Every display driver should implement at least one bitmap class for displayable bitmaps. Normally this class doesn't need to have public ID. In order to use it the driver should pass class pointer as aoHidd_BitMap_ClassPtr value to the graphics base class in its moHidd_Gfx_CreateObject implementation. BitMap base class is in C++ terminology a pure virtual baseclass. It will not allocate any bitmap data at all; that is up to the subclass to do. The main task of the BitMap baseclass is to store some information about the bitmap like its size and pixelformat. A pixelformat is an object of private class which stores the actual information about the format. There are two ways that we can find out the pixfmt in our moHidd_Gfx_CreateObject implementation: Displayable bitmap - The tags will contain a modeid. One can use this modeid to get a pointer to an already registered pixfmt. Non-displayable bitmap - The aoHidd_BitMap_StdPixFmt or aoHidd_BitMap_Friend attribute will always be passed.
[I.G]
Specify number of pixels to align bitmap data width to. This attribute can be added in order to enforce alignment needed for example by blitting hardware. It will have an impact on default aoHidd_BitMap_BytesPerRow value. Direct specification of aoHidd_BitMap_BytesPerRow attribute overrides any value of this attribute.
Default value of this attribute is 16. This alignment is required by graphics.library for AmigaOS(tm) compatibility reasons.
[ISG], ULONG
Specify or query number of bytes per row in the bitmap storage buffer. Setting this attribute doesn't actually cause changing buffer layout, just updates the information about it. Use this only from within subclasses and only if you exactly know why you do this. Specifying this attribute during object creation overrides the value calculated based on aoHidd_BitMap_Width and aoHidd_BitMap_Align values. Useful for wrapping own buffers into bitmap objects, for example, in conjunction with aoHidd_ChunkyBM_Buffer.
The returned value includes possible padding needed for alignment.
[I..]
Explicitly specify bitmap's class ID. The purpose of this attribute is to let graphics driver base class to select a class on which to call OOP_NewObject() in its moHidd_Gfx_CreateObject implementation. If neither this attribute nor aoHidd_BitMap_ClassPtr attribute is provided for moHidd_Gfx_CreateObject, graphics base class will do its best in order to find out the correct class based on aoHidd_StdPixFmt attribute value or aoHidd_BitMap_ClassPtr value of friend bitmap.
The pointer to a given class will not be remembered as aoHidd_BitMap_ClassPtr value.
aoHidd_BitMap_ClassPtr CLID_Hidd_Gfx/moHidd_Gfx_CreateObject
[I..], OOP_Class *
Explicitly specify bitmap's class pointer. This attribute is not actually a bitmap's attribute. Your display driver class can supply it to base class' moHidd_Gfx_CreateObject method in order to select a class on which to call OOP_NewObject(). If neither this attribute nor aoHidd_BitMap_ClassID attribute is provided for moHidd_Gfx_CreateObject, graphics base class will do its best in order to find out the correct class based on aoHidd_StdPixFmt attribute value or friend bitmap.
If a friend bitmap is given, the new bitmap will have the same class, if your driver doesn't override it by supplying explicit class specification (using either aoHidd_BitMap_ClassPtr or aoHidd_BitMap_ClassID attribute).
aoHidd_BitMap_ClassID CLID_Hidd_Gfx/moHidd_Gfx_CreateObject
[..G], OOP_Object *
Return associated colormap (palette) object. By default only displayable bitmaps have colormaps. However a colormap can be attached to any bitmap using moHidd_BitMap_SetColors or moHidd_BitMap_SetColorMap. Note that manual attaching of a colormap to a nondisplayable bitmap may cause undesired side-effects on graphics.library behavior. It's better not to do this at all. The system knows what it does better than you.
moHidd_BitMap_SetColorMap moHidd_BitMap_SetColors.
