socket_readable() returns TRUE if recv() will not block if called
on the specified socket.

socket_writable() returns TRUE if send() will not block if called
on the specified socket.

socket_recvdone() returns TRUE if all data has been recv()ed an
the socket is closed.

"will not block" includes "will return an error".

This matches the three main use-cases for select() of a single socket
in Synchronet.

Use them in the SyncTERM RLogin stuff so they can be tested easily.

MinGW32 is getting long in the tooth and is missing a lot of modern
Windows features as well as having broken headers.  Most people will
be using MinGW-w64 at this point, so add support for it.  Once I
ensure SyncTERM works properly with it, MinGW32 support will be
discontinued.

I suspect this will impact exactly zero people since the reason this
exists is to build the Win32 versions of SyncTERM on FreeBSD.

Changes:
- Explicitly request 32-bit Windows output
- Detect the string "mingw32" anywhere in the hardware description
- Explicitly link with libuuid
- Add a terrible hack to syncterm.c to block wincrypt.h

Do not clear at the current X/Y on scroll (derp)

Scrolling would clear the line to the end of the screen, not the
end of the scrolling region.

Sorry echicken.

Could be useful for knowing if you need to increase MaxClients for typical
usage.

First look for a linked node matching the node address string as passed,
then look for a match using the normalized address (removing .0 and @domain).
The "ALL" wildcard should not be used for password lookups.

This should fix reported issue 240.

Borland hates it.

There has been a sighting of Synchronet exceeding FD_SETSIZE sockets.
This means select() is of less use and we need to start migrating to
either poll() or platform-specific methods to achieve the same end.

This commit is mostly to check that poll() builds on Win32 and that
it actually works.

Hopefully fixes crash in lord when composing a message and using
word-wrap.

There's a comment saying we set key=0 when we consume it... do so
in a couple cases where it wasn't done.

Range check the key before sending to the remote.

This should not actually do anything different, but in the future
if things are changed at least the comments won't be misleading.

This is pretty insane since SDL2 completely ignores the keysym value
of the X11 keyboard event, so here's what we do...

1) When there's a keypress that includes right-alt, store the sym/mod
2) When we get text input, if it's the same as we would get if ALT
   wasn't pressed for the last keydown, parse through the mapping
   (ie: handle ALT keys)
3) If it's different, use that (gets AltGr modified value)

I get that AltGr is hard in a cross-platform way, but just pretending
it doesn't exist at all for key input is a very weird choice.

Caused a SyncTERM hang if it played a FG beep.

Thanks Mr. Executive Chicken esq. for the telephone call and bug report.

2 new configuration params: MaxLogSize (default: 10M) and MaxLogsKept (10)

error: ‘handle_mutex’ undeclared

This reverts commit 6a4fbf28.

This allows the playback thread to open the device and errors to
be checked.  Otherwise, the first sound won't play.

1) user.number immediately after login() remains zero, but changes
   "sometime" after.  Avoid using user.number right after login()
   call to generate filename.  All data/user/0000.imap files can
   be deleted after this change.
2) Saving Seen data would modify the object storing it when binary
   seen values were used.  Use JSON.parse(JSON.stringify() to make
   a copy before saving and restore it afterward.  Should fix issue
   reported by nickshanks1 over IRC.

Addresses issue #239

Also note the spectacularily poorly named WIN_REDRAW and WIN_NODRAW
flags.  WIN_REDRAW indicates that this *is* a redraw of the menu,
which indicates that you don't need to redraw the menu (because it's
already correct on screen).  Weird, but not insane... WIN_NODRAW on
the other hand indicates this is *not* a redraw, and that the window
is incorrect on screen and you therefore *must* redraw the menu.

That is to say:
WIN_REDRAW prevents the window from being redrawn
WIN_NODRAW forces the window to be redrawn

The previous attempt completely disabled pulse audio.

It seems that the "real" ALSA will actually silently fail if you
try to clear errors when no errors have occured.  The FreeBSD
emulation of ALSA does not have this insane requirement, so this
went unnoticed for the 1.1 release.

I suspect that this actually fixes SF bug 24, because ALSA is
preferred over pulseaudio.

The jury is still out on some details...
1) If you press ESC while editing the comment, it currently cancels
   the edit.  Would people expect ESC to exit the program instead?
2) The format/colours of the displayed comment... right now it's
   using the UIFC background colours, and not draing a border or
   window.
3) Just how long should the comment field be, and how should it be
   edited.

While errors on transmit seem to be handled well, errors on receive
do not, especially through js_recv_line() which has been seen to
trigger a large number (hundreds) of ECONNRESET errors.  To prevent
this, simply close the socket when an otherwise unhandled error
occurs.

Almost certainly fixes that issue, but the underlying cause is still
undetermined.  The calling script (imapservice.js) was checking
Socket.is_connected after each recv_line() call, so if the socket
was actually reset, it would be expected to only call it once.

An alternative would be to explicitly handle the error that is
seen (CRYPT_ERROR_PARAM1), but let's try a generic fix first and see
of anything breaks because of it.  Most likely issue would be an
inability to recv() data after calling shutdown(), but I don't think
many people do that except to move the TIME_WAIT to where they want
it.

It fixed an issue in js_socket.c, no reason to expect better
behaviour with TLS.  This may fix SZ YModem-G transfers on
cvs.synchro.net...

Don't include the obsolete unstructured regex in abs.unstrctured,
and use the abnf.unstructured for parsing headers.

This fixes an issue seen running the IMAP server on gratuitously
spammy messages.

- Backtab also exits with K_TABEXIT when the string is selected
- Backspace scrolls the text to the left of the cursor to the right
  if the left of the input is not the start of the string.