qvtd-dev

[qvtd-dev] Welcome Horacio Hoyos as a new modeling.mmt.qvtd Committer

portal-noreply@xxxxxxx (portal on behalf of Ed Willink) — Tue, 18 Jun 2013 13:28:01 GMT

modeling.mmt.qvtd Committers,
This automatically generated message marks the completion of all the legal
paperwork and webmaster provisioning for Horacio Hoyos. Horacio Hoyos is a
new full Committer on the modeling.mmt.qvtd project.

Welcome!

Re: [qvtd-dev] Slow down

ed@xxxxxxx (Ed Willink) — Tue, 04 Jun 2013 10:04:18 GMT

Hi

The intention was that the call 'add'ed all the bound variables to the nested environment, then the loops 'replace'd with the prevailing iteration value. If you really don't have control you can just 'replace' anyway.

I needed to add safeVisit to the QVTiEvaluationVisitor interface to avoid CCEs. At that point it gives the same utility as if it were in the Visitor interface. You can also have a protected overloading undecoratedVisitor2 to avoid the getUndecoratedVisitor() calls and casts.

I need to see a sequence diagram to understand why the loop should be in the visitor interface.

Regards

Ed

On 04/06/2013 10:32, Horacio Hoyos Rodriguez wrote:

Hi Ed,

Finding the root of the IAE is in my to-do list. I think it has to do something with the nested evaluation environment being reused somehow or not being nested correctly.

The retiming of the creation of the nested evaluation visitor was needed to make the safeVisit work with the current visitor implementation, but as you say I think that we should leave it as it simplifies the recursive loop in mapping calls. If the doLoop could somehow be added to the QVTi evaluation interface we could have access the the recursive call through the decorated visitor.

Regards,
_______________________________________________
qvtd-dev mailing list
qvtd-dev@xxxxxxxxxxx
http://dev.eclipse.org/mailman/listinfo/qvtd-dev
No virus found in this message.
Checked by AVG - www.avg.com
Version: 2013.0.3343 / Virus Database: 3184/6381 - Release Date: 06/03/13

Re: [qvtd-dev] Slow down

hhr502@xxxxxxx (Horacio Hoyos Rodriguez) — Tue, 04 Jun 2013 09:32:25 GMT

Hi Ed,

Finding the root of the IAE is in my to-do list. I think it has to do something with the nested evaluation environment being reused somehow or not being nested correctly.

The retiming of the creation of the nested evaluation visitor was needed to make the safeVisit work with the current visitor implementation, but as you say I think that we should leave it as it simplifies the recursive loop in mapping calls. If the doLoop could somehow be added to the QVTi evaluation interface we could have access the the recursive call through the decorated visitor.

Regards,

[qvtd-dev] Slow down

ed@xxxxxxx (Ed Willink) — Tue, 04 Jun 2013 09:16:23 GMT

Hi Horacio

I've been trying to introduce some discipline to the lash up I used to get performance figures and find that the QVTi Interpreter has got 25% slower. The problem is that your retiming of the creation of the nested evaluation visitor means that additions to the environment regularly throw an IAE causing the retry as a replacement. Normal execution should not throw exceptions.

In principle the retiming may be fractionally faster since there is one fewer parameters to pass, but the IAEs need to be fixed.

    Regards

Ed

[qvtd-dev] Committer vote for Horacio Hoyos has been approved by the PMC

portal-noreply@xxxxxxx (portal on behalf of portal-noreply) — Mon, 03 Jun 2013 19:41:01 GMT

modeling.mmt.qvtd Committers,
This automatically generated message marks the PMC's approval of the vote
for Horacio Hoyos's full Committer status on the modeling.mmt.qvtd project.
The next step is for the project lead to return to the portal and fill in
the VCS package and employer information for Horacio Hoyos.

The PMC's comments were: Welcome.

If you have any questions, please do not hesitate to contact your project
lead, PMC member, or the EMO

[qvtd-dev] Committer vote for Horacio Hoyos has concluded successfully

portal-noreply@xxxxxxx (portal on behalf of ) — Mon, 03 Jun 2013 17:11:01 GMT

modeling.mmt.qvtd Committers,
This automatically generated message marks the successful completion of
voting for Horacio Hoyos to receive full Committer status on the
modeling.mmt.qvtd project. The next step is for the PMC to approve this
vote, followed by the EMO processing the paperwork and provisioning the
account.

