CCGLProgram::compileShader() calls abort() when the shader program is invalid.
But it should return false.

What I want to do is this:

・Shader A: Better performance. Compilable on some device, not on the other device.
・Shader B: Worse performance. Compilable on all devices.

In this case, first I want to try using shader A, and on failure, next shader B.

But current behavior make the program aborted when failed in shader A, so I can't try shader B.
To do this, I have to try glCompileShader(shader A) manually and re-compile with CCGLProgram::initWithVertexShaderByteArray(), it is inefficient.
If CCGLProgram::compileShader() returns false on failure, this'll be more simple.
I have only to call CCGLProgram::initWithVertexShaderByteArray() twice.