State of the attempt to natively port love to android

[Robin]

Every game should come with a LOVE binary of its own. No user is going to memorize which games he downloads are for which version of LOVE.

Well, preferably most games will be released when LÖVE has a stable API, so it wouldn't matter. And otherwise, this could still be managed transparently for the user.

[TechnoCat]

Every game should come with a LOVE binary of its own. No user is going to memorize which games he downloads are for which version of LOVE.

Well, preferably most games will be released when LÖVE has a stable API, so it wouldn't matter. And otherwise, this could still be managed transparently for the user.

You aren't getting it.
If I download a .love from the market that was released when 0.5.0 was stable, I'd try to run it with LOVE 0.7.2 and it would fail. Which I would equate to a terrible user experience right off the bat.
Download app, open app, app doesn't open, 0/5 stars, uninstall.

[Robin]

No, I do. I just don't call 0.7.2 stable.

When LÖVE 1.0 is released, we can start worrying about compatibility.

And my other solution still stands: oldgame depends on love050, newgame depends on love072, even if you have a ton of games, you would only need one copy of LÖVE for each version you have at least one game for.

[TechnoCat]

And my other solution still stands: oldgame depends on love050, newgame depends on love072, even if you have a ton of games, you would only need one copy of LÖVE for each version you have at least one game for.

Well, I suppose there could be a LOVE binary pool application. And it auto downloads the binaries required when it opens a .love and caches that specific love version binary.
But to be honest, I would much prefer the fused binary as a developer and as a user.

[T-Bone]

Android users expect apps to be completely stand-alone. Doing anything else would be to struggle against Android's ecosystem, which will only end up hurting the end user.

And my other solution still stands: oldgame depends on love050, newgame depends on love072, even if you have a ton of games, you would only need one copy of LÖVE for each version you have at least one game for.

This is only useful if the end user has plenty of LÖVE games installed on their phone. This is, quite honestly, unlikely, even after a few years. And even if that was common, phone memory will have expanded to the extent that it wouldn't be an issue anyway. One could argue that we're already there, there are tons of Android games out there that are over 10 MB large and nobody think's of that as an issue. And since love-native-android itself is roughly 6 MB, a normal sized game with a few music tracks will be say 7-8 MB which is perfectly normal. And with love-native-android's excellent App2SD support (some of the best I've ever seen in any app), space is even less of an issue.

Just compile and distribute love-native-android games as any other Android app. One shouldn't struggle against the OS one's targeting.

[Moe]

I'd love to integrate it, it allows us to concentrate on bugfixing. Can you upload it somewhere, mail it to me, or make a pull request on github?

After some trouble getting used to how git works, I managed to send a pull request. It should be available on github as I'm speaking

We will review your pull request and integrate the relevant parts only. Don't worry about that.

Also, I want to confirm that all crashes disappear in my game if love.modules.audio is disabled. The game starts every single time.

Good to know. I will investigate and try to fix this as soon as time permits.

Well, I suppose there could be a LOVE binary pool application.

This sounds like an awesome solution to me. Technically it would be possible. I am also a big fan of a pure engine release with an integrated loader and I plan to put it into the Android market (or how it is called at that time), because this would allow to download a love file wherever you want and to run it. No special Android version of the game is required, sounds perfect to me.
You can still create a single game build and do Android optimisations, but I do not want to limit the port to this solution. It will have the feeling of an emulator, but this does not mean that it is bad.

[TechnoCat]

Well, I suppose there could be a LOVE binary pool application.

This sounds like an awesome solution to me. Technically it would be possible. I am also a big fan of a pure engine release with an integrated loader and I plan to put it into the Android market (or how it is called at that time), because this would allow to download a love file wherever you want and to run it. No special Android version of the game is required, sounds perfect to me.
You can still create a single game build and do Android optimisations, but I do not want to limit the port to this solution. It will have the feeling of an emulator, but this does not mean that it is bad.

Oh no! What have I done?!

[miko]

Well, I suppose there could be a LOVE binary pool application.

This sounds like an awesome solution to me. Technically it would be possible. I am also a big fan of a pure engine release with an integrated loader and I plan to put it into the Android market (or how it is called at that time), because this would allow to download a love file wherever you want and to run it. No special Android version of the game is required, sounds perfect to me.
You can still create a single game build and do Android optimisations, but I do not want to limit the port to this solution. It will have the feeling of an emulator, but this does not mean that it is bad.

Oh no! What have I done?!

Don't panic Lets try and see how those 2 solutions work out. It is better to have a choice. Anyways, whichever way it takes, the most important thing is to get some stable love on android. Hope to have some soon.

[T-Bone]

Choice is great, but development should be focused on what's most important. But as Niko states, the most important thing is that we get a stable version of löve for Android. The source code can always be modified to suit every goal. But it is a matter of timing as well. I think that Android needs löve now. And my game needs löve as soon as possible too. And I am going to release it as a stand alone app, so whatever you do, don't make it difficult to do it that way.

because this would allow to download a love file wherever you want and to run it. No special Android version of the game is required, sounds perfect to me.

First of all, if löve games are stand alone apps, then people can download them from Google Play whenever they want, and run it, with the added benefit of doing it in a way that they know.
The second sentence sounds like an outright fabrication. If a game should run on a cell phone, it has to be adapted to the special conditions that apply there, such as limited and different control options, as well as limited performance. Any game intended to run on Android should be developed with that in mind. This should of course not stop games from being multiplatform, but you can't expect a computer game to work well on a phone, at least not without modification. And thus you need specific Android versions of games, unless the game checks at runtime wether it runs on a cell phone or a computer and adapts itself.

[Moe]

We started the native port with the intention of a Löve game launcher, the possibility to attach a single game only was a feature request right here in the forums. We will work on this feature, but my main focus will remain on our original target. And yes, this includes joypad and keyboard games.
So, to make things clear, our targets are:
1. Make Löve stable.
2. Add Android support and add emulation for non-native features, e.g. joypad.
3. Have a cool launcher.
4. Improve performance.
5. Enable developers to release single game releases.
And yes, the list is in the order of my personal preferences. Now, you can hate me