Vote summary: 1/0/0 with 0 not voting 
  +1  Ed Willink

If you have any questions, please do not hesitate to contact your project
lead, PMC member, or the EMO

[qvtd-dev] Vote for Committer status for Horacio Hoyos was started by Ed Willink

portal-noreply@xxxxxxx (portal on behalf of ) — Mon, 03 Jun 2013 17:11:01 GMT

modeling.mmt.qvtd Committers,
This automatically generated message signals that Ed Willink has nominated
Horacio Hoyos as a Committer on the modeling.mmt.qvtd project. The reason
given is as follows:

Horacio is a student working under my supervision at York.

The first phase of his research has led to the contribution of an
interpreter for QVTi (Bug 405987). 

He is now ready to work directly as part of the WQVTd project rather than
remotely on GitHub.

  * https://bugs.eclipse.org/bugs/show_bug.cgi?id=405987

The vote is being held via the MyFoundation portal: voters *must* use the
portal for the votes to be properly recorded.  The voting will continue
until either all 1 existing Committers have voted or until they have been
given enough time to vote, even if they do not do so (defined as at least
one week). Because there are three or fewer existing Committers, they must
all vote and vote +1 for a successful election.

Eligible Committers must cast their votes through their My Foundation
portal page (do NOT just reply to this email; your vote will not be
correctly recorded unless you use the portal):

    http://portal.eclipse.org/

The project Committers eligible to vote are:

    Ed Willink

*NOTE*: Successful elections are left open for a maximum of 120 days to
allow for processing of paperwork.  After that time the election will be
expired, regardless of its current status.  Should papework processing on
the part of the candidate take more time than allowed, a new election will
have to be held.

If you have any questions, please do not hesitate to contact your project
lead, PMC member, or the EMO

Re: [qvtd-dev] [Dev] Pivot model and generated opposite relations

ed@xxxxxxx (Ed Willink) — Tue, 28 May 2013 16:33:45 GMT

Hi

I suspect that the problem is that the opposites appear during QVTi evaluation.

OCL assumes that no changes occur, so perhaps the opposites were analyzed too soon.

Anyway the middle to outer opposites are well worth optimising so a variant of ImplicitNonCompositionProperty should collaborate with a better ModelManager to look up the opposites in a custom cache.

We can discuss further tomorrow.

Regards

Ed

On 28.05.2013 17:15, Horacio Hoyos Rodriguez wrote:

Hi Ed,

Doing more work in the QVTimperativeTracingEvaluationVisitor class I have been getting an exception when using the opposite relations auto generated by the Pivot model.

In line 390 of UML2RDBMS.qvti I use p2n := t.PrimitiveToName; to get a reference to the trace model class used to trace the PrimitiveDataType t. While executing I get:

org.eclipse.ocl.examples.domain.values.impl.InvalidValueException: Failed to evaluate simpleuml::PrimitiveDataType.PrimitiveToName

                at org.eclipse.ocl.examples.pivot.evaluation.EvaluationVisitorImpl.visitPropertyCallExp(EvaluationVisitorImpl.java:786)

                at org.eclipse.ocl.examples.pivot.evaluation.AbstractEvaluationVisitorDecorator.visitPropertyCallExp(AbstractEvaluationVisitorDecorator.java:255)

                at org.eclipse.ocl.examples.pivot.internal.impl.PropertyCallExpImpl.accept(PropertyCallExpImpl.java:391)

                at org.eclipse.ocl.examples.pivot.evaluation.EvaluationVisitorImpl.safeVisit(EvaluationVisitorImpl.java:171)

                at uk.ac.york.qvtd.pivot.qvtimperative.evaluation.QVTimperativeTracingEvaluationVisitor.visitVariableAssignment(QVTimperativeTracingEvaluationVisitor.java:180)

                at org.eclipse.qvtd.pivot.qvtcorebase.impl.VariableAssignmentImpl.accept(VariableAssignmentImpl.java:173)

                at uk.ac.york.qvtd.pivot.qvtimperative.evaluation.QVTimperativeLMEvaluationVisitor.visitBottomPattern(QVTimperativeLMEvaluationVisitor.java:103)

                at uk.ac.york.qvtd.pivot.qvtimperative.evaluation.QVTimperativeTracingEvaluationVisitor.visitBottomPattern(QVTimperativeTracingEvaluationVisitor.java:63)

…

I added print statements to check that the t variable is not null. Since the QVTi editor does not complaint about the reference I am guessing something is not working in my code, probably the PrimitiveToName element is not being properly assigned the primitive association value.

