WinUAE 4.4.0 beta series

[Viceroy]

Well Done Toni ;-)

[DamienD]

Yup, many thanks Toni for your continuous improvements, bug fixes and 24 / 7 support

I really don't know what people into Amiga and emulation would do without you at the helm...

It will be a very sad day indeed if / when you decide to throw in the towel; and I don't think anyone would have the necessary skills / knowledge to follow on.

[Viceroy]

Well said DD!

[freddy]

Toni could you create an entry for the translation of this option.

[jotd]

I'm using beta 5 and I can't get messages displayed when writing to $BFFF04/08

Here's the code I'm using

    movem.l A0,-(a7)
    lea arg_array(pc),a0
    move.l  (A5),(a0)+
    move.l  8(A5),(a0)+
    move.w  12(A5),(a0)+
    lea arg_array(pc),a0
    move.l  a0,$BFFF08
    lea arg_entry_format(pc),a0
    move.l  a0,$BFFF04
    movem.l (a7)+,A0
    RTS

arg_array
    dc.l    0,0,0,0
arg_entry_format:
    dc.b    "Entry: %lx Size: %lx Type: %d",0
    even

the code is executed but

- nothing is written in the console or in the log
- memwatchpoints don't even detect writes to $BFFF04/08

too bad as this feature is just top notch.

[Mclane]

Originally Posted by DamienD

Yup, many thanks Toni for your continuous improvements, bug fixes and 24 / 7 support

I really don't know what people into Amiga and emulation would do without you at the helm...

It will be a very sad day indeed if / when you decide to throw in the towel; and I don't think anyone would have the necessary skills / knowledge to follow on.

We can't let this happen, move him to the Presidential bunker and chain him to a desk!

Just joking Toni, you (and your family) stay safe and well, we all need a rest and you deserve it 1000000%

Thank you...

[Toni Wilen]

Originally Posted by jotd

I'm using beta 5 and I can't get messages displayed when writing to $BFFF04/08

"- Added misc panel enable checkbox for beta 1 introduced debug logging. Note that quickstart mode always disables it." ?

[jotd]

sorry can't find anything related to check in "misc" panel. There's "Debug memory space" but enabling this doesn't do anything.

Is it supposed to write in the debugger console? winuaelog.txt file?

using beta 5.

(and memory watchpoints don't work in that zone, when they work with the same config in BLTDAT)



Back to logging in memory in the meanwhile...

[Toni Wilen]

Your code works for me. More information needed. Some weird or special config? (EDIT: and you did tick the checkbox before starting emulation?)

[jotd]

ok, nvm, I'll try to find a reproducible scenario.

I found another bug in b5, that one is pretty easy to reproduce

- start emulator
- shift+F12 to enter debugger
- start trainer: C 3
- lockup!!

just tested with b3 and it worked. I suppose something went wrong between b3 and b5

[Toni Wilen]

http://www.winuae.net/files/b/winuae_4400b6.7z
http://www.winuae.net/files/b/winuae64_4400b6.7z

Beta 6:

- Small 68030 prefetch mode tweak.
- Enable cycle accurate Paula DMA request emulation if more compatible is ticked. Previously either cycle-exact option was required.
- GUI reset button didn't do full hard reset, for example rom images were not reloaded.
- GUI configuration Delete button can be now used to delete empty configuration directory.
- b5 "Debugger fa and s commands skipped chip ram." fix was incomplete, caused some debugger search commands to hang in infinite loop.
- Fixed *.uae icon, it was most likely accidentally changed long time ago to winuae.exe icon without no one noticing.

[ross]

Fixes greatly appreciated, thank you!

[DamienD]

Awesome as always, many thanks Toni

[DamienD]

Originally Posted by Toni Wilen

http://www.winuae.net/files/b/winuae_4400b6.7z
http://www.winuae.net/files/b/winuae64_4400b6.7z

Beta 6:

- Small 68030 prefetch mode tweak.
- Enable cycle accurate Paula DMA request emulation if more compatible is ticked. Previously either cycle-exact option was required.
- GUI reset button didn't do full hard reset, for example rom images were not reloaded.
- GUI configuration Delete button can be now used to delete empty configuration directory.
- b5 "Debugger fa and s commands skipped chip ram." fix was incomplete, caused some debugger search commands to hang in infinite loop.
- Fixed *.uae icon, it was most likely accidentally changed long time ago to winuae.exe icon without no one noticing.

Oh no, Hexen is back to crashing every single time

...even with my new config using a 68030 CPU, which would only crash 1 in 10 times with previous betas.

The old config was using a 68040 CPU.

Edit: I've now changed to a 68020 CPU and all seems good again...

[Toni Wilen]

Unless you had 68030 + more compatible (which makes no sense with PC ports of programs or with JIT), there is no way b6 updates would have any effect.

[DamienD]

Originally Posted by Toni Wilen

Unless you had 68030 + more compatible (which makes no sense with PC ports of programs or with JIT), there is no way b6 updates would have any effect.

Yup, I had "More compatible" ticked...

Let me try without, thanks

[DamienD]

Originally Posted by Toni Wilen

there is no way b6 updates would have any effect.

To be honest; not sure I tested thoroughly with b5.

Anyway, I've now switched off "More compatible".

Out of 10 tries the game still crashed 5 times; always within the first 10 - 15 seconds of play.

I assume you still have the .HDF I sent you?

I've attached my adjusted configuration file in case there's anything else you recommend that I should adjust?

I do need an FPU though for "Heretic".

[Toni Wilen]

Sorry but I have almost zero interest in running or debugging Amiga programs in emulation if the game is boring straight PC port.

[DamienD]

Ok, that's fine; and I know a lot of changes have occurred.

I'm just trying to work out why:

• The game was stable / rock solid in v4.3.0 and never crashed once.
• In beta 4 it crashes 1 out of 10 times (after changing from a 68040 CPU to a 68030 CPU).
• Now with beta 6 it crashes 5 out of 10 times

[Toni Wilen]

It does not prove anything as usual. If it is the only (similar) game crashing: it probably has some hidden bug. Or it is some weird JIT issue which is not my problem.