<tap, tap, tap> – Is this thing on? Yes, it is, and so it’s time to once again introduce the highlight of the Kodi year – friends and fellow travellers, welcome to DevCon 2025!

This year we’re coming to you from Tirana in Albania (or Tiranë in Shqipëri in the local language†). A country that, I suspect, most people know little about, Albania has a long and … bumpy history, mostly involving other countries and some less-than-friendly regimes: Romans (both western Roman and eastern Byzantine), Ottomans, Austro-Hungarians, Italians, from imperial territory to independence, from international engagement to isolationist communism to modern democracy. Sitting astride both the Aegean and Ionic seas, the country now embraces the promise of full modernity and EU membership as it strives to turn its back on the darker chapters of recent history and take its rightful place in Europe.

As in previous years, we’re going to aim to do a daily update (content-dependent), so let’s see how it goes. Without further ado, then, let’s get straight into it with Day One.

We began with round-table introductions and a typical retrospective: reminiscences for older heads, background for newer ones. We’re coming up to a quarter of a century of XBMC/Kodi, and, as you’d expect, there’s now a significant story of how we got to where we are, and some of the main contributors and contributions along the way.

We continued with an update on the Foundation’s financial position from keithah. We’ve always been very transparent that we’re not a wealthy project, and we’ve never actively pursued financial gain for either Foundation or individuals. That hasn’t changed, although money is gradually getting tighter as e.g. major sponsors fall away, with limited opportunity to replace them in the current climate. We remain solvent, and can both pay our bills and invest in development hardware, server infrastructure and, yes, this conference, but we do need to generate new and more regular income to make sure that remains the case. In terms of priority, then, we wait to see if this is the last DevCon for a while.

Next up, we had a readout on the experience of our most recent release manager, garbear. Getting Kodi tagged, packaged and released on multiple platforms is a significant undertaking, so even point releases involve a good chunk of work. Even apparently minor changes involve merging new translations, for example, plus there are inevitable rule/eligibility checks on software stores and similar. Indeed, our most recent “Omega” 21.2 release actually included a complete update to the entire add-on ecosystem precisely because of translation issues, which means 100+ separate components needed to be built and released alongside the core Kodi application. Even that caused further ripple effects: we needed to update binary add-ons as well, which then, in turn, caused problems for existing 21.1 installations. This led straight into a more practical conversation, with garbear joined primarily by martijn, yol and lrusak to talk about how we could improve things: APIs, ABIs, links, scripts, dependencies, redistributables, SDKs, implications for different platforms – and, yes, much, much more about translations. Releasing new versions of Kodi is most certainly not a trivial task!

This conversation then segued into a long debate about build systems, and the complexity that Kodi brings: there are few, if any, multi-platform applications that use the same code base on quite so many target operating systems. If you think about most other applications that might be available on multiple platforms, the chances are that they use completely different code bases that are then built on specific tools to produce applications that might look similar, and might behave similarly, but are actually very different. Alternatively, you have very many applications that are basically web pages, rendered in a platform-specific wrapper – these can deliver a very consistent experience and rich multimedia, but in a very different way. Kodi doesn’t work like that: it’s a huge amount of C++ and other code that’s custom built to be as consistent as possible across platforms, and, while that resolves very many issues, it also creates some others at the bundling stage on some platforms.

A change of direction next: 78andyp joined to talk us through his recent work – and further plans – around Blu-ray support in Kodi. Currently, Kodi supports three disc playback modes: Kodi can show the Blu-ray menu, with all the navigation overhead; you can ask to play the main movie, and Kodi will just take a guess and play the longest media item on the disc, which may or may not be what you want; or you can go into file view, and wander around until you find that the main movie is labelled as item 636 out of a list of … well, lots. This is obviously all less than ideal, with further implications depending on whether it’s a multi-version disc, an episodic series disc, whether there are extras, and so on. Add to this the complexities of ISO versus physical disc, “playlists” with multiple episodes, or initial scraping versus play-time, and there’s clearly more to be done – and that’s the gauntlet picked up by 78andyp: to make playing a Blu-ray in Kodi as straightforward as playing a media file. This is very much a work-in-progress, but there’s a current 3,000-line pull request that aims to at least start this journey, addressing the simple file view for episodes at play-time. And, like so many things, this process is unveiling other side-quests on the way, ranging from NFO support through ranged episodes to library handling – for both Blu-rays and, potentially, DVDs. There’s clearly huge potential for much more to come, so watch this space, shiny disc fans!

