Passing member functions to C functions?

  • Thread starter Steven T. Hatton
  • Start date
S

Steven T. Hatton

I am trying to convert some basic OpenGL code to an OO form. This is the C
version of the program:

http://www.opengl.org/resources/code/basics/redbook/double.c

You can see what my current effort at converting that code to an OO form
looks like in the code listed below. The problem I'm running into is that
the OpenGL functions that take the names of other functions as arguments
don't like the way I'm passing the member functions of my class. I started
out making all member functions non-static, but couldn't figure out a way
to pass them to functions such as glutDisplayFunc(&GlutMain::display);

Is there a way to accomplish this?

#ifndef GLUTMAIN_HH
#define GLUTMAIN_HH

#include <GL/glut.h>
namespace sth {
namespace vmathGL {
class GlutMain {
public:

GlutMain(int argc, char* argv[]);

virtual ~GlutMain();

void run();

static void init();

static void display();

static void reshape(int w, int h);

static void mouse(int button, int state, int x, int y);

static void key(unsigned char k, int x, int y);

static void spinDisplay();
private:
static GLfloat _spin;
};
}
}
#endif

#include "GlutMain.hh"

#include <cstdlib>

namespace sth {
namespace vmathGL {
GlutMain::GlutMain(int argc, char* argv[])
{
glutInit(&argc, argv);
glutInitDisplayMode(GLUT_DOUBLE | GLUT_RGB);
glutInitWindowSize(250,250);

}

GlutMain::~GlutMain()
{

}

void GlutMain::run()
{
init();
glutDisplayFunc(&GlutMain::display); // compiles but segfaults.

/*
glutReshapeFunc(&reshape);
glutMouseFunc(&mouse);
glutMainLoop();
*/
}

void GlutMain::display()
{
glClear(GL_COLOR_BUFFER_BIT);
glPushMatrix();
glRotatef(_spin, 0.0, 0.0, 1.0);
glColor3f(1.0, 1.0, 1.0);
glRectf(-25.0, -25.0, 25.0, 25.0);
glPopMatrix();
glutSwapBuffers();
}

void GlutMain::init()
{
glClearColor (0.0, 0.0, 0.0, 0.0);
glShadeModel (GL_FLAT);
}

void GlutMain::reshape(int w, int h)
{
glViewport (0, 0, (GLsizei) w, (GLsizei) h);
glMatrixMode(GL_PROJECTION);
glLoadIdentity();
glOrtho(-50.0, 50.0, -50.0, 50.0, -1.0, 1.0);
glMatrixMode(GL_MODELVIEW);
glLoadIdentity();
}

void GlutMain::mouse(int button, int state, int x, int y)
{
switch (button) {
case GLUT_LEFT_BUTTON:
if (state == GLUT_DOWN)
glutIdleFunc(spinDisplay);
break;
case GLUT_MIDDLE_BUTTON:
if (state == GLUT_DOWN)
glutIdleFunc(NULL);
break;
default:
break;
}
}

void GlutMain::key(unsigned char k, int x, int y)
{
switch (k) {
case 27: /* Escape */
exit(0);
break;
default:
return;
}
glutPostRedisplay();
}

void GlutMain::spinDisplay()
{
_spin = _spin + 2.0;
if (_spin > 360.0)
_spin = _spin - 360.0;
glutPostRedisplay();
}

GLfloat GlutMain::_spin = 0.0;

}
}

--
"If our hypothesis is about anything and not about some one or more
particular things, then our deductions constitute mathematics. Thus
mathematics may be defined as the subject in which we never know what we
are talking about, nor whether what we are saying is true." - Bertrand
Russell
 
D

David Hilsee

Steven T. Hatton said:
I am trying to convert some basic OpenGL code to an OO form. This is the C
version of the program:

http://www.opengl.org/resources/code/basics/redbook/double.c

You can see what my current effort at converting that code to an OO form
looks like in the code listed below. The problem I'm running into is that
the OpenGL functions that take the names of other functions as arguments
don't like the way I'm passing the member functions of my class. I started
out making all member functions non-static, but couldn't figure out a way
to pass them to functions such as glutDisplayFunc(&GlutMain::display);

Is there a way to accomplish this?
<snip>

Well, you can't pass non-static member functions as a C callback. To invoke
a non-static member function, you need an object, and C simply doesn't know
how to invoke members on an object. The FAQ discusses this in 33.1, 2, and
3.

I think you might want to take a peek at the GlutMaster implementation
(http://www.duke.edu/~stetten/GlutMaster/GlutMaster.html). It apparently
uses the "window id" as a key for the corresponding callback object, so you
could probably do something similar.
 
