7 replies to this topic

[CrimsonCasio]

Ok, this is my first post to the fourm, but Ive been checking it from time to time to see if there was anything interesting. Here is my idea (no idea if its possible):

could anyone make a C or ASM program to compile BASIC games into binary and run them (or something like that), possiblely saving the compiled file to a BASIC file, or some such thing... Anyway the point would be that they would run as fast an Add-in (at least I think so) so anyone could make cool games.

Anyway, I'd appreciate some feedback on this Im in the process of learning C++ but I just started so I dont know much yet.

Crimson Casio
(crimsoncasio.tripod.com)

[rstweb]

Do you mean a program discussed like this in that thread?

bye

rstweb

[dscoshpe]

no, he means a program that would act as a *faster* interpretor of the casio basic code of some sort.

im sure that this feat is possible in some regard, but i dont know of a programmer in the community that would be willing to do it. i know that there is a bit of a movement to do some sort of compiler or assembler for the calc, but that still seems rather distant.

- dscoshpe -

[Roeoender]

This idea was discussed some time earlier (also at the french Forum @ www.graph100.com, they called it, as I remember, something like BASIC2C). It wold be a good thing to do for a person that has no idea what to do
I mean programming in C is very convinient and almost as fast as ASM (especially when you have most routines made in plain ASM, as I have).
And in BASIC you simply cannot make such things as in C/ASM (example - you are forced to use floating point (very slow) and matrices).
The only good thing about BASIC is that many people knows it, but than again the shortcomeings of BASIC won't let you do anything complex.
The better and easier thing would be make some sort of simplified C compiler (simplier logic, no priority of operands (+,-,*,||,&& etc.).

Anyway BradN was thinking about doing something (as I remember an on-board ASM compiler).
How about it Brad?
Roeoender.

[BradN]

Well, if I would try and get an on-calc assembler, it would have to be NASM, because that's my preference, and I'm always right

That causes a couple problems... I'd have to deal with C source code (or DOS drive emulation, I don't know which would be easier), and we'd have to find a way to get NASM RXE'd in order to run. (it's about 200 KB, and that won't fit in ram).

Even if we can get that far, we might need to tweak it a bit to make it fit within the 58k for ram while it runs, or hack up a way to give it more ram at that point (re-locate the casio basic data temporarily perhaps).

QuickView says nasm16.exe needs 188K of ram to run, and that should include data and code, so there's a definite chance of it working.

[rstweb]

Why don't you use a smaller assembler?
or another chance would be to resize the RAM. i don't know how it is possible at an AFX but it should be possible to remove the old RAM and solder a new bigger RAM-chip in. on the 9850 it seems that you have to make some changes in the OS of the calculator but it could be also possible to do that on tha afx, too.

can somebody send me pictures of an opened afx? thx!

bye

rstweb

[BradN]

If we get NASM into an RXE, there might be enough ram free to run it. If not, we should be able to get enough by tweaking limits in the program and/or relocating all the casio data storage to get an extra (up to) 144k. I haven't done much in depth stuff on the second topic, so hopefully nasm can be made to run with ~58k of data. We can always strip out useless stuff too (thinking extra output formats, etc.)

I think the big problem will be figuring out how to give input to and get output from NASM, or whatever assembler someone might want to use. Also, one would probably want a better editor... the casio basic editor is not well suited to the task.

There should be plenty of linkers to choose from if we get an assembler working, or we could just make a simple binary loader and skip linking altogether (depending on the complexity of the stuff being made... anything that needs a linker is probably not worth editing on the calc anyway).

[Roeoender]

I was looking for the smallest Assembler about a year ago, and the best I've found was A86 made by Eric Isaacson (I tested version 4.05). The compiler (it procuces .COM from .ASM) is only 32Kb small. I already launched it on the calc (using my enter.com to emulate ENTER key), but I hadn't got idea how to change it's output to RAM. (I wanted to dissasemble it and change it's disk operation code, but I hadn't enught time to do it).
I can send A86.zip to anyone interested (or you can search the web for it).

Roeoender