Day Three dawns. We have a compressed schedule today before we break for individual workshops/breakouts - plus some people will be leaving us early in order to get home - so let's get to work.
neo1973 opened the day with a discussion about merge policies, specifically around reviews and self-merging your own PRs. There's a reality that, when you've been deep in a certain area of the code, you may well be the subject matter expert - for the moment, at least - so there's nobody to necessarily critique the what of your code. As such, there's an element of "publish and be damned". There is, though, also the how: does the code itself stand up to inspection? We already use tools to sanity check the code, check for null pointers and undestroyed variables, and similar, so the quality side is already covered to a certain extent. That said, AI tools are falling into place and improving quickly, so there's more that we can do here, and more to investigate.
Onwards into a conversation initiated by yol around some of our hosted infrastructure and some specific issues we have there. That in turn led into a more general conversation around hardware, usage, I/O and capacity, and similar: important to us, but probably less so to the general public.
Next up, chewitt with an update on LibreELEC: user numbers, current platform trends and future plans, dependency/kernel versioning, CI/build, upstreaming fixes, drivers, chipsets, hosting, finances, collaborations, architecture. For what should be obvious reasons, samnazarko weighed in where appropriate with an OSMC perspective.
Finally, lrusak gave an update on a long-desired feature: headless Kodi. It basically exists, it works, but it needs a good chunk of tidying up to make 100% certain that nothing tries to create or access a GUI on a non-existent window manager (e.g. popping up a notification toast). There are headless implementations - such as Docker images - already available from other sources, so this is really just about bringing this into the official release.
And, with that, the main conference comes to an end. Time to huddle together, fire up laptops, and do some coding.
So, that's it for DevCon 2025. Genuine thanks to Tirana for the hospitality, and to everyone who's helped to look after us this year - and thanks to you for reading.
Until the next time ... lamtumirë, dhe faleminderit.
Currently working out a series of videos for theater intros. Using Unreal Engine. Slow going but I hope to make one for every genre. At least I have one. I customized the Regal Cinemas 1990's roller coaster to my theater. Used some AI to change some wording in the narration and Davinci resolve to edit the text on the marque scroller. Gruelling work as unreal Engine is hard to learn. Nothing is straight forward. Everything is a task. But I hope in the end it looks great. Doing an old West town
For the initial release of the trivia metadata I'd like for you to test the basic scoring functionality. Here is the trivia folder to be installed in your PreShow trivia content folder. The trivia testing pack includes the following: Rating G.jpg PG-13.jpg PG.jpg R.jpg Decade 1950s.jpg 1960s.jpg 1970s.jpg 1980s.jpg 1990s.jpg 2000s.jpg 2010s.jpg 2020s.jpg Genres Action.jpg Adventure.jpg Animation.jpg Comedy.jpg Crime.jpg Documentary.jpg Drama.jpg Family.jpg Fantasy.jpg Histor
I have a friend that wants to watch the "scariest movie." I think scary is very subjective, so I figured I'd ask it here. What do you think is the scariest movie?
Day Three dawns. We have a compressed schedule today before we break for individual workshops/breakouts - plus some people will be leaving us early in order to get home - so let's get to work.
neo1973 opened the day with a discussion about merge policies, specifically around reviews and self-merging your own PRs. There's a reality that, when you've been deep in a certain area of the code, you may well be the subject matter expert - for the moment, at least - so there's nobody to necessarily critique the what of your code. As such, there's an element of "publish and be damned". There is, though, also the how: does the code itself stand up to inspection? We already use tools to sanity check the code, check for null pointers and undestroyed variables, and similar, so the quality side is already covered to a certain extent. That said, AI tools are falling into place and improving quickly, so there's more that we can do here, and more to investigate.
Onwards into a conversation initiated by yol around some of our hosted infrastructure and some specific issues we have there. That in turn led into a more general conversation around hardware, usage, I/O and capacity, and similar: important to us, but probably less so to the general public.
Next up, chewitt with an update on LibreELEC: user numbers, current platform trends and future plans, dependency/kernel versioning, CI/build, upstreaming fixes, drivers, chipsets, hosting, finances, collaborations, architecture. For what should be obvious reasons, samnazarko weighed in where appropriate with an OSMC perspective.
Finally, lrusak gave an update on a long-desired feature: headless Kodi. It basically exists, it works, but it needs a good chunk of tidying up to make 100% certain that nothing tries to create or access a GUI on a non-existent window manager (e.g. popping up a notification toast). There are headless implementations - such as Docker images - already available from other sources, so this is really just about bringing this into the official release.
And, with that, the main conference comes to an end. Time to huddle together, fire up laptops, and do some coding.
So, that's it for DevCon 2025. Genuine thanks to Tirana for the hospitality, and to everyone who's helped to look after us this year - and thanks to you for reading.
Until the next time ... lamtumirë, dhe faleminderit.
/images/blog/team-kodi-tirana-2025.webp
Team Kodi 2025, Tirana
View the full article