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.

This falls squarely into the "How did that ever work?" category.

Fix for Altere's logons yesterday list that isn't rotating out because nobody
logged on today.

Previously, the height was left the same, and the width was set to
the next-lowest "standard" size.  Now the cterm window is set to
the size of the requested mode and both top/bottom and side bars
are added to make the virtual screen the same as the requested one.

Using the "current" mode still retains the old behaviour.

Closes SF request 11

With maximized windows, unexpected behaviour occured when the program
resized itself, then the window manager minimized/restored the window
Should address SF bugs #18 and #13

This is used by some BBSs to enable encryption without needing to
integrate the BBS user base into their SSH server (and presumably
so they don't need to run multiple SSH servers).  All users log in
with the same username (ie: "bbs") and no password is requested or
required.

Once the BBS starts, it prompts for the BBS user name and password
as normal.

In SyncTERM, the user/password/syspass fields are redefined as
SSHuser/BBSuser/BBSpassword and they are moved around when you
change the connection type.  This means that if you change a listing
that has a syspass to SSH (no auth) and back, the syspass is lost.

I'm not sure if I plan to fix this or not.

It would be fine if this only warned while building JS, we're used
to ignoring that, but this bugger warns while building Synchronet
stuff.

Only supported on FreeBSD, Linux, Solaris, OS/2, GNU, Cygwin, and Darwin
Should allow OpenBSD to build JS engine.

Not yet tested, so more patches may be incoming.

There are too many options now to fit in an 80x25 screen, so group
all the log options into a separate sub-menu.

While we're here, notice that we're writing paths up to MAX_PATH
into an 80-character buffer and fix that.

Do not get/set these values on other OSs.  Fixes a segfault caused
by default_dosemuconf_path being undefined on FreeBSD, but would
also happen on any other *nix platform (ie: OpenBSD, etc).

Some broken protocol drivers (like sz) don't send a YModem terminate
at the end of a batch (or maybe just after a single file?) so SyncTERM
ends up trying to fall back from G to CRC-16 to 8-bit before ending.

This detects when the last known file was received, and if the YModem
header that follows is not correct, aborts the transfer.

Something weird is happening to the cryptlib session when SZ sends
a lot of data... no real clues left.

This issue may actually be a rekey issue, to clear this, we need
to call cryptPopData(), but we can pop zero bytes.

In this case, re-try the flush.

During YModem-G transfers, the previous 5-second timeout was
exceeded.

Just copied out of the documentation.

It's not clear if this is working properly or not since the only
BBS I know of that supports telnets (fido.beholderbbs.org) doesn't
seem to do any telnet "stuff".

Should fix SF #45 since it appears the input thread started then
exited before SyncTERM realized the connection was successful.

In preperation for telnets support, make conn support TX/RX parsers.
Make telnet use those new hooks, and remove unused and ambiguous bits
of the conn API (conn_recv(), conn_peek()).

All the major bugs I'm aware of are fixed (though RIP is still
incomplete)

Make an exception for all keypad stuff.

Cursor update issue, cursor was only redrawn on a blink status change.
This will redraw the cursor any time the cursor moves now.

Should fix file transfer issues.
(SF bug #38)

for Al (TRMB) and his Russian users, requested via DOVE-Net.

Do an FTN "ASCII 1" -> "US-ASCII" translation of the CHRS->charset too.

Clang 11 throws an error if you do.

Presumably these were part of macros once and never cleaned up?

Just rename ESCDELAY member to escdelay.