WinUAE 1.5.1 beta series

[Toni Wilen]

Originally Posted by wizzel

Hi there (Toni),
i experienced a weird bug in WinUAE 1.5.1 beta 4.
When opening a .uae config-file and changing the following line to:

floppy_volume=25

and loading the config into WinUAE, it says in the 'sound'-tab:

floppy_volume=75

So i guess the internal 'handling' of the 'floppy_volume' variable
is inverted

I'd call it a very minor bug, but maybe Toni wanted it to be that way?
Maybe someone can clear things up?

It is by design. GUI shows "user-friendly" values, configuration file isn't meant to be friendly

btw, this has been since floppy sounds were implemented ages ago..

Okay for me, but from a programmers standpoint it's not ok, because other values in .uae-files are translated 1:1.

[crusader86x]

hq4x filter still crashs on every configuration I choose.
I using only Quickstart options.

[Toni Wilen]

Originally Posted by crusader86x

hq4x filter still crashs on every configuration I choose.
I using only Quickstart options.

Useless without crash dump and display and filter settings. (quickstart only sets emulated hardware, not host settings)

EDIT: include both logs, configuration used and crash dump, thanks.

[Toni Wilen]

Originally Posted by wizzel

Okay for me, but from a programmers standpoint it's not ok, because other values in .uae-files are translated 1:1.

They are not, try ram size settings for example. Some are 1:1, some are not.

hi tony i have some trouble running this demo DMOB-M4A.dms
try it after the initial decrunch the sistem display the adorable request of guru meditation ofcourse in winuae 14xx work fine ............

[Toni Wilen]

Originally Posted by ramkiller

hi tony i have some trouble running this demo DMOB-M4A.dms
try it after the initial decrunch the sistem display the adorable request of guru meditation ofcourse in winuae 14xx work fine ............

Could you upload it to the zone or email it? (I want to be 100% sure I have the exact 100% same file)

Did you use A500 quickstart setting?

Hello! I'm totally new here, but I found one problem in all latest betas and pointer under WinUAE 1.5.1 on desktop in Picasso mode. In short term I'll try to check if this problem occurs on other GFX chipsets, but I raise it right now here, maybe somebody know something about it. Problem is that around mousie pointer I get pink filled rectangle which blinks heavily when I move cursor. Maybe this is some incompatibility with my hardware or something, or maybe some nasty bug inlatest betas of winuae (I'm sure that to version 1.4.6 everything was perfectly fine in that matter, problems starts from 1.5.0). Some tech specs of my machine: Toshiba A135, gfx Intel GMA 945. I cannot prepare screenshot because pointer on screenshots not present, maybe I will take a photo of that. Regards MandiATO

[Toni Wilen]

Originally Posted by MandiATO

Hello! I'm totally new here, but I found one problem in all latest betas and pointer under WinUAE 1.5.1 on desktop in Picasso mode. In short term I'll try to check if this problem occurs on other GFX chipsets, but I raise it right now here, maybe somebody know something about it. Problem is that around mousie pointer I get pink filled rectangle which blinks heavily when I move cursor. Maybe this is some incompatibility with my hardware or something, or maybe some nasty bug inlatest betas of winuae (I'm sure that to version 1.4.6 everything was perfectly fine in that matter, problems starts from 1.5.0). Some tech specs of my machine: Toshiba A135, gfx Intel GMA 945. I cannot prepare screenshot because pointer on screenshots not present, maybe I will take a photo of that. Regards MandiATO

Vista? Intel laptop graphics chips seem to have directdraw destination color key driver bug. You can try adjusting Display Buffer setting in Misc-panel, it may or may not help..

This is confusing problem because directdraw color key support is very basic feature..

Originally Posted by Toni Wilen

Vista? Intel laptop graphics chips seem to have directdraw destination color key driver bug. You can try adjusting Display Buffer setting in Misc-panel, it may or may not help..

This is confusing problem because directdraw color key support is very basic feature..

Yep. This is Vista 32 bit Home Premium. And strangest thing was that that to version 1.4.6 everything works ok. since 1.5.0 it starts to draw that pink-filled rectangle around pointer, and I cannot figure out how to remove it.

I tried different settings right now in misc panel - display buffer - but this niot help. New version of WinUAE is much faster in GFX drawing, but I will be happy if I can help you to nail down that small bug. This is not so important I can live with that, but irritating a little sometimes, anyway I almost all the time use keyboard and pointer is off the screen (hide pointer when typing ewnabled in system). Maybe this is conected some way with Hi-Res pointer in system also? Any ideas what to check, or did you need some more infos?

Regards

MandiATO

Originally Posted by Toni Wilen

This is confusing problem because directdraw color key support is very basic feature..

Strangest thing... Change GFX mode from 32bit BGRA to 16bit PC resolves that issue. Something goes wrong in 32bit modes on Intel GFX chipsets.

