smallEditRam? what are the cases where it is touched? (and yes, my program will fit in there extremely easily, the appvar was 87 bytes + 1 byte per keypress, and that's with the header.)

umm... i wasn't getting ram clears.  just it wasn't running anything, no matter what I tried. I tried with/without padding two bytes for the length, all four page banks, and also incrementing and decrementing the page number! I got nothing. As in, it created the appvar, archived it, and definitely was finding it. but either it refused to set the hook (probably), or it was simply uninstalling it upon the first attempt to run...

now that I think about it, an app wouldn't work anyways. I'd still have to save the keystring somewhere, and that'd be just as easily corrupted as before.

Ok, the actual hook is 68 bytes, plus 2 bytes for string length and current entry, which leaves 37 bytes for the keystring wow that's a long one.

In other news:
 I really just need a way around PTT... all I can think of is IM 2. (which means another hook, depending on where I put the table and interrupt, i'm sure, but it'd be completely secure, anybody know a good place for the table? the interrupt will easily fit in between the four entries...)

 Also, the inclusion of ENTER in keystrings allows for extremely awesome and counter-intuitive keys, like "1/00[enter]" (if you have a buggy os, you can even force the user to crash it to unlock it )

 I'm pretty sure that pulling a battery when the OS has control does not result in a ram clear. I do know that at least 90% of the time it won't from the homescreen. So yay! don't have to worry about that.  Wow. Battery-pull proof... maybe...

 I might want to leave a backdoor. Other than re-sending the OS. Will simply plugging in a cable (without a calc on the other end) call an interrupt? if not, then that seems secure enough... otherwise, I'll have it auto-uninstall when anything is sent to the calc, unless it does that already. (or is the cable plug-in good enough?)

 Who else thinks an accessory program to modify the hook would be nice?  As in, changing which keys it blocks and which key re-enables it. It'd definitely be done in Axe

Either way, here is the program, moved to smallEditRam for security, and keystring upgraded to 37, as well as everything I may have mentioned before.

ok thanks. also I majorly upgraded the post above. too lazy to undo and repost down here.

so yeah scroll up for an update

I found a way past Push-To-Test!

All I did was:

 tell it to not APD (wish it were possible to avoid this)

 tell it to interpret OFF as a keypress, which, when pressed, runs my own power-down code. (still a bit buggy, if you push it too fast it'll turn right back on)

 And, finally, tell it to tell itself these things again when it does much of anything.

But fitting that code in the same 97 bytes was difficult. Keystrings are now limited (in my not-fully-tested version) to 13 due to memory.

The backdoor I've left is:

1: turn the calc 'off'
2: insert a cable, turning it back on
3: wait for it to APD (it should if you don't leave the homescreen)
4: enter any of the 3 PTT combos, then F1 to back out.

... actually I don't have access to a cable, so I'm not sure it works. I'll check it tomorrow.

Wait so this trick prevents PTT from doing its appropriate effect? Or does it let you exit PTT? If it's the later, it would be cool if you found a way to exit PTT without using the conventional way and without having a patch installed.

Well.. not exactly. This just prevents my hooks from being deleted by PTT.

Previously, one could simply open the PTT menu to bypass my whole security system. This whole part blocks the whole PTT menu, thus making it harder to uninstall without the key.

Although a 30 byte program that lets you break out of a full PTT is possible, that's not really my goal here.  I'm just trying to make a practically unbreakable menu blocker.  So I guess it might not be much of a 'suite' after all...

well last night I had my worst crash ever.

My hook for some reason returned k3 no matter what I pressed. (lol stack fail)

found out my backdoor was not working, and after messing with it, my powerdown routine works less reliably than ever, doesn't actually turn the screen off, and is way too large. But it returns the right numbers.

So it looks like this is going to be an app (<500 bytes). any suggestions as to which compiler to download? smaller = better, and I am porting from Mimas, so the less translation I have to do...

Mimas seems pretty good, but if you're gonna code on the computer, most people appear to use Spasm from Revsoft or ZDS

You can take the compiled Mimas ASM code and put it into an Axe program and compile the program as an app.

And in my opinion Brass and ZDS are really good assemblers.

How exactly would I do that? :Asm(prgmSECURITE) doesn't seem to work...

:facepalm:

How about: :[prgmSECURITE]
And I thought I had found a better way... involving appvars and indirect hooking... only to realize that you can't run code directly from appvars...

This might help a bunch, assuming 0.4.5 supports this