Mercurial > sdl-ios-xcode
view docs/man3/SDL_PixelFormat.3 @ 1544:ab1e4c41ab71
Fixed bug #33
Mike Frysinger wrote:
> with libsdl-1.2.9, some games (like bomberclone) started
> segfaulting in Gentoo
[...snip...]
> the last change in the last hunk:
[...snip...]
> if i change the statement to read:
> (table[which].blit_features & GetBlitFeatures()) == GetBlitFeatures()
> bomberclone no longer segfaults on my box
Alex Volkov wrote:
> The test "(table[which].blit_features & GetBlitFeatures()) ==
> table[which].blit_features)" is correct, and the previous
> "(table[which].cpu_mmx == SDL_HasMMX())" was actually broken.
I think there is potentially a slightly different cause of the above problem.
During the introduction of the Altivec code, the blit_table struct field
'alpha' got changed from a straightforward enum to a bitmask, which makes
perfect sense by itself. However, now the table driven blitter selection code
in SDL_CalculateBlitN() can choose the wrong blitters when searching for a
NO_ALPHA blitter because of the following code:
int a_need = 0;
...
(a_need & table[which].alpha) == a_need &&
When searching through the normal_blit_2[] table, a SET_ALPHA blitter (like
Blit_RGB565_ARGB8888) can now be selected instead of a NO_ALPHA one, causing
alpha channel bits to appear in a non-alpha destination surface. I suppose this
could theoretically be an indirect cause of the segfault mentioned above.
I *think* this can be fixed by changing to
int a_need = NO_ALPHA;
author | Sam Lantinga <slouken@libsdl.org> |
---|---|
date | Wed, 15 Mar 2006 15:47:49 +0000 (2006-03-15) |
parents | e867f327aa54 |
children | 546f7c1eb755 |
line wrap: on
line source
.TH "SDL_PixelFormat" "3" "Tue 11 Sep 2001, 23:01" "SDL" "SDL API Reference" .SH "NAME" SDL_PixelFormat\- Stores surface format information .SH "STRUCTURE DEFINITION" .PP .nf \f(CWtypedef struct SDL_PixelFormat { SDL_Palette *palette; Uint8 BitsPerPixel; Uint8 BytesPerPixel; Uint8 Rloss, Gloss, Bloss, Aloss; Uint8 Rshift, Gshift, Bshift, Ashift; Uint32 Rmask, Gmask, Bmask, Amask; Uint32 colorkey; Uint8 alpha; } SDL_PixelFormat;\fR .fi .PP .SH "STRUCTURE DATA" .TP 20 \fBpalette\fR Pointer to the \fIpalette\fR, or \fBNULL\fP if the \fBBitsPerPixel\fR>8 .TP 20 \fBBitsPerPixel\fR The number of bits used to represent each pixel in a surface\&. Usually 8, 16, 24 or 32\&. .TP 20 \fBBytesPerPixel\fR The number of bytes used to represent each pixel in a surface\&. Usually one to four\&. .TP 20 \fB[RGBA]mask\fR Binary mask used to retrieve individual color values .TP 20 \fB[RGBA]loss\fR Precision loss of each color component (2^[RGBA]loss) .TP 20 \fB[RGBA]shift\fR Binary left shift of each color component in the pixel value .TP 20 \fBcolorkey\fR Pixel value of transparent pixels .TP 20 \fBalpha\fR Overall surface alpha value .SH "DESCRIPTION" .PP A \fBSDL_PixelFormat\fR describes the format of the pixel data stored at the \fBpixels\fR field of a \fI\fBSDL_Surface\fR\fR\&. Every surface stores a \fBSDL_PixelFormat\fR in the \fBformat\fR field\&. .PP If you wish to do pixel level modifications on a surface, then understanding how SDL stores its color information is essential\&. .PP 8-bit pixel formats are the easiest to understand\&. Since its an 8-bit format, we have 8 \fBBitsPerPixel\fR and 1 \fBBytesPerPixel\fR\&. Since \fBBytesPerPixel\fR is 1, all pixels are represented by a Uint8 which contains an index into \fBpalette\fR->\fBcolors\fR\&. So, to determine the color of a pixel in a 8-bit surface: we read the color index from \fBsurface\fR->\fBpixels\fR and we use that index to read the \fI\fBSDL_Color\fR\fR structure from \fBsurface\fR->\fBformat\fR->\fBpalette\fR->\fBcolors\fR\&. Like so: .PP .nf \f(CWSDL_Surface *surface; SDL_PixelFormat *fmt; SDL_Color *color; Uint8 index; \&. \&. /* Create surface */ \&. \&. fmt=surface->format; /* Check the bitdepth of the surface */ if(fmt->BitsPerPixel!=8){ fprintf(stderr, "Not an 8-bit surface\&. "); return(-1); } /* Lock the surface */ SDL_LockSurface(surface); /* Get the topleft pixel */ index=*(Uint8 *)surface->pixels; color=fmt->palette->colors[index]; /* Unlock the surface */ SDL_UnlockSurface(surface); printf("Pixel Color-> Red: %d, Green: %d, Blue: %d\&. Index: %d ", color->r, color->g, color->b, index); \&. \&.\fR .fi .PP .PP Pixel formats above 8-bit are an entirely different experience\&. They are considered to be "TrueColor" formats and the color information is stored in the pixels themselves, not in a palette\&. The mask, shift and loss fields tell us how the color information is encoded\&. The mask fields allow us to isolate each color component, the shift fields tell us the number of bits to the right of each component in the pixel value and the loss fields tell us the number of bits lost from each component when packing 8-bit color component in a pixel\&. .PP .nf \f(CW/* Extracting color components from a 32-bit color value */ SDL_PixelFormat *fmt; SDL_Surface *surface; Uint32 temp, pixel; Uint8 red, green, blue, alpha; \&. \&. \&. fmt=surface->format; SDL_LockSurface(surface); pixel=*((Uint32*)surface->pixels); SDL_UnlockSurface(surface); /* Get Red component */ temp=pixel&fmt->Rmask; /* Isolate red component */ temp=temp>>fmt->Rshift;/* Shift it down to 8-bit */ temp=temp<<fmt->Rloss; /* Expand to a full 8-bit number */ red=(Uint8)temp; /* Get Green component */ temp=pixel&fmt->Gmask; /* Isolate green component */ temp=temp>>fmt->Gshift;/* Shift it down to 8-bit */ temp=temp<<fmt->Gloss; /* Expand to a full 8-bit number */ green=(Uint8)temp; /* Get Blue component */ temp=pixel&fmt->Bmask; /* Isolate blue component */ temp=temp>>fmt->Bshift;/* Shift it down to 8-bit */ temp=temp<<fmt->Bloss; /* Expand to a full 8-bit number */ blue=(Uint8)temp; /* Get Alpha component */ temp=pixel&fmt->Amask; /* Isolate alpha component */ temp=temp>>fmt->Ashift;/* Shift it down to 8-bit */ temp=temp<<fmt->Aloss; /* Expand to a full 8-bit number */ alpha=(Uint8)temp; printf("Pixel Color -> R: %d, G: %d, B: %d, A: %d ", red, green, blue, alpha); \&. \&. \&.\fR .fi .PP .SH "SEE ALSO" .PP \fI\fBSDL_Surface\fR\fR, \fI\fBSDL_MapRGB\fP\fR ...\" created by instant / docbook-to-man, Tue 11 Sep 2001, 23:01