[G.I]
Specify or query the actual bitmap depth. This a convenience attribute to simplify handling planar bitmaps, whose actual depth may vary. Default implementation in base class simply returns depth of bitmap's pixelformat, and is ignored during initialization. Planar bitmap class returns the actual depth here. If your specific bitmap class also operates on bitmaps with variable depths, you need to implement this attribute in it.
[I.G], BOOL
The bitmap is displayable. A displayable bitmap is always managed by a display driver and must have valid display mode ID specification. If this attribute is not supplied during bitmap creation, its value defaults to FALSE.
[I.G], BOOL
Specifies that the bitmap is a framebuffer bitmap. A detailed description of a framebuffer is given in CLID_Hidd_Gfx/moHidd_Gfx_CreateObject and in CLID_Hidd_Gfx/moHidd_Gfx_Show documentation. Specifying this attribute causes also implicit setting of aoHidd_BitMap_Displayable to TRUE.
[I.G], OOP_Object *
Specify a friend bitmap. The bitmap will be allocated so that it is optimized for blitting to this bitmap. Display drivers may query this attribute and then query friend bitmap for anything they want (like pixelformat, mode ID, etc). Note that explicit specification of mode ID and/or standard pixelformat should override defaults provided by friend bitmap (i.e. actually breaking the friendship).
[I.G], OOP_Object *
Specify display driver object this bitmap was created with. Normally the user doesn't have to supply this attribute. Instead you should use driver's moHidd_Gfx_CreateObject method in order to create bitmaps. In this case aoHidd_BitMap_GfxHidd attribute will be provided by graphics driver base class with the correct value. It is illegal to manually create bitmap objects with no driver associated. graphics.library maintains at least a memory driver for nondisplayable bitmaps in system RAM without any acceleration.
CLID_Hidd_Gfx/moHidd_Gfx_CreateObject
[ISG], UWORD
Specifies bitmap height in pixels. Setting this attribute does not cause actual bitmap resize, just updates the information about it. Use this only from within subclasses only if you know what you do. For example SDL hosted driver sets it when framebufer changes the resolution.
[..G], BOOL
Check if the bitmap provides linear memory access. This means that bitmap's pixelbuffer is directly addressable by the CPU. Bitmaps with no linear memory may implement moHidd_BitMap_ObtainDirectAccess, but this means that this method will rely on mirrored buffer. In such a case the user must call moHidd_BitMap_UpdateRect after modifying bitmap's contents.
Used by cybergraphics.library/GetCyberMapAttr() for providing CYBRMATTR_ISLINEARMEM value.
Currently no display drivers implement this attribute despite many native mode drivers actually provide linear memory.
[.SG]
Controls horizontal position of a scrollable screen bitmap. Size of displayable bitmaps may differ from actual screen size. In this case the bitmap can be scrolled around the whole display area. If the bitmap is larger than the display, only its part can be visible. Setting this attribute causes changing left origin point of the bitmap. The value of this attribute represents an offset from the physical edge of the display to the logical edge of the bitmap. This means that if a large bitmap scrolls to the left in order to reveal its right part, the offset will be negative. If the bitmap scrolls to the left (possibly revealing another bitmap behind it), the offset will be positive. It's up to the display driver to set scroll limits. If the value of the attribute becomes unacceptable for any reason, the driver should adjust it and provide the real resulting value back.
Implementing screen scrolling does not enforce to implement screen composition, despite the composition is really based on scrolling (in case of composition scrolling a bitmap off-display is expected to reveal another bitmap behind it instead of empty space).
[ISG], HIDDT_ModeID
Specify display mode ID for displayable bitmap. A displayable bitmap must have this attribute supplied with valid value. A nondisplayable one may miss it, however it may remember it if it was created as a friend of displayable one. This way you may create another displayable bitmap as a friend of nondisplayable one which in turn is a friend of displayable one. This attribute can be set on a framebuffer bitmap. Doing so means an explicit request for the driver to change current display mode on the hardware. Dependent parameters (width, height and pixelformat) will be automatically adjusted, if not explicitly specified in the attributes list.
