AtiRageGT issue with 9.1

[Randrianasulu]

Describe the bug
After installing 9.1 window content and drop-down menus sort of transparent, I barely can shut down machine

To Reproduce
Steps to reproduce the behavior:

1. Go to https://winworldpc.com/product/mac-os-9/91 and download 7z compressed iso
2. install 9.1 on 500 mb drive (1 partition)
3. boot like this: bin/dingusppc -b ~/K38_sdcard1/Documents/1997-11\ -\ 78F57389\ -\ Power\ Mac\ G3\ (v3).ROM -r --rambank1_size 64 --hdd_img 9.img
4. See error.

Expected behavior
Image after installation works like in installer, no transparency (?) bug.

Screenshots

`

Device Information (please complete the following information):

• Device Type: Desktop

• Host OS: NetBSD/amd64 10.1

• Host Processor: AMD FX4300

• Host RAM: 16GB

• Guest OS: MacOS 9.1

• Guest Processor: PowerPC G3 (default)

• Guest ROM: Power Mac G3 Beige (autodetect)

• Guest RAM: tried 3*256 and 64 MB

• Build Date: Mar 13 14:06 (selfbuild)

• Version commit 77503df (HEAD -> master, origin/master, origin/HEAD)
  Author: dingusdev 52434309+dingusdev@users.noreply.github.com
  Date: Wed Mar 12 07:58:32 2025 -0700

  Code clean-up

Additional context
I applied patch based on compiler (gcc 10.5.0) advice:

diff --git a/thirdparty/loguru/loguru.cpp b/thirdparty/loguru/loguru.cpp
index 7ebbe6e..8880574 100644
--- a/thirdparty/loguru/loguru.cpp
+++ b/thirdparty/loguru/loguru.cpp
@@ -1101,7 +1101,7 @@ namespace loguru
                                long thread_id;
                                (void)thr_self(&thread_id);
                        #elif LOGURU_PTHREADS
-                               uint64_t thread_id = pthread_self();
+                               uint64_t thread_id = (uint64_t)pthread_self();
                        #else
                                // This ID does not correllate to anything we can get from the OS,
                                // so this is the worst way to get the ID.

Bug exist in all color depths available from control strip, even in Grey256!

[Randrianasulu]

dingusppc.log.txt

log file

[joevt]

ATI 2D acceleration isn't implemented (yet). Move the "ATI Graphics Accelerator" or "Graphics Accelerator" extension from the "Extensions" folder to the "Extensions (Disabled)" folder.

DingusPPC will eventually have a custom graphics controller that can be used instead of ATI. It will support arbitrary resolutions and bit depths and number of displays. It may get its own 2D and 3D acceleration options.

[dingusdev]

Short-term answer: What joevt said

Long-term answer: The Mach64 card emulation is a little unfinished. Judging by the screenshot, we most likely would need to implement the GUI engine. It's especially obvious any time it tries to render a rectangle (that's most likely why it's text floating over a transparent background).

[Randrianasulu]

ATI 2D acceleration isn't implemented (yet). Move the "ATI Graphics Accelerator" or "Graphics Accelerator" extension from the "Extensions" folder to the "Extensions (Disabled)" folder.

DingusPPC will eventually have a custom graphics controller that can be used instead of ATI. It will support arbitrary resolutions and bit depths and number of displays. It may get its own 2D and 3D acceleration options.

ah, thanks. Just right now for some reason this install fails to come up with extensions disabled:( It says "extensions off" if I boot with shift key down, but then just hang at starting up stage ...

[Randrianasulu]

Update: moved all ATI extensions via booting another 9.2 install with this disk attached to qemu, now it works :)

I wonder if I should report Quicktime 4 hang (seemingly) on starting up while trying to play sample movie, or this is too advanced bug for this alpha stage?

Congratulations on reaching this milestone!

[dingusdev]

My guess with QuickTime is that there might be some issue with sound too, but it's tough to say at this point. A lot still needs to be implemented.

[Randrianasulu]

sound from control panel test sounds sort of work .. not clear, but it does not hang.

[JosephM101]

DingusPPC will eventually have a custom graphics controller that can be used instead of ATI. It will support arbitrary resolutions and bit depths and number of displays. It may get its own 2D and 3D acceleration options.

The idea that Mac OS 9 (and hopefully OS X) could be 2D/3D accelerated in the future with a custom graphics controller in a virtual machine is quite exciting! I will be following this project more closely.
The progress might be really slow right now, but at least there is progress.