Developing with OpenGL ES 2.0 - HD2 General

Hi guys!
Since the HD2 doesn't seem to support OpenGL ES 1.0 (at least all the apps run slow as hell) I wanted to aks you how to get OpenGL ES 2.0 going?
Currently I'm working with C# and use this nice wrapper: http://www.koushikdutta.com/2008/08/net-compact-framework-wrapper-for.html.
As mentioned above apps created with this wrapper don't work well.
So: where can I get some info on 2.0? Are there any wrappers for C#? Or do I have to use C++ (which wouldn't be much of a problem )?
Wehere can I find Librarys, tutorials, samples, code etc?
Sydly Google doesn't bring up much (seems the world is switching to iPhone/Android)....
ANY help would be much appreciated!

How complecated is the wrapper? Could it be easily converted to use the new 2.0 calls instead of the old ones?

Good question. Until now I just used it and never took a look "behind the scenes". Maybe it's time to change that
Is there something like an OpenGL ES 2.0 SDK? Any kind of documentation? What are the differences between 1.0 and 2.0?

Related

Programming CE, Book and resource recommendations

Hello,
After a decade of not doing C, I find that alot of things have changed when programming c++ native for CE. Even things like:
string somestring;
no longer function. Wow!.
strcpy has been replaced with StringCchCopy. Wow, MS has done a number here... and so on...
Anyhow I searched xda and found allot of good threads. But what if I'm looking at re-learning C++, with a focus on CE programming only? What books and other resources would you recommend, its almost like starting from scratch. I got vs2005 + emu + been reading allot of MSDN, did some basic apps. etc.
BTW: How the .... do you declare a variable length string???
and in VS2005 is running the debugger the only way to get the app to start automatically ???
Thanks
*bump*, anyone?
Couple of notes:
1) I strongly recommend avoiding VS, it is unbearably slow and balky.
2) strcpy and all the other functions are still supported by c runtime on CE (VS 2005 complains about them but you can ignore it). The problem is all windows API on CE use UNICODE so you have to declare strings as WCHAR (thats MS speak for unsigned short) and use wcscpy, and such.
3) The only book I know on CE specifically is this and I only read the two sample chapters online, so can't really recommend it. (Though the sample chapters were very useful). What you need is to get a good book on Win32 programming that will get you well on your way.
4) There is an MFC class CString which wraps up all string related functions very nicely. If you are not using MFC but clean C/C++ the only thing I can think of is malloc and free or the C++ versions new and delete.
Check out the links in my sig if you want to see code examples of small but functioning apps.
I disagree with levenum's comment on VS, but agree at the same time (weird I know). I wouldn't use VS for native development because it has A LOT of extra features that I think would be useless for native developers. Also, the lighter eVC++ has pretty much everything you need to write native apps.
I had this book called "Programming Windows CE 3.0" or something along those lines a few years ago. I got it for like $3 on Amazon. If you can find an updated version of that (or even the old version would be useful) then you should buy it. It will show you all you need to know to get started. You can learn the new api's from browsing msdn.
If you're only a little rusty with C/C++ I would suggest jumping right in to the Sdk samples and seeing how they do things.
Thanks, the reason why I wanted to use VS is so that maybe later I wouldn't mind going into MFC. So having one platform would be nice.
So far I have done some basic things, like drawing, etc, but I find I'm having a harder time making strings and including the api's to work. Almost all of the msdn examples fail to compile properly.
I have an old book MFC for Windows 95 by MS, is it even worth looking at ?
robp said:
I disagree with levenum's comment on VS, but agree at the same time (weird I know). I wouldn't use VS for native development because it has A LOT of extra features that I think would be useless for native developers. Also, the lighter eVC++ has pretty much everything you need to write native apps.
I had this book called "Programming Windows CE 3.0" or something along those lines a few years ago. I got it for like $3 on Amazon. If you can find an updated version of that (or even the old version would be useful) then you should buy it. It will show you all you need to know to get started. You can learn the new api's from browsing msdn.
If you're only a little rusty with C/C++ I would suggest jumping right in to the Sdk samples and seeing how they do things.
Click to expand...
Click to collapse
could you provide some links for msdn. also i know how to program in c but how do i compile a program for windows ce could you recommend a free one. thanks
The only way to code for Windows Mobile for free is to use embedded Visual C++. You'll have to google it to find a link.
Once you have that downloaded, check out:
http://msdn2.microsoft.com/en-us/library/bb158662.aspx
The new solution files most probably won't open in eVC++ but you can create new projects and add the source files yourself.
You'll want to look at the samples in the Win32 folder.
robp said:
The only way to code for Windows Mobile for free is to use embedded Visual C++.
Click to expand...
Click to collapse
That's not entirely true - there's still ceGCC (it works on linux or windows using cygwin) - however this one lacks resource (window, forms, etc) editor (but if you have resources already created it will compile/link properly). But it works great with apps where you want to handle all the graphics and user interaction by yourself (like games, or programs with 100% custom UI), for example using SDL, pocketHAL, etc.
