mizapf

I suppose I should have a chance to check what is wrong, and if I cannot help, there is still enough time to change the emulator.

What about... Geneve QL* Geneve NT Geneve NG Geneve 10 Geneve 3k 🙂 (*Mind that a quantum leap is the least possible change of energy)

What is the "request" that you intend to make MAME mount a disk image? Or: What do you have in mind how this disk exchange should look like?

OK, I see ... you know, I usually disclaim having ever written those old versions ... If you run X11 (xwindows), why would you not run a desktop environment? You can autostart MAME so you won't see the desktop. I'd recommend to use a lightweight desktop like LXDE. You'll need some libraries like SDL and Qt anyway.

Could you send me the RPK please?

One advantage of using Linux for MAME is that the mouse pointer is not automatically captured. If you use Windows, you lose the pointer if you attach the mouse via "-colorbus mouse" (guess why I added this "deconfiguration" option; previously, the mouse was always connected). Nevertheless, you will need a graphic desktop. We recently discussed the other options of using the framebuffer directly, but it seems to be unsupported for years now. What performance do you get reported when you stop the emulation?

Just recently (it was in the context of copying Adventure files from cassette to disk) I learned that LOAD does not deliver the number of actually loaded bytes, which really surprised me. This would have been trivial to achieve in the DSR; the GeneveOS XOP, however, does return the image size.

I think we need a modernized version of ZeroZap. 😛 "ZeroZap - the way it should have been"

That was really synchronous.

This sounds interesting (page 98)

Not my note, but I read it that way as well.

Here is a copy of the V9938 spec.

What I found particularly handy with the SCSI2SD is that I can take the SD card out of the adapter and plug it in my PC card reader. There, I can access it as a raw device in Linux (using the dd command, which is also available for Windows). Using TIImageTool you can work on this dump directly.

One of the most impressive pieces of code that I saw was Thierry's bootstrap code for his IDE card, using the NVRAM of the real-time clock RTC-65271. The RAM of the clock is organized as 128 pages of 32 bytes, so you really cannot write longer sequences of code, and you always have to map into the next page. Nevertheless, he managed to install a bootloader from several of these 32-byte pages into main memory which downloaded a second-stage bootloader, which in turn pulled the DSR from the IDE drive into the IDE DSR space, and then comes the Master title screen. By that concept, the IDE card does not need any EPROM or FlashROM. See also this posting:

The easiest way would be to download one of the zip files from WHTech (https://ftp.whtech.com/emulators/MAME/div/) for the HSGPL (40col or 80col) and to unpack them in the nvram/ti99_4a folder (40col) and nvram/ti99_4ev (80 col). You must put the HSGPL into slot 5 (or rename the nvram files appropriately).

If you miss your kilos, I seem to have them here, right on my belly. Just give me a call if you want them back.

RPK for the game. skyway.rpk

Please report your issues directly to me, including the operations that you do and the expected outcome. It's on me to fix the issues. All tests that I tried seemed to work correctly, but I may have missed something.

Hmm, we haven't had Car Wars in our monthly competition yet, right? 😛

Hmm, if we're going for a new game every month, we'll have had them all soon. If I remember correctly, Centipede was twice (and I won both 🙂 ). This brings back my idea of another competition (in parallel to this one) where we're playing the same games every year, like a championship. We could have different leagues like "classic pre-1982", "third-party", or "post-2000". You'll get a score for your result (winner = 12 points, follow-up = 10 points etc.) which you have to defend every year (like the Tennis ATP ranking).

R0 = >8304 R1 = >8302 R2 is unchanged On my (real) Geneve:

Interesting! Would you mind sharing this test program as well?

The Editor/Assembler manual states on page 89: If the destination operand is specified as Workspace Register 15, the first word of the destination operand is Workspace register 15 and the second word of the destination operand is the word in memory immediately following the workspace area.

From what I saw in the source code, the operation is a division of R15 by R14. The problem is that the dividend consists of the two words starting at the specified address, which means R15 and the subsequent word (R16, which does not exist). So R15 is at >A09E and contains >0000, the subsequent word at >A0A0 is >1000, and R14 is >0010. I'd read this as a division >00001000 / >0010, and this should be >0100, remainder >0000. So why this is wrong?

This is the microcode that I implemented in MAME: REG_READ, ALU_SETADDR_ADDONE, ALU_ADDONE, REG_WRITE, MEMORY_READ, RET for the data derivation sequence for *Rx+, according to the description in chapter 4 of "9900 Family Systems Design" (see attached copy). This is not a bug, but it may be a surprising effect indeed.