Irrlicht Engine

Should't
#pragma comment(lib, "Irrlicht.lib")
and
#pragma comment (lib, "irrKlang.lib")
be enclosed in #ifdef _MSC_VER instead of #ifdef _IRR_WINDOWS_ ?

In sound.cpp of example 21:
virtual const ik_c8* getFileName() { return file->getFileName().c_str(); }
instead of
virtual const ik_c8* getFileName() { return file->getFileName(); }

about swprintf:
i had problems with this when i compiled with gcc-TDM.
I could solve them by swapping the sequence of some headers
( meaning driverChoice.h on top of irrlicht.h ):
#include "driverChoice.h" // solves problem with TDM.
#include <irrlicht.h>
With other compilers ( e.g. from the mingw-w64 project ) i had no problems.
This is something i hate about C++.
In latest trunk there is swprintf_irr, which is the best solution right now.

Thanks, I've changed it all.

again about s[n,w]printf:

- swprintf was added in C95 while snprintf was added in C99.
- the return values behave differently; with snprintf you can check against the buffer size and "reallocate" if necessary, with swprintf you get '-1' if SOMETHING went wrong, but you don't know what.

I always wondered, why there is no swnprintf or snwprintf, but there IS a snwprintf!
I have tried it with gcc 4.8.x. I am not sure about other compilers (Visual Studio > 2010 should have it), but don't let yourself be confused by this post - leave it as it is know (it works). If there should be any problems in the future, you can still change it if necessary.

I will have a look into other solutions from other libraries.

Also note that all those don't work on Android. Something I run into last year. I probably should go through the source again at some point and check for all those places if any of them matter on that platform and find other solutions there.

They don't even work on Windows, which is the reason for the char-wchar converter functions.