|
Bug in kof96, kof97 and kof98
#379126 - 10/25/18 08:41 AM
|
|
|
Hi Robbbert,
I believe you already know it. In the games listed in the subject and only with MAMEUI, when the game timer is at zero, instead of ending the challenge the time restart from zero.
Are there any chances that this error will be fixed?
Thank you.
|
Site: https://www.progettosnaps.net/
Board: progetto-SNAPS Official Board
Twitter: @progettosnaps
|
|
|
Re: Bug in kof96, kof97 and kof98
[Re: AntoPISA]
#379128 - 10/25/18 10:59 AM
|
|
|
> Hi Robbbert,
> I believe you already know it. In the games listed in the subject and only with
> MAMEUI, when the game timer is at zero instead of ending the challenge the time
> starts from zero.
> Are there any chances that this error will be fixed?
>
> Thank you.
Since this was an issue that was apparently fixed, and based on what you're reporting is occurring again in 202, Robbert *really* needs to explain what the source of the issue is. Transparency is key here.
From the outside it looks like a 68k issue, a very fundamental one of not being able to count / flags being wrong. If that's the case it's going to affect a far broader range of games than these NeoGeo titles in subtle, unwanted ways that completely break the gameplay. It puts MAMEUI builds affected by the issue on a 'do not use' list because the 68k is a widely used processor.
This shouldn't be happening, there should be no reason for this bug, so we need a clear explanation of why it is happening and what was done to fix it before.
If it's a code change specific to MAMEUI it needs pointing out in the MAMEUI source, and reverting (there really shouldn't be changes that affect how drivers work anyway)
If it's a bug that's created by certain compiler versions then those compiler versions should not be trusted as if they're getting something basic like this wrong they're probably doing a lot wrong elsewhere too. Alternatively, if MAMEUI is using the same compiler but different compiler optimizations than baseline, including potentially dangerous ones (options turned too high) that needs to be fixed because again such things can cause widespread issues.
I'm not being a dick here, this kind of bug is genuinely concerning and to have confidence in the product people need to know why it happens and what is being done to address that, it's something that's come out of nowhere after almost 20 years of these games never having such an issue, with them all using emulation code that is mature and well tested.
So, Robbert, why is this happening in MAMEUI?
|
|
|
|
Re: Bug in kof96, kof97 and kof98
[Re: AntoPISA]
#379307 - 11/03/18 05:22 AM
|
|
|
|
|
|
Re: Bug in kof96, kof97 and kof98
[Re: Robbbert]
#379319 - 11/03/18 01:21 PM
|
|
|
> > Hi Robbbert,
> > I believe you already know it. In the games listed in the subject and only with
> > MAMEUI, when the game timer is at zero, instead of ending the challenge the time
> > restart from zero.
> >
> > Are there any chances that this error will be fixed?
> >
> > Thank you.
>
> From the reports I've received, it's fixed in 0.203
from what I'm told, that is the case. what change was needed to fix it?
|
|
|
|
Re: Bug in kof96, kof97 and kof98
[Re: Haze]
#379323 - 11/03/18 02:20 PM
|
|
|
|
|
|
Re: Bug in kof96, kof97 and kof98
[Re: Robbbert]
#379324 - 11/03/18 02:23 PM
|
|
|
|
|
|
Re: Bug in kof96, kof97 and kof98
[Re: Robbbert]
#379327 - 11/03/18 02:36 PM
|
|
|
> > > > Hi Robbbert,
> > > > I believe you already know it. In the games listed in the subject and only with
> > > > MAMEUI, when the game timer is at zero, instead of ending the challenge the
> time
> > > > restart from zero.
> > > >
> > > > Are there any chances that this error will be fixed?
> > > >
> > > > Thank you.
> > >
> > > From the reports I've received, it's fixed in 0.203
> >
> > from what I'm told, that is the case. what change was needed to fix it?
>
> I used the fc command to look for differences between mameui's m68k and mame's. But
> fc seems to get confused by line endings so it showed heaps of stuff. Rather than
> trying to figure out what might be a real difference, I just copied the whole m68000
> folder over from mame to mameui. I'm no good at fighting games, so I had a great deal
> of difficulty staying alive in the kof games, so I had to presume it was now ok. I
> couldn't think of anything else that could cause the problem though.
Why would it differ in the first place? The m68k core is part of the main MAME distribution and has had many important fixes (mostly to the 020+ support) in the last year or two, those files shouldn't even be changing from baseline in MAMEUI? eg. the recent Time Crisis fix is because of a change to the m68k core, had you not synced it sounds like that would have remained broken in MAMEUI too?
are there other parts of MAME that differ because of MAMEUI that we need to keep an eye on?
I'm glad it's fixed, because it seems I was understating just how serious the issue was based on what you've said, but I don't understand how this situation came to be in the first place.
|
|
|
|
Re: Bug in kof96, kof97 and kof98
[Re: Haze]
#379331 - 11/03/18 04:37 PM
|
|
|
|
|
|
Re: Bug in kof96, kof97 and kof98
[Re: Robbbert]
#379333 - 11/03/18 05:10 PM
|
|
|
> > Why would it differ in the first place?
>
> I assume I fucked up somehow. Anyway it's over now, it's dealt with, and I'm not
> going to dwell on it any further.
>
> >
> > are there other parts of MAME that differ because of MAMEUI that we need to keep an
> > eye on?
> >
> Depends what you want to keep an eye on. There won't be any emulation differences (or
> shouldn't be). But, due to the large number of bugs with options handling in standard
> MAME, there's a number of fixes and adjustments in that area.
>
> However, I don't propose to justify to you why they are needed, they just are.
well option handling is option handling, if your fixes are legitimate they should be made into PRs tho
my main concern is with the emulation tho, because that's what I work on, so as long as there are no differences that pertain to emulation I'm not overly concerned; what amounts to a bolted on frontend should not be changing how things actually look and feel once the emulation is running.
|
|
|
Index 
Threaded 
Bug in kof96, kof97 and kof98
Reply
[Re: 
