Telegram for Firefox OS?

[mautematico]

I'll just post it here and see what happen next... :)

Try it! just as http://loqui.im/ is trying. Do you have any plan on it?

[dannluciano]

I want to develop something.
But I fear have to rewrite all c libs in javascript.
In my vacations after the new year I hope to start the project.

[mautematico]

Thank you!

I'm pretty sure you can find some help (if needed) on the Mozilla's list:
http://lists.mozilla.org/

Also, @loqui developers maybe interested on this.

[sebasmagri]

I don't have much experience with emscripten, but would it work for this?

Best Regards,

[dannluciano]

Hi guys I recommend this e-book https://leanpub.com/quickguidefirefoxosdevelopment/ for want learn the first steps in Firefox OS

[dannluciano]

I have many busy in work but I don't gave up of the ideia.
I am a teacher and I have a few students in a research group that could take the first steps.
I will meet in the coming week.

[iamricard]

I'm working on it (learning the API first).

[iamricard]

Would you guys be OK if I create an organization (Telegram FFOS) and added you guys to it so we can collaborate?

[dannluciano]

No problem.
Me and 'my' students, we are learning the api of FFOS.
The goal for the semester is develop the user interface.

[iamricard]

@dannluciano cool, I've added you to the org. I was planning on doing it as my final school project. I'm also looking into Emscripten

edit: @sebasmagri about Emscripten I think it should work, we would only have to work on the UI then. Except for the part where we have to extract what is just the API from this project because we don't need the whole CLI stuff. So In a way (I think) we need to turn this into a library, not a client (which would be really good since it would open up possibilities for clients in other languages). So far I'm trying to comment and document the code because it's terribly commented.

[iamricard]

So I tried building with Emscripten and (as expected) did not work out of the box.

[sebasmagri]

@rcsole Do you think there is some reusable out of the Emscripten output? We'd need to determine if it's worth the effort or if we'd go with a new API upon FxOS specific TCPSocket API as referenced at loqui/im#161 (Spanish).

[iamricard]

@sebasmagri Honestly, I have no clue. The Emscripten doesn't not get far enough to output anything. I am, however, trying to isolate the API from this project, comment it, and then run Emscripten on it because this does seem like a viable project and having a base on C instead of JavaScript does sound like a better plan.

Either way I think the way to go is write some C (or C++) code which afterwards let's us run Emscripten on it, because that's be helpful if anyone were to try and port it to say python or lua (the C part, not the JS part).

I'll add you to the org @sebasmagri

[sebasmagri]

@rcsole having a base C library does sound great. It would be extracted from tg's code and tg would be able to use it also.

[iamricard]

@sebasmagri exactly my idea. But I am going rather slow because this code, while great, is poorly commented (actually not commented whatsoever) and it does have a couple thousand lines in total.

[sebasmagri]

It would be amazing to get some support from tg for this... ;-)

[iamricard]

Indeed, it would be really great. I don't think it's going to happen though. I'll start by commenting the code on the Telegram-FFOS fork, and I guess I'll send pull requests every now and then. Ideally I'll create a new repo containing the API only. I'll keep in touch!

[novia713]

i like the idea
i know javascript but no idea of C
count on me, please!

[vysheng]

@sebasmagri What do you want from this C library? In particular, how do you see it's interface?

[sebasmagri]

@vysheng a pretty straight forward library with methods for common API operations needed for a client would suffice. It doesn't need to deal with UI, just API operations.

[vysheng]

@sebasmagri Problem is with data structures that are used to represent user, message and so on.

[sebasmagri]

@vysheng I have not had time to deeply look at the code. Could you please elaborate?

Thanks for showing up!

Regards,

[vysheng]

@sebasmagri I use some big structures where I can put everithing, that server can tell me about this peer (user, for example). It is easy to use in app, because you have all information about peer in one structure, but for library it is not so good, because it will use more memory, then user wants.

[sebasmagri]

@vysheng a good starting point would be to break such big structures in smaller types for each entity exposed in the API. You could use such types and keep them in your big structure later if you want to.

[pravi]

http://zhukov.github.io/webogram may be this is helpful. Someone already wrote a web based client for telegram.

[iamricard]

@pravi thank you!

[jcatumba]

I would like to join the Telegram FFOS team! Thanks.

[gitaarik]

The webogram project looks like it is already going good. I think it would be wise to join this project for a FFOS version, because FFOS runs webapps anyway.

I don't completely understand the idea of having a C library as a base, you can't run any C in web apps. Besides that, the app will comunicate with the Telegram API, so why would you need some shared codebase anyway?