WinUAE 3.5.0 beta series (was 3.4.1)

[Toni Wilen]

Yeah, "new" FPU mode isn't fully working yet and there can be accidental side-effects with "normal" mode.

Hello.

Please add Visual Studio 2017 support, Because Visual Studio 2017 is already released.

Thanks

[Toni Wilen]

Hmm.. It has been 2017 compatible since about first 2017 test version. (and I don't remember there being any major issues)

[rbm]

Hello.
Starting from beta 2 my os39 setup does not boot and crashes. Works fine with beta 1 and old versions.
Attached dump file, logs and config file.

[Toni Wilen]

It probably is FPU related. (If you have software that needs FPU). Does winuae.7z work again?

Okay. Thanks.

[rbm]

Originally Posted by Toni Wilen

It probably is FPU related. (If you have software that needs FPU). Does winuae.7z work again?

Not works. Nothing to do with FPU software because the same config crashes exacly the same without hard disks.

Attached the configuration file without hds, dmp crash file and logs.

[Toni Wilen]

Should be fixed now. Multiple GFX boards got broken in b2 update.

[Toni Wilen]

http://www.winuae.net/files/b/winuae_3410b3.7z
http://www.winuae.net/files/b/winuae64_3410b3.7z

Beta 3:

- M-Tec AT-500 Megabody v1.33 ROM added. Slightly different than M-Tec AT-500 v1.33w ROM.
- Debugger now shows softfloat mode FPU register contents correctly, also both internal binary format and decimal value is shown.
- More FPU updates and fixes. Arithmetic/unimplemented instruction/datatype exceptions should be mostly complete now. (AG)
- FPU packed decimal format conversions are now accurately emulated (AG)
- JIT FPU works properly again.
- Bitplane overrun emulation was wrong if line didn't have any custom register changes.
- Tablet/touch screen lightpen emulation. (Game Ports panel -> Insert Touchscreen light pen -virtual device in either port) Only listed if Windows touch API or wintab compatible tablet or touch screen is detected.
- Added lightpen_crosshair config entry, set it to false (true is default) to disable yellow lightpen targeting cursor.
- Multiple gfxboards didn't work (b2).
- Fastest possible mode was very slow in some situations (b2)
- Slirp updates merged from Basilisk II.
- Final MSVC 2017 compiled.

M-Tec AT-500 v1.33 vs M-TEC AT-500 v1.33w:
- v1.33w is from external M-Tec AT-500, v1.33 from Megabody.
- w = word? v1.33w uses MOVE.W, v1.33 uses MOVE.L in unrolled transfer loop code.
- ROM has 2 16k banks, one for autoboot off mode, other for autoboot on mode.
- Oddly enough v1.33 and v1.33w have same autoboot off 16k bank. Unused driver code and version strings are 100% identical ("v1.33w").

[ross]

Originally Posted by Toni Wilen

Beta 3:
..
- Bitplane overrun emulation was wrong if line didn't have any custom register changes.
..

Hi Toni, sorry to bother you but 'overrun' act like before (copper operation required in same line).

Thanks for all.

EDIT: work if MMU disabled (how is this related?)

[Toni Wilen]

Originally Posted by ross

Hi Toni, sorry to bother you but 'overrun' act like before (copper operation required in same line).

Thanks for all.

EDIT: work if MMU disabled (how is this related?)

I am quite sure it has nothing to do with MMU but you have some option(s) enabled that don't support weird custom chipset features.

Also: Always include config file!

[ross]

Originally Posted by Toni Wilen

I am quite sure it has nothing to do with MMU but you have some option(s) enabled that don't support weird custom chipset features.

Also: Always include config file!

Thanks Toni.

[MagikGimp]

Hi Toni,

I'm determined to play Pool of Radiance but am encountering a crash occasionally, usually the first time I boot up WinUAE when selecting disk images to load. It doesn't always seem to happen a second time after that and Adobe Acrobat Reader DC (which I'm reading my manuals from) then ALSO crashes soon after too so that's what's making me think this is a Windows problem. I thought I'd mention it anyway though as you never know.

I recently attempted to install ADFView which as you probably know is very old software now and although the files copied, the installer hanged and I wasn't able to use the Uninstaller to then remove it either. It did seem to be causing me problems such as "This PC" (Explorer) also hanging but browsing other directories gave no problems. I've since removed the contents of the C:\Windows\System32\ShellExt folder and haven't had problems after doing so, aside from those crashes I describe above.

I'm running Windows 10 (x64) and attached three log files and the modified Windowed config file I use. I'm running from Quickstart on a standard A500 with 2 floppy drives.

Many thanks.

[Toni Wilen]

Sounds like this http://eab.abime.net/showpost.php?p=1049606&postcount=9
(and I don't think has nothing to do with 3.4.1 betas anyway)

[MagikGimp]

Originally Posted by Toni Wilen

Sounds like this http://eab.abime.net/showpost.php?p=1049606&postcount=9
(and I don't think has nothing to do with 3.4.1 betas anyway)

OK thanks; I'll try that. I was suspecting it was this but as nothing else was crashing apart from Acrobat (that one has me stumped) I figured there could have been a remote chance that it was a beta issue.
I can't replicate the crash today but I was also having trouble with a Task Held software error with the save disk that I thought I might also mention but I'll start another thread about that particular problem as that most certainly isn't a beta issue!

EDIT!: It crashed after around ten minutes of idleness after I went to check something! Just sitting at the software error screen as an unfocused task. Thought you might want to know.

[ross]

Hi Toni, a replicable crash.

An illegal custom setting:

	move.w	#$0018,$092(a5)		;DDFSTRT
	move.w	#$00e0,$094(a5)		;DDFSTOP  <---
	move.w	#$000f,$1fc(a5)		;FMODE

BPLCON0 shres bit, 8 plane.
Config the same as http://eab.abime.net/showpost.php?p=...4&postcount=32

I've to force closure from TaskManager, the error window is not responsive.

[for the casual tester, do not try on real machine...]

Bye!

[MetalliC]

Originally Posted by Toni Wilen

- Final MSVC 2017 compiled.

please don't.
2017 compiler is full of bugs, I've tried it in few my projects then reverted back to 2015 because of weird issues introduced. heard same feedback from other developers as well.
its not even RTM, it calls itself RTW - Release To Web.

[Toni Wilen]

I don't care what others say, I used prereleases since the beginning, no problems with winuae [most previous version betas were compiled with it!]. (above crash has nothing to do with compiler)

I hear same complaints every time new MSVC is released. I ignore them unless there are hard facts.

[MetalliC]

you are the boss here ) I just tried to warn and save some your time,
hard facts - number of reported (C-compiler related) bugs/issues at MS/technet site, so IMHO it is not (yet) production quality software.

[Toni Wilen]

Originally Posted by MetalliC

you are the boss here ) I just tried to warn and save some your time,
hard facts - number of reported (C-compiler related) bugs/issues at MS/technet site, so IMHO it is not (yet) production quality software.

Yes but exact same thing has happened every time new MSVC has been released and I haven't seen any problems so far. Also usually optimizer gets more and more aggressive, easily breaking almost correct code that has previously worked fine.

And finally, I want to live dangerously, I want to use bleeding edge software, even including operating system

Of course, if it really breaks winuae, I'll use previous version because it won't just affect me only.