some requests for clarification from Deuce.

and context. This allows the event thread to perform its own JS cleanup without
deleting its sbbs_t instance (causing the exception on sbbs/sbbsctrl shutdown).
This replaces the change committed in rev 501.

date is not the current date (thus resetting all daily stats like logons today,
posts today, etc.) - but resetdailyuserdat() failed to save the reset stats
to the user.dat file because user->number was 0 at this point in getuserdat().

This bug (reported by Finnigann) only effected non-BBS servers (like FTP, HTTP,
etc.) since the BBS has its own checking of a new day and resetting of the
daily stats for the user during logon (in logon.cpp).

PortAudioCallback types.

<threadwrap.h> must be included AFTER <jsapi.h>
<threadwrap.h> redefines DLLCALL via <wrapdll.h>
The other alternative would be including <threadwrap.h> from js_rtpool.h
which should not be required.

running).

bugs.

with debug JS libraries, Synchronet may now exist with an assertion.  If
this happens, please let me know exactly what was being accessed at the time
so I can hopefully reproduce it.  A backtrace isn't as helpfull as one would
hope in the case of a JS assertion.

(keyword "should")

Fixed game status menu to show your turn if appropriate in a newly started game.

under certain circumstances.

Enabling shared runtimes passes the initial touch testing.

test shared runtimes (which still crashes)

I suspect this is due to the script compilation/startup not being inside
a BeginRequest() block.

semwrap.h which is pulled in by threadwrap.h anyway.

This patch assumes that smb_hfield*() does not potentially block...