CEGUI (Crazy Eddie's GUI) Forums

any plan to migrate the libs on google android ?

Hi,

It's not something that I've personally considered so far. This does not mean that such an idea is not of interest, because IMO the more platforms we can support, the better (world domination, and all that ).

If you, or anybody, wants to get a discussion going about this, including any limitations, potential pitfalls, and other considerations, I think it could achieve a couple of aims - we would all get a better idea of what was needed and if it were possible, and maybe someone would take on the job

CE.

i am saying that because we are migrating to cegui especially because android support openGL light , so an application developped on ceGUI/openGL/SDL shall be easily rewrote to work on android .. i mean this is my guess, howver android is java. but this is no big deal movingfrom c++ to java, if the logic is there.

what u think?

Oh, wouldn't that basically mean porting the entire CEGUI library to java? I can't see that happening to be honest - I think I'd rather cut my left testicle off

Perhaps I misunderstood...

CE.

noo dont cut it

yes it would be this ... at the moment it is not possible to dev c++ app on mobile android .

i know its a pain in the ass; but it worth it! how long it would take?

CrazyEddie wrote:I think I'd rather cut my left testicle off

Is the right one already gone because of CeGuiSharp ?

^^

Pompei2 wrote:Is the right one already gone because of CeGuiSharp ?

Hehe, yeah

Ok, to be a little more serious for a moment, there are various factors to consider with regards to a java port of the lib, some of this is actually drawn from my observations of, and limited participation in, the development of CeGuiSharp (for those not in the know, this was written in C# which is another language that's 'easy' to port to from c++, because they have similar syntax).

Obviously the biggest question would have to be, who is going to do this work? It's almost certain that we - the CEGUI devs - could not undertake such a project. We're stretched really thin as it is, sometimes with virtually no time for CEGUI (C++ version) work (I say we here. I know this applies to myself, and am pretty sure it does scriptkid also). So somebody else would have to step up. Which leads on to the next potential issue.

Synchronisation. One of the largest bug-bears of CeGuiSharp was that the guys had such a hard time keeping that version in line with the C++ version. While there are not quite so many radical changes happening within CEGUI these days, it would obviously remain an issue that the two halves of one project could effectively produce a schism (which is basically what happened with CeGuiSharp, especially when Chris left).

As for the how long would it take question, the above points should make it clear that it's not a one-time deal, but an ongoing process. As for how long to do an initial port? Errrm... 4 months, 1 week, 3 days and 7 hours? Seriously, it depends on the amount of time people could put in - speaking for myself, I have very little time.

Those are the core issues as I see them. Then you come on to issues of interest. As I said, I'm very eager to see CEGUI spread far and wide, however, while I know a lot of people use CEGUI in the course of their work, for me, CEGUI is something done in my free time for pleasure purposes (no, really!). I like C++. I like the challenges it poses, the power it offers, the infinite trouble it allows you to get yourself into. Java is not something I have extensive experience with, although I have used it somewhat on occasion when necessary.

This post might sound overly negative, that's not particularly the intention, so please continue the discussion

CE.

yes but think about the achievment if cegui is running on mobile ... wow this would be a hugh extension!

for who is going to do it right u need more staff, any students around here?

Tiresias wrote:yes but think about the achievment if cegui is running on mobile ... wow this would be a hugh extension!

Oh yes, I agree. It would feel really great to have this

Tiresias wrote:for who is going to do it right u need more staff, any students around here?

Yeah, what we need is a couple of guys with some free time, who perhaps would like something really cool to add to their C.V / resume

CE.

as well,

what would be great is that a cegui guru (Eddie ?) could quantify the work to do ? i mean how big is the cegui is term of c++ line of code?

how difficult u think it will be ? is it just replacing opengl c++ apis with java opengl api + rewrite high level stuff in java ? this shall not be that hard ? or yes ?

in fact i can give money if project is open; little for sure but at least if project is started i can give something, we know all that time is money ... if all people interested give money that good!
there is a site where u can post a project and people start giving money, i will find the name for you, i am sure here u can raise lot of money as project is cool.

site:
http://micropledge.com/custom

I'm not sure how many lines of code there are at present, though I believe it's fairly substantial (I'll post the figure later, perhaps).

The difficulty of the task is something that's hard to quantify. My own general approach would likely to be to port at the function level, rather than trying to take the c++ code and java-ize it. So basically take the CEGUI classes and interfaces and effectively re-code as per the documented purpose. Obviously the C++ code could / should be used as a guide where implementation is not obvious. I think I would have to say the porting project would not be easy, but "challenging".

The only monies that we - the CEGUI devs - can solicit on this site (sf.net hosted) are via the Sourceforge project donations system (which is something we are likely to look at again in the future). I'm not sure how Sourceforge feel about others trying to get donations for a spin-off project via third-party means (meaning it might be best not to push our luck in this area).

CE.

the android forum about c++ is a good start ...

http://groups.google.com/group/android- ... ff3dd8de32

a lot of stuff i dont have a clue but looks interesting,

in fact just wondering if it would not be even simpler to use a c++ to java bytecode compiler ...

after all ceGUI is not really hardware related, it is just using openGL drivers, so maybe it would be easier? i have nooooooo idea

My guess is you would be better off trying to convince Google/Android to offer c++ bindings for their SDK. If apple can do it for the iphone, the only thing stopping google is their shallow views about java being more developer friendly (lawl?); at least that's the jist I got from reading a couple of threads regarding the issue.

Another thing, some C/C++ apps have already been updated to run on the Android like quake3 i think? I know they didn't port all that code to java, they probably used some c++->java bytecode compiler as you mentioned - you may be better off investigating that route.