Path: utzoo!attcan!uunet!convex!killer!letni!merch!cpe!tif
From: tif@cpe.UUCP
Newsgroups: comp.sources.games.bugs
Subject: Re: tetrix INIT_PAUSE
Message-ID: <138600001@cpe>
Date: 10 Mar 89 13:22:00 GMT
References: <204@sabin.UUCP>
Lines: 30
Nf-ID: #R:sabin.UUCP:204:cpe:138600001:000:1394
Nf-From: cpe.UUCP!tif    Mar 10 07:22:00 1989


Written  4:09 pm  Mar  7, 1989 by sco.UUCP!jonl in cpe:comp.src.gam.b
>+-I seem to recall bjm@sabin.UUCP (Brendan J. McMahon) writing:
>|How do I adjust the speed without this nonexistant constant?
>you change the value of CurrentPause on around line 140 or so in tet.c
>to whatever you want.

First, to increase the speed tetrix just decrements this value.  This
means that if you change this to 1000, you'll never even notice the
decrement.  I think the decrement should be something like

	CurrentPause -= 10 /* some fraction of original CurrentPause value */

Next, anyone got any ideas on how to do this without the busy-loop stuff?

How about doing all reads with a VTIM =< the desired pause time and VMIN
of 1.  Then using timeb (is that the right name for the time function that
gives fractional seconds) to actually time the game and decide when to move
pieces.  But the way read works if you did several moves in one tick they'd
all get done at the end of the tick (is that acceptable?).

What we really need is a read that causes interrupts but the only way I
can see doing that is with two processes.  One that does nothing but wait
for keys and send signals to the other.  Is this feasible?  What about
the old signal bug (too many too fast)?  Would you have to use sigset
and family?

			Paul Chamberlain
			Computer Product Engineering, Tandy Corp.
			{killer | texbell}!cpe!tif