Rob Ede

Having some threshold on the spawn might be a good idea. I can see this current patch being quite risky for high concurrency situations.

> We might need to do a major version bump For 1.60 -> 1.72? Surely 1.72 (Aug 2023) is old enough to not warrant that. > how aggressively people are...

It's slightly surprising that /formats hopped on 2024 so quickly but I assume it had a good reason and was basically waiting for it. Still, lockfiles exist to solve this...

I expect (and indeed have experienced) the same vocal minority objecting to (cascading) major library releases. The solutions are much simpler for the MSRV case. I'll leave it at that.

I think there'd be less misuse of this type if it was called `http::RequestTarget` which is more descriptive of it's actual purpose (using the [specs' BNF language](https://datatracker.ietf.org/doc/html/rfc9112#name-request-line), see ). In...

Appreciate the look into it. For sure this is non-trivial to solve and I understand that lib consumers aren't your primary use case. Really the problem lies in serde_json and...

`cargo-make` looks incredibly verbose tbh, both in terms of task definitions and in its output.

To address the original point: As `SessionKey` is a wrapper over a string, not bytes, makes the generated session key more accessible for the storage backends. The thing we give...

Providing customisable session key generation is probably quite easy to work into this library, I'll consider this a feature request issue for that.

OWASP only says: > Session identifiers must have at least 64 bits of entropy It's a load-bearing "at least", which is satisfied by our current generator. And when they change...