Psilon

So, the problem in a nutshell: 1. Current implementation limit requests _on client side_ with `u16::MAX`, although some servers are fine with such requests 2. All rust requests libraries (such...

> Many many servers would not allow a 10MB request-target normally (it's preferable to send that data in the body, and use Expect: 100-continue to prevent waste bandwidth). I agree...

> The size limit was placed to reserve the ability to shrink the Uri struct, which is quite large. Are we talking about 2 extra bytes per Uri instance (which...

In this case I think u16 is fine for storing anything but query string since in most real cases it's the main source of big URIs. So it's possible to...

So, any progress on that?

Hello. Currently [we patched this for ourselves](https://github.com/seravkin/http/commit/adce775e798ab18064dc4adb174b278dba0852c6), but we'd like to have some feature gate merged in upstream to remove need in patching deps of our deps. It's a critical...

Ping?.. Could we make it a feature gate like `len-u32` which is disabled by default? We'd really like to use the upstream instead of our own fork. But we just...

Guys, are you serious? It doesn't matter if VS starts in 10 more seconds or takes 0.5 sec more time for loading some file until it's way way more productive,...

I actually tried to implement superpi using this crate. However, it wasn't a big success: https://github.com/Pzixel/superpi-rs It gives 3.1410... for 10 iterations (should give 10^2 correct digits). It's not directly...

> That's exactly the ⌊√a⌋/⌊√b⌋ which I said would be lossy. Yep, I know, but I don't know any better approach. Error is too high to make this algorithm work...