I think socketaddr_in is also a bit different on Vita and we need to add sin_len and sin_vport fields under a cfg flag
NVM, I think it's related to 0.4.9 and hyper using it

No official support is fine for me, it's a tier 3 platform after all :)

Any way we can set up CI for this? Running cargo test or cargo check?

AFAIK there's currently no way to run tests right now. Check works using:

cargo +nightly check -Z build-std=std --target armv7-sony-vita-newlibeabihf

So we could add check to Github CI using nightly and -Z build-std.

No official support is fine for me, it's a tier 3 platform after all :)

Any way we can set up CI for this? Running cargo test or cargo check?

AFAIK there's currently no way to run tests right now. Check works using:

cargo +nightly check -Z build-std=std --target armv7-sony-vita-newlibeabihf

So we could add check to Github CI using nightly and -Z build-std.

That sounds good. You can copy what I removed in 14be1bc

Some tests are failing, because on vita there's no fcntl for non-blocking sockets, need to do it via setsockopt like in std:

setsockopt(fd, libc::SOL_SOCKET, libc::SO_NONBLOCK, nonblocking as libc::c_int)

Converted this PR to draft as we (vita-rust community) are still figuring some things out

At this point PR can be undrafted when rust-lang/libc#3284 is merged and released (we will need to bump libc version here for a green pipeline).

So, we're finally ready 😇

• Added check and docs for VIta to CI pipeline
• Nonblocking flag is set in via setsockopt and not fcntl on Vita (fcntl doesn't exist in Vita newlib, though at some point we probably will implement it least for that particular operation)
• There is no ipv6 stack on Vita, so all tests using that are disabled for the target
• All tests that are enabled for Vita are passing

Fixed all the comments, please check. With the API change I shouldn't have trusted clippy --fix, reverted and added clippy allow there.

Thanks @nikarh