Katerpillar Attack Cartridge for the Tandy CoCo

[zephyr]

Here is a copy of the Katerpillar Attack cartridge for the Tandy Color Computer.

Katerpillar Attack on YouTube:
http://www.youtube.com/watch?v=DQHCCuWxTkY

[Rolo]

Sure, that this is a cartridge? Reading the assembly-code I would think, it's running somewhere between $2500 and $4000
I would assume it's a disk file?!

[zephyr]

Sure, that this is a cartridge?

Yes.

[Rolo]

Thanks for the info!
Is it still the raw rom image or somehow converted/processed?

[zephyr]

Raw ROM image.

[Rolo]

Yes, I see it now. I can launch it with XRoar (Dragon32 and Dragon64), but real Dragon crashes with garbage on screen. Don't know why, yet. It's the first programme I found, that runs in the emulator, but not on the machine...

[sixxie]

Are you sure? Under emulation I see it do some sanity checks, then when the RESET vector doesn't compare to what's expected on a CoCo, it jumps to an infinite loop (c01b BRA $c01b).

Runs fine on CoCo (only tried emulation), but I wouldn't expect garbage on a real Dragon, just a blank green screen.

[Rolo]

Yes, I did not mention, I'm not working with the original rom anymore, but I'm changing the rom step by step until it (hopefully) works. I've already resolved the endless-loop at the start of the programme. But obviously not right, yet:-) Unfortunately I do not have time now...

[sixxie]

Ah right, should have got that from the context of your other posts.

[Rolo]

Still fighting with this one! I've analyzed the code a bit.
They start with some checks (cold start, top of ram at least 16k, CoCo reset vectors). If there is anything, that looks like a copy or not a CoCo, then they simply jump into an endless loop . Then there is a block copy (c03f...d539 to 2906...3e00). So almost the entire cartridge is copied into ram. That's why it looks like a non cartridge game, what fooled me before. After enabling interrupts CA1/CB1 the program jumps into ram.
They jump around a lot. It starts with clearing video ram ($1000...$2800). Then they set PMode3/green with PIA2 Port A (11100000) and SAM bit 1 and 2. After settings some program internal values, they clear the video ram a second time. Now they set the base address of video ram to $1000 (SAM: ffc6,ffc8,ffca,ffcd,ffce). Then, I believe, they jump to the displaying routine of the title screen. I can't see any mistakes, and the program is working in XRoar.
In the real world, my dragon crashes with garbage characters on the screen. It's alphanumeric, not pixels, so switching to the graphics mode seems not to work properly. The screen is not cleared, so setting the base address is not working too (, or clearing the memory). No sound and nothing changing on the screen. I wonder what the processor is doing...

Maybe I'll hook up a logic analyzer (quite some work) or I copy parts of the code, put it on "floppy disk" and load directly into ram. Then I do not have to erase and burn an eprom for each test...