in the headers anymore (yay!)

It is NO LONGER POSSIBLE to use Cryptlib from upstream, you MUST use the
included one now, so it really needs to be statically linked.

Thanks Digital Man!

the .iss file are specific to my build system, so not of immediate use to
anyone else. Deuce might want to review the info (like the Publisher).

file. Previously, any threshold set by manually editing the .cfg file would be
lost when saving changes using echocfg.

USEPACKER...
Basically, if there are more than one USEPACKER lines of the same archive
type as the new one, the user will be added to the first instance of the
USEPACKER, and every other USEPACKER line of the same type will be deleted.

We likely don't want to do that either.

all AREAFIX lines except for the one that was updated.

This is apparently "bad".  Stop doing this bad thing (We'll still delete
AREAFIX lines without an address though).

eventual failure to open any files or even log the failure in the error.log.

destroy the ssh_mutex (which was not initialized).

not the same behaviour as in genwrap.c

compiler.

FreeBSD to fix builds for OSs without alloca() (ie: OpenBSD) that define
__GNUC__

headers are rife with.  Re-enables cross-building with mingw32 (at least for
SyncTERM).

for the static 64-byte sbbs->telnet_cmd buffer, it will bork the session.
(Seen with a long user-name and DISPLAY when using the OS X telnet client).