OSDev.org

Yes, you're right. I was being too kind to him.

I just had my F*CK YEAH moment a few seconds ago, when I saw this:


Bigger: http://oi59.tinypic.com/wtyqsh.jpg

YES!!! My VM86 is working, and I was able to retrieve the VBE mode list and set the graphics mode.
Some might think this is just a black window - but its a window to my soul, revealing images of so many endless hours of fighting with my scheduler. This is such a wonderful, beautiful day.

Umm... that was in the days before I learnt how to compile Bochs with the built-in debugger support. Nowadays whenever I get a triple fault I track it down in a few minutes.

In any case, I don't ask questions until I've tried anything I can by myself (and that includes exception handlers if I need to).

if(hardwork == success)
{
invoke(AWW_YEAH_MOMENT);
}
Not to derail the thread but just sharing some info:
If you ever wish to support x64 (or get bored with v86) just grab a copy of Real-Mode Emulator or x86emu, for mode-switching I just wrote a simple program that captures arguments and switches to relevant video mode and is linked against Real Mode Emulator. Whenever I want to switch I just execute that program with arguments. IIRC both RME and x86emu don't require any sort of library unless you wan't to debug in which both will require a fairly standard printf.

It's good to see that you are now relying upon someone else's exception handlers even if you can't write your own. That still won't help much when you try to run on real hardware but it's a start.

I've had a few recently:
-When I realised my OS was complex enough to require a disk-flush-before-shutdown procedure (although I haven't yet got one)
-When I got hardware multitasking working
-Not quite on-topic, but when I switched to writing a test shell and realised I could use any label I want because it wasn't used already elsewhere.

I just had one recently - my assembler (I use software isolation so my applications are written in bytecode) now has the ability to insert debugging information such as variable names and the exact column/line of code each bytecode was derived from.

A small victory but it opens up a lot of doors. I am hoping to one day have a fully integrated IDE that can singlestep through code. Technically, my IDE should be able to single step through any language with this information (assuming the compiler emitted debugging data to the assembler.)

My other small one was the introduction of 8/16/32 bit operands for jumps, loading integers, etc. and the assembler being able to detect the most appropriate size to use. This can save a lot of space (if you're loading the number "10", there's no reason it has to be stored with the full 8 bytes.)

this

That's a crash! It goes in the "When your OS goes crazy" thread. This thread is for good things!

Just kidding... really, what on earth is that?

That isn't a crash, it's Peterbjornx showing off his objdump which is obviously an AWWW YEAH! moment as I think objdump does require a standard C library implementation with file I/O. The process exits and there is nothing for the kernel to do now, so it halts.

Sorry, I just saw some sort of a dump followed by the words "PANIC!" and thought it must have crashed .

Anyway, why does everyone post such big screenshots?

Everyone? Then it's possible that not the screenshots are big but you are the _one_ who just have small display. No offence, just joking.

Actually yes, now that I think about it I have got a small screen. Having screenshots where you can only see 3/4 of the width at a time is not unreasonable for someone viewing them on a 1024x768 display.

By the way, I like old computers. The older the better!

reading that post on my 2880x1800 display..

While my choice of sample payload may have been questionable this was definitely the highlight of my last few weeks of work.

With a better payload: