HTTP/2: Consider Resurrecting Original Functionality

[dominic-p]

With HTTP/2 gaining popularity some of the optimization techniques developed for HTTP/1 are becoming counter-productive. Specifically, because connections are multiplexed, HTTP requests are cheap in HTTP/2. So, combining unlike files can actually hurt performance by minimizing the effectiveness of caching.

It strikes me that the original purpose and design of this script fits quite well in an HTTP/2 environment. But, I'm curious to hear what the developers think.

Here's a whitepaper on performance in HTTP/2 that suggests unwinding some HTTP/1 based optimizations going forward.