Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re:[higgins-dev]changesforhttps://bugs.eclipse.org/bugs/show_bug.cgi?id=197665

On second thought, I think I'll go ahead and check in the change for 197665 and address 190594 separately.


I'll send another mail with my questions about 190594


Jim

>>> "Jim Sermersheim" <jimse@xxxxxxxxxx> 01/04/08 12:36 PM >>>

I need to deal with the URI dataType param here as well as in the same IAttribute values (the latter is captured in https://bugs.eclipse.org/bugs/show_bug.cgi?id=190594)


There is one CP which needs this information, but I agree it's in the wrong place.  The XMLFile CP uses what I would call a "dynamic model".  It has no model to begin with, and as subjects, attributes, and their values are added, it builds a model on the fly.  So it does need to know the data type for attribute and metadata values until someone creates one.


So, if we remove these parameters, I wonder if we need to offer an optional way for the caller to specify the value datatype for a newly added attribute.


I'll look further into this and try to address bug #190594 at the same time.


Jim

>>> "Sergey Lyakhov" <slyakhov@xxxxxxxxxxxxxx> 01/04/08 12:05 PM >>>

Jim,

                 

We fixed the problem with jena2 CP nightly build and also do not see any problems with your changes for complex metadata. There is only question we have now: why do we need the "URI dataType" parameter in IMetadata.addValue(..), IMetadata.addSimpleValue(..) and addComplexValue(..). The type of metadata automatically defines the type of its value, and it looks this parameter is unnecessary here.

                 

Thanks,
Sergey Lyakhov

----- Original Message -----

Sent: Friday, January 04, 2008 7:23 PM

Subject: Re: [higgins-dev] changesforhttps://bugs.eclipse.org/bugs/show_bug.cgi?id=197665


Well, I'd rather get 1.0 destabilizing changes in sooner than later, and I just noticed that according to the nightly build, the jena CP is currently failing with 925 errors (and has been for some time) anyway.


So in absence of any other comments, I'm going to commit this change.


Jim

>>> "Jim Sermersheim" <jimse@xxxxxxxxxx> 01/02/08 1:18 PM >>>

<resend>


Attached are the changes for bugzilla #197665.  Basically, I took the cut/paste route.  Now metadata may be complex and multi-valued, just like attributes.  The only difference between metadata and attributes now is that metadata cannot have metadata on it.


With this change, there are 5 remaining errors as well as 7 method stubs in the jena2 context provider.  

I'll commit within 24 hours unless someone doesn't want me to.


Jim





_______________________________________________
higgins-dev mailing list
higgins-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/higgins-dev


Back to the top