This function is used to set the orientation of a stream to either
byte-oriented or wchar-oriented.

More info on this function is here:
https://man7.org/linux/man-pages/man3/fwide.3p.html



This implementation only impmlemnts the manual switching and does
not yet guard against using a byte-oriented stream with wchar
functions and vice versa. Those step will come in additional
commits.

Signed-off-by: Wren Turkal <wt@penguintechs.org>

I don't really know for sure what all these silly rules are, but I think
I got it now...

Rust's TcpListener fails because of the address format being wrong. The format
comes from `accept`, but there we internally use this function.

This patch makes use of the data structures and functions impelemented
in the last patch to enable RTLD debugging protocol as per SVR4

These files needs formating by the auto formatter and It keeps popping
up every time I format my own code.

As with the previous commit, accept() was calling inner_get_name() and
assuming only "tcp:" or "udp:" addresses would be received. Thus, in
order to support AF_UNIX sockets, inner_get_name() was split into two,
inner_af_inet() and inner_af_unix() - where the former keeps the
previous logic, dealing with "tcp:" and "udp:" addresses, and the latter
deals now with "chan:" addresses and filling in the sockaddr_un
appropriately.

Previously, domain AF_INET was assumed while processing bind() /
connect(), which end up calling bind_or_connect!. Instead, match on the
domain type and process the path for AF_UNIX domains.

To add support for UNIX sockets (AF_UNIX), of SOCK_STREAM type, the
"chan:" scheme is used, which will be supportedby the ipcd running in
userspace.

Later commits add similar AF_UNIX support for the rest of the methods in
impl PalSocket.

The sigaction handler called map on an option, creating a pointer to a
move value. This in turned caused UB for signal handlers. Avoid using
pointers directly, and instead prefer references.

I didn't think it'd be this useful first, but thank god for `cargo fix --edition`!

Not sure if I should add a RwLock for the ptrace state too...

It's happening...