Thanks gang

Hello, I’m not sure if this has been posted already (I didn’t bother reading through all 7 pages of posts), but because this bug isn’t fixed, I’m pretty sure it hasn’t. I don’t know if somebody has experienced this bug before, but whatever.
Basically, the icon hook setting in ceashell 2.0 messes up probability simulator. It’s kind of complicated, but when I turn the setting on and launch probability simulator, part of the top bar with the icon gets erased. the different experiments have really weird quirks, I’ll probably explain them later. If anybody else has this glitch if they try it, then please add a new post here.


I have a ti84 plus ce python, with os 5.8.0 (it came with that preinstalled when I bought it sadly). Please try this out if you see it, I don’t know if it’s only icon hook that causes the glitch, or maybe it’s other apps or settings, but I know that icon hook will toggle the glitch for sure.

maybe it’s because my calculator is a python edition? I’m not sure. I might upload some pictures (if i find out how).


Edit: idk how to put an image on, I’m on my uncles iPhone right now, I think I need to upload it somewhere because it needs an url…


but basically I’m able to make there be 7 marble types in pick marbles by setting it to 5 types? And in adv settings the marble count for the extra 2 contains a semicolon. It’s as weird as it sounds. That was the image I was trying to upload, but random numbers experiment is even weirder. Please try it out.

Thanks for reporting this and I think you gave great information on what the bug is. We'll definitely be sure to test this out as soon as possible and hopefully get it fixed. In fact I already have an idea of what could possibly be going on at least with the status bar glitch.

so now im on my aunts phone which isnt even apple and im literally using wechat as a browser somehow idk if this post will even go thru

anyways, i made several discoveries about the prob sim glitch. it seemed.weirs, but its like really really weird. it doesnt like completely break the app and crash it (sometimes it does tho), it just makes the app do really weird things.

simple summary for those who dont want to read the rest (even tho u should):

u can make prob sim have 7 marble types, 4 dice, 4 coins, and 7 or 8 random numbers. also you can make prob sim just crash, it will say "memory error prob sim will exit now" or smth.

ok now the main part of the post
so basically with icon hook on, top bar will be covered a bit so the part of the icon and some of the words will disappear (sometimes randomly, sometimes only when you encounter a glitch).

so this is a more generic thing, if you select certain numbers for the setting that is below trial set (or the first setting if there is no trial set) it will add 2 to it. hard to explain, (and hard to type right now because this phone uses an abcdef not qwerty keyboard) but basically if you select 1 coin or die it will make you roll 3 even though the bar graph acts theres only 1, but you can visually see 3 coins or dice. selecting 2 gives 4, (visually only 3 coins tho), but 3 coins is normal.
also marble work the same way but only on some numbers. random numvers work the same way but 8 numbers freezes your calculator and you need to press reset button.

oh yeah also you can the memory error thing by making 4 20 side dice and rolling them hundreds of times. and the bars in thw bar graph will break into pieces literally.

oh yeah also saving data in corrupted experiments may instantly crash your calculator and reset ram or only do it when u try to read from list.

this is too annoying im gonna stop coreecting my typo

sot heres also a lpt more stu.f f. but i cant explain 。。 it lal because thittus is too hard uo type.

i mighr aed more later

Thanks for your help!

We were actually already able to correct this bug, and we'll be pushing it in the small bug fix version hopefully within the next few weeks.

For anyone curious, the odd behavior was mostly caused by conflicting use of appData by both our hook and the app.

More things:
* In mathprint mode with the edit archived programs hook active. The calculator RAM resets when I try to edit an archived program with a lot of mathprint on the screen (eg. a 16x16 matrix).
* In classic mode with the edit archived programs hook active. The cursor is in the wrong position after displaying a large (16x16) matrix, editing a program, and quitting to the home screen.
* Both classic and mathprint modes with the edit archived programs hook active. Error messages are printed to home screen in certain cases (eg. Break out of a “While 1:Disp 1:End” program and select quit).

There is a tool to downgrade your OS, I downgraded from 5.8.1 to 5.3.0. (It is Called "CERMASTR 2.0", it can only be found in a youtube video discription/comment, but i dont have the link to the video)

ok thanks for the help I’ll probably search for it


Edit: apparently the video doesn’t exist anymore

Both issues with the matrix have been fixed. It was a pretty obscure bug to track down, but we noticed that it seemed the main cause of the issue was when Ans would contain a value that could scroll. After a lot of messing around, we realized we needed to switch the order of ReleaseBuffer and NewContext0 to get it working.

The final issue is a result of trying to preserve the text on the screen after a program launches. We settled on clearing the screen if an error was triggered when exiting a program to TI-OS as a way to improve this.

Aside from correcting these issues, we've also added something to detect if a program has exited displaying something on the home screen when returning to CEaShell, to allow it to pause for a key so the user can read it without being immediately sent back into the shell.

All previously mentioned fixes have been published, along with the addition of showing file types in the TI-OS programs menu in place of a description if the file being selected does not have one:


Version 2.0.1 is available to download in both English and French from GitHub, Cemetech, and TI-Planet, and will be available on ticalc.org once the upload is approved.

If you're interested in the details of this update, feel free to check out the full changelog here!

If you already permanently borrowed the error handler what about a USB handler https://github.com/mateoconlechuga/cesium/blob/master/src/usb.asm

Version 2.0.2 is now available on GitHub, Cemetech , and TI-Planet, and will be available on ticalc soon.