Also, wxWidgets seems to have support for windowsCE, and might be made to work with ceGCC, but i haven't tested it yet.
Sorry, I actually meant to put "(I think)" after that sentence lol.
Actually I think you should be able to use any C++ compiler as long as you have the necessary libraries, and I think all the libs you need are in the sdk download. But the easiest route for beginners in native coding is probably eVC++
Well, now i can agree with you
That is, unless you want to create games - in that case ceGCC+XFlib might be a good way to go for someone with general programming knowledge, but not familiar with winCE-specific coding. Some games coded with it (and their sources) can be found on XFlib homepage: www.xflib.net.
thanks for the info
Ok, as I wouldn't mind at a later time going into MFC, and building dll's for mobile. How much "harder" would it be for me to learn VS2005 native for ce instead of using evc++ ?
Just trying to weight the cons and the pros. BTW where can I get eVC ?
Thank you.
Two more free options: PellesC and Lazarus + WinCE add-on
I don't think it would much harder. Once you get the hang of it you'll find that it makes life much easier. If you buy a boxed retail version it comes with a nice book that will take you from n00b to master in a few hours (not really but it's still a great reference).
I think the hardest part of upgrading to newer versions of VS is learning the new project file setup. Atleast that's what it looks like from watching other students at my school go from VS6.0 to VS2k5 and 2k8.
Also, I should note that I prefer VS2008 hands-down, but I don't disagree with other people who say eVC++ is faster. While I haven't done any real native development since eVC++ 4 was new, I still dabble a bit (especially since getting active on this forum) and VS2005 works great to me.
So, my advice would be to use the newest tools available to you, unless money is an issue. But don't be worried about upgrading. Your life will only get easier once you do.
robp said:
Also, I should note that I prefer VS2008 hands-down, but I don't disagree with other people who say eVC++ is faster. While I haven't done any real native development since eVC++ 4 was new, I still dabble a bit (especially since getting active on this forum) and VS2005 works great to me.
So, my advice would be to use the newest tools available to you, unless money is an issue. But don't be worried about upgrading. Your life will only get easier once you do.
Click to expand...
Click to collapse
Well I just got eVC and will give it a try too. But I am pretty much used to vs2005. I do not think I can't afford the vs2008 right now. How much difference is there between 2005 and 2008 when it comes to programming for WM?
I'm looking to create some commercial applications for WM. For the last 10 years I'v been doing .com apps, now I figured I would do some for the mobile platform. We have all the graphics/design/marketing guys for .com, so it should work out just fine for WM.
One thing we noticed when working with exec type people is their frustration, I see exactly how we can help them resolve their frustration with WM, Symb, and java type phones...
Thanks
marek101 said:
Well I just got eVC and will give it a try too. But I am pretty much used to vs2005. I do not think I can't afford the vs2008 right now. How much difference is there between 2005 and 2008 when it comes to programming for WM?
I'm looking to create some commercial applications for WM. For the last 10 years I'v been doing .com apps, now I figured I would do some for the mobile platform. We have all the graphics/design/marketing guys for .com, so it should work out just fine for WM.
One thing we noticed when working with exec type people is their frustration, I see exactly how we can help them resolve their frustration with WM, Symb, and java type phones...
Thanks
Click to expand...
Click to collapse
Honestly I wouldn't rush to upgrade. VS2008 is better, but doesn't add much to the WM development arena.
And if you already have VS2005 and are used to it, I would stick with that and forget about eVC++. You should give the compact framework a try. I promise you'll love it
compact framework, that's a part of the .NET isn't it?.
For the time being, for learning purposes, I think I will stick with the native for now. Once I get a grip, I will explore .net. The thing is I really don't like bulky apps for WM. Most of the apps I came across that required the .net framework were slow, any apps that use the .net with a today plug in seem to really slow down my phone.
BTW is HTC Home plug in written with the .net?
Thanks
marek, it seems to me that you are under the impression that you need VS 2005 to use MFC.
eVC fully supports MFC as well as creating DLLs and MFC DLLs.
It admit the reason I don't like VS is because I only write in native and don't like .NET languages a lot.
Since you want to create commercial apps you should note that any app created with WM 5 SKD will not be backwards compatible with WM 2003 which reduces your clients range.
You should compile with WM 2003 SDK which is forward compatible.
Our primary target audience would be WM6. We will make every effort to make WM5 compatible apps. Yes it does reduces significantly the client base, but also there are many positives...
We are still looking for more programmers for the simple concepts we have developed. I believe this will work out just fine. Having WM5 support is a "would be nice" but not required. WM5 has a bag of problems, 2003 I don't even want to touch it.
Thanks
I have to admit that I love VS because I love using .net
But VS is also a much more powerful editor in general than eVC is. It's still nice but the experience isn't as smooth as it is in newer versions, which is expected since the newer versions build on the previous ones.

