Latest WinUAE: yellow H2 in CPU % and hangs

[rsn8887]

When I load my .hdf from my old Amiga in the latest WinUAE, I always get a Yellow "H2" in the CPU % indicator and the system hangs during bootup. This is using a standard A1200 quickstart config, both in cycle exact and non-cycle exact modes.

This might not sound interesting by itself, but here comes the weird part:
I then go into WinUAE settings, select reset, and use the boot menu to "boot without startup-sequence" then in the shell I enter "execute s:startup-sequence" the system hangs in the same way, yellow H2.

I then perform a soft-reset, and the boot goes through without problems!!!!

This is exactly reproducible and always works.

This does not make any sense to me and I think this behavior hints at some emulator problem. The problem with obtaining more information is that it is a major pain to debug startup sequences on the Amiga, I would have to add "echo" statements to about 60 lines to figure out when the crash happens etc. I might do it later tonight due to post-xmas enthusiasm.
EDIT: I found out it is MCP. I cannot live without MCP, it does all my assigns (about 60 or so). What can I do?
EDIT2: I found the culprit: It was the "NoGURU" patch in MCP. After disabling that, the crashes are now history.


In addition (unrelated), when I display a hires interlace image during boot-up using syspic before the WB is displayed, then the mousepointer is distorted (see attachment). This distortion does NOT happen on real Amiga. This was introduced into WinUAE years back and I have had this problem ever since. The distortion of the pointer does not happen if I disable syspic. It is somehow related to showing a hi-res interlace startup screen during boot-up. But it does not happen on real Amiga.

EDIT: This might not be unrelated at all. If I open MCP Prefs, and click "save" without changing anythings, the mouse pointer goes back to normal. Afer a re-boot, it is distorted again.

[Toni Wilen]

First problem: need logs and config, as usual.

Second: It looks like what happens when sprite resolution is superhires but emulation "base" screen mode is hires: 2 shres pixels gets converted to single hires pixel. (can't fit 2 emulated pixels in place of 1 physical pixel)

You can enable superhires resolution mode but it does not make much sense in WB use except to slow down emulation. (only exception is WB doublescan hires modes)

[rsn8887]

If I enable superhires, then the mousepointer does not have missing pixels anymore. However it is still displayed as a superhires sprite (squished to half the size along the horizontal) when it should just be lo-res.

[Toni Wilen]

100% exact same HDF, 100% exact same ROM, chipset etc? (Not even tiny difference allowed).

[rsn8887]

Originally Posted by Toni Wilen

100% exact same HDF, 100% exact same ROM, chipset etc? (Not even tiny difference allowed).

Yes all I did was enable superhires. I will do some testing once I figure this prescale shader out completely.

[Toni Wilen]

I meant exact same as your real hardware. Including exact same HD contents. Image looks like mismatched chipset or KS ROM. (OS trying to show shres sprite without AGA)