[OpenLayers-Dev] Some questions about WFS-T

[Christopher Schmidt]

On Mon, Apr 02, 2007 at 08:11:07PM +0200, Andrea Aime wrote:
> Hi,
> today I was trying Openlayers trunk, WFS-T and Geoserver; well, guess 
> what, I have some questions :-)
> * there is no example I can see that allows to edit the attributes of
>    a geometry, or edit an existing geometry. The WFS-T one seems to be
>    dealing with just adding geometry. Missing functionality, or missing
>    example?

Editing attributes of a feature (features combine attributes, geometry,
and style) doesn't have a UI, but is available programmatically:
attributes are available via the 'attributes' property of the feature,
and are (hopefully) properly saved.

Editing geometry is not available in this release.

> * in the wfs-t example, once I select a tool, I cannot deactivate it
>    anymore. I can commit transactions all right, but the editing tool
>    remains active.

This looks like a recently introduced bug -- I know it worked at one
point ;) I'll poke at this a bit. 

> Oh, that was with firefox. If I use IE7 instead, I cannot even commit
> the transaction, getting an error instead, whose rough translation to
> english is "line 50, column 9, property or method not supported by the 
> object, code 0, URL: http://localhost:8080/geoserver/ol/gs-test/wfs-t.html".
> Oh the wfs-t.html example is just the standard wfs-t page that's using
> locahost:8080/geoserver instead of the online one.

"Internet Explorer does not have a built in XML Serializer, so it can
not save WFS-T payloads to a server via the 'commit' method on a WFS
layer." -- http://trac.openlayers.org/wiki/Release/2.4/Notes

http://trac.openlayers.org/ticket/535 is tracking for this, though it
appears this is actually more complex, since IE doesn't have support for
a number of methods used by the WFS format writer.

> Another thing I noticed, this time in the wfs-states.html sample, is
> that while the states layer is loading, firefox is frozen, and 
> performing a CPU intensive task (I guess it's GML parsing).
> The same goes with IE7, with the aggravation that IE suggests killing
> the script that's blocking it (with a popup).

Yes. GML parsing is slow. 

Regards,
-- 
Christopher Schmidt
MetaCarta