Packed ScreamTracker 3 issue is fixed in last revision. Great job, Sarah. Looks like it should fix a whole bunch of stuff.
On that note, I just disabled emm386 and tested Fast Tracker 2. And PC speaker works fine . With enabled emm386 it hangs emulator. So can you please check FT2 with emm386 enabled ? It might be a similar problem, that you couldn't reproduce because of it.

The Incredible Machine 1. Problem with gfx. The bottom part is wrong and blinking, I saw it in a few other games as well, but here its the most obvious.

EDIT: Attachment deleted. Do NOT post copyrighted software here.

Still can't reproduce the FT2 issue, even with EMM386 loaded. More details about your setup please.

Regarding FT2. Host processor Core 2 Quad 6600 2.4Ghz. I guess its some timing issue. Another thing, for some reason when you setup soundblaster alongside GUS in FT2, GUS is sounds considerably louder. Maybe its supposed to be that way, i don't know. FT2 doesn't work properly in DOSbox, so i have nothing to test against.
Incredible machine 1 is fixed, nice! There is a similar problem with Dangerous Dave 2, which is not completely fixed. Its abit harder to reproduce, screen starts blinking after you jump a few times in the start.

I meant the setup in PCem. I don't really care what your host machine is.

About the cdaudio problem I have talked about earlier, here it is:
Hope it can be fixed as soon as possible.

from 00:00, it goes directly to 05:54 skipping the other seconds.

SarahWalker wrote:

I meant the setup in PCem. I don't really care what your host machine is.

I've removed soundblaster and gus from setup, to make it cleaner. This way FT2 defaults to PC speaker and mouse pointer hangs right after start up. Without emm its working fine.

Attached autoexec.bat and config.sys just in case. Using ctmouse.exe as mouse driver.

I have noticed that MS (not IBM) OS/2 1.0, 1.1 and 1.2 (not 1.21) don't work properly, same for IBM OS/2 2.0, which resets itself after inserting disk 1 in the setup prompt.

24bit/32bit color has some minor glitches in the right side of the desktop on Windows 3.1x and Windows 9x (ATI Mach64GX)

PCem should not use the host RTC as base time and date anymore because most betas of operating systems require changing the BIOS time and date.

some beta builds, notably 1500 of Win98, makes the emulator die for some reason when finishing setup (shortly before you get to the timezone config).

Betas.

With the stock drivers for both S3 Trio64 and S3 ViRGE on Windows 95 OSR2, I can't get dxdiag to run tests on DirectDraw or Direct3D! DirectDrawCreate() fails with HRESULT = 0x88760231

Alegend45, for some reason, 1024x768x32bit (same for 1280x1024x256) isn't available in the S3 on PCem and 1600x1200 has the mouse a bit busted.

Which S3? There are three different S3 chips/boards emulated now.

S3 Trio64 (N9) and Vision864 (Paradise)
EDIT: the mouse pointer in the S3 Virge in 1024x768x32bit looks strange.

1024x768x32 isn't available on the Vision864 or Trio64 as those cards don't have enough memory bandwidth to implement it.

thanks for these fixes and info Sarah.
Well, 1024x768x32bit, 1152x864x32bit (Mach64 GX only), on windows 3.1x or 9x, show some mess in the right side of the desktop, here's an example:

the DOS/V problem isn't fixed yet, but the OS works fine with /HS=ON but not without any parameters in jdisp.sys.
Considering that DOS/V 6.20 has the default setting in config.sys after it is installed, the part about the scrolling should be documented, so, is it possible to fix it too?

SarahWalker wrote:

1024x768x32 isn't available on the Vision864 or Trio64 as those cards don't have enough memory bandwidth to implement it.

It defintiely is, the original Windows 3.1 drivers for it list the mode, as does Windows 9x when you install the included S3 Trio64 driver. True though, by default it was sold with 2 MB VRAM or less, and therefore did not support that mode, but it was upgradable up to 4 MB VRAM, and at 4 MB VRAM it definitely supported that mode. The fact the VESA mode list returned by the BIOS, as well as the drivers themselves, all list his mode, clearly tells me the mode is supported.

Edit: FPM DRAM or EDO DRAM, not VRAM.

I also second the request to fix the problem with DOS/V (Japanese DOS). Basically, when JDISP.SYS (the DBCS display driver) is run in the default configuration (that is, without parameters), PCem shows a black screen. I think that's because it uses a 640x475 mode (since it uses a 8x19 font for the single-byte SBCS characters) that PCem doesn't recognize. Therefore, it seems PCem's hardware scrolling doesn't work well in that mode.