BasiliskII: Win32 build can crash on startup

[DavidLudwig]

A user on the http://emaculation.com forums reported that Prerelease 3 crashed for them, on their Windows 10 machine.

Via http://www.emaculation.com/forum/viewtopic.php?f=6&t=9423 ,

"The windows build just shuts down for me - open it, it closes.

Using Windows 10 64b - just dropped you version over my old working one."

[c-hri-s]

Hi,

That was me :-)

I get the same behaviour with the new Pre-release 4.

I dropped the new BasiliskII.exe and SDL2.dll over the top of my previously-working version of BasiliskII - open it and the window opens and closes.

Video here, but it really doesn't add much I'm afraid: https://youtu.be/SRlEDm_BN80

[DavidLudwig]

Might you be able to try running Basilisk II from a command prompt window? If so, there is a chance that it might output something related to why the app closes.

[DavidLudwig]

Also, might you be able to post a copy of your prefs file? If not, no worries.

[c-hri-s]

Tried running it from a command prompt, it didn't output anything when it closed.

My prefs file is attached (renamed to .txt for Github's benefit)

BasiliskII_prefs.txt

[DavidLudwig]

I've posted a new build, Prerelease 5, which on Windows, will attempt to write diagnostic information to a text file, BasiliskII_log.txt. This file will get created in the current working directory, which'll probably be the same directory that the .exe is in.

Can you give Prerelease 5 a try, and if it still fails to boot, send me a copy of the log file? If so, I make no guarantees that it'll pinpoint the issue, but hopefully, will narrow it down a bit.

[c-hri-s]

Hi,

Thanks - log attached.

BasiliskII_log.txt

[cat7]

Hi c-hri-s,

The prefs file you tried to use originates from the old Basilisk 142 for windows. It contains a lot of prefs unknown to later BasiliskII builds.

[c-hri-s]

Hi,

That's true, and I saw those errors - but even with a basic preferences file I still get the error unfortunately (new prefs file attached, and new log)

BasiliskII_prefs.txt

BasiliskII_log.txt

[c-hri-s]

I've also tried copying only the necessary files to a new folder, with the same result:

---

-a---- 15/08/2017 22:14 1741824 BasiliskII.exe
-a---- 19/08/2017 21:21 356 BasiliskII_log.txt
-a---- 19/08/2017 21:17 470 BasiliskII_prefs
-a---- 19/08/2017 21:21 256 BasiliskII_xpram.dat
-a---- 11/08/2017 08:55 104857600 Cutie.dsk
-a---- 11/08/2017 08:55 524288000 Games.dsk
-a---- 11/08/2017 08:55 524288000 MacDrive.dsk
-a---- 11/08/2017 08:55 209715200 MacOS8.1.dsk
-a---- 06/12/1999 09:39 1048576 MacOSROM
-a---- 15/08/2017 22:02 791552 SDL2.dll

[DavidLudwig]

It looks a bit like an unimplemented opcode is attempted to get executed, which leads to the crash. I'm wondering what that is, if anything. Is there any chance that you might have a separate, bootable disk around, with which you could try booting off of, just to see if it works? Preferably, something with an older system version.

…

On Sat, Aug 19, 2017 at 4:23 PM, c-hri-s ***@***.***> wrote: I've also tried copying only the necessary files to a new folder, with the same result: `Mode LastWriteTime Length Name ------------------------------ -a---- 15/08/2017 22:14 1741824 BasiliskII.exe -a---- 19/08/2017 21:21 356 BasiliskII_log.txt -a---- 19/08/2017 21:17 470 BasiliskII_prefs -a---- 19/08/2017 21:21 256 BasiliskII_xpram.dat -a---- 11/08/2017 08:55 104857600 Cutie.dsk -a---- 11/08/2017 08:55 524288000 Games.dsk -a---- 11/08/2017 08:55 524288000 MacDrive.dsk -a---- 11/08/2017 08:55 209715200 MacOS8.1.dsk -a---- 06/12/1999 09:39 1048576 MacOSROM -a---- 15/08/2017 22:02 791552 SDL2.dll` — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#6 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAdC0Jb_TM7D5Xq1pS5-yaBg7Hp8UQxNks5sZ0QlgaJpZM4OzuFY> .

[c-hri-s]

If I use a blank prefs file (only with a link to the ROM):

rom D:\Downloads\Mac Emulation\PERFORMA.ROM

Then it still crashes with:

Basilisk II V1.0 by Christian Bauer et al.
Reading ROM file...
Using SDL/directsound audio output
FATAL: EMUL_OP called with bogus opcode 00005243
d0 000001ba d1 000068b8 d2 00000000 d3 0005001a
d4 00000000 d5 00000000 d6 00000000 d7 00000202
a0 00803378 a1 00803760 a2 fe118000 a3 008b34ce
a4 00809ae6 a5 008b35a2 a6 008b358c a7 0017ff7e
sr 2708

I tried a new ROM from http://www.redundantrobot.com/#/sheepshaver in case there was something wrong with mine (this one: http://www.redundantrobot.com/sheepshaver_files/roms/1mbMacrom.zip)

If I fire up "old" BasiliskII then it boots to a ?Mac icon as you'd expect.

[c-hri-s]

I tried it in a clean Windows 10 x64 VM, just to compare.
That errors as well:

#Basilisk II V1.0 by Christian Bauer et al.
Reading ROM file...
Using SDL/directsound audio output
FATAL: EMUL_OP called with bogus opcode 00005243
d0 000001ba d1 000068b8 d2 00000000 d3 0005001a
d4 00000000 d5 00000000 d6 00000000 d7 00000202
a0 00803378 a1 00803760 a2 ffbd8000 a3 008b34ce
a4 00809ae6 a5 008b35a2 a6 008b358c a7 0017ff7e
sr 2708

[DavidLudwig]

Thanks for trying the narrowed down setup! I'll try giving that rom file a go within the next week or two, and see if I can repro it.

[DavidLudwig]

I am now able to duplicate this bug, however, if I turn JIT off (by adding "jit false" to the BasiliskII_prefs file), the bug no longer reproduces.

I suspect this is a bug, or a missing feature, in the JIT code itself, rather than the SDL2 port. I haven't had a chance to try it using the non-SDL2 code in cebix's repository, though.

[c-hri-s]

Yep, confirm that works for me.
I guess my "old" basiliskII wasn't a JIT-enabled build.

[DavidLudwig]

It still sounds, to me, like a bit of a bug, or at least, an unimplemented and possibly worthwhile feature. Perhaps this is worth adding an issue to https://github.com/cebix/macemu ?