(There was a sidebar conversation during this presentation in which ksooo raised some very valid questions about vision and the natural constraints on Kodi: what are the design principles we are using that ultimately drive what should and shouldn’t be included in Kodi versus, say, a media manager, and thus what are the dialogue boxes or other user interactions that are acceptable? Similarly, how effective could a heuristic be, when will it need user intervention, and what should that intervention look like? Does it need some external “priming”, or metadata information?).

And that’s it for Day One – it’s late, and dinner calls. We’ll be back tomorrow with more, so please, join us then!

† Fun fact: Albanian is the only surviving member of the Albanoid group of languages, and so stands alone, with no direct modern relatives. It’s the official language of Albania and Kosovo, one of the official languages in North Macedonia and Montenegro, and a widespread minority language in Italy, Croatia, Romania, Serbia and Greece. Add in the Albanian diaspora across the Americas, Europe and Oceania, and there are three or four times more native speakers worldwide than there are inhabitants of modern Albania – including singers Dua Lipa and Rita Ora.

Source link

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.

Team Kodi 2025, Tirana

Source link

So, a slightly different topic for a change.

We talk a lot about the Kodi software, its development and releases, and that’s fine, as that’s what we’re all about and that’s what everyone ultimately sees. However, we thought we’d change the subject a little and explore what goes on behind the scenes: how “Team Kodi” is structured, and what else is going on in the background in support of that glorious code.

The Kodi Foundation and the Board of Directors

Kodi has a legal personality. We’re incorporated in the US as the Kodi Foundation, a 501(c)(3) non-profit – or a type of NGO, in common European parlance. This is the “company”, it’s the entity that receives donations, signs hosting agreements, protects our trademarks, and owns any assets and intellectual property. However, we don’t have any of the trappings of a regular company: we have no employees, minimal assets, and no property. We’re wholly run by volunteers and a few carefully-chosen outside agents for e.g. legal work.

The Kodi Foundation is then made up of members of Team Kodi. It’s not mandatory to join – you can be a Team member and not join – but it’s the Foundation members who get a vote on major decisions. Probably the most regular of those decisions is the appointment of our Board of Directors – again, Team members, all volunteers, each of whom serves a two-year term as part of a Board of five members. The Board then manages Kodi: funds, publicity, taxes, lawyers, copyright bodies, banks, sponsors, non-profit oversight, and so on. There’s also a whole raft of internal administrivia – member lists, permissions, access rights, what’s filed where, who’s responding to what, why are we getting these emails, no we don’t want your adverts.

As both a Team and as a Board, we hold regular Zoom meetings to discuss progress, schedules, and any related issues that may crop up from time to time. As you can imagine, this can be quite a challenge given the timezones we cover, so everything is available afterwards for Team members to catch up on.

The Greater Team Kodi

It’s not just paperwork, though. It’s easy to think of Kodi as a bunch of lines of code, hammered out in the small hours by pizza-fuelled developers, and that’s not necessarily too far wide of the mark in some ways. But there’s also so much more: even to maintain that code needs discipline, a release manager, code reviews, Github admin. We have people who look after build servers and DNS records, who administer our email and collaboration tools, who feed and water everything from backup servers to the server that hosts the forums: software upgrades, patches, hardware replacement, migrations. And we have people to write posts like this, to design our graphics for each release, to modify skins, to compile and build new releases.

Funding it All

As mentioned above, this is all voluntary: nobody at Kodi gets paid anything for what they do. We all do this as a hobby, because we enjoy it in some way. And what makes that possible is primarily you, our users.

We generate a few tens of thousands of dollars a year. Most of that comes directly from donations – that little orange button in the top right; some comes from merchandise sales such as T-shirts or royalties on branded Flirc goods; and the rest comes from corporate sponsorship, which is an ephemeral beast at best. You’ll perhaps notice that we’ve always steered clear of advertising, so there are no popups on this site, no banner ads, and nothing within the application itself: that’s very deliberate, and we work hard to keep it like that.

