Epsilon Forum

That seems fine according to
the book. In our case we use
[%=out.startPreserve("
;--",
"--",entity.identifi
er + " Additional
Descriptions",
true)%]

[%=out.stopPreserve()%]
;

But I now recall I have a
report of preserved regions
being overwritten as in our
case. Any one has any other
ideas?

Section 7.6 of the Epsilon
book talks about formatters
and how to use them from
within EGL! 

I'm trying to use protected
regions with ETL.  So, I put
the following code in the
template:

[%=out.preserve("/*&q
uot;, "*/",
name+"-Prop",
true," 
//TODO")%]

This generates the following
code:
  /* protected region
Local-Prop on begin */
    //TODO
  /* protected region
Local-Prop end */

which I think is correct.
Then I changed the previous
code to this:
  /* protected region
Local-Prop on begin */
    test();
  /* protected region
Local-Prop end */

I was hopping that after
generating the code again I
was going to keep the
'test();', but that content is
lost, and it is replaced with
'//TODO'.

Am I doing anything wrong?  Is
there anything else I need to
do, besides using the
"preserve" method?

I was starting to work on some
validations and from the error
message it seems either I am
doing something wrong, or
satisfiesAll is buggy . This
is the code, the metamodel is
Papyrus SysML.


context
SysML!sysml::portandflows::Flo
wPort
{
    // The base port must be
typed
    constraint IsTyped
    {
        check :
self.base_Port.type.isDefined(
)
		
	message : 
"FlowPort" +
self.base_Port.name + "
is not typed"
    }	
	
    constraint
IsTypedWithHiLeS
    {
        guard :
self.satisfies("IsTyped&q
uot;) and 
		self.satisfiesOne("IsBo
oleanOrLogic",
"IsTerminalOrNode")
		
	check : true
		
    }
	
    // Logic and Boolean are
not generalizations
    @lazy
    constraint
IsBooleanOrLogic
    {
	guard :
self.satisfies("IsTyped&q
uot;)
		
	check : self.type.name =
"Logic"
		or self.type.name =
"Boolean"
    }
	
    @lazy
    constraint
HasGeneralization
    {
		
	guard :
self.satisfies("IsTyped&q
uot;)
		
	check :
self.base_Port.type.generaliza
tion.isDefined()
		
	message : 
"FlowPort" +
self.base_Port.name + "
type is not a generalization
"
			+ "of another type
(All HiLeS library types are
generalizations)."
    }

    @lazy
    constraint
IsTerminalOrNode
    {  
	guard :
self.satisfies("HasGenera
lization")
		
	check :
self.base_Port.type.generaliza
tion.general.name =
"Terminal" or
		self.base_Port.type.generali
zation.general.name =
"Node"
		
	message : "Flow Port
" + self.base_Port.name +
" NOT" 
    }
}

I was looking for this post so
I thought I'd bump it. Is
there any plans on working on
a debuger in the Epsilon
Roadmap?

Regards, 

Sorry for finding all these
bugs... 

Bug 380692

I suppose fixing this problem
could be complex... Personally
I do not use the resolutions
(yet), so for me it is not a
show-stopper.

Hi Maarten,

Yep. I suppose yet another bug
fix is due  Could you please
file a bug report so that we
can keep track of this?

Cheers,
Dimitris

In the EvlValidator#validate()
the following line is
present:

EvlMarkerResolutionGenerator.I
NSTANCE.removeFixesFor(eObject
);

It removes all 'old' fixes for
the given object.

I was wondering what happens
if a single object is checked
by multiple EVL validators (eg
in a CompositeEValidator
situation): Are only the fixes
of the last validate()
preserved?

If so, I suppose this needs to
be fixed?

I created bug #380597.

If I may add, not only the
editors, but also the wizards
and all other UI/user friendly
code/plugins. As Maarten my
application runs epsilon
transformations in the
"back" and the users
don't really need to know that
it is epsilon what I am using.
So with a more fine separation
we could control more easily
what editors, views, menus,
wizards, etc., we want to
deploy.

Regards,