Two nice improvements have been made in Tilecache:
- TileCache can now convert the image send to the browser from the original format (defined in the configuration file) to a format specified in the WMS request. This feature has been developed for IE6 because this old browser can't handle transparent png files. If you want to support this feature, configure your OpenLayers WMS layer to something like:
new OpenLayers.Layer.WMS("admin", "http://www.example.com", {
layers: ['A', 'B', 'C'],
format: OpenLayers.Util.alphaHack() ? 'image/gif' : 'image/png'
});
And don't set the alpha option to true.
Remember that this feature is time consuming if a format conversion has to be made
- If several layers need to be merged, OpenLayers can do that. But, this merge process is time consuming on the client side. Tilecache is now able to do that on the server side and let you build performant application.
3 comments:
"Tilecache is now able to do that on the server side and let you build performant application. "
You have a url that describes that ? I gues it's on SVN only right ?
Cheers
Hi,
Yes it's only on SVN trunk.
You can find the feature request here and the changeset here
Regards
feature request says : "The odification is quite simple: the layer's images are extracted from the cache and merged with the Python Imaging Library."
So I understand that instead of having N WMS layers with one tilecached layer each you have one WMS layer with N tilecached layers ? I suspect PIL is faster to combine the images than N calls to the WMS layer ? What is the typical use case behind the request ?
Cheers,
Yves Moisan
Post a Comment