a single byte into a special buffer.  Note that peek() and nread will never
return/indicate more than a single byte, so scripts which rely on these may
be inefficient with TLS (websocket stuff appears to use nread heavily).

used, that requires a socket lib too. util builds (that don't need the .ini
file functions), they #define NO_SOCKET_SUPPORT, so exclude the user/####.ini
support functions in that case too.

since it's not defined otherwise. Not sure if this ways intended to be
included in Windows builds or not.

NULL

Always store the prot and user with the client_t to avoid even worse
headaches.

This makes the client update a bit trickier, but it seems to be good now.

Windows

example client.

Will tie into sbbscon.c after DigitalManian approval.

Fixes recvline() returning an zero-length "line" on timeouts, which causes
infinite loops and 100% CPU utilization with IMAPv4-TLS service (and likely
any other service that uses recvline with a short timeout).

Added "Node # SSH" to a lot of SSH-related log messages.
Lowered the severity of the SSH "attempt to use channel" msg from ERROR
to WARNING.
Removed "last:" from temp-ban message when there is no last-used username.

when creating a new netmail message. Not exactly an "address book", but a
convenience nonetheless.

I'm calling these "user properties", but that might change.

literals (e.g. won't detect and escape them automatically). If you set
a key value with carriage-returns and line-feeds, those will end up in the
.ini file as-is, so use with caution.

exists in a .QWK packet. Also, delete the other control/metadata files
case-insensitively (just in case).

the return value if recv() for non-TLS sockets.

TLS, use MSG_WAITALL for reading integers, add a time() based timeout,
return as soon as any bytes are read (including zero), and generally
behave more closely to how recv() itself behaves.

first one.

reason.

to be changed to something the individual clients can frob in the near
future.

bytes have been sent!

"Sometimes" the write timeout value is used for read timeouts.  Since we
use a read timeout of zero, and a mutex, this can cause some serious delays
in SSH processing.  As a workaround, we set the write timeout to zero.
However, a flush failure has historically been fatal.

So, we now change the write timeout to one second just before a flush, and
have a flush timeout not be fatal anymore... we set it back to zero after
the flush.

The fiddling with the write timouts will be removed once the bug is fixed,
but the flush timout not being fatal may stay forever... add some comments
to remind us to remove the hack in the future.

attempt.