Re: A92: line 1111 & Raven

To: assembly-92@lists.ticalc.org
Subject: Re: A92: line 1111 & Raven
From: "Bryan Christopher Turner" <bcturner@eos.ncsu.edu>
Date: Mon, 11 Aug 1997 11:12:07 -0400
In-Reply-To: Hmlacage@aol.com "A92: line 1111 & Raven" (Aug 11, 5:54am)
References: <970811055432_805925863@emout18.mail.aol.com>
Reply-To: assembly-92@lists.ticalc.org

On Aug 11,  5:54am, Hmlacage@aol.com wrote:
> Subject: A92: line 1111  & Raven
> i am truly fascinated by this thing i talked about in last e mail: the line
> 1010 error
> doesn' t work ! ... There IS something.... I happened to discover that many
> vectors
> had adresses bigger than 24 bits and that these adresses : the numbers bigger
> than 24 bits are in fact the vector numbers ( as explained in the raven doc )
>
> Is this due to Raven ? ( raven redirect these vectors to the same adress but
> allows to know the vector triggered just by looking at the "PC": What raven
> represents is not the PC: it is the PC plus one byte to the left used for
> that purpose ? )
>
> If all this is true, then, why is line 1010 not redirected in the same manner
> ?
>
> i hope that someone will answer .... even if i haven' t made myself clear
> enough....
>
> mathieu, <hm lacage@aol.com>
>-- End of excerpt from Hmlacage@aol.com

Hello, I'm sorry I havn't been keeping up with the lists, my calc is in for
service right now..

   In order to make Raven as small as possible, I chose to write just one
exception processing routine.  But the processing of the exceptions is slightly
different based on which vector is called.  This information is placed in the
high byte of the vector address for each vector.  (You noticed this in your
explorations)

   Those vectors which are used often by the system are NOT overridden, as this
would cause significant performance penalties.  Therefore, if the high byte is
zero, chances are that vector is used by the system often enough to discourage
Raven from overriding it.  The A-Trap (line 1010 exception) is one such vector.
 It is used for some mode switching and is called quite regularly.

Does this answer your questions?  Please clarify if you need more info, Raven
may be easily modifyed to fit your needs (which is the whole purpose for
including fully commented souce!)  :)

   For those interested parties, I have solved the problem of Raven causing
Fargo's Shift-On combo to stop working, it is due to the Shell program, and not
an error within Raven.  I have a patch for Shell if anyone would like it.

   Also, breakpoint setting and tracing of code is easily possible including
step, step into, step over and exit.  Unfortunately you'll need some expertise
and experiece to perform them right now.. I havn't decided on the best user
interface for this yet.  Also searching memory and setting breakpoints is
planned.  Please let me know if you use this program, without any support
there's no point in continuing development.

Thanks!

--Bryan
bcturner@eos.ncsu.edu

References:

A92: line 1111 & Raven

From: Hmlacage@aol.com