Re: ocl conformance test



Date view Thread view Subject view Author view Attachment view

From: Stefan Haustein (haustein@kimo.cs.uni-dortmund.de)
Date: Wed 25 Feb 2004 - 16:02:36 GMT


Hi,

thanks a lot for the scripts, they helped to identify several issues in our 
implementation (unfortunately, some are still unfixed...)

Some remarks:

- The scripts use the infix syntax for div and mod, but both are
   not listed as operators in the 2.0 and 1.5 version of the spec. So
   I changed "4 div 5" and "4 mod 5" to "4.div(5)" and "4.mod(5)"

- The scripts seem to assume that a division by 0 results in
   OclUndefined. I did not find explicit support for that in the
   specification, so perhaps it is better to use a different
   Expression to create OclUndefined....

- I have added results for most expressions to an XML file,
   but I am not yet sure how to evaluate some expressions correctly.
   For instace, can OclUndefined be contained in Sequences or Sets?

- Some tests still need modifications to be able to add a result value,
   for example the transformation of a Set{1,2}->asSequence() may yield
   Sequence{2,1} or Sequence{1,2}

The XML file I have created from the scripts is available at

http://kobjects.org/ocltest/

Best Regards,
Stefan



D.H.Akehurst wrote:
> 
> We realise that they are not a complete set of sets,
> but we use the statements in the enclosed files to test
> our OCL tool against (http://www.cs.kent.ac.uk/projects/ocl/)
> 
> They are based on some statements I found on the web somewhere used to test
> some other OCL tool. I forget which one, thanks to the original authors
> though.
> 
> 
> 
> 
> -----Original Message-----
> From: puml-list-request@cs.york.ac.uk
> [mailto:puml-list-request@cs.york.ac.uk] On Behalf Of Stefan Haustein
> Sent: 11 February 2004 00:13
> To: puml-list@cs.york.ac.uk
> Subject: ocl conformance test
> 
> 
> Hi,
> 
> does anybody know if there is an OCL conformance test? Something like a 
> simple class and object diagram and a set of queries and their expected 
> results? If not, wouldn't it make sense to build one? This would 
> probably help to find implementation bugs and encourage the addition of 
> missing functionality.
> 
> Best regards
> Stefan Haustein
> 
> 
> 
> 
> 
> 
> To remove yourself from this list please mail
> puml-list-request@cs.york.ac.uk with a message containing the word
> "unsubscribe".
> 

Date view Thread view Subject view Author view Attachment view