Eclipse Community Forums
Forum Search:

Search      Help    Register    Login    Home
Home » Modeling » QVT-OML » Concat EString and String - 3.1 -> 3.3 regress ?(Concatenation and plugin versioning)
Concat EString and String - 3.1 -> 3.3 regress ? [message #905374] Thu, 30 August 2012 09:08 Go to next message
thomas legrand is currently offline thomas legrandFriend
Messages: 10
Registered: September 2010
Junior Member
Hi everybody,

I recently installed QVTo in Juno and the concatenation of a EString (from EMF) and a String dose not work properly anymore.

Test case:

modeltype Ecore uses ecore('http://www.eclipse.org/emf/2002/Ecore');

transformation concatTestCase();


main()
{
	var myFirstname : String ;
	myFirstname := 'thomas';
	log('CONCAT STRING',myFirstname + ' moore');
	
	var myPseudo : EString ;
	myPseudo := 'franky';
	log('CONCAT ESTRING',myPseudo + ' moore');
}


Gives:

CONCAT STRING, data: thomas moore
CONCAT ESTRING, data: <Invalid>


Is this a regress or is it the new expected behaviour ?


Furthermore, I feel lost with the versionning: following the dates, it seems the Version should be 3.3.0 in the plugin descriptions. Here is my plugin version:

Operational QVT SDK 3.1.0.v201206041614-7U7A-CL5V0ETk4K7wL253-yIi90n org.eclipse.m2m.qvt.oml.sdk.feature.group Eclipse Modeling Project

Thanks!,
Tom
Re: Concat EString and String - 3.1 -&gt; 3.3 regress ? [message #905406 is a reply to message #905374] Thu, 30 August 2012 10:33 Go to previous messageGo to next message
Ed Willink is currently offline Ed WillinkFriend
Messages: 4188
Registered: July 2009
Senior Member
Hi

Apologies for the confusion. At one point it looked as if QVTo would be
dropped from the Juno release. This was avoided, but the contribution
was a little rushed. There was only an M4, RC1 and RC3. 3.2.0RC3 was not
renamed to appear as just 3.2.0. Since there were 7 benign JUnit test
failures, the P2 repositories were not correctly published.

The Juno release was 3.2.0, but Eclipse versioning policies allow
unchanged plugins to retain unchanged versions. So, since very little
changed in 3.2, much of the 3.2 content has 3.1 versions. New plugins
start at 1.0, so there are also some 1.1 plugins.

The builds have now migrated to Buckminster and the code is now under GIT.

Before using the new regime to build 3.2.1RC1, I built a 3.2.0a that is
in principle just a rebuild or 3., but I spotted that while changed
plugin versions had been updated to 3.2.0, the corresponding features
had not been, so your puzzlement about "Operational QVT SDK
3.1.0.v201206041614-7U7A-CL5V0ETk4K7wL253-yIi90n" should be resolved by
installing 3.2.0a.

(You can see the successful 3.2.1RC1 at
https://hudson.eclipse.org/hudson/job/buckminster-mmt-qvto-3.2/.)

===========

The Juno release of Eclipse OCL introduced a number of library functions
such as String::+(), String::characters() from OCL 2.2. These had
previously been defined by QVTo as workarounds for the Eclipse OCL
deficiency. Consequently QVTo 3.2.0 has ambiguous definitions which show
up as warnings in the editor. Since both routines are 'the same' it
doesn't matter which is chosen. The duplicate QVTo workarounds have been
removed in 3.2.1 and 3.3.0 development branches and so should disappear
in 3.2.1RC2 and 3.3.0M2.

Your regression suggests that the OCL and QVTo versions of String::+ are
not 'the same' in the case of EString + String.

Please raise a Bugzilla and maybe we can fix it for 3.2.1RC2.

Regards

Ed Willink


On 30/08/2012 10:08, thomas legrand wrote:
> Hi everybody,
>
> I recently installed QVTo in Juno and the concatenation of a EString
> (from EMF) and a String dose not work properly anymore.
>
> Test case:
>
> modeltype Ecore uses ecore('http://www.eclipse.org/emf/2002/Ecore');
>
> transformation concatTestCase();
>
>
> main()
> {
> var myFirstname : String ;
> myFirstname := 'thomas';
> log('CONCAT STRING',myFirstname + ' moore');
>
> var myPseudo : EString ;
> myPseudo := 'franky';
> log('CONCAT ESTRING',myPseudo + ' moore');
> }
>
>
> Gives:
>
> CONCAT STRING, data: thomas moore
> CONCAT ESTRING, data: <Invalid>
>
> Is this a regress or is it the new expected behaviour ?
>
>
> Furthermore, I feel lost with the versionning: following the dates, it
> seems the Version should be 3.3.0 in the plugin descriptions. Here is
> my plugin version:
>
> Operational QVT SDK 3.1.0.v201206041614-7U7A-CL5V0ETk4K7wL253-yIi90n
> org.eclipse.m2m.qvt.oml.sdk.feature.group Eclipse Modeling Project
>
> Thanks!,
> Tom
Re: Concat EString and String - 3.1 -&gt; 3.3 regress ? [message #905834 is a reply to message #905406] Fri, 31 August 2012 07:01 Go to previous messageGo to next message
Eclipse UserFriend
On 30.08.2012 12:33, Ed Willink wrote:
> Apologies for the confusion. At one point it looked as if QVTo would be
> dropped from the Juno release. This was avoided, but the contribution

Slightly off-topic: thanks a lot averting that!

Regards
Marius
Re: Concat EString and String - 3.1 -&gt; 3.3 regress ? [message #906487 is a reply to message #905406] Sat, 01 September 2012 16:36 Go to previous message
Ed Willink is currently offline Ed WillinkFriend
Messages: 4188
Registered: July 2009
Senior Member
Hi
On 30/08/2012 11:33, Ed Willink wrote:
> Your regression suggests that the OCL and QVTo versions of String::+
> are not 'the same' in the case of EString + String.
>
> Please raise a Bugzilla and maybe we can fix it for 3.2.1RC2.
>
https://bugs.eclipse.org/bugs/show_bug.cgi?id=388595 raised against OCL,
since it appears that any use of an EDataType as an operation call
source fails.

Fix in the pipeline hopefully for OCL 4.0.1 RC2 tomorrow.

Regards

Ed Willink
Previous Topic:Part of Transformations in Java
Next Topic:applyStereotype on input model elements
Goto Forum:
  


Current Time: Sun Dec 21 07:09:53 GMT 2014

Powered by FUDForum. Page generated in 0.01585 seconds
.:: Contact :: Home ::.

Powered by: FUDforum 3.0.2.
Copyright ©2001-2010 FUDforum Bulletin Board Software