Posted By
Csabo
on 2007-04-10
23:45:11
|  SCN News 04
What's the deal with this one? Unreal and Moldi both sent me a "good" copy of this mag, but it crashes. Unreal mentioned that due to Steve's defender, it only works like 1 time out of 10 on a real Plus/4. My Plussy is not hooked up right now (video cable problems), can anyone check this on a real machine?
I debugged it, but the defender is actually pretty straightforward, and doesn't use FF1E or FF1D or anything like that. (Meaning: it should be solid code, very like there's no emulation problems.)
Maybe it was a joke from the SCN team? I could imagine them releasing a bugged mag, and then everyone just speads it... Which they did...
Which reminds me, are any of the remaining members of this team still around, or did we lose contact with them all? I recall Steve being around a few years back.
|
|
Posted By
Lavina
on 2007-04-11
04:20:29
 | Re: SCN News 04
I can reach Radula and Steve any time. I'll ask them.
|
|
Posted By
Luca
on 2007-04-11
07:22:41
 | Re: SCN News 04
...and kick'em to code in the meantime!
|
|
Posted By
Moldi
on 2007-04-11
19:17:32
| Re: SCN News 04
I repacking this issue,I send it to you again  )) (thx from YAPE!)
|
|
Posted By
Lavina
on 2007-04-12
03:25:10
| Re: SCN News 04
Steve has replied already. He downloaded the issue, tried it with Minus4 two times, and both times it worked just fine. So, maybe it's a YAPE bug?
|
|
Posted By
siz
on 2007-04-12
06:51:43
| Re: SCN News 04
Just to make sure, I tried it on Yape (0.73), VICE xplus4 (1.21), István's plus4 emu and minus4. On minus4 it works fine but on the others it fails to run. 
|
|
Posted By
Csabo
on 2007-04-12
12:06:13
| Re: SCN News 04
Moldi sent me a "fixed" version, as soon as I saw the intro screen I remembered this mag. So it's definitely supposed to be working. This is actually good news, we have a chance to improve emulation
|
|
Posted By
Gaia
on 2007-04-12
12:33:28
| Re: SCN News 04
Hm, I seem to deduct from Moldi's post that he used Yape for depacking this? Doesn't it do JMP ($00FF) or something akin? I seem to recall from early experiments, that the address $0000 then acts like a normal RAM address and not like the usual CPU port direction register (when you're accessing it directly). Sounds like a cool challenge to sort this out. But heck, we could just rip some code from Minus/4, too 
|
|
Posted By
Luca
on 2007-04-12
12:54:58
| Re: SCN News 04
Hahaha! "As usual", we could add!
|
|
Posted By
Csabo
on 2007-04-12
14:04:31
| Re: SCN News 04
The defender runs properly without a hitch, you can set a breakpoint to $1010 to stop it. Moldi mentioned that it's the packer (HCS in this case?) that causes the problem. I'm not sure why yet, but I don't think I saw that JMP. At $01AA there's a branch which jumps to $01B8 - but from $01B8 there's only "broken code". Maybe it shouldn't get to that point already.
|
|
Posted By
Gaia
on 2007-04-12
15:33:04
| Re: SCN News 04
Actually, after the defender has run, it is already "wrong"... I found the reason: it simply expects that the RAM memory is filled with zeroes (I haven't sought for the exact location) and only Minus/4 does that (but you can configure the powerup RAM pattern in VICE and then it runs, too). Optionally, you can set the memory in the external monitor to zero (F 1000 FD00 0) in Yape and load the PRG with F8. Maybe I'll add a config dialog for this. It's not the first game that requires a certain powerup pattern (remember "Land It" and "Menekülés"?)
|
|
Posted By
Csabo
on 2007-04-12
19:49:00
| Re: SCN News 04
Right! Actually it's a single byte at the end of the program that causes the problem. If you enter >9C75 00 in the emulator monitor (even after the defender has already started), it will run properly. Also, by adding a $00 byte to the end of the file, it will run fine, regardless of what byte follows. So the missing $00 could be considered a file error.
Can someone check that on the real machine, any non-zero value on $9C75 will cause a crash? If so, then it's definitely a program error: that byte is required, therefore it should be part of the program.
|
|