If the given ModeID is not supported, the operation causes an error. You can check for this by checking return value of OOP_SetAttrs() function. It will be TRUE in case of success and FALSE upon failure. In case of failure none of bitmap attributes will be changed.
[I.G], OOP_Object *
Specify or query pixelformat descriptor object associated with the bitmap. Every bitmap has some associated pixelformat object. Pixelformat objects are shared data storages, so many bitmaps may refer to the same pixelformat objects.
This attribute is internally specified during bitmap creation, but it's illegal to do this for the user. CreateObject method of graphics driver performs an explicit check against this. It's up to graphics base classes to figure out its value.
[...]
Private, very obsolete and currently has no function. Considered reserved.
[I..], HIDDT_StdPixFmt
Specify standard pixelformat code (one of vHidd_StdPixFmt_... values) for the bitmap. Values less than num_Hidd_PseudoStdPixFmt are illegal for this attribute. The bitmap class itself ignores this attribute. It is processed by CLID_Hidd_Gfx/moHidd_Gfx_CreateObject method in order to look up a corresponding pixelformat object in the system's database.
Bitmaps with this attribute set should be created as RAM bitmaps with direct CPU access. It is not recommended to replace them with, for example, virtual surfaces on hosted AROS. Such bitmaps are expected to be directly addressable and breaking this may cause undesired side effects.
aoHidd_BitMap_PixFmt CLID_Hidd_Gfx/moHidd_Gfx_CreateObject
[.SG]
Controls vertical position of a scrollable screen bitmap. Size of displayable bitmaps may differ from actual screen size. In this case the bitmap can be scrolled around the whole display area. If the bitmap is larger than the display, only its part can be visible. Setting this attribute causes changing top origin point of the bitmap. The value of this attribute represents an offset from the physical edge of the display to the logical edge of the bitmap. This means that if a large bitmap scrolls upwards in order to reveal its bottom part, the offset will be negative. If the bitmap scrolls downwards (possibly revealing another bitmap behind it), the offset will be positive. It's up to the display driver to set scroll limits. If the value of the attribute becomes unacceptable for any reason, the driver should adjust it and provide the real resulting value back.
Implementing screen scrolling does not enforce to implement screen composition, despite the composition is really based on scrolling (in case of composition scrolling a bitmap off-display is expected to reveal another bitmap behind it instead of empty space).
[..G], BOOL
Check if the bitmap is currently visible on screen
Not all display drivers implement this attribute. No AROS components currently rely on its value.
[ISG], UWORD
Specifies bitmap width in pixels. Setting this attribute does not cause actual bitmap resize, just updates the information about it. Use this only from within subclasses only if you know what you do. For example SDL hosted driver sets it when framebufer changes the resolution.
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_BitMapScale * msg); VOID HIDD_BM_BitMapScale(OOP_Object *obj, OOP_Object *src, OOP_Object *dest, struct BitScaleArgs * bsa, OOP_Object *gc);
obj - src - dest - bsa - gc -
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_BlitColorExpansion *msg); VOID HIDD_BM_BlitColorExpansion (OOP_Object *obj, OOP_Object *gc, OOP_Object *srcBitMap, WORD srcX, WORD srcY, WORD destX, WORD destY, UWORD width, UWORD height);
Perform a color expansion of the mask in srcBitMap according to foreground and background colors and expansion mode specified by the supplied GC. Pixels which are set to zero in the mask bitmap will be either painted by background (in opaque mode) or left without change (in transparent mode). Pixels which are set to nonzero in the mask will be painted by foreground color. The result of expansion is blitted onto the destination bitmap accorging to GC's draw mode.
obj - A bitmap to draw on gc - A GC object to use for drawing srcBitMap - A bitmap object containing mask image. srcX, srcY - A top-left coordinate of the used rectangle in the source bitmap destX, destY - A top-left coordinate of the destination rectangle to draw in width, height - A size of the rectangle to blit
None.
