[Geoserver-devel] Re: R: [Geotools-devel] (no subject)

Paolo,

Go ahead and mark those 2 bugs as 'resolved' - we'll close them later
when we've done more testing.

If you make a bug report, make sure that you set a "fixed version" -
this will make sure that it will get tracked and fixed. You can always
set it to the next release - I'll move it forward if I dont have time
to fix it.

If you go to:
http://jira.codehaus.org/browse/GEOS?report=com.atlassian.jira.plugin.system.project:roadmap-panel

you should see your bug listed as "pending" for the next release.

When I do a release, I make sure all the bugs are resolved or i move
them to the next release. Either way, your bug gets tracked way better
here than anywhere else. This goes for everyone whos submitting bugs!

You can edit any old bugs and set their fixed version.

dave

P.Rizzi Ag.Mobilità Ambiente wrote:

Paolo;

I've fixed:
(GEOT-480) PostgisDataStore error for table with no Geometry.
(GEOT-378) PostgisDataStore doesn't always correctly handle
mixed-case
table and field names

Could you verify that they're actually fixed? I did some small tests,
but I'd appreciate it if you could check and close the bug.

Great!!! I checked both fixes with my data and they work.

I would like to test them more extensively, but I can't right now,
because we're transiting to GeoTools 2.1 and I don't have a stable
environment.
Do you prefer I close the bugs anyway and, eventually, reopen them
if and when I'll found other errors, or I better leave them open
until after a very deep test???

I changed the <boundedby> tag to emit <gml:null>unknown</gml:null> if
the bounding box is null (ie. no features returned or features w/o
geometry).

This I haven't tested because of the transition said above.
ALso I'm not a GML expert either, so I can't tell what is actually

better to

emit.
I only can think that maybe the two cases should be distinguished.
When there's no Geometry column at all, a <gml:null>unknown</gml:null>

could

be OK.
When a Geometry column exists, but is empty, maybe it would be better
to emit something different, for example the CRS info could be emitted
anyway???

I also found (and fixed) a few other places where columns/table names
were not being double-quoted.

Yes, there were a few...

dave

Thank you!!!

Bye
Paolo Rizzi

P.S: While we're at it, I noticed other faults with the

PostgisDataStore,

now I'll check in the JIRA if they are already there and eventually

post

them.
For example the isNillable/min/max info seems not to be transferred

from the

DB table schema to the AttributeTypes.

----------------------------------------------------------
This mail sent through IMP: https://webmail.limegroup.com/