User:Vladbogolin/GSoC2014/Logs

Project Info

Project Name	Embedding a framebuffer window
Student Name	Bogolin Simion Vlad
IRC nick	vladbogo
Link to proposal	http://brlcad.org/wiki/User:Vladbogolin/Proposal/EmbeddingFrameBuffer

Milestones

• Implementing a new Qt framebuffer
• Raytracing
• Embedding a framebuffer window

Progress reports

• Midterm evaluation report can be found here

Logs

Community Bonding

• Studied the existing framebuffer implementation
• Researched to see exactly what framebuffer functions are called when using the rt command

Week 1

• Weekly milestone:
  • Start working at the "open" function.
• 19 May:
  • Started creating a new Qt framebuffer: I started by creating a txt framebuffer that just logs function calls.
• 20 May:
  • Mostly a break day
• 21 May - 25 May:
  • In this interval I had my final exams

Week 2

• Weekly milestone:
  • Make up for the lost time due to exams
  • Implement the open/close functions
• 26 May:
  • I had my last exam so from now on I will focus on the project.
  • Worked a little bit more on the txt framebuffer
• 27 May:
  • Worked at attaching the txt framebuffer to the display manager - the txt framebuffer is almost done.
  • Studied the existing "open" functions.
• 28 May:
  • Attached the qt framebuffer to the display manager.
  • Started working at the "open" function.
• 29 May:
  • Continued working at the "open" function - it should be done shortly
• 30 May:
  • Researched how the raytracer uses the framebuffer and attached the new framebuffer to it.
• 31 May:
  • Mostly a break day.
• 1 June:
  • Break day.

Week 3

• Weekly milestone:
  • Finish implementing open/close and start working at the write function.
• 2 June:
  • Continued working at the open function: created an empty window - the function should be done tomorrow.
• 3 June:
  • Basically finished the open function - some changes might be needed during the implementation of the other functions.
• 4 June:
  • Break day.
• 5 June:
  • Researched the write function and the existing framebuffer implementation.
• 6 June:
  • Break day.
• 7 June:
  • Continued researching the write function and how the content is displayed.
• 8 June:
  • More research, but mostly a break day.
• Conclusion:
  • Finished implementing the open function and studied existing implementation of the write function

Week 4

• Weekly milestone:
  • Have content displayed in the window
• 9 June:
  • Solved a issue regarding the cmake build - I forgot to add Qt_INCLUDE_DIRS.
  • Created a custom window so that it can be easily added content to it.
• 10 June:
  • Released memory so that the qt_close function is done.
  • Started working at content displaying. Since this is the main aspect I want to do as much research as possible on how to represent the backing store and how to display its content.
• 11 June:
  • Break day.
• 12 June:
  • Created and allocated the backing store for the framebuffer.
  • Started working at the write function.
  • Managed to get some content displayed - even though it's far from done, it is a good start

• 13 June:
  • Continued working at the write function.
• 14 June:
  • Managed to fix the content display:

• 15 June:
  • Mostly a break day.

Week 5

• Weekly milestone:
  • Finish the write function.
• 16 June:
  • Finished a first version of the write function. Also, tested if more complex models are displayed correctly.

• 17 June:
  • Started working at the open_existing function.
• 18 June:
  • Break day.
• 19 June:
  • Searched for the best way on how to create the open_existing function. Since until now, the functions were declared in fb.h and all framebuffers use the same header, after adding the Qt includes a lot of conflicts appear so a new method of dealing with the open_existing functions is necessary.
• 20 June:
  • Continued studying how to integrate the open_existing function.
• 21 June:
  • Break day.
• 22 June:
  • Worked at the open_existing function.

Week 6

• Weekly milestone:
  • Finish the open_existing function
• 23 June:
  • Worked on raytracing from the display manager.
• 24 June:
  • Break day.
• 25 June:
  • Created the midterm report that can be found here
• 26 June:
  • Processed events so that the output is displayed line by line
  • Fixed the upside down drawing problem:

• 27 June:
  • Break day.
• 28 June:
  • Started modifying the fbio interface in order to add a generic open_existing entry.
• 29 June:
  • Mostly a break day.

