Re: [cc65-devel] c64 - getting seemingly random chars on screen

[Michael P. <mp...@bl...>]

On Tue, 2 Mar 2021, Michael Parson wrote:
> On 2021-03-02 17:44, Michael Parson wrote:
>> On Tue, 2 Mar 2021, gr...@gm... wrote:
>>> I'd use watchpoints on the entire screen, like
>>> 
>>> . watch store 0400 07e7
>>> . x
>>> 
>>> (assuming your screen memory is at 0400)
>>> 
>>> then when the first wrong character pops up break into the monitor
>>> again, and you'll see a list of writes to the screen. find out which
>>> one wrote the wrong character, and you know where it happens.
>> 
>> Thanks, I'll give that a shot.
>
> Took a few tries, but here's where it happens:
>
> #1 (Stop on store 057c)  006 020
> .C:222d  91 D1       STA ($D1),Y    - A:4A X:09 Y:14 SP:ee ..-....C 
> 84639134
> .C:222f  AD 86 02    LDA $0286      - A:4A X:09 Y:14 SP:ee ..-....C 
> 84639134
>
> Right when the piece I clicked on turned red, I also got a red 'I'
>
> 222d is in the middle of the subroutine starting at 2229:
>
> .C:2229  05 C7       ORA $C7
> .C:222b  A4 D3       LDY $D3
> .C:222d  91 D1       STA ($D1),Y
> .C:222f  AD 86 02    LDA $0286
> .C:2232  91 F3       STA ($F3),Y
> .C:2234  60          RTS
>
> Which, according to the generated label file, is '.putchar'
>
> There should only be 1 character being put to the screen at that time,
> the red version of whatever character I clicked on, which is happening,
> just that the extra & erroneous character shows up too.

OK, put together a stripped down version.  No score keeping, no stdio.h
calls, no filesystem access, just the basic gameplay.  All screen
manipulation should be going through conio. Since this is a C-64
test, I even removed the asm bits and used a POKE to kick back into
uppercase/PETSCII mode.  I'd still prefer a cross-platform 'safe' way to
do that.

https://github.com/mparson/gridgame/blob/main/c64/minimal.c

Still getting the same behavior.

For grins, I compiled it for C-128, also got the errant chars.

-- 
Michael Parson
Pflugerville, TX