94 replies to this topic

[huhn_m]

Hello there!

Because I now use the 4th set of batteries in 12 weeks (gets expensive ...)
i decided to delay my office to write a kind of new OS.
I've done ~40% already

Commandline 90%
Filesystem 70%
Hardwaredetec 100%

This system will be able to create several ramdisks (? 60KB) with
their own filesystem. Later when the programming system is done
I'll implement a "swap to flash" system that will be able to load ramdisks
to flash where they are stored as readonly.
Till now the FS supports 1 63K image with a 3K Fat. It is able to create
and to search and list files (dir / cf / format).
It can reboot the system and turn the calc off.

There will be a debug system that reports all registers and parts of the
ram through the comport to the PC. (but this is future)
Till now only the console works. No graphics but they'll come as
soon as I start the development program.

So now my question is would you like to have a new language and if
yes what functions (commands) should it have??
I should say first that in the first versions there will be no complex math
support. The only functions will be the ones know
of assembler (ADD / SUB / MUL / DIV / INC / DEC).

Waiting for your opinions ...

PS: also write if you want to be a beta tester

credits for the librarys used go to:

BiTwhise (cmm and bcd2num)
2072 (memzones and touche)
superna (casio2.h)

[BiTwhise]

Sure, I'de like to betatest

Was actually thinking of making a new Basic like language myself, but I can't find time for that now

Anyway, I'll be happy to test yours

As for commands in this language, I'll expect most ppl would like more advanced graphic routines, for games and such.
This would have to include:
- Sprites, constant and animated
- Polygons / Polylines
- Doublebuffer
- Tilemaps perhaps..
- Scroll
etc

well, good luck

[X-thunder28]

lol Good luck !

[X-thunder28]

Floating point fonction like cos, sin, tan, root, square, etc !
Some fonction for reading, writing, deleting files of your RAMdisk !
... Screen capture, basic 3D fonctions and GREYSCALE !!


I was thinking also for making my own language, but I have stopped it !
Your idea sound more interresting !

[huhn_m]

As I said I'll probably NOT inmplement anything that needs floating
point calculations if the first versions. SO cos,sin, tan , root will definately
not be included. The first versions also will only include integer (0...32767)
support. If anyone can do that he could write a floating point library.

As for doublebuffer and tilemaps .. What do these mean?? Please explain it.
What about a window mannager??

It should be possible to create a virtual desktop with a resulution up to
480 * 640 by using some scroll functions.

(Just a thought far from doing any graphic works now .. First the Filesystem
and the dev-center have to be completed.)

Expect first Beta version ~20.04.03
I'll do an daily updated beta on my site ... (hope I can manage it )

EDITED:

The URL for downloads is:

linux.mh-development.de

[X-thunder28]

For Cos for sample you can do it this :
cos100[90]=
{
100,
100,
99,
98,
96,
...
}

Cos100 is an array representativ of cos, but with 4 per 4 degres !

It should be possible to create a virtual desktop with a resulution up to
480 * 640 by using some scroll functions.

lol download TCX2 here and you'll see that idea has been already exploited !
http://teamg100.free...ode=jeux&type=0

480*640 = 294400 pixel = 36800 byte
a bit enormous
And for scrolling, you'll need a cursor or a virtual mouse
I know somethink of it

[X-thunder28]

I'm sending linuxs on my calc ...
So i test it
...
The listing of AFX hardware caracteristik scroll too quickly lol
Loggin and password : ???

I read the executabl and I found some information on hardware who are already writed lol

[BiTwhise]

Double buffer:
Two video buffers, so you can draw in one while displaying the other

Tilemaps
A map/background defined as a list of references to tiles/sprites
For drawing backgrounds in games

Your beta download
It crashes, and corrupts memory..

[X-thunder28]

No crash for me !

loggin : ROOT
PassWord : ROOT
Found in the executable !

