Omnimaga
Calculator Community => TI Calculators => TI-BASIC => Topic started by: jsj795 on April 12, 2010, 07:03:10 pm
-
Okay, I was trying to purposely make a program larger than ram. So what I did was to make a super long string until the ans variable is biggest and paste it into the program. I used 'F' because it's 2 bytes :P
If you see the screen shot, during the second paste, the calculator pastes in other tokens, seemingly randomized. I am guessing these are random hex tokens? I first found this out on calculator. There's no way to stop this unless you take out the battery, which causes ram clear. Leaving it for like 5 minutes, the calculator turns off, then when you turn it back on, it gives you ram clear.
Usually, I thought that what you are trying to paste is larger than the program can support, it doesn't let you paste, but I guess in this case I was wrong. May be because I was trying to paste string, not program?
It just baffled me, so just throwing out there. And I backed up all my stuff before doing this, because I knew something bad might happen :P It only gave me ram clear, so... I guess it isn't too bad
-
Huh, I just tried it and it didn't happen to me. I just ran out of RAM and it stopped and put the dotted pattern for the cursor. I'm on a 84+ though, if that makes a difference.
-
I did this on TI 84+ SE OS version 2.43...
May be it's only for TI84+SE?
Or may be, it's because it doesn't have extra ram (my hardware doesn't have the extra ram, because it was made later than others)
-
Hmm, odd. Someone else should try it with the 84+SE.
-
im doing it now
-
I tried it on wabbitemu with a TI-84+ SE ROM running 2.43 and it didn't work, although it was acting a little strange.
-
TI-84 Plus OS v2.43
Stops pasting and gives dotted cursor.
-
I also got it on my Ti84+ SE. And i DO have the extra RAM pages, so it cant be that.
-
hmmm... weird? Mine works, but on some people's it doesn't?
My hardware version found on the back: ends with H
Boot Code 1.02
Base Code 2.43
ID: 0A337-583AA-E61F
don't know if these will help on finding the differences?
-
I would try it, but my sis won't let me borrow her calc.
-
Strange I don't think I ever experienced this before. I had a similar thing happen with an old version of MirageOs or Omnicalc I think (maybe the Entry menu) but never something like this done without ASM. What program did you ran on your calc before this happens? I also hope you didn't lose anything x.x
It could be a 84+only bug. I can,t try since I don't own a 84 and I don't feel like trying on my Nspire (which would require me to take out the keyboard, remove a battery, then put them back since it has the power managment bug on OS 1.1
-
I first backed up all the files, archived all the programs in ram, then I did this. I was purposely experimenting with this thing, to see what will happen. And tada! this happened X.X I didn't lose anything (was prepared for all memory clear, but it only gave ram clear)
Did anyone try it on regular 83+?
-
I haven't tried it yet but I got a bunch of stuff on my calc, so idk if I will right now. I used one for almost 9 years, though, and never had this happen, even if I Rcl'ed a lot of large stuff before. In my case, it always stops pasting stuff when I run out of RAM and I see a checkered patten cursor.
-
Ya, I tried on a 83+ and nothing happened. Just the same checkered pattern for the cursor.
-
So only Builderboy's and mine did this. ???
I guess it can't be extra ram pages tho, because I did it in wabbitemu (I can't run TI-Boy on calc, but I can on wabbitemu so wabbitemu has extra ram pages)
May be it's the ROM file on 84+SE?
Although for some reason it didn't happen to Eeems' and ACagliano's...
-
I still suspect something else than the TI OS could have caused it. Can you recreate this bug after a fresh RAM reset? Even the most stable ASM programs could be enough to trigger weird things. Even CODEX did.
-
Well, I only tried it once on calculator, but on WabbitEmu, I was able to recreate it a lot of times, even with ram clears, all memory clear, etc.
Right after all memory clear (in wabbitemu):
RAM:24317
After you get memory error from too big of Ans variable:
RAM:16129
After making program named 'A' and recalling the answer 1 time:
RAM:7925
After recalling the answer 2nd time, weird stuff happens.
-
wow, so I guess it's a new bug then x.x
I wonder if OS 2.53 MP fixes it?
