Re: Integer BASIC & Applesoft BASIC Do Not Work

[Rubywand <rubywand@swbell.net>]

Bryan Parkoff writes ...
> 
>     I did an experiment to switch from ROM to RAM at $D000 through $FFFF.  I
> mean that $D000 through $FFFF are set to read/write instead of read-only
> mode.  I can be able to modify monitor routines and autostart routines.  For
> example, I change from $A5 + $22 (LDA $22) to $60 (RTS) at $FC58 (HOME) by
> typing "*FC58: 60 (CR)" using Apple II+'s keyboard.  You type, "CALL - 936"
> or "HOME" to clear the screen, but it will never clear the screen.
>     I start to test Integer BASIC and Applesoft BASIC by typing, "*E003G
> (CR)".  "> 65535 " appears for Integer BASIC and Applesoft BASIC crashed.
>     I did discover that $F000 through $F022 routine after I typed "*E000G
> (CR)" is used to modify from $00 to $FF each page ($0000 through $FFFF).  I
> suspect that this routine is used to test RAM available like 4K, 8K, 16K
> through 48K.  If it is only 8K on RAM, it will stop scanning until it
> reaches $1FFF.  If it is only 48K on RAM, it will stop scanning until it
> reaches $BFFF.  If it is only 64K on RAM, it will stop scanning until it
> reaches $FFFF.
>     It is how Integer BASIC and Applesoft BASIC fail to work if it attempts
> to overwrite or modify from $00 to $FF on $D000 through $FFFF because it can
> modify monitor routines and autostart routines.  It is interesting.  How do
> software understand memory available without hardware's help?  Who is doing
> clever by writing 6502 assembly language that is $F000 through $F022
> routine.  It can have bug by overwriting Integer BASIC routines or Applesoft
> BASIC routines on $D000 through $F7FF, but it should be designed to prevent
> overwriting like using branch.
>     I switched from RAM back to ROM on $D000 through $FFFF so it can only
> scan by reaching 48K before it does not modify monitor routines and
> autostart routines.  Integer BASIC and Applesoft BASIC start to work fine.
>     Have you done the same experiment?  I am so curious.
> 
> --
> Bryan Parkoff


     Don't think I've ever tried that experiment. When doing modifications to
our II+ $F800-$FFFF ROM, my approach was to write an EPROM and plug it in for
testing.

     The result is a modified Monitor ROM with extensions in two $C8xx blocks
in Slot 7 (another EPROM). A switch allows selecting the modified Monitor ROMs
or the original ROM.

     One way you might keep a Monitor RAM from getting messed up is to install
a mod which allows switching OFF Writes to $F800-$FFFF.



Rubywand