Test seems to create a file in your ramdisk !
But some fonction not working :
PAUSE
HALT
ECHO
CF
DF
Always found in the executable!(thanks to sysxpl40)

lol :
Loading System [WAIT]
Loading Setup [WAIT]
Graphic installed :
Processor instlled : ( fault ! check your sources)
[...]
<{GNULINUX}> has found some Hardware that is not know ! Please connectPC<->CALC and start the PC-debug program!
[...]
Bios version:
Bios revision:
Processor:
Model:
Date:
[...]
0.60 (corresponding to bios version)
1.10 (corresponding to bios revision)
V30MX (corresponding to Processor)
ZX-950 (corresponding to model)
11/28/97 (corresponding to date)
So you don't detect hardware, you have write already all information !!

But the idea is really good !!
If you want some help, tell me

[huhn_m]

I detect the hardware but compare it to the one of my calc.
The commands that don't work are not yet implmented.

[BiTwhise]

You need roedisk to run it.. <_<

[huhn_m]

No I do not!

I use my own ramdisk in basic files. (Thanks to your cmm)
The programs won't be compiled but
interpreted.

[BiTwhise]

Strange.. it crashes

(launched from Touche, ROM1.01)

It displays:
Loading <{GNULINUX}> [WAIT]
Bios Version:

and then it halts..
after a while, it starts corrupting the screen, and sometimes the lcd starts behaving very strangly with shades flimmering across it..

[huhn_m]

Oh!

Please look at your mem and send me the Bios info on
the processor etc. (~CB60:6)
May it be caused by the rom ver?
Should I send you the sources?
Do you have basic files?
Reset Rom and try again.

[BiTwhise]

Seems the screen corruption disapeared after reseting the memory (and I did a ROM test as well.. don't think that makes any difference though..)

The program still halts...

If you don't mind, please send me your sources, I can have a look at them

Read the memory at the address CB60:6, and it read: 00,FF, FF,20,20,20,"OFF:ESC"
(seems this is in the middle of some error messages.. (correct address?)

[huhn_m]

Searcho for "BIOS" and tell me the adresse of this.
Theere is definately a failure in hardware detection.
Send you the sources at 6:00 pm (no time now).

Bye huhn_m

[BiTwhise]

"BIOS", found nothing

"Bios", found:
6191:7
"Bios Version: "
"Bios Revision: "
"Processor: "
"Model: "
"Date: "
(all NULL terminated, and pre linefeed(D,A) (excep the first one))
"0.60"
"1.10"
"V30MX"
"ZX-950"
"11/28/97"
(all NULL terminated)

Guess this it the Test Screen Data

[huhn_m]

It's strange that it halts ...
It should at least display an incompatible warning.

I'll send you the sources.
The things that were found are not the ones I want. These are the ones of
my executable but I mean the whole memory (TOUCHE->6->F2)

The information I need are also displayed by sysexplorer (info->Calcinfo)
I need the exact adresse in ram of these.

[BiTwhise]

I used Touche, and it found no matches, except this one (that's in flash mem.. as you said, your executable)..
Searched both 'BIOS' and 'Bios'


edit:
Strange though...
I mean, it must be there somewhere..
at least TOUCHE3.80 doesn't find it

[BiTwhise]

Looked at your source

Think I fixed some stuff

Sent you an email about it

bye,
BiTwhise

[huhn_m]

I'll put up a modified version this evening that identifies your
rom version and does resume after the unknown Hardware message.

But please look in SysExplorer 4.0 / 3.? and go to info -> Calcinfo
there. Does this display the calcs infos??

Try searching for V30MX etc ... (see my sources)

Thanx for you modifications.

[BiTwhise]

heh

it displays:
CASIO Z945A G100..(which I assume is hard coded into the program)
and then tons of random signs

guess it collects it from the same place as you do..(lookes like it)

the testmenu displays it correctly (so my ROM should be just fine)
don't know why I can't find it when searching (does TOUCHE search the whole ROM?)

[huhn_m]

Yes, i know sysexplorer does so. I've the sources and therefor know it
uses nearly the same method as I do. I'd have wondered if it had displayed
anything.

