Well I have a patch to allow inner class names in OQL names, but I have
another issue that is more troublesome. So right now one can use an
inner class name in a mapping and in OQL, but one cannot use it in a
relationship. The reason is that the XSD for the mapping file defines
'type' as NMTOKEN ', but the class 'name' is ID. These are actually
incompatable since NMTOKEN doesn't allow '$'. While I can change the
XSD to ID, I'm not sure of the full ramifications. I *think* it is a
backward compatable change since it appears that ID containse NMTOKEN
sets. Any thoughts?
-Brian
Bruce Snyder wrote:
>This one time, at band camp, bikehead said:
>
>b>I can't use db.load() because I want to iterate over every persistent
>b>instance of a class and there doesn't seem to be any other way to do
>b>this. db.load requires an identifier which I don't have.
>
>I see.
>
>b>I'm working on a fix for this in org.exolab.castor.jdo.oql.Parser and
>b>ParserTreeWalker. It appears to be a simple case of getting the parser
>b>and expression builder to recognize '$' as a valid delimiter in a
>b>iterator specification.
>
>This is exactly why I wanted the line numbers in the stack trace. I know
>it's in the ParserTreeWalker somewhere. Eventually I'd like to rework
>the OQL implementation to make use of ANTLR. It's modeled after ANTLR,
>but when using ANTLR, a lot of the code can be generated during the
>build process. Makes for very easy changes to the implementation ;-)
>
>Let me know if you create a patch for this issue.
>
>BTW, thanks for you hard work.
>
>Bruce
>
>
--
__o
_-\<,_ Brian
(_)/ (_) [EMAIL PROTECTED] x503
-----------------------------------------------------------
If you wish to unsubscribe from this mailing, send mail to
[EMAIL PROTECTED] with a subject of:
unsubscribe castor-dev
