crash on swapbuffers

Started by
6 comments, last by benjamin bunny 19 years, 7 months ago
hey there everyone! my program seems to be crashing on swapbuffers whenever my 'layer' (array of pointers to tiles) is drawn above/on/near(hard to tell because my system totally locks up) ONLY the top of the client rect of my window. all other sides work fine. layer.y is 10.2 in this situation, i will check if its the same with further away z values but i dont want to know cause it might crash and ill loose this post lol. it also crashes on glFlush() if i add that in there before my swapbuffers call... heres the code for my drawing, which is done right before the swap buffers call... (btw layer is vector<c_layer> with just one element as far as i have tested)

glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT);			// Clear color buffer (what you see) and depth buffer

		//glPushMatrix();
		glLoadIdentity();
		glTranslatef(0.0f, 0.0f, -20.0f);

		for (uint i = 0; i < layer.size(); i++)
		{
			if(gl_mode == GL_SELECT)
				glLoadName(i + 1);

			glPushMatrix();
			glTranslatef(layer.pos.x, layer.pos.y, layer.pos.z);

			glRotatef(layer.rot.x, 1.0f, 0.0f, 0.0f);
			glRotatef(layer.rot.y, 0.0f, 1.0f, 0.0f);
			glRotatef(layer.rot.z, 0.0f, 0.0f, 1.0f);

			// tile widths
			int width = 1;		// half 
			int height = 1;	// half

			if (cur_layer == &layer)
			{
				glColor3f(1.0f, 1.0f, 1.0f);
				glBegin(GL_QUADS);
					glVertex3f(-0.5f - layer.dimensions.x / 2, -0.5f - layer.dimensions.y / 2, 0.0f);
					glVertex3f(0.5f - layer.dimensions.x / 2, -0.5f - layer.dimensions.y / 2, 0.0f);
					glVertex3f(0.5f - layer.dimensions.x / 2, 0.5f - layer.dimensions.y / 2, 0.0f);
					glVertex3f(-0.5f - layer.dimensions.x / 2, 0.5f - layer.dimensions.y / 2, 0.0f);
				glEnd();
			}

			for (uint n1 = 0; n1 < layer.dimensions.y; n1++)
			{
				for (uint n2 = 0; n2 < layer.dimensions.x; n2++)
				{
					glBegin(GL_QUADS);

						color_f *color = layer.p_gettile(n2, n1);
						glColor4f(color->r, color->g, color->b, 1.0f);

						glVertex3f(0.0f - (float) layer.dimensions.x / 2 + n2, 
							height - (float) layer.dimensions.y / 2 + n1, 0.0f);
						
						glVertex3f(width - (float) layer.dimensions.x / 2 + n2, 
							height - (float) layer.dimensions.y / 2 + n1, 0.0f);
						
						glVertex3f(width - (float) layer.dimensions.x / 2 + n2, 
							-1.0f * (float) layer.dimensions.y / 2 + n1, 0.0f);
						
						glVertex3f(0.0f - (float) layer.dimensions.x / 2 + n2, 
							-1.0f * (float) layer.dimensions.y / 2 + n1, 0.0f);

					glEnd();
				}
			}

			glPopMatrix();
		}

		return true;


it could be my graphics drivers because my geforce2 (lol sad i know...) fried and im using a matrox millenium g400 (lol even saddeerrrr) which of course has old drivers. but, if perhaps you see something that could make swapbuffers flip out or i dont know... lol im still relatively new to opengl *hugs red book* thanks for any input!
-plasmicsoup
Advertisement
Sounds like maybe you're writing some memory to the layer vector that hasn't been allocated. For example, maybe you're setting layer[0] when layer.size()? Do some more debugging, perhaps put an

assert(i<layer.size());

before setting the layer in any code. If the assertion fails then you've found the problem.

Oh and make sure you're using debug mode.

____________________________________________________________www.elf-stone.com | Automated GL Extension Loading: GLee 5.00 for Win32 and Linux

Ive stepped through the drawing code layer.size() is returning the correct value and i stays under 1 when it crashes... though i still wonder if its something to do with some crazy funk in vector's mem management

but then again i dunno something tells me that something else would've failed first!

oh, and it tested it with different z values... its still happens... so basically, layer.y is > around 7 or so (ill find out exact value)... swapbuffers fails. its only on swapbuffers call that is crashes.

why oh why lol

im beginning to think its graphics drivers being retarded... but how could such fundementally simple drawing crash it?

*shrug*

maybe i didnt pay enough homage to the opengl gods this week?

thanks tho :P
-plasmicsoup
It could be drivers, but I still think it's more likely you've overwritten some area of memory you shouldn't have. Such bugs can manifest themselves in odd ways. More debugging is probably the answer. Comment out parts of the code until you find the cause. Experiment. It'll be something really obvious, I bet :).

BTW, it's unlikely to be a bug in vector, but perhaps you could make a wrapper for it that does bounds checking. It could help locate the problem. I know my bounds checked vector equivilant class has saved me some headaches in the past. If you use assertions then it doesn't affect the performance in release mode.

____________________________________________________________www.elf-stone.com | Automated GL Extension Loading: GLee 5.00 for Win32 and Linux

youre quite astute! it definetly makes more sense that ive gone and recklessly chucked data about my memory, which being relatively new to the whole thing i find thats easy to do. oh man what a headache lol :P

but perhaps the only headaches that i can enjoy!

thanks! im scanning my code now (this post is basically just delay til the next time i feel like crashing and rebooting haha)
-plasmicsoup
aaaaaaactually

(and i do hope this is a pointful post)

ive been getting error messages on reboot that say my graphics hardware failed, was stuck in an infinite loop actually. listed one of my dll's for the g400. though i suppose once you start f'in hardcore with memory that could happen still but, well, if that changes your outlook on the problem, let me know.

also, it crashes on open if i set the gltranslatef's y coord to a constant like say 4 or something (ive determined 2, roughly, or above makes it crash).

-plasmicsoup
ok so i got around to testing it on my bro's computer finally

and....



drum roll...

it works.
fancy that, eh?

well im guessing its the vid card then.

could it still be the memory? i know his computer has much more ram then mine has, i guess then with his there is less chance of randomly overwriting something important, and more of achance of writing over clear memory. but its quite consistent with the way its crashed.

oh well lol i feel like im beating a dead horse here so thanks a lot for your input! *showers you with love* haha
-plasmicsoup
It could still be memory, but it does sound like a problem with your system. Could be a cooling issue I suppose. Do you have the problems with other 3D apps, in particular OpenGL ones? Have you tried updating the drivers?

____________________________________________________________www.elf-stone.com | Automated GL Extension Loading: GLee 5.00 for Win32 and Linux

This topic is closed to new replies.

Advertisement