Regards
MandiATO

[Toni Wilen]

[edited]I was too confused, WinUAE uses Source Color Key and it is reported properly by the driver[/edited]

I guess it is best to use "software emulation" when drawing cursor. It will fix this problem without measurable performance difference, even with drivers that have bugs..

[Toni Wilen]

Does attached version fix the color key problem? It has color key test that hopefully notices the bug (it is still possible the bug does not happen in test blit..) and falls back to software mode, there should be message in winuaelog.txt.

yes i use the quickstart mode 1.3 ocs 512k slow most common
and all other quick settings but the demo refuse to start .....
mhhh.... now i try to undms the disk if find the utils ....
take a look to my attachment ........

see ya soon dudez.............

hi tony i search all the net ....
and download a new one it seems to be work .....
the other have a strange mistake i test it whit dms pro and pass the test
but when i decrunch it the dms break on the track 13 and show me a crc error ...
the second disk work ''load'' better
sorry for my lame english ......

there is a bug in dms manager routines in winuae or the disk image is corrupt ???


ciao ....

Hi Toni!

Originally Posted by Toni Wilen

Does attached version fix the color key problem? It has color key test that hopefully notices the bug (it is still possible the bug does not happen in test blit..) and falls back to software mode, there should be message in winuaelog.txt.

Partially, to be exact. I can switch to 32 bit mode and there is no pointer bug anymore. But bug remain in other case. I will try to describe it with details. When I have goto opened 1280x800x16PC mode and in screenmode prefs I selected 1280x800x32 BGRA, and use Test, on test screen everything was OK, but when system backs to WB, there,s blue rectangle at mousepointer. But switching to 32RBGA mode and using it goes without problems. (winuaelog.zip).

But to this work I need to switch between 2 screenmodes. Saved prefs with 32 BGRA on first boot results with pointer bug. (winuaelog_firstboot.zip). It starting working as it should after switching resolution. And using "Reset" when it working OK don't bring back error.

Logs attached.

I know that resolving so virtual problems might be problematic, so if you need any more hints, pointer anything feel free to ask. I want to help with that because I use it heavily to develop translations on my laptop (OS4 and so on).

Regards
MandiATO

[Toni Wilen]

Originally Posted by MandiATO

Saved prefs with 32 BGRA on first boot results with pointer bug. (winuaelog_firstboot.zip)

Color key blit seems to work properly at start (when the test is run) but it breaks later... I moved the test in different place.

Attempt 2: test modified, 16-bit software color key fixed.

(Intel most likely releases fixed driver when workaround is working properly..)

[crusader86x]

Originally Posted by Toni Wilen

Useless without crash dump and display and filter settings. (quickstart only sets emulated hardware, not host settings)

EDIT: include both logs, configuration used and crash dump, thanks.

Here are the files you requested

When using hq3x my PC really slows down, but I don´t know why - I have got an AMD Athlon 6000+
And when choosing hq4x WinUAE crashs

[Toni Wilen]

Originally Posted by crusader86x

When using hq3x my PC really slows down

Not again
Filters were originally created for 8/16-bit console emulators with small resolutions, like 320x200, not for full overscan hires Amiga resolutions. Filters in "high" resolutions (especially if it is >2x) require extreme CPU power and huge memory bandwidth.

Anyway, I can duplicate the crash now. btw, >2x is not normally needed.

ADDED:

Note that most filters appear to do nothing if Amiga-side mode is in lores and hires+doubling is enabled because filtering is done after uae's internal doubling (filter sees 2x2 single colored pixel blocks that mess up filter's edge detection logic). Filters are more useful if lores is enabled and doubling is disabled. (but now this is getting out of topic already..)

ADDED even more:

2x filter when 800x600 with hires and doubling = 1600x1200 internal resolution. 4x would be 3200x2400.. Brute-forcing with CPU isn't going to make it

Originally Posted by Toni Wilen

Color key blit seems to work properly at start (when the test is run) but it breaks later... I moved the test in different place.

Attempt 2: test modified, 16-bit software color key fixed.

Sorry, but bug still remains. And it shows only on 32 bit screens on 16 bit ones it always was OK. Temporary solution for me is to stay with 16 bit screen (what is not big problem for me). Boting to 32 bit BGRA mode results with pink filled rectangle around pointer qwhich heavily flicks when I move mouse pointer. If I switch to 16 bit mode and back to 32 bit one, everything works OK. Next run of WinUAE and pointer backs to pink-rectangle state. (log dfrom booting to 32-bit mode attached).

Originally Posted by Toni Wilen

(Intel most likely releases fixed driver when workaround is working properly..)

So maybe all that effort is not worth of... Maybe only put in readme info about troubles on some intel GFX chipsets under Vista and advice to temporarilly set 16-bit screenmodes.

Regards
MandiATO