an out model extent is
not bound
to a resource during transformation execution /* for performance reasons
*/.
Every newly created object has to be added to an extent but we do not add
it to a resource during execution.
It is done during serialization phase.
I will check what we can do ASAP.
Thanks for the help offer, it would be great if you can provide a simple
testcase scenario provoking that problem (model, transf,...), so we can add
it to our test-suite once it's resolved.
Regards,
/Radek
On Wed, 18 Feb 2009 13:13:49 +0100, Matt Fomr <matt.fomr@gmail.com> wrote:
> Hi I've noticed that current qvto implementation lacks proper
> serializing of out model with applied stereotypes to its elements:
>
> https://bugs.eclipse.org/bugs/show_bug.cgi?id=238805
>
> Because it's a 'must have' feature for my purposes my question is:
> is there a chance for that this bug to be fixed in near future?, how can
> I help (provide test cases, etc.)
>
> If not - what's recommended 'clean' work around for this issue? My only
> 'not clean' idea is to use keywords as a storage for 'serialized'
> stereotypes (for example 'stereotype_name
> [steretype_property_name=value]' and add transformation postprocessing
> in pure java which would convert keywords to stereotypes...
>
> Anyone with a better idea (or any other idea :))?
>
> Regards
>
> Matt
>
|
|
|
|
Powered by
FUDForum. Page generated in 0.02819 seconds