This method was previously used by graphics.library/Text() to draw fonts with no styles specified. Currently graphics.library always uses BltTemplate() and this method is considered obsolete.
ULONG OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_BytesPerLine *msg); ULONG HIDD_BM_BytesPerLine(OOP_Object *obj, HIDDT_StdPixFmt pixFmt, UWORD width);
This method is currently not used and reserved.
obj - pixFmt - width -
OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_Clear *msg); VOID HIDD_BM_Clear (OOP_Object *obj, OOP_Object *gc);
Sets all pixels of the drawing area to the background color.
obj - A bitmap to clear. gc - A GC object, specifies background color value
This method is not used by the system and considered reserved. However it can be useful for display driver's own needs.
Default implementation in the base class sets all pixels to zero color instead of the background color from GC
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_DrawEllipse *msg); VOID HIDD_BM_DrawEllipse (OOP_Object *obj, OOP_Object *gc, WORD x, WORD y, WORD rx, WORD ry);
Draws a hollow ellipse from the center point (x,y) with the radii rx and ry in the specified bitmap. The function does not clip the ellipse against the drawing area.
obj - A bitmap to draw on gc - A GC object to use for drawing x,y - Coordinates of center point in pixels rx,ry - ry and ry radius in pixels
None.
Because of overflow the current code do not work with big values of rx and ry.
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_DrawPixel *msg); VOID HIDD_BM_DrawLine(OOP_Object *obj, OOP_Object *gc, WORD x1, WORD y1, WORD x2, WORD y2);
Draws a line from (x1,y1) to (x2,y2) in the specified gc. The function does not clip the line against the drawing area.
obj - A bitmap to draw on gc - A graphics context object to use x1,y1 - start point of the line in pixels x2,y2 - end point of the line in pixels
None.
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_DrawPixel *msg); VOID HIDD_BM_DrawPixel(OOP_Object *obj, OOP_Object *gc, WORD x, WORD y);
Changes the pixel at (x,y). The color of the pixel depends on the attributes of gc, eg. colors, drawmode, colormask etc. This function does not check the coordinates.
obj - A bitmap to draw on gc - A GC (graphics context) object to use for drawing x, y - Coordinates of the pixel to draw
None.
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_DrawPolygon *msg); VOID HIDD_BM_DrawPolygon (OOP_Object *obj, OOP_Object *gc, UWORD n, WORD *coords);
Draws a hollow polygon from the list of coordinates in coords[]. The function does not clip the polygon against the drawing area.
obj - A bitmap to draw on gc - A GC object to use for drawing n - number of coordinate pairs coords - array of n (x, y) coordinates in pixels
None.
This method is not used by the system and considered reserved.
OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_DrawRect *msg); VOID HIDD_BM_DrawRect (OOP_Object *obj, OOP_Object *gc, WORD minX, WORD minY, WORD maxX, WORD maxY);
Draws a hollow rectangle. minX and minY specifies the upper left corner of the rectangle. minY and maxY specifies the lower right corner of the rectangle. The function does not clip the rectangle against the drawing area.
obj - A bitmap to draw on gc - A GC object to use for drawing minX, minY - upper left corner of the rectangle in pixels maxX, maxY - lower right corner of the rectangle in pixels
None.
This method is not used by the system and considered reserved.
OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_DrawText *msg); VOID HIDD_BM_DrawText (OOP_Object *obj, OOP_Object *gc, WORD x, WORD y, STRPTR text, UWORD length);
Draws the first length characters of text at (x, y). The function does not clip the text against the drawing area.
obj - A bitmap to draw on gc - A GC object to use for drawing and font specification x, y - Position to start drawing in pixels. The x coordinate is relativ to the left side of the first character. The y coordinate is relative to the baseline of the font. text - Pointer to a Latin 1 string length - Number of characters to draw
None.
At the moment text drawing is processed entirely by graphics.library using BltTemplate(), which in turn uses moHodd_BitMap_PutTemplate. This method is considered obsolete.