My question is, can you recommend a smart way of printing the Variable values to get useful information? Perhaps a printer that prints the variable name and the value of all its attributes?

Regards,

Horacio Hoyos Rodríguez

EngD Student

University of York

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

[qvtd-dev] [Dev] Pivot model and generated opposite relations

hhr502@xxxxxxx (Horacio Hoyos Rodriguez) — Tue, 28 May 2013 16:15:43 GMT

Hi Ed,

Doing more work in the QVTimperativeTracingEvaluationVisitor class I have been getting an exception when using the opposite relations auto generated by the Pivot model.

In line 390 of UML2RDBMS.qvti I use p2n := t.PrimitiveToName; to get a reference to the trace model class used to trace the PrimitiveDataType t. While executing I get:

org.eclipse.ocl.examples.domain.values.impl.InvalidValueException: Failed to evaluate simpleuml::PrimitiveDataType.PrimitiveToName

at org.eclipse.ocl.examples.pivot.evaluation.EvaluationVisitorImpl.visitPropertyCallExp(EvaluationVisitorImpl.java:786)

at org.eclipse.ocl.examples.pivot.evaluation.AbstractEvaluationVisitorDecorator.visitPropertyCallExp(AbstractEvaluationVisitorDecorator.java:255)

at org.eclipse.ocl.examples.pivot.internal.impl.PropertyCallExpImpl.accept(PropertyCallExpImpl.java:391)

at org.eclipse.ocl.examples.pivot.evaluation.EvaluationVisitorImpl.safeVisit(EvaluationVisitorImpl.java:171)

at uk.ac.york.qvtd.pivot.qvtimperative.evaluation.QVTimperativeTracingEvaluationVisitor.visitVariableAssignment(QVTimperativeTracingEvaluationVisitor.java:180)

at org.eclipse.qvtd.pivot.qvtcorebase.impl.VariableAssignmentImpl.accept(VariableAssignmentImpl.java:173)

at uk.ac.york.qvtd.pivot.qvtimperative.evaluation.QVTimperativeLMEvaluationVisitor.visitBottomPattern(QVTimperativeLMEvaluationVisitor.java:103)

at uk.ac.york.qvtd.pivot.qvtimperative.evaluation.QVTimperativeTracingEvaluationVisitor.visitBottomPattern(QVTimperativeTracingEvaluationVisitor.java:63)

…

I added print statements to check that the t variable is not null. Since the QVTi editor does not complaint about the reference I am guessing something is not working in my code, probably the PrimitiveToName element is not being properly assigned the primitive association value.

My question is, can you recommend a smart way of printing the Variable values to get useful information? Perhaps a printer that prints the variable name and the value of all its attributes?

Regards,

Horacio Hoyos Rodríguez

EngD Student

University of York

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

[qvtd-dev] [Dev] OCL Sets and java Util Lists

hhr502@xxxxxxx (Horacio Hoyos Rodriguez) — Thu, 23 May 2013 16:42:29 GMT

Hi Ed,

I have been running the UMLtoRDBMSSchedule.qvti transformation to find bugs in the transformation and in the engine. I am getting the following exception

java.lang.ClassCastException: org.eclipse.ocl.examples.domain.values.impl.SetValueImpl cannot be cast to java.util.List

when executing line 425 of the transformation:

atc.leafs := Set(AttributeToColumn) {atc};

In the QVTimperativeAbstractEvaluationVisitor class, the exception is thrown by the initValue method call (line 186).

I am guessing "Set(AttributeToColumn) {atc}" in the transformation returns a SetValueImpl when visited, and then the initValue tries to assign this value to the leafs attribute which is a java.util.List.

Is there some cast or helper method invocation missing in the initValue method, or does the Visitor must handle this?

(I have attached the complete trace).

Regards,

Horacio Hoyos Rodríguez
EngD Student
University of York

Attachment: qvti.trace
Description: Binary data