 No.05655
Tafoid Administrator
Feb 7, 2010, 23:24
|
As with your other submission (03232), I'd request an .INP file given it's a hard to obtain or difficult to reach bug. Such a file would also help developers quicker determine the problem and possibly fix it. Thank you. |
|---|---|
|
No.05657
Tafoid Administrator
Feb 8, 2010, 00:14
edited on: Feb 8, 2010, 00:18
|
I was able to confirm the issue - there is a definite hang going on shortly after 2-4 starts, regardless of what you are doing. In debug, there is mention of a "Stopped at temporary breakpoint FFFFFFFF on cpu 'maincpu'" message which is puzzling unless the source has this set up to allow analysis of problems. When I attempt to "go" again via debugger, the music plays but the rest of the game has hanged up. |
 No.05658
Haze Senior Tester
Feb 8, 2010, 07:33
|
sounds like the mainCPU has ended up in lala land, and is wrapping over the top address, hence the breakpoint, and crash. as for why, I don't know. a proper regression version might indicate more, the game isn't encrypted so it can't be a decryption issue. |
|
No.05659
Tafoid Administrator
Feb 8, 2010, 14:25
|
I had determined a regression version through testing - 0.135u3. Pugsy was nice enough to help with a level select cheat (on his site) making validating the bug and confirming regression so much easier. |
|
No.05661
Haze Senior Tester
Feb 8, 2010, 14:42
|
yeah sorry, I missed that field, was looking at version instead for some reason. Interesting that it's a recent regression, I wonder what significant thing changed enough to affect this driver in that release. CPU crashing bugs are pretty serious and usually caused by something significant. |
|
No.05662
Tafoid Administrator
Feb 8, 2010, 14:57
|
It was the version which Aaron did a lot of 'bank' and memory changes. It could simply be a mistype or something similar. |
|
No.05673
Haze Senior Tester
Feb 9, 2010, 11:13
|
hmm yeah.. and the System16/18 games have the whole funky reconfigurable memory map thing, which, if the behavior was broken in a subtle way, could cause the whole thing to crash I guess.. |
 No.06471
abelenki Tester
Aug 1, 2010, 07:55
|
the bug with freezing on stage 2-4 is still there (MAME 0.139). i can confirm it on shdancerj set. |
 No.06576
Phil Bennett Developer
Aug 21, 2010, 19:40
|
The game was writing to ROM, which in turn was being treated as a write to the memory-mapping registers. Bad. Why it's writing to ROM, I have no idea. I'd suspect a bug but this seems quite deliberate (protection?). The function at $18a46 is where the problem lies (this function appears to be called only during stage 2-4). The offending instruction is: $18a98: ori.b #$2e, ($31, A3) This is actually part of the instruction at $18a94 - the code jumps into the middle of it. Wacky. |
|
No.09052
Tafoid Administrator
Oct 25, 2012, 22:23
|
Recent rework of the Sega drivers caused this to be an issue again. Reopening. |
|
No.09130
Phil Bennett Developer
Dec 9, 2012, 18:07
|
Same issue again, the fix was removed during the recent segaic16.c refactor |