ICE Compiler

→ **MyCalcs** profile · de **Ti64CLi++** » 03 Déc 2016, 16:07

I have re-download ICE, and I send it into my calculator, and I compile Pong, it works, thank you very much

→ **MyCalcs** profile · de **PT_** » 03 Déc 2016, 16:13

Ehm, did I it right?

→ **MyCalcs** profile · de **STaa** » 03 Déc 2016, 16:19

Nope, you have to compile it another time. I'll try to reproduce it when I'll be home and send a video

→ **MyCalcs** profile · de **PT_** » 03 Déc 2016, 16:59

Okay, I've figured out why they won't work together. PHASM uses a hook to run something before ICE starts, and loads the byte 1 into (pixelShadow2). Then ICE runs, but whatever program you have, after ICE is ready with compiling, it clears pixelShadow, pixelShadow2 and some more RAM areas, which means the 1 disappears. Then the OS switches back to PHASM, and PHASM checks if (pixelShadow) is equal to 1, and if not, it displays a *random* error. So either PHASM need to save the byte somewhere else (use a bit maybe), or I need less cleanup

. For now, I can fix it, and I will fix this in the new version I'm currently working on, ICE v1.5

Many many thanks for seeing this

EDIT: no, this is not the bug. I'm still trying to figure out what goes wrong

→ **MyCalcs** profile · de **PT_** » 03 Déc 2016, 18:42

Heh, it seems I kind of know what the bug is. It's not the fault of PHASM, nor ICE, and maybe the OS

When the OS wants to execute an ASM program, it shifts X bytes X places up from UserMem, where X is the ASM program size, just to make place for the ASM program. There is nothing wrong with it, because after execution, these bytes are shifted back. But it can happen that the ASM program also shifts these bytes, and then you get an error! To explain it simple: what you entered at the homescreen, is pointed by some pointers, called curPC and endPC. When the OS moves the bytes forward AND back, these pointers points to the right equation, which was entered at the homescreen. But if the ASM program (ICE) also moves the bytes a bit, then the pointers don't point to the homescreen equation anymore, but instead to *random* stuff, which will very likely generate an error. I'm not sure who needs to solve it, but the OS can't. I will think about it, but for now, don't run "prgmICE", but "Asm(prgmICE"

(See also http://pastebin.com/K8mvTjU2 for my talk with TheMachine02

)

→ **MyCalcs** profile · de **STaa** » 03 Déc 2016, 20:10

Ok, fine, thabks for taking your time trying to fix this weird bug

→ **MyCalcs** profile · de **PT_** » 03 Déc 2016, 23:24

New update, ICE v1.5. I've added direct key inputting, so very fast checking if a key is pressed, small For loops, which are independant of a variable, and much faster, and CompilePrgm( which compiles a subprogram. Additionally, the graphics library version 3 is supported. Maybe more, but I can't remember

Download ICE Compiler

→ **MyCalcs** profile · de **Ti64CLi++** » 05 Déc 2016, 17:54

Thank you for this update.
I have an unknown error, on the line remainder(L1(theta),16). ICE don't accept most than one character in the first option of remainder ^^

→ **MyCalcs** profile · de **PT_** » 06 Déc 2016, 00:24

Ti64CLi++ a écrit:Thank you for this update.
I have an unknown error, on the line remainder(L1(theta),16). ICE don't accept most than one character in the first option of remainder ^^

Found, fixed, updated, thanks