D

David Hilsee

Markus Elfring said:
[...]
Well, you can't pass non-static member functions as a C callback. To invoke
a non-static member function, you need an object, and C simply doesn't know
how to invoke members on an object. [...]

http://parashift.com/c++-faq-lite/pointers-to-members.html#faq-33.2

Can function objects/functors also be used for the requested application?
http://en.wikipedia.org/wiki/Function_object

If a library asks for a function pointer as a callback, then, no, you can't
pass it a functor. C libraries often require function pointers and have no
understanding of operator overloading, member functions, etc. The wikipedia
entry is a little misleading because it doesn't mention that the sort
function in the C example and the sort function in the C++ example are
actually two different functions. It should have a declaration of sort() in
each example.
 
M

Markus Elfring

http://parashift.com/c++-faq-lite/pointers-to-members.html#faq-33.2
If a library asks for a function pointer as a callback, then, no, you can't
pass it a functor. C libraries often require function pointers and have no
understanding of operator overloading, member functions, etc. The wikipedia
entry is a little misleading because it doesn't mention that the sort
function in the C example and the sort function in the C++ example are
actually two different functions. It should have a declaration of sort() in
each example.

Would you like to improve the article?

Are the descriptions about "functionoids" relevant here?
http://parashift.com/c++-faq-lite/pointers-to-members.html#faq-33.10
 
S

Steven T. Hatton

David said:
Well, you can't pass non-static member functions as a C callback. To
invoke a non-static member function, you need an object, and C simply
doesn't know how to invoke members on an object. The FAQ discusses this in
33.1, 2, and 3.

Oh my! I got FAQed. For shame! ;)

I actually had some problems with using static members as well, but don't
recall the details. What I ended up doing was to create a global <blush>
GlutMain* gmPtr and instantiating a corresponding object at the top of
main(). Additionally, I replaced the bodies of the corresponding C
functions with calls to gmPtr->functionName(/*args...*/). In the case where
I had to deal with function pointers, I did this:

void GlutMain::mouse(int button, int state, int x, int y, void(*spinDisp)())
{
switch (button) {
case GLUT_LEFT_BUTTON:
if (state == GLUT_DOWN)
glutIdleFunc(spinDisp);
break;
case GLUT_MIDDLE_BUTTON:
if (state == GLUT_DOWN)
glutIdleFunc(NULL);
break;
default:
break;
}
}

My CVS server had a HD crash last week, so (alack and alas) I have not been
keeping snapshots. I therefore don't have the corresponding version of
spinDisp, but it was very similar to

void spinDisp()
{
gmPtr->spinDisplay();
}

So I pass as an argument to a C++ member function, a pointer to a C
(compatable under current conditions) function that invokes a C++ member
function on the same object. The main() I ended up with looked very similar
to the original redbook example with the addition of the object creation.

How's /that/ for elegant simplicity? :D


--
"If our hypothesis is about anything and not about some one or more
particular things, then our deductions constitute mathematics. Thus
mathematics may be defined as the subject in which we never know what we
are talking about, nor whether what we are saying is true." - Bertrand
Russell
 
D

David Hilsee

Markus Elfring said:
Would you like to improve the article?

Are the descriptions about "functionoids" relevant here?
http://parashift.com/c++-faq-lite/pointers-to-members.html#faq-33.10

I added a few extra lines to the example to clarify. The functionoid is
certainly relevant. The C++ example's actually pretty worthless because
there is no way to supply something that implements the comparison function
differently. Perhaps that should be the next enhancement (polymorphism or
templates).
 
D

David Hilsee

Steven T. Hatton said:
Oh my! I got FAQed. For shame! ;)

Ah, there's no shame in being FAQed. I FAQ everyone. (Better make sure I
pronounce that correctly)
I actually had some problems with using static members as well, but don't
recall the details. What I ended up doing was to create a global <blush>
GlutMain* gmPtr and instantiating a corresponding object at the top of
main(). Additionally, I replaced the bodies of the corresponding C
functions with calls to gmPtr->functionName(/*args...*/). In the case where
I had to deal with function pointers, I did this:

void GlutMain::mouse(int button, int state, int x, int y, void(*spinDisp)())
{
switch (button) {
case GLUT_LEFT_BUTTON:
if (state == GLUT_DOWN)
glutIdleFunc(spinDisp);
break;
case GLUT_MIDDLE_BUTTON:
if (state == GLUT_DOWN)
glutIdleFunc(NULL);
break;
default:
break;
}
}

My CVS server had a HD crash last week, so (alack and alas) I have not been
keeping snapshots. I therefore don't have the corresponding version of
spinDisp, but it was very similar to

void spinDisp()
{
gmPtr->spinDisplay();
}

So I pass as an argument to a C++ member function, a pointer to a C
(compatable under current conditions) function that invokes a C++ member
function on the same object. The main() I ended up with looked very similar
to the original redbook example with the addition of the object creation.

How's /that/ for elegant simplicity? :D

That works. GlutMaster's approach replaced your global with a table, which
is a little better, but in the end, it's about the same, really. I'm
somewhat surprised that the library didn't allow you to register a void* and
a function pointer. That's a common practice that is very friendly to C++
applications.
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Members online

No members online now.

Forum statistics

Threads
473,755
Messages
2,569,536
Members
45,009
Latest member
GidgetGamb

Latest Threads

Top