Threaded HTML Parser should send preloads directly to the network thread
The catch is that we'd need to check the MemoryCache before sending out the preload. Which would likely require a lock to allow access from the networking thread. I would expect little contention for such a lock.
Others who are more familiar with the loader could speculate better I'm sure.
Should we have a meta bug for these sorts threaded parser improvements?
(In reply to comment #1)
> Should we have a meta bug for these sorts threaded parser improvements?
Created bug 111645.