OSDev.org

Posted: **Sat Apr 19, 2014 2:14 pm**

Nable wrote:to klange
Why did `ls' show only vim's swp file but not the 'test.c' entry?

He wrote the file in one terminal, compiled and ran it in another. The directory listing clearly preceded this.

I'd say it's a good demonstration of multi-terminal support.

Posted: **Sat Apr 19, 2014 6:00 pm**

I added UTF-8 decoding support to the kernel terminal and modified my kernel console to store characters as wchar_t's rather than simple chars, such that Unicode is supported and not just ASCII. I mapped each Code Page 437 character present in the VGA font to Unicode characters, which means I have ASCII and 128 other characters available and usable through UTF-8 sequences (I need to add more font data to the kernel, so it has more available unicode characters, but this is a font issue and not a technical restriction). I then improved my editor with UTF-8 support:

Posted: **Sat Apr 19, 2014 9:33 pm**

Nable wrote:to klange
Why did `ls' show only vim's swp file but not the 'test.c' entry?

I opened the (non-existent) file in vim before `ls` was run in the first terminal, but saved it afterwards, so the swap file is there but the actual file hadn't been created yet.

And while I'm replying, here's a video showing the animation system I just ported from my old compositor into Yutani.

Posted: **Mon Apr 21, 2014 2:11 pm**

Not much, But I am proud of it. Finally got FAT12 support (working on sub-directories now)

Currently Working:
Keyboard
80X25 Text Display
Physical Memory Manager
FDC
Paging
Virtual Memory Manager
Virtual File System
FAT12 (Root Directory Access Only)

In Development:
FAT12

Posted: **Tue Apr 22, 2014 4:48 am**

Not much, But I am proud of it. Finally got FAT12 support (working on sub-directories now)

Sub-Directories isn't tough, they're like files, except an extra bit is set to tell you it's a directory, provided you are not going for write support (which blew my sanity off), but it was fun.

And while I'm posting here, not sure if I should post to "When your OS goes crazy?", 'cause it's a problem with my userspace. A "Hello World" program is 1083 bytes.

Also, I fixed a bug with my FAT driver which did not permit 8.3 file names but 7.1.3 files, (A file name can't be 7(+)-bytes, which was a bad experience), I still have to find the cause for not supporting 2-byte or 1-byte extensions.
A Hello World on my OS is just:

Code: Select all

include '../znu/libasm/applib.s'
;; Include STDAPI Interface
include '../znu/libasm/stdapi/interface.s'
_start:
        ;; Declare mystr
	mystr string "Hello World"
   ;; Display it.
	call mystr.display
   ;; Damn Integer
	myint int32_t 3000
	call myint.display
	call stdapi_exit

My `cat` command has a problem though, it quits as soon as it finds a null character (0x00). GFX.ZNX probably had a NULL char.
Here's testing out my primitive graphics stuff: (Line Drawing, Blocks, Switching to proper graphics mode)

Yes. I'm not klange, you don't see awesome Window compositors with rollin' windows here and there, with HQ wallpaper, and some cool terminals.

Posted: **Tue Apr 22, 2014 6:45 am**

Bender wrote:A Hello World on my OS is just:

Nice work

But what happens if a process does not call stdapi_exit when it's finished?

Posted: **Tue Apr 22, 2014 7:32 am**

max wrote:
Bender wrote:A Hello World on my OS is just:
Nice work
But what happens if a process does not call stdapi_exit when it's finished?

I'm puzzled that I do not have an answer to that question. Maybe we could split this thread and I'd like to know how do you guys handle when your program doesn't call exit() or equivalent?
EDIT: For now, the user can press ALT+Q to exit an application forcefully.

Posted: **Tue Apr 22, 2014 7:43 am**

Bender wrote:
max wrote:
Bender wrote:A Hello World on my OS is just:
Nice work
But what happens if a process does not call stdapi_exit when it's finished?
I'm puzzled that I do not have an answer to that question. Maybe we could split this thread and I'd like to know how do you guys handle when your program doesn't call exit() or equivalent?
EDIT: For now, the user can press ALT+Q to exit an application forcefully.

Yep you could do this, and then link it here.

Posted: **Tue Apr 22, 2014 7:58 am**

I don't think there's a lot to discuss in such a thread. A program that doesn't call this syscall at the end, is an invalid program and causes undefined behaviour. In practice this means that it will try to execute whatever comes next in memory, possibly data, and will likely segfault and be killed by the kernel in the end. Except if it doesn't.

Posted: **Tue Apr 22, 2014 8:32 am**

Kevin wrote:I don't think there's a lot to discuss in such a thread. A program that doesn't call this syscall at the end, is an invalid program and causes undefined behaviour. In practice this means that it will try to execute whatever comes next in memory, possibly data, and will likely segfault and be killed by the kernel in the end. Except if it doesn't.

Now how about:

Code: Select all

while(1)
{
     
}

Or better:

Code: Select all

   mov ax,7
asmstyle007:
     cmp ax, 7
     je asmstyle007

I highly doubt there is any way to handle this kind of situation (Unless there is third party interference like User Input, deBUGGER etc.). I remember reading about a page on Wikipedia that covers this concept, all I can remember is it said, Whether a program infinitely loops....... I can't remember the exact words, but I'm sure there is some research done in this field.
EDIT:
Found it. (hopefully it's this one)

The Halting Problem
"Given a description of an arbitrary computer program, decide whether the program finishes running or continues to run forever"

we already had a discussion in os design and theory started by Brendan or Rusky (?).

Posted: **Tue Apr 22, 2014 12:23 pm**

Bender wrote:Now how about:
[...]
Or better:
[...]

That's the "except if it doesn't" part.

Yes, you can't compute the halting problem, but that's not really a problem here. It's not even the job of the OS to catch such situations as long as the program still behaves in a reasonable enough way that allows to keep it running. The endless loop might even be intentional.

Posted: **Tue Apr 22, 2014 5:01 pm**

Bender wrote:
Not much, But I am proud of it. Finally got FAT12 support (working on sub-directories now)
Sub-Directories isn't tough, they're like files, except an extra bit is set to tell you it's a directory, provided you are not going for write support (which blew my sanity off), but it was fun.

Yea, an hour or two after posting I got sub-dirs up, have to get writing soon though

But my next step is to get a usermode program running (possibly a calculator)

Posted: **Thu Apr 24, 2014 5:53 am**

It´s a simple start, but Digos Operating is borning...

Posted: **Thu Apr 24, 2014 6:12 am**

It´s a simple start, but Digos Operating is borning...

2005? It's 9 years now, you may wish to call that a rebirth.

Posted: **Thu Apr 24, 2014 6:29 am**

OSDev.org

What does your OS look like? (Screen Shots..)

Re: What does your OS look like? (Screen Shots..)

Re: What does your OS look like? (Screen Shots..)

Re: What does your OS look like? (Screen Shots..)

Re: What does your OS look like? (Screen Shots..)

Re: What does your OS look like? (Screen Shots..)

Re: What does your OS look like? (Screen Shots..)

Re: What does your OS look like? (Screen Shots..)

Re: What does your OS look like? (Screen Shots..)

Re: What does your OS look like? (Screen Shots..)

Re: What does your OS look like? (Screen Shots..)

Re: What does your OS look like? (Screen Shots..)

Re: What does your OS look like? (Screen Shots..)

Re: What does your OS look like? (Screen Shots..)

Re: What does your OS look like? (Screen Shots..)

Re: What does your OS look like? (Screen Shots..)