mamesick

IIRC that neogeo raster hack was implemented upon request to "fix" some gfx glitches that happens on real hardware. The sengoku2 check can go away, yes, but the rest should stay, mostly the "screen().update_partial(scanline + 1);" is important because fixes the position of the raster beam in a lot of games like for example Zed Blade and golf games. The check if raster irq is enabled (m_neogeo_raster_hack) could go away too but in this case the glitches mentioned above will appear again, mostly noticeables in Shock Troopers 2nd Squad intro. It's up to you, Robert, there's no problem at all if you want to remove the code, I simply explained what it does.

It won't happen. Too much bad things happened when I was the mantainer of MAMEUIFX and the haters are still there. So feel free to leave if you wish. It's an hobby, not a paid job.

Revert modified drivers to original state is actually not a good practice, the update to new standards can be easily done with programs like WinMerge or others side-by-side diff tools. Though I don't want to criticize, I perfectly know how much time-consuming is keep alive ARCADE. Another change that seems gone away is the correct refresh rate for CAVE.CPP games, it's reverted back to the 271.5 divisor, instead of 272.0. SSV.CPP resolutions as already pointed out by haynor666 are reverted too. IIRC there are 23 resolution changes for that file, maybe I'll do the job by myself and upload the updated file on my mediafire account.

I haven't understood how the new mechanism works too but this should work for example for thunderx.cpp subdevice<screen_device>("screen")->set_visarea(14*8, (64-14)*8-1, 2*8, 30*8-1); You must put this in the derived machine_config you want to modify. In thunderx.cpp for example you should find SCONTRA(CONFIG) in the derived machine_configs and add the line. In this way you can also change refresh rate and so on. Hope this helps.

Just did some tests on my side here, the new patch is fine. I don't see any real breakage in games that uses this CPU.

The fix for darkseal and vaportra is no more valid. The function that is commented out is now declared internally in the H6280 CPU core and is always enabled. So those extra lines listed in 202 diff are useless now. In file h6280.cpp infact there is: void h6280_device::internal_map(address_map &map) { map(0x1fe800, 0x1fe80f).mirror(0x3f0).rw(FUNC(h6280_device::io_buffer_r), FUNC(h6280_device::psg_w)); map(0x1fec00, 0x1fec01).mirror(0x3fe).rw(FUNC(h6280_device::timer_r), FUNC(h6280_device::timer_w)); map(0x1ff000, 0x1ff000).mirror(0x3ff).rw(FUNC(h6280_device::port_r), FUNC(h6280_device::port_w)); map(0x1ff400, 0x1ff403).mirror(0x3fc).rw(FUNC(h6280_device::irq_status_r), FUNC(h6280_device::irq_status_w)); } So the only way to enable again the fix would be comment out the line here but with the serious risk to break a lot of other games that uses the same audio CPU. Robbbert should remove both VAPORTRA.CPP and DARKSEAL.CPP files from his source. The code there is no more valid.

Remove that useless NPLAYERS.INI because all you need is MULTIPLAYER.INI as Robbbert already told you.

This is good.

Nobody fixed these bugs yet in the daily builds. I posted here for Robbbert, it will be up to him decide what to do. For reasons I cannot (and don't want) to explain, I don't contribute to the official MAME Development.

http://mametesters.org/view.php?id=7075 http://mametesters.org/view.php?id=6970 http://mametesters.org/view.php?id=6932

Linked an M72.cpp file which contains the fixes for the MAMETesters bug listed in topic title. They are very simple and small changes, easy to be mantained if you'll decide to not submit them to the official tree. http://www.mediafire.com/file/bn8ebnpgqlmbnmc/m72.7z/file

To compile ARCADE 0.200 you need to add #include "romload.h" in the MAME headers listed in WINUI.H (add it just after pool.h)

Life (and releasing ARCADE build) must be easier so go ahead with official code. Maybe you could keep a copy of old code for future use, who knows what will happen.

In my opinion Kale fix is a nasty hack. ARCADE code for ironhors is more accurate emulation and it's based on gberet code which was verified to be correct. Hardware is pretty the same, memory maps, etc. etc.

About the VR/VF bug: http://mametesters.org/view.php?id=7015