Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [qvtd-dev] [Dev] MTCBorker creates GraphML representations

Hi

"Didn't work" again. You make it very difficult to help you.

Rebase worked fine for me both ways round. See for example ewillink/fb90.

Just let me know if they're ready for master and I'll sort it out.

    Ed

On 22/04/2015 17:11, Horacio Hoyos Rodriguez wrote:

Hi Ed,

 

I had to cherry pick hhoyos/463966 and hhoyos/461806 on fb90 because a rebase didn’t work.  Is this correct, or do we want to keep track of the origin of the comits?

 

Regards,

 

Horacio Hoyos Rodríguez

EngD Student

University of York

 

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

 

From: qvtd-dev-bounces@xxxxxxxxxxx [mailto:qvtd-dev-bounces@xxxxxxxxxxx] On Behalf Of Ed Willink
Sent: 22 April 2015 11:55
To: qvtd-dev@xxxxxxxxxxx
Subject: Re: [qvtd-dev] [Dev] MTCBorker creates GraphML representations

 

Hi Horacio

You have two branches that look useful. They rebase happily.

Are they ready to be promoted to master?

Do you want to rebase one n the other or shall I?

    Regards

        Ed

On 21/04/2015 14:56, Horacio Hoyos Rodriguez wrote:

Hi,

 

I have pushed changes to fb90 to test for domains being null and correctly handle the case.  Datums without a domain are not assigned one of the used domain colors to make them distinguishable.

 

Graphml files are now generated in a separate debug package.

 

 

Regards,

 

Horacio Hoyos Rodríguez

EngD Student

University of York

 

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

 

From: qvtd-dev-bounces@xxxxxxxxxxx [mailto:qvtd-dev-bounces@xxxxxxxxxxx] On Behalf Of Ed Willink
Sent: 21 April 2015 10:40
To: qvtd-dev@xxxxxxxxxxx
Subject: Re: [qvtd-dev] [Dev] MTCBorker creates GraphML representations

 

Hi

mtc.executeQvtsToGraphML() is a debug tool. It should be as robust as possible so that it will always do its best to draw a graph. Bad names can be displayed as <<null>>. Bad classes as red boxes. The 'bad' graph can still be very instructive. In the same way I endeavour to ensure that ToStringVisitor and PrettyPrintVisitor are robust. Sadly I seem to keep having to dig out more assumptions about eResource() being non-null.

    Regards

        Ed

   

On 21/04/2015 10:10, Horacio Hoyos Rodriguez wrote:

Hi Ed,

 

Yes, unfortunately this example suffers from the current issue of proper domain identification (homogenous domains). However, I fixed the Schedule model manually. So if you use the UpperToLowerSchedule.xmi in my branch it should work.

 

The other caveat I forgot to mention is that in fb90 I am only testing the graph generation, so I am not running the complete MTC and hence after graph generation the test will fail when trying to assert that all models in the mtc exists. I could probably create separate test cases for graph generation, in the form of either only running mtc.executeQvtsToGraphML(), or running the whole MTC with the graph generation option on/off.

 

Regards,

 

Horacio Hoyos Rodríguez

EngD Student

University of York

 

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

 

From: qvtd-dev-bounces@xxxxxxxxxxx [mailto:qvtd-dev-bounces@xxxxxxxxxxx] On Behalf Of Ed Willink
Sent: 21 April 2015 10:04
To: qvtd-dev@xxxxxxxxxxx
Subject: Re: [qvtd-dev] [Dev] MTCBorker creates GraphML representations

 

Hi

Just tried fb90. Unfortunately the one testUpperToLower test that uses it fails with an NPE. The problem is that a ClassDatum has a null domain and when passed to

    public MutiNamedElementKeyImpl(@NonNull NamedElement... values) {

the internal code barfs on the null.

NB the @NonNull in the declaration is really

    public MutiNamedElementKeyImpl(@NonNull NamedElement[] values) {

and so asserts that the array is non-null with no promises about the content. Array/List content @NonNull declarations must wait till we move to Java 8, probably for Mars+1.

    Regards

        Ed

On 20/04/2015 20:09, Horacio Hoyos Rodriguez wrote:

Hi,

 

Branch hhoyos/fb90 has the added functionality for automatic generation of the dependency graph, currently graph supported is yEd’s GraphML.

 

·         The QVT Schedule has a ScheduleToDependecyGraphVisitor which is responsible for the transformation

·         GraphMLBuilder implements the GraphBuilder, tailored for yEd’s GraphML format.

·         Current colors, lines,  shapes , etc., are specified in ScheduleToDependecyGraphVisitor and I think I have made it customizable enough to allow easy personalization (to be added in the future, e.g. eclipse color options). Current color selection is based on [1], with the intention to facilitate easy change between light and dark themed environments.

 

Regards,

 

Horacio Hoyos Rodríguez

EngD Student

University of York

 

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

 

From: Ed Willink [mailto:ed@xxxxxxxxxxxxx]
Sent: 16 April 2015 12:48
To: Horacio Hoyos Rodriguez
Subject: Re: [Dev] MTCBorker creates GraphML representations

 

Hi

On 16/04/2015 12:24, Ed Willink wrote:


It's probably quicker for me to do it.

Give ewillink/464714 a try.

    Regards

        Ed

 

_______________________________________________
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

 

No virus found in this message.
Checked by AVG - www.avg.com
Version: 2015.0.5863 / Virus Database: 4331/9581 - Release Date: 04/20/15

 



_______________________________________________
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



No virus found in this message.
Checked by AVG - www.avg.com
Version: 2015.0.5863 / Virus Database: 4334/9589 - Release Date: 04/21/15

 




_______________________________________________
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




No virus found in this message.
Checked by AVG - www.avg.com
Version: 2015.0.5863 / Virus Database: 4334/9589 - Release Date: 04/21/15

 



_______________________________________________
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


No virus found in this message.
Checked by AVG - www.avg.com
Version: 2015.0.5863 / Virus Database: 4334/9596 - Release Date: 04/22/15



Back to the top