I think it would be a good idea to make it optional and still throw a
null value when this particular option is selected "off", like I first
suggested. Question is only, at which point in time and in which manner
this option is selected. Is it something that should be left completely
to the users? For example in the case of app-schema, the problem was
that non-existing columns in the mapping file can cause unclear
exceptions. In this particular case it would be advised always to check
on the existence of columns, while this might be different for other
uses of expression evaluation.
Remark that the patch also fixes something else that could cause bugs in
the future.
The way the XPath Property Accessor worked, it could easily prevent
other property accessors from being used (when they are stored later in
the queue - since xpath property accessor always returns true to
canHandle), but still fail. (Refer to the earlier discussion with Jody.)
That's why Jody suggested to actually try multiple property accessors. I
think that is an important change that definitely should be kept.
With Regards,
Niels
On 21/09/10 09:18, Justin Deoliveira wrote:
Hi all,
Recently this patch has been proposed:
http://jira.codehaus.org/browse/GEOT-3066
And this issue was discussed in detail on the mailing list some time
ago in this thread:
http://osgeo-org.1803224.n2.nabble.com/evaluating-invalid-attribute-names-td5489979.html
Apologies for not weighing in then. But I think this has some pretty
serious backward compatibility issues. The issue/proposal is to throw
an exception vs return null when a type has no such attribute
described by an attribute name or xpath, rather than return null. The
thread above proposes to just change the behaviour and update any test
cases that rely on it. I could not disagree more with his approach.
This breaks backward compatibility and i think requires more
discussion. There is much more use of geotools beyond test cases and
even beyond geoserver and udig. If there is one thing i learned at
foss4g it is that there are lots of people using geotools, even though
they don't take parts in discussions on the developer list.
So i think this change requires more discussion and probably a
proposal. And it certainly can't take place on stable branch imo.
Is there any way we can have this behaviour engage only when the user
explicitly requests it? That has been a pattern that has worked very
well in the past for such changes.
-Justin
--
Justin Deoliveira
OpenGeo - http://opengeo.org
Enterprise support for open source geospatial.
--
*Niels Charlier*
Software Engineer
CSIRO Earth Science and Resource Engineering
Phone: +61 8 6436 8914
Australian Resources Research Centre
26 Dick Perry Avenue, Kensington WA 6151
------------------------------------------------------------------------------
Start uncovering the many advantages of virtual appliances
and start using them to simplify application deployment and
accelerate your shift to cloud computing.
http://p.sf.net/sfu/novell-sfdev2dev
_______________________________________________
Geotools-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/geotools-devel