Week 7

• Time spent preparing my final project

Week 8

• Weekly milestone:
  • Finish modifying the interface in order to use the open_existing function
• 7-8 July:
  • Had my final paper presentation. Now I can focus only on GSoC for the rest of the summer.
• 9 July:
  • Added a new entry in struct dm so that the framebuffer can be opened from libdm. This is necessary in order to avoid using qt code in mged/dm-qt and also can be used for refactoring dm related code from mged folder.
  • Added the _qt_open_existing function in fb.h. This is a temporary solution until a final form of the function is implemented. In the end a new entry will be added in the fbio interface.
  • Opened the framebuffer from the display manager, but at the moment there are two different windows.
• 10 July:
  • Continued working at the open_existing function. Hopefully it will be done by the end of the week.
• 11 July:
  • Struggled more with the embedding of a framebuffer window.
• 12 July:
  • More struggle regarding the display of raytraced objects from the display manager.
• 13 July:
  • Mostly a break day.

Week 9

• Weekly milestone:
  • Finish the embedding of a framebuffer window
• 14 July:
  • Worked at the actual embedding of a framebuffer window - still couldn't find a working solution
• 15 July:
  • Did a deeper research on how the existing implementation is done. Still couldn't find a way do display content in the embedded window.
• 16 July:
  • Break day
• 17 July:
  • After a lot of struggle, I have finally found a solution that seems to solve all the problems. Hopefully tomorrow there would be content displayed in the embedded framebuffer window.
• 18 July:
  • After using a QPainter to draw the raytraced content in the existing window, content can be displayed so I am on the right track with the embedding. There still are some fixes that need to be done: when the raytracing is done the drawn content disappears and also a few fixes regarding the position of the drawn image are necessary.
• 19 July:
  • Tried to solve the problems regarding the embedding so I did a little more research on why the content disappears but with no luck yet.
• 20 July:
  • Tried to solve the problem regarding the position of the drawn image.

Week 10

• Weekly milestone:
  • Have a working framebuffer.
• 21 July:
  • Implemented the configureWindow function.
  • Continued trying to solve the problem regarding the position of the drawn image - still no luck.
  • Redesigned the write function in order to solve the display problem - there still are some fixes that need to be made.
• 22 July:
  • Continued working at the new write function.
• 23 July:
  • Solved the display problem - now correct drawings can be made in the embedded framebuffer window:

• 24 July:
  • Tried to find out why the image disappears after is drawn in the embedded framebuffer window - more research needs to be done since I haven't found the problem.
• 25 July:
  • Break day
• 26 July:
  • Break day
• 27 July:
  • More research regarding the disappearing of the drawings - found the problem (after the raytracing the dm redraws the entire window and the raytraced content erases in this way) and tried to come up with a solution.

Week 11

• 28 July:
  • Continued trying to solve the disappearing problem.
• 29 July:
  • Solved the disappearing problem. The framebuffer is almost done - there still are some fixes and tests that need to be made.
• 30 July:
  • Break day.
• 31 July:
  • Mostly a break day.
• 1 August:
  • Made some fixes regarding the embedding of a framebuffer window.
  • Tested the embedding with different models - found some problems that need some further investigation (at some point the display manager crashes)

• 2 August:
  • Investigated to see why the display manager sometimes crashes using valgrind.
  • Started working at solving the invalid memory access.
• 3 August:
  • Mostly a break day.

Week 12

• 4 August:
  • Continued working at the invalid memory access problems - now everything works as expected.
• 5 August:
  • Implemented the qt_clear function.
  • Made the display manager show only non-empty images (after the raytraced content is added to the image)
• 6 August:
  • Integrated the fb in archer:

• 7 August:
  • Did some testing to see that everything works as expected including the underlay, interlay and overlay modes
• 8 August:
  • Implemented the qt_read, qt_readrect and qt_writerect functions.
  • Integrated mouse events.