Home » Modeling » EMF » [EMF] ResourceSetImpl.getResource does not behave in the same way calling it twice
| [EMF] ResourceSetImpl.getResource does not behave in the same way calling it twice [message #1691639] |
Wed, 08 April 2015 16:37  |
|
Hello Ed
When calling ResourceSetImpl.getResource, I am facing with what it looks like an incoherence.
Given an URI which does not correspond to anything (for example "platform:/resource/testproject/NonExisting.ecore"), i am calling ResourceSetImpl.getResource(uri, true); with loadOnDemand=true
The resource is not found but, as loadOnDemand=true, the resource is created but then when attempting to load it, a WrappedException is raised.
At second call to ResourceSetImpl.getResource(uri, true);, a resource is returned.
The reason is that at first call the created resource has been set on resourceSet, but the second call just return that resource.
Whatever how it works, in my opinion, both calls should have the same behavior either returning the same resource or raising an exception.
Proposition: Maybe, at first call, when internally catching the ResourceException, the resource could be removed from the resourceSet, so the second call could behave in the same way.
Looking forward to your answer.
Laurent Fasani - Obeo
Need training or professional services for Sirius?
http://www.obeodesigner.com/sirius
|
|
|
| Re: [EMF] ResourceSetImpl.getResource does not behave in the same way calling it twice [message #1691684 is a reply to message #1691639] |
Thu, 09 April 2015 04:27  |
 Ed Merks
Messages: 33140
Registered: July 2009 |
Senior Member |
|
|
Laurent,
Comments below.
On 08/04/2015 6:37 PM, Laurent Fasani wrote:
> Hello Ed
>
> When calling ResourceSetImpl.getResource, I am facing with what it
> looks like an incoherence.
>
> Given an URI which does not correspond to anything (for example
> "platform:/resource/testproject/NonExisting.ecore"), i am calling
> ResourceSetImpl.getResource(uri, true); with loadOnDemand=true
> The resource is not found but, as loadOnDemand=true, the resource is
> created but then when attempting to load it, a WrappedException is
> raised.
>
> At second call to ResourceSetImpl.getResource(uri, true);, a resource
> is returned.
Yes, that's expected.
>
> The reason is that at first call the created resource has been set on
> resourceSet, but the second call just return that resource.
>
> Whatever how it works, in my opinion, both calls should have the same
> behavior either returning the same resource or raising an exception.
This is long standing behavior and behaves as intended. A resource,
once loaded, will not be loaded again until unloaded.
>
> Proposition: Maybe, at first call, when internally catching the
> ResourceException, the resource could be removed from the resourceSet,
> so the second call could behave in the same way.
Absolutely not, a load failure is generally expensive, and it's often
the case that an exception it thrown, but the resource is mostly
loaded. I.e., if some attribute that's supposed to be a number contains
something that's not a number, an exception will be thrown, but the
entire resource is loaded and only that one attribute's value is not
populated.
>
> Looking forward to your answer.
Note that failures to load are recorded in Resource.getErrors() so you
can examine those to determine which resources have had major or minor
problems when they were loaded.
Ed Merks
Professional Support: https://www.macromodeling.com/
|
|
| |
Goto Forum:
Current Time: Thu Apr 25 12:02:11 GMT 2024
Powered by FUDForum. Page generated in 0.03130 seconds |
] 
Search
Help
Register
Login
Home
