Pitivi GSoC: Final Report

3 min read | Sun Sep 11 2022


/images/uploads/gnome-gif.gif

Pitivi GSoC Final Report

I Love Pitivi

This project was aimed at porting Pitivi to GTK4, Pitivi used the GTK3+ library and required the change.

The whole project is confined to a single MR: gitlab.gnome.org/GNOME/pitivi/-/merge_requests/442

The port required extensive changes and brainstorming. There were namely three categories of changes -

  1. Component renamed
  2. Component gone but ideal replacement given
  3. Component gone with no real/simple replacement

Some code needed refactoring, with varying complexity and length, some were small but some were very extensive, for example, GtkDialog run() removal, some parts were easy to refactor to the callback-based system but some places still remain to be ported over due to the extent of their reliance on the return process.

After getting suggestions, I have also listed basic things I've done in the MR for easy reference and tracking.

The last commit I did during the GSoC period is: "Fix effectslibrary.ui search" "gitlab.gnome.org/GNOME/pitivi/-/commit/0ad9503df5f49e3a3dbbf1f71bc3b87b706ad213?merge_request_iid=442"

Work Done -

Till now I have been able to port a lot of the stuff to GTK4, most files are ported over and with some local changes I can open the application, this has given me a chance to understand what I should focus on. Pitivi is hard to port due to its massive size, but I'm really happy that I was able to port it to the extent I did.

Progress and hardships in chronologial order -

During the first update -

During the second update -

During the third update -

During the fourth update -

Is the port complete?

No, there is still work remaining, but it is not much, most things are done, and the things which are remaining, are so because of the amount of refactoring they would require and some cosmetic fixes.

Most of the application is ported over, but if you open it, you won't be able to do much, the reason for which is that if some places fail to run, then it will prevent the rest of the file from loading further, making a lot more things useless. Currently, you can open the application, open, import, or create a new project, access the effects and medialibrary (some work is required), and other miscellaneous things.

The timeline still remains to be properly ported due to the removal of GtkLayout and drag & drop event controller.

You cannot hope to do extensive refactoring work during GSoC because that will limit how much work you can do. In the start I was pushing commits with great quality, i.e. no change required later on, but that results in less time remaining thus it is better to get a working version and push that. If you still have remaining time before the deadline and other work is done then make those commits better and better, else just continue and after the work is done, focus on how you can improve it further to get the best result.

At least that's what I've come to believe.

But it depends on the project's length, complexity and expectations.

My plans after GSoC -

GSoC never made me join Open Source, it is something I believe in and love to be a part of. GSoC gave me an opportunity to work on such a big project along with the support of mentors and I appreciate all the love I've got. Thus, to conclude, this was just the beginning, you will keep hearing from me, in a good way :D

GUADEC -

I also had an amazing experience presenting my work during GUADEC, it gave me great confidence and was an awesome experience. Thanks, GNOME for providing me the opportunity.

You can watch me present at - GUADEC Youtube

GUADEC

That will be it for this one, have a nice time everyone :)