autocannon icon indicating copy to clipboard operation
autocannon copied to clipboard

benchmark chunked encoding

Open yoshuawuyts opened this issue 4 years ago • 6 comments

A lot of traffic uses chunked encoding; this is common for streams that don't have a known size up front. This often takes a different route in HTTP parsers and encoders that isn't covered when benchmarking requests with a known length. It'd be fantastic if autocannon was able to issue repeated chunked requests.

Motivation for this is @dignifiedquire wanted to bench our chunked impl in https://github.com/http-rs/async-h1, and I didn't see an option for it in autocannon yet. Thanks!

yoshuawuyts avatar Apr 10 '20 12:04 yoshuawuyts

That would be pretty interesting to do. A PR would be mostly welcomed.

mcollina avatar Apr 10 '20 13:04 mcollina

I am currently catching up on my DB course. @mcollina Is it possible you guide me through? I would like to try a stab at this after the Easter break. That is from Tuesday 14th. Till then I am going to read the source code. Hopefully, I can post my questions here.

evanxg852000 avatar Apr 11 '20 14:04 evanxg852000

Sure, give a shout whenever you are ready!

mcollina avatar Apr 11 '20 15:04 mcollina

I would also like to put in some effort too. Any way I can help?

nimit95 avatar Apr 11 '20 20:04 nimit95

I think this is a task for a single developer. Let's see how far we get along and then regroup, ok?

This repo has some issues - if you want to work on one feel free to pick!

mcollina avatar Apr 12 '20 09:04 mcollina

@mcollina I have a WIP PR here to discuss around the API before going too far https://github.com/mcollina/autocannon/pull/258

evanxg852000 avatar Apr 14 '20 18:04 evanxg852000