Re: [Development] Unsatisfied symbols for QAbstractOpenGLFunctions

2014-11-26 Thread Christophe de Dinechin
> On 26 Nov 2014, at 19:31, Thiago Macieira wrote: > > On Wednesday 26 November 2014 10:10:15 Christophe de Dinechin wrote: >> It compiles OK, but it fails to link stating that it does not find the >> vtable for QAbstractOpenGLFunctions, referenced from the compiler-generated >> copy constructor

Re: [Development] Unsatisfied symbols for QAbstractOpenGLFunctions

2014-11-26 Thread Thiago Macieira
On Wednesday 26 November 2014 10:10:15 Christophe de Dinechin wrote: > It compiles OK, but it fails to link stating that it does not find the > vtable for QAbstractOpenGLFunctions, referenced from the compiler-generated > copy constructor for QAbstractOpenGLFunctions. Maybe the problem is that > th

Re: [Development] Unsatisfied symbols for QAbstractOpenGLFunctions

2014-11-26 Thread Christophe de Dinechin
> On 26 Nov 2014, at 10:21, Konstantin Ritt wrote: > > struct OpenGLState : GraphicState { TaoOpenGLFunctions *funcs; ... } > voi-la Nope: I don't inherit the GL functions that way, so I have to rewrite every glFoo as funcs->glFoo, which is double plus annoying. And then, there are enough leve

Re: [Development] Unsatisfied symbols for QAbstractOpenGLFunctions

2014-11-26 Thread Konstantin Ritt
struct OpenGLState : GraphicState { TaoOpenGLFunctions *funcs; ... } voi-la Regards, Konstantin 2014-11-26 13:15 GMT+04:00 Christophe de Dinechin : > > Try with a class that has a copy constructor or with a usage pattern > that causes the compiler to generate a copy constructor. > > Which gives

Re: [Development] Unsatisfied symbols for QAbstractOpenGLFunctions

2014-11-26 Thread Christophe de Dinechin
> Try with a class that has a copy constructor or with a usage pattern that > causes the compiler to generate a copy constructor. Which gives me a possible fix: in my TaoOpenGLFunctions, I can add a copy constructor that invokes the default constructor instead of the copy constructor. So I can

Re: [Development] Unsatisfied symbols for QAbstractOpenGLFunctions

2014-11-26 Thread Christophe de Dinechin
On 26 Nov 2014, at 09:46, Sean Harmer wrote: > > On Wednesday 26 Nov 2014 09:13:38 Christophe de Dinechin wrote: >>> On 26 Nov 2014, at 09:04, Konstantin Ritt wrote: >>> >>> class QAbstractOpenGLFunctions >>> (note missing Q_GUI_EXPORT) >> >> Do you mean this is a bug, or that I'm using it wro

Re: [Development] Unsatisfied symbols for QAbstractOpenGLFunctions

2014-11-26 Thread Sean Harmer
On Wednesday 26 Nov 2014 09:13:38 Christophe de Dinechin wrote: > > On 26 Nov 2014, at 09:04, Konstantin Ritt wrote: > > > > class QAbstractOpenGLFunctions > > (note missing Q_GUI_EXPORT) > > Do you mean this is a bug, or that I'm using it wrong and it's not supposed > to be visible in the libra

Re: [Development] Unsatisfied symbols for QAbstractOpenGLFunctions

2014-11-26 Thread Christophe de Dinechin
> On 26 Nov 2014, at 09:04, Konstantin Ritt wrote: > > class QAbstractOpenGLFunctions > (note missing Q_GUI_EXPORT) Do you mean this is a bug, or that I'm using it wrong and it's not supposed to be visible in the library? If not, why is its destructor virtual (implying that the symbol for dto

Re: [Development] Unsatisfied symbols for QAbstractOpenGLFunctions

2014-11-26 Thread Konstantin Ritt
class QAbstractOpenGLFunctions (note missing Q_GUI_EXPORT) Regards, Konstantin 2014-11-26 11:42 GMT+04:00 Christophe de Dinechin : > I'm trying to adapt Tao3D (http://tao3d.sourceforge.net) to the "new" way > of doing OpenGL on Qt. I run into unsatisfied symbols for > QAbstractOpenGLFunctions (s

[Development] Unsatisfied symbols for QAbstractOpenGLFunctions

2014-11-25 Thread Christophe de Dinechin
I'm trying to adapt Tao3D (http://tao3d.sourceforge.net) to the "new" way of doing OpenGL on Qt. I run into unsatisfied symbols for QAbstractOpenGLFunctions (specifically, the vtable and the destructor), despite the fact that QOpenGLFunctions_3_0 (which derives from QAbstractOpenGLFunctions) is