PONG: 4A A4 4B A5 4C
RushHour: 55 AF 56 B0 57
FONT: 57 B1 58 B2 59
GrayMine: 48 A2 49 A3 4A
mt3: 36 95 37 96 38
PI: 4C A6 4D A7 4E

So these are all following the same pattern:
05 1F 1F 1F 1F 1F 1F 1F 1F 01 XX XX+5A 01 XX+01 XX+5B 01 XX+02

*bump* At last, progress! When Beta7 was walking me through the exact replication process, he mentioned that he always ON-PRGMs to launch Doors CS the second time. Lo and behold, SUCCESS! In reproducing the bug, at any rate. No one mentioned that the cursor ended up in pixel-increment mode rather than character-increment mode, that is key. I don't get the same character sequences, but that's irrelevant. Now, for finding a fix!

Edit: On 2.53MP, the following manual fixes seem to do the trick. The main problem appears to be that MathPrint is still active during the Input command when ON-PRGM is used to launch Doors CS. Solutions:
- reset bit 2, iy+68
- reset bit 5, iy+69
- set bit 7, iy+73

Final EDIT: Please please please try this out before tomorrow afternoon, regardless of whether you have MP or not. I'd like to release 7.1 by Tuesday at the latest.

http://www.cemetech.net/files/dcs71b3.8xk

Yay! New beta!

Kerm, I don't know what more to tell you about the prgmCAVE glitch I had. It occurred two or three times so I deleted the game. I don't really want to resend it.

It triggered a TIOS error when I pressed Up. prgmescs was always in RAM, and pressing Del on it from the Memory Management/Delete menu gave me a RAM Clear.

New beta!

Transfering to calculator already

I can't. I have a problematic password protect on my calculator how do you RAM clear? (yes i realize this sounds stupid.)

[2nd] [+] [7] [1] [2]

Oh, huh! Thx Beta7!

Well, I already found a glitch, but it's not a big deal, the program says it's Beta 2, but that doesn't matter xD

It doesn't sound stupid; we're not all born knowing how to make the calculators do our bidding. [2nd][+] brings up the Memory menu, [7] is Reset, [1] is All RAM, and then [2] confirms.

@ztrumpet: I suspect that that prgmCAVE may have a glitch in it.

@ScoutDavid: I look forward to hearing your results. I was in a hurry, because I want to release Doors CS 7.1 today or tomorrow if I get the green light from you guys, so I didn't bother fixing that.

Me too, but I don't like the fact that Doors can leave a program in RAM that RAM Clears my calc when it gets deleted.

It's not the program; it's the fact that your RAM structure is wrong because DCS didn't get to swap back from $9d95. The RAM program is just a symptom of that.

Has anyone tested this yet?

I can confirm that this fixes the BASIC/ASM program running/crashing bug.

The evil program corruption bug isn't fixed yet, but I won't demand anything until I get time to record a video of the bug in action, plus post it as well!

I've had more than one user now confirm replicating that bug without Doors CS, so I'm breathing a little easier now. I'm going to push out Doors CS 7.1 without a fix for that, but I'll make a note in the readme about 2.53MP and how much it fails, and I'll keep searching for a fix or a workaround once I see your video. Thanks for testing this, I appreciate your effort at what I'm sure is as busy a time for you as it is for me academically.

Edit: For everyone still concerned about the Hourglass "bug", I have now institute a proper fix using IconSpace32b, which I hope to hell I'm not trying to use elsewhere. Unfortunately, I had to optimize out another 15 bytes or so to jam this in, so I have the following delightful stats at 3am:


Code:

Wow, Kerm. It's pretty amazing you're able to cram that all in. I think that your next thing should simply be to convince TI to fix the OS so that it doesn't cause errors.

Which parts of DCS are currently in which pages?

Page 0 is the "main" code, especially program running, writeback, properties, CALCnet, and all that. Page 1 is almost entirely GUI libraries, menus, and data. Page 2 is taken up by all the different BASIC libraries.

And thanks, c.sprinkle; I've been trying, but TI-Cares most certainly does not.

Did you get time to look into that 'left margin of file menu'-constant mismatch yet, or is that something that has to wait till right before the final release of the next version?

It's not that it's not critical, but such small details can sometimes really annoy me.

Yeah, I fixed that right after you mentioned it. I can certainly agree how little tiny details like that can be annoying.