Arcade Game Selector 2

[StatMat]

Indeed. I certainly hope it'll run a little quicker on real hardware such that it can be useful to everyone.

[StatMat]

I am still tinkering away on the frontend and the compilation builder when I can find time. I'm happy to see there's continued interest in this little project.

Anyway, I have uploaded a video of the latest build running on my XD here:




Hopefully this'll give you all a better feel for how it operates. I will try to get a beta build released soon.

[Retrofan]

Hey: I've just tried AGS and it doesn't work right if you are using CyberBugFixAga. That's surely for the screenmode used. Cyberbug changes the names to be CGXPAL instead of PAL for example.

I wonder if that could be fixed.

[MagerValp]

What's CyberBugFixAGA and in what way doesn't it work right?

[Retrofan]

CyberBugFixAga is a patch to use Fast memory as Chip one. It uses the fast as if you had a Graphic board. You can find it included in CWB OS39. Using it in an OS you don't even need to run AGS from booting, you could run it perfectly well from WB with plenty of chip mem.

It gets a distorted image surely because of the screenmode used or the cgxsystem.library used. KGLoad had this same problem and he fixed it. http://eab.abime.net/showthread.php?t=66086 from post 46 or so.

[Retrofan]

I see ags101 didn't have this problem, it was looking right:



The Aga version of TomCat666 also works:

[MagerValp]

I looked around a bit, is it this one? http://aminet.net/package/driver/video/CGX-AGA? I don't have a CyberGraphX system so it's hard for me to test.

AGS2 does all its rendering through gfx.library, so there's nothing obvious that needs fixing. Is the CGX-AGA author on EAB? Maybe he can shed some light on what's happening.

Edit: Found the CyberBugFixAGA how to enable thread. It does not instill confidence and I'm inclined to agree with NovaCoder.

[Retrofan]

You can install Classic Workbench OS39 to try it, it's included there.

Novacoder hates it because he has one of the rare programs that don't work with it, ScummVM. He also hates Delfinas and Fast Atas but people use them.

[MagerValp]

I'm sorry, but I'm in the US and I did not bring any of my Amiga stuff with me. It would take hours for me to source an install and set up a debug environment, and I wouldn't even know where to start since all I'm doing is calling system friendly routines from gfx.library. I'd be more than happy to fix AGS2 if I'm doing something wrong, but since the patch appears to be unstable it seems like a better idea to fix the patch or simply not enable it in the first place.

[Retrofan]

Ok, but then I can't include your program in AKReal. Thanks anyway.

Edit: These are the hits with Enforcer anyway. Attached. I can try whatever you want to try.

Ah, BTW CyberbugFix was fixed years ago by StingRay: http://eab.abime.net/showpost.php?p=732358&postcount=29

Novacoder still is needing some fixings

[MagerValp]

Sorry for the late reply, I didn't get a notification when you edited the post!

Enforcer hits are a clear sign that something is wrong, thank you for that. I can't seem to reproduce them though - I enabled Enforcer in FS-UAE 2.8.1u3, and WhichAmiga shows the MMU as being active, yet LawBreaker (and AGS2) doesn't generate any Enforcer hits.

After reading the Enforcer manual I managed to reproduce it, I'll make a debug build and see if I can figure out what I did wrong.

[Retrofan]

Yes, sorry unfortunately I like to edit posts.

Thanks a lot, it would be fantastic if it works here.

[MagerValp]

Enforcer hits found and fixed, see if it works better for you now: AGS2-20161130.lha

[Retrofan]

Ah. What a pity. Yes, now there aren't enforcer hits (good for your program).

But unfortunately it doesn't work with CyberbugFix yet. The only way you could make a working version is searching for the small differences between you version and the old working ones.

I only can promise you that I will try as much as you want to try, if you prefer by PM or here. It would be awesome to include it on AKReal for everybody with music and so.

Edit: Maybe you changed some setting in the screenmode or forced size or something?

[MagerValp]

AGS2 is a complete rewrite, there is no code used from the old AGS version.

For CyberBugFix, is there a complete archive with the bugfixed version and all support files, and working installation instructions somewhere? The discussion thread is a mess...

[Retrofan]

If you (please) could contact StingRay that fixed it, then surely this problem could be fixed. He is the one that knows more actually.

Edit: I only can tell you what it does:

CyberBugfiX does following things:

1) fixes memory leak
2) fixes OS (not CGX) bug with, for example, trashing of IBrowse's
graphics buttons.
3) fixes compatibility with blitter-using software like Prayer2.
4) moves many bitmaps to FastRAM
5) disables automatic fallback from BMF_INTERLEAVED mode - it will be
useful for software like MooVId, that uses direct moving to
interleaved bitmaps and may trash memory and hang system if
non-interleaved bitmap was allocated.
6) disables CyberGraphX detection for selected software like MooVId
PRO 2, that may be confused by absence of gfx card.

[Retrofan]

GOT IT !!

Using gfxroute on Aminet http://aminet.net/package/util/misc/gfxroute
and typing "gfxroute bitmap chip" it is working now with CyberBugFixAga

I am using his fblit config too. Edit: Not necessary

I have to investigate because I've got to open first the OCS version to get the AGA one working right. Edit: Not any more.

[MagerValp]

Nice, glad you tracked it down!

4) moves many bitmaps to FastRAM

OK, well that's obviously what's breaking things, as something isn't successfully patching BltBitMapRastPort to make up for it. But at the same time this is where the chip ram savings are.

Using gfxroute on Aminet http://aminet.net/package/util/misc/gfxroute and typing "gfxroute bitmap chip" it is working now with CyberBugFixAga

Doesn't this negate the chip ram savings, as it moves the bitmaps back into chip?

[Retrofan]

Originally Posted by MagerValp

Nice, glad you tracked it down!

OK, well that's obviously what's breaking things, as something isn't successfully patching BltBitMapRastPort to make up for it. But at the same time this is where the chip ram savings are.

Doesn't this negate the chip ram savings, as it moves the bitmaps back into chip?

I can't tell about the chip. It is saving it, then I use a script to launch your program and then I can not see the chip used, but I guess it doesn't decrease. Well, it doesn't decrease after I close AGS.

BTW I've seen there was a mistake in the icon.info (the tools I mean) of the AGS AGA setup and now it works perfect.

I will continue tomorrow. Thanks for your help

[MagerValp]

Ah, yeah, if you temporarily run gfxroute when starting AGS2, you'd still get the savings on Workbench. Glad you found a workaround!