Yes Touche does!
This means there are no info in ram ...
Strange ...
I'll not test hardware on a 1.01 rom but this could lead to
malfunctions ...
You'll have to enable debug mode manually ...

[BiTwhise]

The information you're looking for is located in ROM, not in RAM right..

So, You have to map the right part of ROM into address space, cause the information is there, only a question of where..

[huhn_m]

A question to rom detection:

I do the folowing

MOV BX,0C000h
MOV ES,BX
MOV DI,0Dh
MOV AL,ES:[DI]

then there is only a zero in AL why ?? I think there have to be either
41h,42h or 43h ???

--EDITED--

OK! Problem fixed!

[BiTwhise]

Maybe you haven't mapped the ROM correctly to the address space..

mov al, n
out 5Ah, al

where n is a number C0h - DFh, where C0h corresponds to the start of ROM, and DF to the end, incrementing in chunks of 128k (32 * 128K = 4M)
port 5Ah corresponds to what data should be mapped in the segment address space C000h - DFFFh


--EDIT--
ok

[huhn_m]

The new version on my site should work.

[BiTwhise]

yup, it works..

just a few small things:

wait for keypress after error message
clear the screen after the error message. all the text is unalligned

[huhn_m]

Ooops! Seems to be a missing linebrake!
I didn't test the error message.

I'll fix it today. Expect a new version this evening or tomorrow
with the CF function and the DF function working.

[X-thunder28]

... this project is stopped ?

[huhn_m]

No I completed the file create routine and will do the delete routine
today. Expect a new download this evening (GMT).

I had no time!
At monday final test of the theatre piece till 6:00
At tuesday the theatre piece till 21:00
At wendsday I was not allowed to use the PC :-(
At Thursday I got my new PC and I had sime trouble with it (HD is defect)
At Friday I had school till 16:00 and from 19:00 till 22:00 I had no time
either.

So you see! No time for coding! But I will complete this OS for sure!

Good to see anyone is interested! Thanx

[X-thunder28]

this is comforting!

I'm waiting for another news and/or updates !

[huhn_m]

Hi there!

I did the following:

improve the format function
correct bug in command line (compare_string proc)
correct bug in create file
renamed test to CF (CreateFile)
added DF (deletefile)
added THANX.

Expect the version to be up at 18:00 GMT

needs only some debugging in the DF function (fat is not assigned correctly
after memory has been moved!) Hope to fix this till evening!

PS: Isn't someone able to fix this silly mistake with the QUOTE / CODE ...
buttons! It's annoying

[BiTwhise]

thats kool

if you need any assistans/more elaborate write functions (with rep ie) for the memoru stuff (and your formating), just let me know

[huhn_m]

Hi!

I did the Update!
DF still isn't working

@BiTwhise

I need a function that moves a given number of bytes (in a register)
exactly 13 bytes downwards in the memory overwriting the older bytes.
The data is stored in the @DATA segment. The maximum number of bytes
is 3315 (size of LFAT). (It's for deleting an LFAT entry!)
If you could do this it would be great! My function simply doesn't work!!

[BiTwhise]

;ds:[si] = start of block to move
;cx = size of block
mov ax, ds
mov es, ax
mov di, si
sub di, 13
rep movsb

Think this should do it, taken that the block is moved down, si > 13, direction flag is cleared, block is within ds (pointer is normalized)
Obviously you can optimize it to move words instead of bytes, and so on..

[X-thunder28]

I'll upload this update on my calc

[huhn_m]

Updated again!

(Will not do any work the next 2 days 'cause I'm taking part in
a workshop and am therefor not at home)

Have a lot of fun!

[2072]

If you want very improved memory moving functions you could look at memzones sources these functions are in C/ASM but mainly in ASM. Look at the code of the function "my_movedata" for word/word move.

[huhn_m]

Thanks 2072! Will have a look at it soon

Updated <{GNULINUX}>!

Linux