On that note, we’d like to give a huge shout out to Flirc right now, as they’ve just started selling a Raspberry Pi 5 branded Kodi case. This is the only company authorised to use our brand on a product, and the only company who gives us royalties: we get a small amount for every case sold. They’ve been consistently helping us out financially for the last few years and, without them, we would be attending many fewer conferences.

And where does the money go? We have hosting costs, professional service fees (lawyers, copyright specialists, accountants), banking fees. While we sometimes get sponsored hardware, we have to pay for servers, storage, bandwidth. If someone is working on a feature and they need a particular piece of hardware that we can’t get donated, we’ll buy them something (e.g. to assist with porting Kodi to a new platform). Perhaps our biggest costs are conferences: the Open Source world thrives on community, so we have tried to get people representing us at FOSDEM, Embedded Linux Conference, OSS Summit, SCALE, and the Open Source Leadership Summit – although the majority of funds goes to getting as many members as possible to attend our annual developer conference, Kodi Devcon.

We think we do pretty well on limited funds, and we’ve had some big donations in the past, but unfortunately those are drying up, so we’re starting to have to make hard decisions on not covering our volunteers costs to represent us at some of the conferences.

Once 2024 ends, we will publish a detailed financial blog post, outlining our expenses this past year, but we’d love help from readers and users like you!

What’s Needed

• More hosting is always appreciated. We get by nearly entirely from free hosting and can’t thank our sponsors such as OSUOSL and Leaseweb enough, but can always use more diversity.
• We always need volunteers to work on stuff! Social Media and someone to manage our store are two that come to mind, but platform-specific development experience (e.g. Android) is always good.
• Sponsorship! if you think you have a good fit for a sponsor who aligns with our interests, please, reach out!
• If you have any corporate matching, Kodi is registered as a corporate nonprofit beneficiary with the Benevity platform.
• And lastly, monetary donations go a long way too. As stated, we are a 501(c)(3), so these donations are tax deductible in the US. If you have a few dollars, pounds, euros or equivalent going spare, you can really help us out. Thank you!

We always want to keep some money in the bank, and we have a rule: we need to be able to cover the next Devcon before we hold this one and, unfortunately, 2025 may be the last year for a couple years at the current rate of donations. We spend less than US$30k for the entire event, including flying our team from around the world to attend plus all hotel/food/conference room costs, but our current finances just won’t support this after next year’s conference unless something changes, such as a new sponsor (or the RPi5 case becomes exceedingly popular and sells out!).

Thank you very much in advance for anything you can do to help.

— Team Kodi

Source link

Okay, a late announcement – apologies, you can blame the summer holiday season. Or, if it’s easier, you can just blame me as the bloke with the keyboard, that works equally well. Either way, many of you will already have noticed this, but, in case you haven’t …

Team Kodi is here with the first point release of the 21.x “Omega” series.

A range of bugfixes have been collected since the original release of Omega 21.0. Rather than list them all here, you can see the full changelog on Github, as usual.

We always want to thank everyone who has helped us track down and fix any issues. We endeavour to minimise the issues everyone experiences, but with such a large project, and the fact everyone contributing to Kodi is a volunteer, issues and bugs are a part of life. If you happen to experience any bugs/issues, don’t hesitate to reach out on the forums, or raise an issue on Github.

As this is a point release, there are no major changes since the previous version, and you should be fine to install this straight over the top of any existing Kodi 21.0 installation – indeed, this will happen, and probably already has happened, automatically on many platforms. However, as for all software installations, back up your userdata beforehand if you’ve any doubts or have anything you can’t afford to lose (and definitely do this if you’re going for a major version upgrade).

A list of all changes for this version can be found at GitHub. If you want to read back on the full history of v21 itself, or of previous versions, you can find the corresponding articles in the relevant blog posts.

Application deployment on different platforms (notably the Microsoft Store) can vary considerably due to circumstances outside of our control, so just be patient, and the update will inevitably find its way through.

Android Deployment

We always do what’s known as a “staged” deployment roll out for Android. So if you didn’t receive the update immediately, it should have come through over the following week or so, and rollout should now be complete.

Source link