[Geoserver-devel] KMPlacemark: Removing TextSymbolizer vs Disabling LabelCache

A little bit of background:
http://geoserver.org/display/GEOS/KMZ+KMSCORE+feature+addition
http://jira.codehaus.org/browse/GEOS-2030

Crucible review:
http://crucible.opengeo.org/cru/geoserver-10/review

The idea is that I'm preventing the rendering of text labels in the raster image, because that text is served in a more useful way with the Placemarks. On suggestion from Jody, it was initially implemented using a 'dummy' LabelCache that did nothing with labels, so labels are ignored when rendering. Andrea has suggested that removing TextSymbolizers is better, as then the renderer doesn't have to process the labels at all. Both approaches are simple implementations and a quick, incomplete test didn't show any noticeable difference.

Any thoughts?

Thanks,
Wayne Fang

--
Refractions Research
Suite 300 - 1207 Douglas St
Victoria, BC, V8W 2E7, Canada
ph: (250) 383-3022
fax:(250) 383-2140

Wayne Fang ha scritto:

A little bit of background:
http://geoserver.org/display/GEOS/KMZ+KMSCORE+feature+addition
http://jira.codehaus.org/browse/GEOS-2030

Crucible review:
http://crucible.opengeo.org/cru/geoserver-10/review

The idea is that I'm preventing the rendering of text labels in the raster image, because that text is served in a more useful way with the Placemarks. On suggestion from Jody, it was initially implemented using a 'dummy' LabelCache that did nothing with labels, so labels are ignored when rendering. Andrea has suggested that removing TextSymbolizers is better, as then the renderer doesn't have to process the labels at all. Both approaches are simple implementations and a quick, incomplete test didn't show any noticeable difference.

This comment of mine is more of a optimization idea, and one of
cleanness, it seems more elegant to just avoid doing stuff than pulling
levers and pushing buttons against an fake machine that will do
nothing, but I'm not very much concerned about it,
one way or the other is ok.

Cheers
Andrea