-
Hmm, I tried this, but I just got the dotted cursor. I have TI-84+SE, with OS 2.53 MP.
I tried using a normal letter for a string, instead of the serifed F in the screenshot. I wonder if that makes a difference?
-
I tried this too, but I just got the dotted cursor. I have TI-84+SE, with OS 2.43. :)
-
Weird...I just did this again and the glitch happened to me this time ???
-
I'm doing it on a fresh install of nspire os 1.1, after an all mem clear.
It's taking a while tho....
Edit: Nothing for the nspire
-
I got the dotted cursor, OS 2.53 with no extra ram pages.
-
I got something new: even after I ClrEnt, I'm still 4000 short, after deleting EVERYTHING
Ram Clear fixed it tho. :)
-
what I don't understand is how you got that F to show up. I don't get that in the window vars
-
It's under Vars>Statistics>Test.
-
it worked! cal'c crashed. I had to paste mine 3 time before it did random stuff. Its still going as I type this message right now. TI-83+ OS 1.19
-
Uhhh... Another glitch I have to fix. I just did it on my calc. Recall it once, you get about 2000 F's, recall it twice you get about 500 F's then unlimited random charaters.
I'll look into it.
Oh, and 100% guarantee it has nothing to do with extra ram pages as those aren't used for this kind of thing.
Edit #1:
It only works with two byte characters. The OS checks to see if the total ram used is equal to the total ram available. If you count by twos however, 999!=1000, 1001!=1000, the OS misses it's chance and guess what happens.
Edit #2:
Found it. I'm not going to go through the trouble of fixing it because 1) I'm lazy, and 2) I'm not sure this has ever caused someone an accidental ram clear. But for future reference here is the problem:
In bcall($4942), (editTail) is checked with (editBtm) to see if they are equal, and if they're not, it figures that there is enough space to put another character. But if it is putting a two byte character, it goes past the end of the editBuffer and copies into oblivion. The way to fix it would be to either stop it once it goes over, or check for two byte characters and handle them accordingly.
For all those who couldn't get the recall glitch, delete your second recall. Add in one more byte, pressing enter will work. And then recall it again.
-
That's crazy. Nice find! :)
-
Very nice, Penguin. Glad you could shed some light on what was going on. Not that I completely understand it :P But cool :)
-
Wow nice find ThePenguin77 :)
It's under Vars>Statistics>Test.
That F is awesome. I used it often in my old ASCII RPG maps ;D
-
Wow, interesting. It fails so epically ;D
Does this occur in all OS versions? I've gotten it to work on OS 2.43 and on a friend's 83+ (I don't know what OS version)
-
Well, it works on the flash debugger as well as 2.53. I would say that's a pretty good indication that it is Ti's mistake from the beginning and is on all OS's.
-
That's what I like about having an open, easily reviewable piece of software. Bugs can be found much more quickly.
And yeah, I guess so. Who has an 82/83 to test this on?
-
Yes, it also crashes on a TI-83 using Wabbitemu.
-
When I get a chance, I'll try it on my TI-82 and possibly my TI-81.
-
The TI-81 has no two-byte tokens, so that wouldn't work ;D
We'll wait for the TI-82 test results, though. :)
-
OK. I didn't know since I haven't messed with it too much.
-
What would be a fast way to trigger this glitch on a TI-82? The TI-82 doesn't support strings.
Note: I am unable to reproduce something similar on my TI-85. I guess it might just support 1 byte chars or just 2 (when detokenized)
EDIT: I just reproduced that bug on my old TI-83+ now. It is running OS 1.19.
EDIT: It happens regardless of the size of the string. It can also be recreated in the Y= menu, the EQU solver (found in the MATH menu), in the list editor (when selecting a list and entering a formula there) and the home screen. Bug summed up at http://ourl.ca/3687, last bug listed in the first post there, which I just edited.
-
What would be a fast way to trigger this glitch on a TI-82? The TI-82 doesn't support strings.
Send a large program to your calc that fills up almost all your memory and then type the characters in manually?
-
now to find a large enough prog :P
-
Have fun searching, I guess. :P