| [QVTo] Exceptions and warnings [message #510728] |
Thu, 28 January 2010 13:24  |
 PBarendrecht
Messages: 36
Registered: November 2009
Location: Eindhoven, Netherlands |
Member |
|
|
Hi all,
To prevent incomplete transformations, I'm using xmap and assert at several places in my transformation. However, they don't work as I expected them to:
- According to the MOF QVT specs, xmap should return an exception if the when-clause doesn't hold. I don't see any exception, even though my "example"-when is pretty rigid: when {1 = 2} { ... } --should result in an exception.
Do I have to configure something to be able to see these exceptions?
- I couldn't find this in the specs, but does xmap also throw an exception if there is no right disjunct mapping? The specs state that in this case the value "null" is returned, this also happens when a when-clause doesn't hold.
If not, is there some kind of "default" for the disjuncts option?
- When using assert, one can choose "warning", "error" or "fatal". The latter one terminates the entire transformation while the first two just display a message and continue the transformation. It would be nice if the "error" assertion would terminate the current mapping. At the moment I don't see a difference between the "warning" and "error".
---
As promised I'm working on a QVTo tutorial, but my own QVTo project has of course a higher priority. I think it will take a few weeks until my basic tutorial is finished, when people find it useful I'll expand the tutorial.
[Updated on: Thu, 28 January 2010 13:25] Report message to a moderator |
|
|
| Re: [QVTo] Exceptions and warnings [message #510749 is a reply to message #510728] |
Thu, 28 January 2010 14:04  |
Eclipse User |
|
|
|
Originally posted by: dvorak.radek.gmail.com
Hi,
> - According to the MOF QVT specs, xmap should return an exception if the
> when-clause doesn't hold. I don't see any exception, even though my when
> is pretty rigid: when {1 = 2} { ... } --should result in an exception.
> Do I have to configure something to be able to see these exceptions? - I
> couldn't find this in the specs, but does xmap also throw an exception
> if there is no right disjunct mapping? The specs state that in this case
> the value "null" is returned, this also happens when a when-clause
> doesn't hold. If not, is there some kind of "default" for the disjuncts
> option?
The spec does not explicitly mentions this, however I assume it would be
logical
that xmap on disjuncts works the same as on normal mappings.
Could you please raise bugzilla on that?
https://bugs.eclipse.org/bugs/enter_bug.cgi?bug_status=NEW&a mp;component=QVT_OML-Engine&form_name=enter_bug&prod uct=M2M&version=2.0
>
> - When using assert, one can choose "warning", "error" or "fatal". The
> latter one terminates the entire transformation while the first two just
> display a message and continue the transformation. It would be nice if
> the "error" assertion would terminate the current mapping. At the moment
> I don't see a difference between the "warning" and "error".
>
Only assert fatal terminates execution, the other variants produces just a
log
record. I agree it is not very useful, perhaps a debugger might make use
of it
and suspend there. I would just use fatal everywhere.
The worse thing is that it can't be disabled, so it always has to execute.
Not sure why the OMG spec tried to reinvent the wheels here, if Java
assertion
semantics works fine.
> ---
> As promised I'm working on a QVTo tutorial, but my own QVTo project has
> of course a higher priority. I think it will take a few weeks until my
> basic tutorial is finished, when people find it useful I'll expand the
> tutorial.
Thanks!
Regards,
/Radek
|
|
|
Powered by
FUDForum. Page generated in 0.02930 seconds
] 
Search
Help
Register
Login
Home
