General Category > LisaList2
LisaEm 1.2.7-RC4 support bug reports
stepleton:
D'oh, you called it! I just had half of a ROM sitting there in that file. No wonder! It works fine with the ROM from my Linux installation. Sorry for wasting your time!
(using it now) This is brilliant! It's nice to have the emulator working on this other machine.
--- Quote from: rayarachelian on February 20, 2021, 02:20:48 pm ---Yeah, no worries, I understand not wanting to release vaporware/unstable things, and also the need to secrecy. ;)
--- End quote ---
Much more boringly --- it's just because I need to get an IP release from my employer first. Shouldn't take too long, I hope!
rayarachelian:
No worries. :)
rayarachelian:
So I think there may have been some unicode fuckery that snuck into that class somehow, maybe an invisible non-space space char or something similar. I retyped the method name a few days ago by hand, exactly as it was before, but no copy paste, and into both the .h as well as the .cpp files and it compiled.
More importantly I found a couple more bugs around UniPlus and the vertical retrace. UniPlus very early on tries to run a 68010 only MOVEC opcode which needed special handling in LisaEm to work properly with the illegal opcode. Once that was fixed, I saw the kernel locking up in a very tight loop turning on the VTIR, then checking to see if the vertical retrace happened.
Originally, because it said so in the HWG, when the VTIR is enabled, I reset the video state machine timing, so it would always be in the same state when the VTIR was set, however, this was causing uniplus to hang. So I added a guard variable around it - so if you set VTIR repeatedly within something like 100 CPU cycles again, it will ignore it and let the state machine run its course.
That allows UniPlus's kernel to get a bit further, but now it fails with a profile state error. So looks like it doesn't like my profile emulation. Oh well.
Anyway for now I'll finish off the terminal stuff and testing and then release RC4. I'll come back to the profile emulation in 1.2.8 and hopefully that will get UniPlus going. Not sure what's up with Xenix after the kernel starts up. It sort of works and shows the console and cursor, but doesn't get any further than that. If you type stuff in, it will echo it as it should, but that's it. So it's waiting for something, not sure what. Not VTIR I expect.
Now with UniPlus since the kernel source is available, it will be somewhat easier to debug, and also a lot easier to intercept the pro.c code in the kernel with high level emulation code.
So that's where we are for now.
jamesdenton:
Awesome progress, and I appreciate your attentiveness to some of the issues I brought up recently. Thanks!
rayarachelian:
For the impatient, I've copied the current code into an unstable branch here: https://github.com/rayarachelian/lisaem/tree/unstable
This isn't a release, it's a nightmare with big nasty sharp teeth and big claws.
Do the usual git clone and then do
--- Code: ---git pull --all; git checkout unstable
--- End code ---
and then build it yourself.
It's more of a checkpoint before I start working on getting the UniPlus HLE stuff started on. After that's done, I'll come back to the TerminalWx stuff (which you shouldn't expect to work though it will compile and you can use as a console to talk to LisaTerminal.)
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version