Re: [Geoserver-users] SLD

[Andrea A. <aa...@op...>]

Ross Elliott ha scritto:
> That's exactly how UMN Mapserver works, you change the size of the
> buffered pixel for each query layer, not ideal but it does work
> providing the buffer size takes account of the biggest symbol size in
> the layer and the placement of it. Ideally (for me at least) the WMS
> spec would allow a BBOX for a query area rather than a single point
> (like UMN Mapserver's cgi-mode query) this wouldn't solve the problem
> where a symbol is placed far away from the actual point but I suspect
> that's a rare occurance anyway.

This reminds me a lot about the meta-buffer issue with tiled rendering.
If the SLD is relatively static, we can spot the right buffer size by
inspecting it. If not, we could have a new parameter in the request
that allows the user to force a specific search area.

Ross, can you open a jira issue about this? Maybe we won't fix this
overnight, at least a jira issue will make us avoid forgetting about it.

Cheers
Andrea