Sun, 22 Nov 2009 07:00:26 +0000 |
Sam Lantinga |
More fixes to compile under Visual C++
|
Sun, 22 Nov 2009 06:42:58 +0000 |
Sam Lantinga |
Fixed include paths for Visual C++
|
Sun, 22 Nov 2009 06:37:43 +0000 |
Sam Lantinga |
The Direct3D Read/Write pixels interface is in progress.
|
Sun, 22 Nov 2009 06:34:45 +0000 |
Sam Lantinga |
Whoops, actually set the SDL error, don't just print the error.
|
Sat, 21 Nov 2009 08:42:42 +0000 |
Mike Gorchak |
Added support for QNX default font. Backspace and Return keys now handled.
|
Sat, 21 Nov 2009 07:59:19 +0000 |
Sam Lantinga |
We want to be strict on software renderer tests and opaque tests, but give a decent margin for blending inaccuracy for the blended tests.
|
Sat, 21 Nov 2009 07:46:12 +0000 |
Sam Lantinga |
Increased tolerance a little bit more for multiple blending passes accumulating error.
|
Sat, 21 Nov 2009 07:26:52 +0000 |
Sam Lantinga |
Added comment for pixel-perfect line workaround.
|
Sat, 21 Nov 2009 07:22:59 +0000 |
Sam Lantinga |
Of COURSE that trick wouldn't work on all renderers. Fall back to something for now, hopefully figure out a better way to do this later.
|
Sat, 21 Nov 2009 07:14:21 +0000 |
Sam Lantinga |
This is terrible, but the OpenGL standard says that lines are half open, which means that one endpoint is not covered so adjoining lines don't overlap. It also doesn't define which end is open, and indeed Mac OS X and Linux differ. Mac OS X seems to leave the second endpoint open, but Linux uses the right-most endpoint for x major lines and the bottom-most endpoint for y major lines.
|
Sat, 21 Nov 2009 06:34:43 +0000 |
Sam Lantinga |
Fixed the coordinates for pixel coverage in blits
|
Sat, 21 Nov 2009 06:28:25 +0000 |
Sam Lantinga |
Use 45 degree lines for the diagonal test to avoid aliasing errors in line drawing.
|
Sat, 21 Nov 2009 06:19:34 +0000 |
Sam Lantinga |
It's not the last pixel, it's the rightmost pixel, or if they're both the same x coordinate, the bottommost pixel.
|
Sat, 21 Nov 2009 06:17:50 +0000 |
Mike Gorchak |
RenderReadPixels and RenderWritePixels functions work with back buffer now and all asynchronous operations are flushed before reading or writing to backbuffer. Thanks Sam for clarification of this.
|
Sat, 21 Nov 2009 05:29:31 +0000 |
Sam Lantinga |
My first OpenGL shader! Momma will be so proud!
|
Sat, 21 Nov 2009 05:25:08 +0000 |
Sam Lantinga |
pixels don't need to be dynamically allocated
|
Sat, 21 Nov 2009 05:05:19 +0000 |
Sam Lantinga |
SDL_RenderReadPixels() needs to flush asynchronous operations before it reads.
|
Fri, 20 Nov 2009 14:42:40 +0000 |
Mike Gorchak |
Support for UTF-8 text input has been added.
|
Fri, 20 Nov 2009 07:11:29 +0000 |
Mike Gorchak |
1. SDL_RenderPresent() call has been added after each test to be sure, that all graphics output is flushed in case if it was asynchronous.
|
Fri, 20 Nov 2009 07:08:28 +0000 |
Mike Gorchak |
Output last SDL error in case of test was failed.
|
Fri, 20 Nov 2009 07:07:41 +0000 |
Mike Gorchak |
RenderReadPixels and RenderWritePixels now work properly.
|
Fri, 20 Nov 2009 07:06:50 +0000 |
Mike Gorchak |
Deinitialization fixes, in case if QNXGF driver is not initialized properly.
|
Thu, 19 Nov 2009 09:07:09 +0000 |
Mike Gorchak |
Added support for generic getopt() function instead of getopt_long(). Because not all platforms have getopt_long().
|
Thu, 19 Nov 2009 08:44:07 +0000 |
Mike Gorchak |
Support for RendererReadPixels and RendererWritePixels has been added to photon renderer.
|
Thu, 19 Nov 2009 08:02:00 +0000 |
Sam Lantinga |
Mike Gorchak to Sam
|
Thu, 19 Nov 2009 05:33:41 +0000 |
Sam Lantinga |
Found a way to implement mask semantics in OpenGL
|
Thu, 19 Nov 2009 05:06:01 +0000 |
Sam Lantinga |
Allow some variation in the pixel values to account for blending accuracy differences.
|
Thu, 19 Nov 2009 04:59:19 +0000 |
Sam Lantinga |
Include the endpoint in the line we're drawing
|