Embedding a framebuffer window[edit]

In the past month I started working at the Embedding a framebuffer window project. Like last year, I was really excited about participating in Google Summer of Code for BRL-CAD.

The project can be split into two main parts:

  • Creating a new Qt framebuffer
  • Embedding a framebuffer window

As expected, I started by creating a new cross-platform Qt framebuffer. Firstly, I studied the existing framebuffer implementation, especially the X24 framebuffer. The second step was to create, just like in the display manager's case, a text framebuffer that just logs function calls. Being familiarized with the code and with BRL-CAD's libraries made this task a lot easier than it was last year.

After having a new framebuffer interface, even though it was a simple one, I was able to integrate the framebuffer so that in can be used by the rt command. At this point, it was the time to start implementing the open and close functions. Having last year's experience, made the integration of Qt in cmake a lot easier and also the creation of an empty Qt window (which basically is what the open function does). I was really excited when the window was created, but having an empty window isn't really helpful. So, I started working at the write function so that it can display raytraced content. After some struggle with some bad drawing, I managed to successfully display content:

BqYS7Mr.png kFaMQk9.png

As seen from the above pictures, the new framebuffer is integrated with the raytracer and it can successfully display raytraced models. However, there still are things that need to be done such as using, reading and writing a colormap or adding the keyboard and mouse functionality. In order to implement this, I have to do a little more research so that I could find the best approach.

Since at this time I am a little bit busy with my final paper and I can't say I am really eager to do more research, I decided to change a little bit the original schedule and start working at embedding a framebuffer window. I have done this because I have a more clear idea on what to do and the missing functionality of the framebuffer shouldn't interfere with this task.

So, in the following week or two I hope to have the embedding done so I could start working at the missing features. Even though, I changed a bit the implementation order, I am on track with the original schedule and everything should be done as expected.

As it comes to the actual code, until now I've written about 750 lines of code and almost all of it can be found in libfb/if_qt.cpp. In order to have a fully functional framebuffer, I think it would take another 700-800 lines. What I really like at this project and what keeps me motivated is the fact that I can have visual feedback, which is different from other projects I have recently worked at.