Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utcs!mnetor!seismo!brl-adm!caip!topaz!nike!ucbcad!ucbvax!UHUPVM1.BITNET!ACS19
From: ACS19@UHUPVM1.BITNET.UUCP
Newsgroups: net.micro.atari16
Subject: megamax bug, YES!!
Message-ID: <8607200204.AA19611@ucbvax.Berkeley.EDU>
Date: Sat, 19-Jul-86 20:09:59 EDT
Article-I.D.: ucbvax.8607200204.AA19611
Posted: Sat Jul 19 20:09:59 1986
Date-Received: Sun, 20-Jul-86 07:57:41 EDT
Sender: daemon@ucbvax.BERKELEY.EDU
Reply-To: ACS19%UHUPVM1.BITNET@WISCVM.ARPA
Organization: The ARPA Internet
Lines: 22

Well, I hate to say it, but there is a definite bug in Megamax C.  The relative
addressing did not work.  When I was replacing a system vector with the
address of a routine, the relative offset from the PC was not correct, and the
system would crash every time I would call the routine via (*ptr)(var1,var2,..)

Example of what I am doing...

long   *vector= .....    /*vector address*/
....
...
xbios(38,install);
....
...
install()
*vector=(long)routine_name;
....
...

basically, that is it.
The relative offset (as I examined the routine in SID) was incorrect by about
24 bytes.  If anyone from Megamax is out there, I can send the source to you,
so you can try it for yourself.