[stella] behold, i am become shiva, destroyer of kernals!

Subject: [stella] behold, i am become shiva, destroyer of kernals!
From: KirkIsrael@xxxxxxxxxxxxx
Date: 11 Mar 2004 05:00:00 -0000
So, I am learning some basic page boundary issue detection 
skills, but this one has me stumped.  (Yeah, I know, big surprise.)
I am getting better at isolating and noticing this kind of thing,
honest...and for every bit that seems likely to be timing-sensitive,
I align 256 it or at least use a CHECKPAGE macro around it.

So, like usual with this kind of thing (at least, with the kind of 
thing I think this might be, I'm not quite certain :-/   )
the problem showed up when fixing something else, a causal
relationship is assumed, disproved when it messes up even with 
NOPs in the recently worked on code...then I started shoving NOPs 
around, figuring if I moved the NOP pair to after a certain properly 
bit and it started to work, then maybe that was the bit that 
needed page boudary attention. 

So, I found a crucial point, a single STA insruction, two NOPs immediatley
or some distance before kills everything, two NOPs immediately or some 
distance after, no problem.  And it seems to be 
NOWHERE NEAR anything time related...looking at the .lst file, 
it's not near any particular page boundary. (And of course, this
is only on a (supercharged) atari...it's fine and dandy on Z26
and PCAEwin.)

Here's the working and broken source code, respectively:

search for the term "shiva", it indicates where I empirically
determined (the? a?) breaking/no-breaking point to be.  

I am so grateful for this group.  This game NEVER could have 
happened, ever, without you guys.  You've let a semi-ok high level
programmer have a great time bringing his 2600 dreams to life.
I'm just praying for one more miracle from you guys , because I am 
SO close to being done...I found this while fixing a bug where
I forgot to change the criteria for flashing P2 score during 
Poorlords, so there's that, there was one slightly scary case 
I MIGHTa saw, where the ball bounced off a wall but no chunk
of wall disappeared (I was actually quite proud of how I isolated
that bug before, but maybe some case of it was missed or re-emerged),
and MAYBE I'll try touching up the physics and effective AI, but 
that's IT, I'm so close to ready for Philly...
(Oh, and maybe I should look into eliminating flash when reset is 
hit. But that's it!)

Oh, Eckhard: re the title screen.
It's Z26 (2.09) -- June 22, 2003 -- you mentioned it was 265, but
according to my version of Z26 it was 266...but that wasn't the more
standard 262, so I fixed it...and then that seemed to help with the 
problem I was previously describing, wher it'd say 1 Line, 1300+
FPS, I haven't seen or heard that problem since.  So maybe at some
point it might be interesting to see what Z26 does w/ odd scanline 
counts? I'm not sure.

A bounty of a free t-shirt to anyone who can tell me what
this is and how to fix it!

Thanks always,

Archives (includes files) at http://www.biglist.com/lists/stella/archives/
Unsub & more at http://www.biglist.com/lists/stella/

Current Thread