Porting a game from PC to ppc

Hi, I never did that, but I was playing to Anthelion 2 on my PPC and I thank that it could be great to port Homeworld 1 or Cataclysm to the PPC, and I would like to know how to do that, I think that I could "recompiler" it, but I don't know how, and I would like to know if a tutorial has been created somewhere on the net?...
I'm not sure if it would be even technically possible. Well, the newest pdas _might_ be powerful enough to run something like HW1 but i'm not sure if it's such a good idea. Did you try to run homeworld in 640x480 resolution? Most of the time you'll see ships as groups of two to eight pixels. Now imagine it all squished on a phone three-inch screen: try ordering your corvettes to smash that annoying bomber on a screen that small I think that a bit better idea would be porting really old games, that were designed to run in VGA or even lower resolutions (SubCulture, Command&Conquer, I-War, Dark Forces are some of the titles i'd pay for ).
Anyway, back to porting subject.
First of all you would need the source code of the program you want to port - in case of homewrld1 it's not a problem.
Secondly, you would have to make sure that all libraries (graphic, sound, input, etc) used by a game have windows mobile/windowsCE versions. Again, homeworld1 seems lucky since it has been ported to SDL - a multiplatform opensource graphic/sound/input library.
But that's where good news end. Porting a game is not just a matter of grabbing the PC version source and recompiling it. If it was as easy, we would have hundreds of PC games already ported You need considerable programming skills to actually create a port because usualy not all libraries used by a game a compatible with WindowsCE. An example - the opensource version of homeworld uses OpenGL for graphic rendering. The pocket version would have to use OpenGL's "little brother" - OpenGLES. As far as i know, they're not 100% identical, so to put it simply, you would have to make the game talk in OGLES language, instead of standard OGL. And doing changes in graphic rendering routines usually breaks something else, so you'd have to go and fix it.
I'm not trying to discourage you here but i'd suggest learning to program for WindowsCE (or at least for PC) _before_ attempting to port anything - doing it the other way around will be just a waste of time and a source of frustration.
There are some development resources that can help start the adventure with programming here on xda. You could also search for some general C/C++ tutorials targeting PC's. If you consider getting into programming, i suggest checking out SDL - Many games use it, and thanks to this library you can skip the OS-specific part of coding and get right to the fun stuff - a program that actually does/displays anything For an even easier start, you might want to check out QuickCG - a SDL wrapper simplifying the coding even further.
Oki, thanks for your answer, I've a friend who is learning to program in C++, so, I'll ask him if he can help me to do that, it would be great to have this game on a PPC (perhaps the Diamond, because it has D3D and OpenGL Drivers, or of the iPhone, but I guess that the programming language is not the same as the PPC...
[EDIT] STARCRAFT would be great to, and easier to port on PPC, because of his age and that he uses 2d Graphisms...
You should look into the stratagus engine.
antrak said:
You should look into the stratagus engine.
Click to expand...
Click to collapse
Nice Engine, but it's not Starcraft, but I can't find the Source Code on the net, they could give the source code with the game when you buy it
Psycho said:
Hey, it's Calvin, I found him
Click to expand...
Click to collapse
Stargus a starcraft's clone, I'm trying to download it, but I don't know if it works for PPC...
You might want to check this thread:
http://forum.xda-developers.com/showthread.php?t=497086&highlight=starcraft

