Can anyone see what's wrong with this?

[hogsy]

[szo]

Your use of sizeof(parttex) is wrong: you need something like sizeof(parttex)/sizeof(void*) to match your 32.

[hogsy]

That might be so but it's not the cause of the problem.

[Baker]

This is probably a 5 second problem in the debugger (set a breakpoint or browser the variables).

I am guessing that you are using a FitzQuake code base and Windows.

1) What FitzQuake did you use as a basis for the project? Was it the official FitzQuake 0.85?
2) What version of Visual Studio are you using (if that is what you are using)?

[hogsy]

1 - Yes.
2 - Studio 2010.

[Baker]

[mh]

Creating and uploading a new texture is illegal between a glBegin/glEnd pair, as is changing the current texture.

http://www.opengl.org/sdk/docs/man/xhtm ... mage2D.xml
"GL_INVALID_OPERATION is generated if glTexImage2D is executed between the execution of glBegin and the corresponding execution of glEnd."

http://www.opengl.org/sdk/docs/man/xhtm ... exture.xml
"GL_INVALID_OPERATION is generated if glBindTexture is executed between the execution of glBegin and the corresponding execution of glEnd."

The right solution is not to put your glBegin/glEnd inside the loop; that will cause your performance to plummet like a stone. What you want instead is to check if the texture has changed from the last one used, if so you call glEnd, then switch texture, then call glBegin again. (you can optimize out the case where no particles have yet been drawn later on).

It's also not a good idea to load textures at runtime like this, you'll want to load them upfront during initialization, otherwise it will also hurt performance.

[hogsy]

@Baker
None of that would fix it. I've already stated I'm aware it's messy

@mh
I'll sort that out and yes I'm aware it's a bad idea but we're doing it anyway

Edit:
Yup that fixed it, thanks mh!

[Baker]

[hogsy]

"Catch me later and I'll buy you a beer"

Thanks for your suggestions anyway