Eclipse Community Forums: EMF » Calling eGet for invalid structural feature

Home » Modeling » EMF » Calling eGet for invalid structural feature

Calling eGet for invalid structural feature [message #1421404]

Thu, 11 September 2014 06:18

Eclipse User

I want to eGet a given feature on a collection of EObjects, but not all
of the EObjects are instances of the EClass that defines the feature.

So which is preferable?

// This is the shortest, but is this safe? I guess if the feature isn't
defined on the objects (super)type(s) it will just return null..
for (EObject e : someEObjects){
Object o = e.eGet(feature)
}

// check if all structural features include the interesting feature
for (EObject e : someEObjects){
if (e.eClass().getEAllStructuralFeatures().contains(feature)){
Object o = e.eGet(feature)
}
}

// check if the object is an instance of the type that defines the feature
for (EObject e : someEObjects){
if (feature.eContainingClass().isInstance(e)){
e.eGet(feature)
}
}

Re: Calling eGet for invalid structural feature [message #1421505 is a reply to message #1421404]

Thu, 11 September 2014 09:03

Eclipse User

Hi, Felix,

I express my own opinions, below. Probably not the canonical answer :-)

Christian

On 2014-09-11 10:18:03 +0000, Felix Dorner said:

> I want to eGet a given feature on a collection of EObjects, but not all
> of the EObjects are instances of the EClass that defines the feature.
>
> So which is preferable?
>
> // This is the shortest, but is this safe? I guess if the feature isn't
> defined on the objects (super)type(s) it will just return null..
> for (EObject e : someEObjects){
> Object o = e.eGet(feature)
> }

This throws some kind of RuntimeException.

> // check if all structural features include the interesting feature
> for (EObject e : someEObjects){
> if (e.eClass().getEAllStructuralFeatures().contains(feature)){
> Object o = e.eGet(feature)
> }
> }

That works.

> // check if the object is an instance of the type that defines the feature
> for (EObject e : someEObjects){
> if (feature.eContainingClass().isInstance(e)){
> e.eGet(feature)
> }
> }

This feels best IMO because one really ought to be able to rely on the
EMF Run-time actually making the feature available if the instance
conforms to the feature's defining class.

Re: Calling eGet for invalid structural feature [message #1421543 is a reply to message #1421505]

Thu, 11 September 2014 10:03

Eclipse User

On 11/09/2014 15:03, Christian W. Damus wrote:

>> // This is the shortest, but is this safe? I guess if the feature
>> isn't defined on the objects (super)type(s) it will just return null..
>> for (EObject e : someEObjects){
>> Object o = e.eGet(feature)
>> }
>
> This throws some kind of RuntimeException.

Thanks, I was being too lazy to check or look. Still, it's the sexiest
approach - until you run it :)

>
>> // check if the object is an instance of the type that defines the
>> feature
>> for (EObject e : someEObjects){
>> if (feature.eContainingClass().isInstance(e)){
>> e.eGet(feature)
>> }
>> }
>
> This feels best IMO because one really ought to be able to rely on the
> EMF Run-time actually making the feature available if the instance
> conforms to the feature's defining class.
>
I also think this one is better then.

Re: Calling eGet for invalid structural feature [message #1423837 is a reply to message #1421543]

Mon, 15 September 2014 00:11

Eclipse User

Comments below.

On 11/09/2014 4:03 PM, Felix Dorner wrote:
> On 11/09/2014 15:03, Christian W. Damus wrote:
>
>>> // This is the shortest, but is this safe? I guess if the feature
>>> isn't defined on the objects (super)type(s) it will just return null..
>>> for (EObject e : someEObjects){
>>> Object o = e.eGet(feature)
>>> }
>>
>> This throws some kind of RuntimeException.
>
> Thanks, I was being too lazy to check or look. Still, it's the sexiest
> approach - until you run it :)
>
>>
>>> // check if the object is an instance of the type that defines the
>>> feature
>>> for (EObject e : someEObjects){
>>> if (feature.eContainingClass().isInstance(e)){
>>> e.eGet(feature)
>>> }
>>> }
>>
>> This feels best IMO because one really ought to be able to rely on the
>> EMF Run-time actually making the feature available if the instance
>> conforms to the feature's defining class.
>>
> I also think this one is better then.

Yes, or alternatively checking that
e.eClass().getEAllStructuralFeatures().contains(feature) or for even
better performance, using eObject.eGet(e.eClass().getFeatureID(feature))
*but* checking that the feature ID >= 0.

Re: Calling eGet for invalid structural feature [message #1424103 is a reply to message #1423837]

Mon, 15 September 2014 09:09

Eclipse User

On 2014-09-15 04:11:54 +0000, Ed Merks said:

>
> Yes, or alternatively checking that
> e.eClass().getEAllStructuralFeatures().contains(feature) or for even
> better performance, using
> eObject.eGet(e.eClass().getFeatureID(feature)) *but* checking that the
> feature ID >= 0.

Whoa, that one rather blows my mind.

:-)

Previous Topic:	CDORemoteSessionMessage for accessing a CDOModel
Next Topic:	Problem with comparing models

Goto Forum:

-=] Back to Top [=-

Current Time: Fri Apr 18 16:20:38 EDT 2025

.:: Contact :: Home ::.

Breadcrumbs

Sign up to our Newsletter