You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The http crate is pretty much the default to use in every HTTP stack and its Request/Response types have generic body types that could be something implementing embedded_io::asynch::{Read, Write} for example. It's not no_std friendly yet but that could change(hyperium/http#563) although the PR still relies on alloc which I'm not sure if it's acceptable for this client.
The text was updated successfully, but these errors were encountered:
We want to avoid alloc for sure, do you think it's feasible to change http to support that? In any case, I think #2 (using httparse) will at least remove some of the duplication we have in reqwless.
It doesn't look trivial to remove the dependency on alloc from http, it relies a lot on Bytes and has custom types like the HeaderMap which does allocation, it would likely be a mayor refactor that probably will introduce breaking changes.
The
http
crate is pretty much the default to use in every HTTP stack and its Request/Response types have generic body types that could be something implementingembedded_io::asynch::{Read, Write}
for example. It's notno_std
friendly yet but that could change(hyperium/http#563) although the PR still relies onalloc
which I'm not sure if it's acceptable for this client.The text was updated successfully, but these errors were encountered: