GUI / GUI Framework

[WebFreak001]

needs to be there in the future

[kamirr]

Maybe port some most important parts of Qt? Actually, it's developped enough to be part of system's standard library. What do you think? Imo it would save much effort.

[M4GNV5]

Qt? sounds like C++
gtk masterrace

[kamirr]

Gtk is mainly gui, Qt has… everything. And as it's OO D wrapper would be quite good.
//edit: Developpers of PowerNex would have to implement low-level operations and nice D wrapper but pretty much work would be done by other people. I mean – no need of implementing high-level stuff (it's hard in big projects), which is nice.
//edit2: Aaand Qt has epic Qt Creator, so Qt masterrace.

[Vild]

I don't see why I should use qt, it seems like it would be too big. I want just a gui library, not a alternative standard library.

My goal is also to have everything written in D, so if I would take a preexisting library I would take dlangui. But my current goal is to write one from scratch.

[rjmcguire]

dlangui is the right approach to get something up quick even if it doesn't quite fit your code style or whatever. ... I went and looked at the code now, and there is a lot of stuff in that library.

Perhaps it would be interesting to just port the console and dml parts to PowerNex for now? I think that would help UI people to start playing around with PowerNex.

Is there a equivalent to epoll / kqueue in PowerNex yet? Any pointer into the code for event driven programming?

[ZILtoid1991]

I wrote a GUI framework (which I named Concrete) for my game engine for various reasons (editor, etc.), if you want to then you can recycle those parts of the engine. Link
Things that might need to be changed and other issues:

• Rewrite the draw functions to be 32bit natively instead of 16bit paletted. I can do that after my finals or just rewrite the blitter algorithms in the BitmapDrawer class to 32bit. You might want to port the code from Assembly to make use of the built-in SIMD commands of Phobos.

• The skin is made to mimic old systems and thus are designed for low-DPI rasters. Either needs a new skin, or deal with it.

• Drop-down menus are currently unimplemented, as the main goal for version 0.9.1 is to get the collision detection working again as I tried to optimize it. As I have to completely write it by myself (so I could use the engine for my thesis if I find a consultant), I cannot allow external people to make it work. Either wait for me to get it working or consult about me on changes.

• There's no hardware acceleration at the time, my whole engine will need a custom code in OpenCL due to the emulation of fixed point rendering. Currently I cannot work on that.

• Windows are handled as sprites. It can waste a lot of memory to store the prerendered windows, also I currently don't have any way to limit the rendering area to the places changed, thus wastes a lot of CPU time.

[ZILtoid1991]

My finals have ended, if needed I can work on the GUI (as well as some standard graphics drivers to make it work). I even came up with ideas for a DirectX-esque framework containing stuff for input handling, audio, and 2D graphics (3D should use a port of OpenGL and/or Vulkan).

[Vild]

Currently, I'm in the process of refactoring and planning big parts of the kernel, so I wouldn't really recommend working on a GUI framework yet.
I think it will take ½-1 year before it would be ready for one.

So you could work on one if you want, but I can't guarantee if it will be used, just because I want the framework to feel natural in the OS, both API wise and infrastructure wise.
I don't know how the kernel infrastructure will look when it is time due to the code is being changed a lot.

I will set up some sort of planning place for planning different parts of the OS, in the near future. But until then I can't really recommend that you plan too much stuff for PowerNex.

[ZILtoid1991]

How's the OS now? If those things are finished, I can start the development of Concrete for PowerNex.

[Vild]

Ehh, sadly school took a lot of time and energy from me. Can really been seen here https://github.com/PowerNex/PowerNex/graphs/commit-activity

I still have two reports I need to write before I have time to really work on PowerNex again.
But I have done a lot of the refactoring already, about 50-75% I would guess.
I mostly just need to move a few things, restore the functionality that the latest release had, and design a less shitty API for framebuffer access (and a few more userspace APIs)