there's no data waiting.
KnK v4.4 (using QBSERIAL 3.10) apparently keeps reading the base I/O register
for input data until it reads 0.

the sound card on Win32 (if available).

Fixed typos in ANSI music help.

default note length for successive notes.  Used by KnK (and probobly others)

scrolled down.

no longer need to re-display previous WIN_SAV|WIN_ACT menus for them to be
erased properly.

"auto-savnum" feature isn't working 100%. Deuce: If you monitor api->savnum
going in and out of options from xtrn_cfg() you'll see what I mean.

debug output.
Added optional debugging of FOSSIL functions (counts number of calls to each).

debug output.

(and it's yield interval, in milliseconds, now).

assuming that select()ing for write can detect the condition before a
sendsocket() if more than a single byte, attempt to fix the wedged session
bug.  I give this about a 27% probability of being a fix.

apparently some Linux's clib doesn't define CLK_TCK. <shrug>

potential forced yield. This "tames" BRE down to single-digit CPU utilization!

if the passed value contains the y variable)

  to be destroyed in the internal IRCd nickname list.

This script should probably use the 'sbbs -d' option to force daemon mode.
This script could probably be merged with the Redhat/SUSE version.

paranoia - also prevents foot-shooting (masks access()).

(see docs/tcpip_faq.html for details).

(if it exists) over-ride those settings.

(the "exec" directory) - used if there's no sbbsexec.ini in the current working
directory (e.g. xtrn/somedoor).

properly checked on dorectory requests which didn't end with a /

(Pops up a calendar for selecting now)
Removed useless glade project file from CVS.

We'll need to fix this in ciolib.