Re: MCL 2.01 and the Radius Rocket 25 (IIcx)

[Bao Chau Ha <habaoch@Eng.Auburn.EDU>]

On Fri, 5 Aug 1994, Kalman Reti wrote:

> At 12:21 PM 8/5/94 +0000, Bao Chau Ha wrote:
> >I have just gotten the Radius Rocket 25 for my Mac IIcx.  MCL 2.01 has
> >been running fine with the IIcx for a year now.  However, with the
> >Radius Rocket 25, I keep getting an unimplemented trap error when I
> >try to start it.  I patched the MCL 2.01 with the MMU4-for-IIci using
> >ResEdit and still get the same problem.
> >
> >The System software is 7.1.  And the Radius Rocketware is 1.5.
> >
> >I would appreciate any suggestions and comments to resolve this
> >problem.  The reason I got the Rocket Radius is to run MCL, and it
> >has been a very frustrated experience.  Everything else seems to run
> >fine.
> >
> 
> The problem is that the Rocket's configuration and the ROM in the
> Rocket don't quite match; when dispatching on the floating point
> set halt vector trap, the ROM code tries to decide what to translate
> that into.  On a real quadra, it goes to one set of code, on the Rocket
> it decides to issue an FPPRIV trap, which doesn't exist any  more.
> 
> Another customer and I have both tried to explain this problem to
> Radius on more than one occasion, and have never succeeded in talking
> to anyone who gave any signs of understanding it.
> 
> You can get around THIS particular problem by applying a patch (which I'll
> send under separate cover) which causes MCL to not even attempt to set the
> floating point halt vector.
> 
> I had another customer with a rocket that he wanted to run MCL on; he was
> trying to use Rocketshare.  I diagnosed a Rocketshare problem that clobbered
> low memory occasionally, specifically part of the CHK trap dispatch.  This
> killed MCL (which heavily uses the CHK trap) but every other application I
> tried worked fine.  Again, multiple attempts to explain the problem (and the
> fix) to Radius got nowhere.  The customer finally gave up the attempt.
> 
> Anyway, even though you aren't running Rocketshare, I wouldn't be surprised if
> there were more such subtle problems lurking in the wings.
> 
> Anyway, the patch should let you get somewhat farther....
> 
Thanks for the patch.  At least it will let me to work with MCL 2.01
for a while.  

I checked with Radius, and they know about the problem.  However,
they are not going to do anything about it.  RocketWare 1.5 will
be the last version.  I guess it is the price I pay for going the
accelerator route instead of getting a faster Quadra.  I have
been holding out until MCL is ported to PowerMac.  Oh well, it
is just like everything else in the Lisp world nowaday. :-(

Bao