OSDev.org

Double buffering... Right

Oh dear lord, what have you done to my rectangles?

I decided that screwing around with the kernel memory allocator is probably a good time to make the kernel memory allocator able to report fatal errors. Fifteen minutes after implementing an error handler, I suddenly learned that it worked...

Who knows what the hell did octacone done... Anyways you need some math to draw that-like rectangle.
Maybe is Octacone testing his math library?

Drawing rectangular axis aligned shapes involves really complex math I think.

Lukand wrote:Who knows what the hell did octacone done... Anyways you need some math to draw that-like rectangle.
Maybe is Octacone testing his math library?

jojo wrote:Oh dear lord, what have you done to my rectangles?

Nah, not your rectangles, they are all fine (except for the color format).
That is my double buffering implementation, that was not supposed to happen. I really have no idea why is that happening, I did everything as suggested.

Math library? Not dealing with that sh*t. (I hate maths)
That was an error produced by my double buffering implementation.
It looks "cool" maybe but that was like pure randomness. (Captured in between frames)

Haha, no, I assumed as much. Just messing with you.

jojo wrote:Haha, no, I assumed as much. Just messing with you.

xD

@Octacone YOU DON'T LIKE MATHEMATICS!?

Everybody knows you're supposed to go through every pixel on the screen and check if ((x - x1) / abs(x - x1) + (x2 - x) / abs(x2 - x) + (y - y1) / abs(y - y1) + (y2 - y) / abs(y2 - y)) * 0.25 - 1 is equal to 0.

octacone wrote:(I hate maths)

Now I understand why that mouse pointer math gave you so much trouble.

Octocontrabass wrote:

octacone wrote:(I hate maths)

Now I understand why that mouse pointer math gave you so much trouble.

Exactly!

Well, I'm working with VGA text mode, and I found my bug: I was using instead of &ing foreColor.

Maybe next time I'll get it right the first time :

Cheers,
trekos

How did you manage to pull off tha--

• Shift 7 → &
• Shift 8 → *

...oh

In the 'Forum Software Update' thread, I recently mentioned the Discopædia (warning, lots of NSFW language) glossary of, well, insults and terms of abuse cataloged about Discourse (and more recently, NodeBB). It just occurred to me that one of the Discopædia terms might be of use here in describing the failure modes found in these screenshots: 'jellypotato' (warning, more NSFW language).

It originally was a catch-all term for bizarre bugs that no one seems to understand the mechanism of, but came to refer primarily to screen rendering glitches. A typical example found in both DiscoHorse and NoseBBleed is one where scrolling a forum thread causes a sort of wobbling where the page seems to jump past the target point for about half a second, and then back past it in the other direction, possibly doing this several times before settling in a spot that may or may not be the intended one. It is particularly frustrating glitch for both the devs, who can't seem to figure out the cause after a number of attempts to fix it have failed, and the users, because the initial jump lasts just long enough for the user to try and respond to it, but by the time they actually try to correct it, it has moved again.

It has been applied to many other problems, but mostly to things where there is something vaguely spooky or enigmatic about it - the sort of thing which makes you say, "how is that even possible?" Bonus points if it manages to look silly at the same time.

Anyway, it seems that more than a few of the oddities in this thread would be well described by this term. Any thoughts on this?