Here's the third version of Doors CS 5.8 Beta, with a bunch of improvements and revisions. Major improvements include:

o Minor bug in BAU feature fixed with foamy3's help.
o Error messages for battery and memory warnings masked over filename text.
o Version number added in About display.
o About display now appears at AppVar creation.
o Routine created and debugged to save AppVar data in an archives appvar.
o Routine to detect backed-up appvar and restore from backup created and debugged.
o Routine to display message after crash completed.
o Include files organized to put language-specific data together.

Same as usual: download it below, report bugs to me, have fun.

Download:
Doors CS 5.8 Beta 3

Ah, including that great feature of Windoze- crashing for no apparent reason.
Nice, although I'm still using 5.8 beta 1. I should probably update it...

I was trying to make it detect whether crashes were DCS's fault or faulty ASM programs' fault, but that would wear out the Flash too much.

Maybe you could add an interrupt tasker like in MirageOS programs?

The interrupt tasker for MOS programs can be started by MOS or DCS programs running under DCS, as long as the Cn2 interrupt isn't active.

*Still key debounce when closing with 2nd or alpha on the lower right 'X'
*When unchecking 'use tios lowercase', you have to RAM clear for it to work.
*HUGE lag when trying to close the DCS menu. (~3seconds until it starts to close)

*Key debounce: *sigh* Drat you, TI-OS. Where is your hidden key buffer?!
*TI-OS lowercase: should be a piece of cake
*~3sec close:

I got this really weird bug. Here's the steps to reproduce it:
1) Check the "Launch DCS at startup" option. Make sure that "Enable TI-OS lowercase" and "Allow background" are checked.
2) Quit out of DCS, and close the calc
3) Go back into DCS, and uncheck the "Launch DCS at startup."
Then Click Ok.
4) Close the DCS menu, and open the options window again
5) For me, the "Enable TI-OS lowercase" and "Allow background" are unchecked now

You don't get that? It happens to me EVERY time. I don't think I've had the beta3 menu close normally yet. I keep thinking my calc crashed

And this is with 5.8 Beta 3?! I'll have to investigate it further. I only have a ~0.5sec lag on my version.

Oh, I figured it was a known issue >_<. TI-84+SE 2.41

kk, I'll test it under 2.40 on my TI-84+ and see if I get the same issue. Question: do you have CALCnet2 enabled?

Erm, let me check. Only if it's by default.

No, it's not enabled.

EDIT: I just enabled it. I hit okay and then closed the menu. It still had the 3 second lag. Plus, as soon as the menu finished closing, there was a RAM clear.

EDIT2: Woah... Major bug. After the ram clear, I got that message saying I ram cleared and it loaded the archived appvar. The mouse moves really slowly and leaves pixels everywhere.


EDIT3: Trying to move the mouse more, my RAM cleared again. Now it froze at the "Your calc crashed" screen.

The RAM Clear from Cn2 enable is a known issue. And yeah, clearly Cn2 still has some bugs to be worked out.

Idea: Autobackup. It could compress your programs in RAM into an AppVar and then if your calc crashes, it could extract the archive back into RAM. And it would be useful, too, if you screwed up one of your programs doing something and you need to get a working copy back.

Interesting idea, I like it. Unfortunately, you'd run into the same problem as with detecting whether crashes are DCS's fault or not: ungodly amounts of Flash wear.

Maybe something can be implemented so that when a program is about to crash, DCS intercepts it and gives you a nice little message instead of a RAM clear, giving you a chance to go back up?

That would be genius, yet I'm afraid I have no idea how to implement that. I can make an interrupt tasker, I can make it stop ASM execution, but I know of no way to detect infinite loops or proactively block accidental execution into the $c000 block, not to mention all the other creative ways ASM programmers find of making programs crash.

can't you just do a bcall(_getkeycsc) loop before exiting? (say, 5 times, just to be safe) and have menu options be 'clicked' on the button release rather than the button press, as that is how it is on desktop GUIs, which is what DCS is trying to mimic, no?

That's precisely what I'm already doing. Including the getkeycsc loop and the buttonup. *sighs again*