This update features a number of minor bug fixes, along with a significant size optimization (~200+ bytes) from using the asm_data_ptr1 RAM location to store data pointers when moving RAM around (creating files or inserting memory), rather than recalculating them later. For a full changelog, see here.

I installed CEaShell on my CE last week. Here are a few thoughts:
•The shell's worked perfectly and hasn't triggered any resets, which is huge.
•The shell is very responsive and quick.
•It doesn't display Basic icons the same way Cesium does. In Cesium, a ton of single-byte characters have unique non-OS colors, while in CEaShell, they're almost all the same shade of light blue, except for a few (e.g. : ) that are white and, for some reason, H, which is more turquoise.
•Pressing a letter jumps to programs that begin with that letter. However, I'd prefer those programs to start on the left (i.e. instead of H, H, H, I..., it should be I, I, I, J...). I apologize for the inevitable confusion of those trying to process what the heck that means, but I don't know how else to put it, so here we are.
•The "About" text for Transition Speed has a typo: "animaitons" instead of "animations."
•I'd like the ability to change the default icons for programs that have none. I haven't seen other shells with this capability, but why not?
•It would be cool to see Celtic integrated in the shell, with a toggle for enabling it built in.
Overall, this shell is amazing! 10/10 as is; a few small changes would take it to 11

Nice! It's great to hear you're enjoying it and we appreciate the feedback a lot!

What you've encountered with Cesium is actually unintended behavior. The only legal colors for TI-BASIC icons are 0-9 and A-F, and anything beyond that is essentially just random colors being grabbed from memory. It just happens that the colors which are grabbed are different depending on the shell, but neither is standard behavior.

We actually attempted to adjust the alpha jump behavior at one point in the past and it caused a lot of issues and cursor problems. As it turns out, the algorithm was a lot more complicated and we decided against it for both speed and stability.

We've fixed the text, thanks for pointing it out!

Icon adding is actually a lot more challenging than it might seem, as not only does it involve adding an icon editor, but also involves handling a wide variety of icon formats, BASIC and assembly files, and properly resizing programs with the included data. While it is doable, it's kind of outside of the scope of the shell and if someone made it it would probably be better off as a standalone program.

We decided a while back (also after asking people on Cemetech) that it made sense to keep the two programs separate. In some cases, people only wish to use Celtic, and in other cases, they only want to use CEaShell, and having them combined would greatly increase the size of the app. Combining also wouldn't save space but might actually add some, which is why we decided against it.

TIny_Hacker wrote a lot of things


I figured that wasn't supposed to happen, but it's interesting nonetheless.


True, but what about multiple default options?


Makes sense. Would it be difficult to implement a toggle for those who have Celtic anyway? Would that require an update to both apps?

Are you talking about being able to change what the default icon for a C program or a TI-BASIC program, etc would end up looking like inside the shell?



It would require both CEaShell and Celtic to be updated in order for them to not conflict with each other. Unless you're talking about just having CEaShell install Celtic from within itself only if the Celtic app exists. In that case I still feel like that would add unnecessary space to the shell when you can just open Celtic from the CEaShell already at that point.

Yes, basically. I was thinking along the lines of multiple styles for icon-less programs.

I have a few other observations after using CEaShell some more.
▪Typing within the shell (e.g. when naming a program) is a bit too responsive, making it somewhat difficult to do accurately. I think it could be improved with a small built-in delay of maybe .1 second. Also, delete deletes the letter to the right of the cursor, which is a bit unintuitive but by no means a big deal.
▪For some reason, you can't exit a Basic program by pressing [2nd] [quit] during an Input/Prompt command, which is kind of annoying.
▪Some of the load times are kind of slow.
▪Having to press [zoom] et al. in order to see if a program is archived and/or locked is somewhat inconvenient.
▪I'm not usually worried about resets, but a backup RAM feature like Cesium's would offer some peace of mind.
Don't think I dislike this shell - the fact that I bothered mentioning these fairly trivial things shows how good CEaShell is overall. It's definitely worth checking out if you haven't already.

They arent implementing ram backups due to the archive existing.

Our icons are actually not sprites, but scalable patterns made from shapes, so this isn't really that feasible.

We'll look into adjusting the delay a bit. This isn't too tricky (it's just a constant value in the program code), but the difficult part is finding something that the majority of people find comfortable. We need to find some way to survey a large group of users and find the perfect delay As for the behavior of the [del] key, this is intended and in the README. It's a little confusing if you're used to Cesium, but we did it this way to match the behavior in TI-OS, and if you want to backspace you can use the [mode] key.

We do this to prevent crashes and memory leaks because TI-OS does not make context switching easy.

Any chance you could go into a little more detail?

We've thought about different ways of showing this information, but weren't able to find a way to fit it into CEaShell's interface effectively (along with not wanting to add a significant amount of code at the same time). Hopefully you don't need to check the file info that often, and one keypress to open the menu isn't too much work. You can always speed up the transition animations or turn them off completely if you find that the menus open too slowly.

We've had a few people bring this up in the past, but since CEaShell lets you edit archived programs, you shouldn't really need to rely on RAM that much.

We definitely appreciate your feedback and it's great to hear that you're still enjoying using the shell!

I accidentally found a bug. If you use the [on] shortcut to turn the calc off with CEaShell running, then run an assembly program from within a Basic program using det(70, APD will still be active, and it'll turn the calc off while that assembly program is still running. When you turn the calc back on, the program crashes and the RAM is cleared.