The default implementation in the base class does not process styles, color and alpha-blended fonts.
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_DrawEllipse *msg); VOID HIDD_BM_FillEllipse (OOP_Object *obj, OOP_Object *gc, WORD x, WORD y, WORD ry, WORD rx);
Draws a solid ellipse from the center point (x,y) with the radii rx and ry in the specified bitmap. The function does not clip the ellipse against the drawing area.
obj - A bitmap to draw on gc - A GC object to use for drawing x,y - Coordinates of center point in pixels rx,ry - ry and ry radius in pixels
None.
Because of overflow the current code do not work with big values of rx and ry.
This method is not used by the system and considered reserved.
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_DrawPolygon *msg); VOID HIDD_BM_FillPolygon (OOP_Object *obj, OOP_Object *gc, UWORD n, WORD *coords);
This method was initially designed for drawing solid polygons, however it was never used and implemented. At the moment it is considered reserved, its synopsis and semantics may change in future.
obj - A bitmap to draw on gc - A GC object to use for drawing n - number of coordinate pairs coords - array of n (x, y) coordinates in pixels
None
Never used and implemented
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_DrawRect *msg); VOID HIDD_BM_FillRect (OOP_Object *obj, OOP_Object *gc, WORD minX, WORD minY, WORD maxX, WORD maxY);
Draws a solid rectangle. minX and minY specifies the upper left corner of the rectangle. maxX and maxY specifies the lower right corner of the rectangle. The function does not clip the rectangle against the drawing area.
obj - A bitmap to draw on gc - A GC object to use for drawing minX, minY - upper left corner of the rectangle in pixels maxX, maxY - lower right corner of the rectangle in pixels
None.
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_DrawText *msg);
Reserved, never implemented method. The definition will change in future.
None.
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_DrawText *msg); VOID HIDD_BM_FillText (OOP_Object *obj, OOP_Object *gc, WORD x, WORD y, STRPTR text, UWORD length);
Historically this method was designed to draw a text with background. It was never implemented. Currently this method is considered reserved. Its synopsis and semantics may change in future.
obj - A bitmap to draw on gc - A GC object to use for drawing x, y - Position to start drawing in pixels. The x coordinate is relative to the left side of the first character. The y coordinate is relative to the baseline of the font. text - Pointer to a Latin 1 string length - Number of characters to draw
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_GetImage *msg); VOID HIDD_BM_GetImage (OOP_Object *obj, UBYTE *pixels, ULONG modulo, WORD x, WORD y, WORD width, WORD height, HIDDT_StdPixFmt pixFmt);
obj - pixels - modulo - x, y - width - height - pixFmt -
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_GetImageLUT *msg); VOID HIDD_BM_GetImageLUT (OOP_Object *obj, UBYTE *pixels, ULONG modulo, WORD x, WORD y, WORD width, WORD height, HIDDT_PixelLUT *pixlut);
obj - pixels - modulo - x, y - width - height - pixlut -
HIDDT_Pixel OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_MapColor *msg); HIDDT_Pixel HIDD_BM_MapColor(OOP_Object *obj, HIDDT_Color *color);
obj - color -
BOOL OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_ObtainDirectAccess *msg); BOOL HIDD_BM_ObtainDirectAccess(OOP_Object *obj, UBYTE **addressReturn, ULONG *widthReturn, ULONG *heightReturn, ULONG *bankSizeReturn, ULONG *memSizeReturn);
obj - addressReturn - widthReturn - heightReturn - bankSizeReturn - memSizeReturn -
BOOL
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_PutAlphaImage *msg); VOID HIDD_BM_PutAlphaImage (OOP_Object *obj, OOP_Object *gc, UBYTE *pixels, ULONG modulo, WORD x, WORD y, WORD width, WORD height);
Perform an alpha-blending operation between a bitmap and ARGB pixel array.
obj - A bitmap to operate on gc - A GC object, internally needed to perform the operation. All its attributes are ignored. pixels - A pointer to an array of pixels modulo - Number of bytes per row in pixel array x, y - Top-left corner of affected bitmap's region width - Width of the modified rectangle. height - Height of the modified rectangle.
