|
Notes |
|
|
(0004161)
|
|
Wepl
|
|
2015-01-02 17:15
|
|
|
Are you using lastest slave version? Because there were fixes regarding the highscores. |
|
|
|
(0004166)
|
|
Axel
|
|
2015-01-03 10:03
|
|
|
I'm using the latest WHDLoad version (18) and the latest slave version (3.3) on both WinUAE and on real hardware. So I was surprised to notice that it works on one but not on the other. I changed WinUAE's chipset settings, memory settings, everything - the highscore list stays broken. |
|
|
|
(0004168)
|
|
Wepl
|
|
2015-01-03 12:08
|
|
|
|
|
(0004170)
|
|
Axel
|
|
2015-01-03 15:29
|
|
|
|
|
(0006188)
|
|
JOTD
|
2018-04-30 23:57
(Last edited: 2018-04-30 23:58) |
|
I've tested on WinUAE and I confirm the issues.
- hiscore positionning is off by 4 chars
- disk seen as write protected
- also stars are not that random and the effect is lame
works on the real machine (what the hell!!)
|
|
|
|
(0006190)
|
|
twilen
|
|
2018-05-06 17:58
|
|
(Apparently I was never contacted)
Title screen has multiple problems:
- Advantec logo is horizontally shifted if OCS Agnus (Original game works exactly opposite, it is only correct if OCS Agnus. So "Advantec logo recentered" broke it if OCS Agnus)
- small text writer is 1 scanline too low if ECS/AGA.
- small text writer is horizontally shifted if ECS and only if ECS Agnus.
All problems are caused by buggy copperlist that enables/disables bitplane DMA when DMA is about to start. This causes side-effects and OCS/ECS behave differently in this situation.
Fix probably is as simple as moving DMACON bitplane disable copper MOVEs horizontally to earlier position. WinUAE handles these 100% correctly, just change chipset type on the fly to see the difference.
And finally, AGA side-effects apparently are not exactly same as ECS because real AGA hardware shows small text writer in horizontally correct "OCS" position (but still 1 line too low). Unfortunately there are multiple reasons that can cause this difference and I can't be sure which exactly causes it (without breaking anything that shouldn't break) without more testing first.
--
Write protection issue is almost guaranteed to not be directly emulation related.
Stars: use more accurate CPU mode? |
|
|
|
(0006191)
|
|
JOTD
|
|
2018-05-06 22:08
|
|
Toni I was about to post something on EAB about this, but in the meanwhile I ran the cracked version with WinUAE and no star issue (didn't test the rest)
Your explanations are invaluable, and I realize now that at least the "advantec center logo" fix shouldn't be applied when OCS Agnus
found some code on EAB: http://eab.abime.net/showthread.php?t=72300 to detect AGA.
The stars are (more) random once you played a game and lost. I'd prefer fixing this in the whdload slave, but that's probably too much work. |
|
|
|
(0006194)
|
|
twilen
|
|
2018-05-07 09:09
|
|
It probably is best to fix the copper list so that it works with all chipsets without separate chipset-specific fixes. I can do it while trying to understand whats the (previously not known) difference between ECS and AGA. This game is good test case :)
Star field probably uses usual stupid read $dff006 for generating "random" numbers which returns very static numbers if CPU is very fast without chipset DMA slowing it down. |
|
|
|
(0014297)
|
|
Hexaae
|
|
2024-12-15 10:47
|
|
|