Fastest/Lightest way to code WinMo?

Hi,
I'm the author of a fairly popular j2me chat application, and I've been pondering making a native WinMo version for a while now. I've thought about a .net port, as it would probably be easiest, but all the .net programs I've used seem dead slow. Or am I totally off base with that?
Is there another environment I should look at as an alternative? Keep in mind, I don't have visual studio.
jonnycat26 said:
Hi,
I'm the author of a fairly popular j2me chat application, and I've been pondering making a native WinMo version for a while now. I've thought about a .net port, as it would probably be easiest, but all the .net programs I've used seem dead slow. Or am I totally off base with that?
Is there another environment I should look at as an alternative? Keep in mind, I don't have visual studio.
Click to expand...
Click to collapse
You can still stick with j2me, you'll just have to change a few things to make the program compatible with the java midlet managers for WinMo. You should read up on the Mildet Bible: http://forum.xda-developers.com/showthread.php?t=339579
Hope this helps
From testing open source apps and they ones I have made I found that C++ is faster and c# is easier but a little slower. I used c++ for along time but have moved on to C# and love it.
u can just develop in any .net language, then use tool like SmartAssembly to optimize and convert the code to binary.
jonnycat26 said:
Hi,
I'm the author of a fairly popular j2me chat application, and I've been pondering making a native WinMo version for a while now. I've thought about a .net port, as it would probably be easiest, but all the .net programs I've used seem dead slow. Or am I totally off base with that?
Is there another environment I should look at as an alternative? Keep in mind, I don't have visual studio.
Click to expand...
Click to collapse
Hello, I strongly recomend to use c++ for winmo dev.
cheers,
Hlov
The fastest and leanest method is C++, available in Visual Studio Professional, or as a standalone free download as Embedded C++ version 4. The development models are MFC, (Microsoft Foundation Classes), ATL (Active Template Library), or good old fashioned WIN32.
Of these WIN32 is the leanest, all the others, including .NET are wrappers around it. If you haven't used or seen WIN32 before, expect a pretty steep learning curve. Look at the Hello World example it generates, to try and figure out what the hell is going on. Basically you respond to WM_XXXXXX messages fired at your application by the operating system. Limit the main menu to two items, and WinMo 5/6 will display it correctly, with the main menu options either side of the keyboard/SIP icon, and accessible by the left and right menu buttons. Otherwise it will switch to a WM 2002/3 menubar, which you then have to click on to use.
If you want to go any lower down than this, you will need to brush up your ARM assembly language.
Unless you are developing graphics intensive applications, using c# is fine for applications. C++ is necessary if you are writing games and the like but the SLIGHT performance decrease is not that big of a deal otherwise.
Soul_Est said:
You can still stick with j2me, you'll just have to change a few things to make the program compatible with the java midlet managers for WinMo. You should read up on the Mildet Bible: http://forum.xda-developers.com/showthread.php?t=339579
Hope this helps
Click to expand...
Click to collapse
I'm using J2ME polish, so I can (and have) built a build specifically for WinMo, but I want a native version so I can multitask (can't do that with Esmertec) and I'd also like to develop a homescreen plugin.
My app is here:
http://www.buildhigh.com/jonnychat/
If anyone has any suggestions on what I should do with a port, I'd love to hear them.
stephj said:
The fastest and leanest method is C++, available in Visual Studio Professional, or as a standalone free download as Embedded C++ version 4. The development models are MFC, (Microsoft Foundation Classes), ATL (Active Template Library), or good old fashioned WIN32.
Click to expand...
Click to collapse
It's been a long time since I've done standard Win32 development, and I really don't plan on going back down that road again. For starters, I imagine Visual Studio would be a dog in a VM (I only run windows in a VM these days, won't run it on actual hardware if I can avoid it).
It's looking like I"m going to try C++ or C# for this. I'd like to do C# because that's pretty close to Java, but I also have moral qualms about using a language designed to usurp a good and popular language just because MSFT didn't develop it.
Decisions, Decisions....
You can also try using Ruby with the Rhodes framework: http://www.rhomobile.com/home
Or you can try using Mysaifu which may allow multitasking (haven't tried any of my programs on it yet): http://www2s.biglobe.ne.jp/~dat/java/project/jvm/index_en.html