None.
Do not rely on 'gc' parameter being valid when implementing this method in own display driver. This parameter is actually obsolete, and will be set to NULL in future AROS versions. Current base class implementation ignores it.
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_PutAlphaTemplate *msg); VOID HIDD_BM_PutAlphaTemplate (OOP_Object *obj, OOP_Object *gc, UBYTE *alpha, ULONG modulo, WORD x, WORD y, WORD width, WORD height, BOOL invertalpha);
Perform a drawing with current foreground color, using 8-bit alpha channel mask. The following GC attributes are considered: Foreground - a foreground color Background - a background color DrawMode - if set to Invert, foreground and background colors will be ignored. Instead, pixels, for which alpha channel value is greater than 127, will be inverted. Other pixels will be left unchanged. ColorExpansion - if set to Opaque, alpha blending will happen between foreground and background colors, instead of between foreground color and old bitmap contents.
obj - A bitmap to draw on gc - A GC object specifying drawing parameters alpha - A pointer to an 8-bit per pixel alpha channel mask modulo - Number of bytes per line in the mask x, y - Top-left corner of the affected bitmap's region width - Width of the affected bitmap's region height - Height of the affected bitmap's region invertalpha - If set to TRUE, alpha mask values will be treated in inverted form
None
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_PutImage *msg); VOID HIDD_BM_PutImage (OOP_Object *obj, OOP_Object *gc, UBYTE *pixels, ULONG modulo, WORD x, WORD y, WORD width, WORD height, HIDDT_StdPixFmt pixFmt);
obj - gc - pixels - modulo - x, y - width - height - pixFmt -
VOID OOP_DoMethod(OOP_Object *o, struct pHidd_BitMap_PutImageLUT *msg); VOID HIDD_BM_PutImageLUT (OOP_Object *obj, OOP_Object *gc, UBYTE *pixels, ULONG modulo, WORD x, WORD y, WORD width, WORD height, HIDDT_PixelLUT *pixlut);
obj - gc - pixels - modulo - x, y - width - height - pixlut -
VOID OOP_DoMethod(OOP_Object *o, struct pHidd_BitMap_PutPattern *msg); VOID HIDD_BM_PutPattern(OOP_Object *obj, OOP_Object *gc, UBYTE *pattern, WORD patternsrcx, WORD patternsrcy, WORD patternheight, WORD patterndepth, HIDDT_PixelLUT *patternlut, BOOL invertpattern, UBYTE *mask, ULONG maskmodulo, WORD masksrcx, WORD x, WORD y, WORD width, WORD height);
obj - A bitmap to draw on gc - A GC object to use for drawing pattern - patternsrcx - patternsrcy - patternheight - patterndepth - patternlut - invertpattern - mask - maskmodulo - masksrcx - x, y - width - height -
None
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_PutPixel *msg); VOID HIDD_BM_PutPixel(OOP_Object *obj, WORD x, WORD y, HIDDT_Pixel pixel);
Sets a new color value for the pixel at (x,y). The actual color stored may be an approximation, due to the limited color depth or palette size of the bitmap. This function does not check the coordinates.
obj - bitmap to write to. x, y - coordinates of the pixel to write. pixel - the pixel's new color value.
None.
