Christopher Schmidt wrote: > On Tue, Jul 15, 2008 at 08:27:16AM -0400, Paul Spencer wrote: >> Steve, >> >> I think you would need a specific layer type that supports S3. If you >> are using kaMap to generate the tiles (and the path to the tiles) then >> you would probably want to start with the kaMap layer and modify or >> subclass it to make the appropriate request directly from S3. >> Conceptually, this is fairly trivial to do, you would just overload >> the getURL function with something that generates a url to S3. > > In general, the problem here is just that the ka-Map Layer doesn't > actually support the ka-Map disk-style caches (if it did, you could just > cause it to read from the URL). > >> TileCache supports S3 as a cache, but I'm not sure if the TileCache >> layer in OpenLayers can directly read them. > > No; the TileCache S3 support needs to be reworked. It was sort of a > proof of concept that I didn't carry far enough. > > Regards,
Thanks. I saw Chris' blogs on the TileCache stuff, but nothing on the OpenLayers side. For ka-map caches I use mod_perl to rewrite the tile.php request directly to a file and avoid tile.php completely, which is very fast if the cache is pre-populated - but that will not work for S3 unless I use my server to proxy all the requests to S3 which I do not think is desirable. I will look into Paul's idea of sub-classing the kaMap layer and reworking the getURL() to do what I need. You'll get a patch if I can get something to work and probably some questions along the way :) I appreciate the suggestions. -Steve _______________________________________________ Users mailing list [email protected] http://openlayers.org/mailman/listinfo/users
