This is exactly the kind of work we should all be doing together! We'll have a look right away at this one. Also, there are going to be a number of texture pipeline changes checked in by Bao in the next few days that will be important to examine.
Also, if anyone wants to work on automated/instrumented testing of whether we are optimally downloading and displaying textures, go for it. It would be so great, for example, to have a repeatable test where we login to a known sim/content and measure the timing and performance of the texture system. P Robin Cornelius wrote: > As discussed at Rob's hour yesterday, i've detected an issue with the > texture caching that is preventing saving very small textures in to > the cache. > > I've opened a JIRA and attached a patch to > http://jira.secondlife.com/browse/VWR-12686. > > I'm well aware that the first 600 bytes are stored in an index file. > but the code path i have identified will not make any attempt to save > any data if the data is smaller than TEXTURE_CACHE_ENTRY_SIZE. It will > just delete the WriteResponder and return. After the llTextureFetch > has finished running and decoded it moves to state cache. this invokes > LLtextureCache::WriteToCache() which is the only path of saving to the > disk cache. > > I would quite like to commit this to http-texture as it seems very > relevant to that branch currently as textures are the focus. So Merov, > i think this is your area currently? > > Regards > > Robin > _______________________________________________ > Policies and (un)subscribe information available here: > http://wiki.secondlife.com/wiki/SLDev > Please read the policies before posting to keep unmoderated posting privileges > _______________________________________________ Policies and (un)subscribe information available here: http://wiki.secondlife.com/wiki/SLDev Please read the policies before posting to keep unmoderated posting privileges