Jython in Android?

Hey Folks,
I was wondering if anyone knew if it was possible to use Jython instead of plane Java for android development? I don't know much about Java development but I'm with Python. I was figuring that this might ease me into it better. Any thoughts?
Thanks,
-=GB=-
Have you had a look at the Android scripting environment?
http://code.google.com/p/android-scripting/
http://google-opensource.blogspot.com/2009/06/introducing-android-scripting.html
http://code.google.com/p/android-scripting/wiki/PythonAndroidAPI
Regards,
Akshay
Thanks aksd,
I don't know how I didn't find that before. Perhaps I should have just searched for Python and Android as opposed to just Jython and androind.
Anyway, this is very cool for prototyping and simple scripting. However, the main thing I was hoping for was writing apps in Python (via Jython) that could then be distributed on the Marketplace. It's kind of to bad. It doesn't look like anyone is really trying to port Jython to Android because ASE is sort of Google's official response to the whole thing. On the whole, a really great thing for people who just want to script Android but not really for App development.
In all of this searching I did find some interesting information about how to link in C libs so you could code some of the hardcore speed sucking elements in C (like a game engine) and then make calls into it via Java. I didn't know that was even possible so... Pretty cool stuff.
Hey ngrava,
I'm not much into scripts, so dont really follow whats happening with scripting and Android.
If you wanted a bit of a performance gain you can use the NDK and call the JNI to execute native libraries or code. been there for awhile but few use them. Java is just so much easier than C .
Theres a lot of interesting stuff going on on the Android platform, Mono being ported to Android, Scala being ported to Android, I'm actually currently working on profiling the performance of scala code on android.
Regards,
Akshay
On the Google DevFest on Argentina, Google´s ingeneers afirmed that they are working to make more languages availeable on Gingerbread, they specifically mentioned Ruby. We just have to wait to see if this is true and if we are going to be able to ship apps of other languages to the market, and I´m not sure if they will be backwards compatible with earlier versions of android as well.
PD: Oops, sorry for reviving this old post, I just realised i clicked last page instead of next

NDK porting question - AlmostTI

Hello,
I am interested in trying to port this project:
http://fms.komkon.org/ATI85/
to Android. This is a TI calculator emulator.
I would like some advice as to where to partition the code between C and Java.
My thinking right now is that I should try to replace the X calls in the Linux version of the emulator (which actually is core dumping on me, but that is another story) with OpenGL calls. Then use OpenGL for all graphical display in Android and only use Java for touchscreen input.
Is this the best way to go about this? Or do I really need to move more to Java?
Thanks for any advise. If someone with more experience is interested in this, that would be great - I don't have much experience in this space.
xnappo

Categories

Resources