Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[protege-discussion] problems with projects in server mode

Tania Tudorache tudorache at stanford.edu
Wed Jan 10 09:44:49 PST 2007


Hi Henk-Jan,

Thanks for the bug reports. We have fixed the 
ConcurrentModificationException which occurred in different situations 
involving instances with multiple types in client-server mode. This was 
a bug that did not affect the content of the ontology, but only the 
display of the values of a slot widget.

I have also added a check in the Types Panel that will prevent the user 
from adding an already existing type to an instance. However this is a 
check that is done by the GUI but it should be enforced at the API level.

The fixes are already committed in our SVN repository and will be 
available in the next beta build.

Tania


Henk-Jan Lebbink wrote:
> Hi Tania,
>
> On Thu Dec 14 14:21:20 PST 2006 Tania Tudorache wrote:
> (quoting from:
> https://mailman.stanford.edu/pipermail/protege-discussion/2006-December/000414.html)
>
>   
>> There is also another error (the concurrent modification error) that we
>>     
> are going to look at.
>
> I have pinpointed the concurrent modification error in the 365 build. It
> can be reproduced with the following steps.
>
> 1] Create three classes in thing with name: class1, class2 and class3. 2]
> Create one slot with name: slot1, value type: instance, allowed classes:
> class3, domain: class1.
> 3] Create one instance in class3.
> 4] Create one instance in class1; add in slot1 the instance from class3.
> 5] Add to the types of the instance in class1 the class2.
> 6] Somehow, you are not paying attention; add to the types of the instance
> in class1 the class1.
>
> WARNING: java.util.ConcurrentModificationException --
> AbstractSlotWidget.setWidgetValues()
>
> The warning is not that bad; however, the content of slot1 is not
> displayed. We can circumvent the bug by manually checking whether classes
> occur more than twice (and delete superfluous classes).
>
> Wouldn't it be reasonable that types should only allow one instance of a
> class?
>
> With regards,
> Henk-Jan.
>
>
>
> _______________________________________________
> protege-discussion mailing list
> protege-discussion at lists.stanford.edu
> https://mailman.stanford.edu/mailman/listinfo/protege-discussion
>
> Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03 
>
>   




More information about the protege-discussion mailing list