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_PutTemplate *msg); VOID HIDD_BM_PutTemplate (OOP_Object *obj, OOP_Object *gc, UBYTE *masktemplate, ULONG modulo, WORD srcx, WORD x, WORD y, WORD width, WORD height, BOOL inverttemplate);
Apply a single-bit mask to the given portion of the bitmap. Pixels set to 1 in the mask will be filled by foreground color. Pixels set to 0 in the mask will be filled by background color or left unchanged, according to the following GC attributes: Foreground - a foreground color Background - a background color DrawMode - if set to Invert, foreground and background colors will be ignored. Instead, pixels which are set to 1 in the mask, will be inverted. Other pixels will be left unchanged. ColorExpansion - if set to Transparent, only pixels which are set to 1 in the mask, will be modified. Other pixels will not be changed (background color will be ignored).
obj - A bitmap to draw on gc - A GC object, holding operation parameters masktemplate - A pointer to a bit mask modulo - Number of bytes per line in the mask srcx - Horizontal offset of the mask x, y - Top-left corner of the bitmap's region to affect width - Width of the affected region height - Height of the affected region inverttemplate - If set to TRUE, bit mask will be interpreted in inverted form
None
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_PutTranspImageLUT *msg); VOID HIDD_BM_PutTranspImageLUT (OOP_Object *obj, OOP_Object *gc, UBYTE *pixels, ULONG modulo, WORD x, WORD y, WORD width, WORD height, HIDDT_PixelLUT *pixlut, UBYTE transparent);
Copy an array of 8-bit LUT pixels to the bitmap at the specified position making one of colors transparent. Pixels are converted to bitmap's native format using either user-supplied LUT (if given) or bitmap's own colormap. Draw mode of the supplied GC is ignored, the operation is always bulk copy.
obj - A bitmap to draw image on gc - A GC used for drawing pixels - A pointer to source pixel array modulo - Total number of bytes per line in the source array x, y - Top-left corner of the destination rectangle width - Width of the image to draw height - Height of the image to draw pixlut - An optional pointer to a LUT to use. NULL means using bitmap's own colormap (if available) transparent - Value of pixels in the source array which will be made transparent
None
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_ReleaseDirectAccess *msg); VOID HIDD_BM_ReleaseDirectAccess(OOP_Object *obj);
obj -
OOP_Object * OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_SetColorMap *msg); OOP_Object * HIDD_BM_SetColorMap(OOP_Object *obj, OOP_Object *colorMap);
obj - colorMap -
BOOL OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_SetColors *msg); BOOL HIDD_BM_SetColors (OOP_Object *obj, HIDDT_Color *colors, UWORD firstColor, UWORD numColors);
Sets values for one or more colors in the colormap object associated with the bitmap. The colormap will be created if it does not exist. Only ARGB values from the source array are taken into account. pixval member is updated with the real pixel value for every color.
obj - A bitmap object whose colormap needs to be set colors - A pointer to source data array firstColor - Number of the first color to set numColors - Number of subsequent colors to set
TRUE on success, FALSE in case of some error (like out of memory)
CLID_Hidd_ColorMap/moHidd_ColorMap_SetColors
HIDDT_RGBConversionFunction OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_SetRGBConversionFunction *msg); HIDDT_RGBConversionFunction HIDD_BM_SetRGBConversionFunction(OOP_Object *obj, HIDDT_StdPixFmt srcPixFmt, HIDDT_StdPixFmt dstPixFmt, HIDDT_RGBConversionFunction function);
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_UnmapPixel *msg); VOID HIDD_BM_UnmapPixel(OOP_Object *obj, HIDDT_Pixel pixel, HIDDT_Color *color);
obj - pixel - color -
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_BitMap_UpdateRect *msg); VOID HIDD_BM_UpdateRect(OOP_Object *obj, WORD x, WORD y, WORD width, WORD height);
Update displayed image of the given rectangle. Some drivers (like VGA and VESA) may work not with VRAM directly, but with a mirrored copy of it. Usually it is done in case if VRAM reading is slow. This method is called by the system after it completes any drawing operation, in order to make sure that changes made are visible on the actual screen. If your driver uses mirroring, this method should copy the given rectangle (at least) from the mirror buffer to the actual VRAM. This method is also called after changing currently visible bitmap (after moHidd_Gfx_Show method call) in order to allow the mirroring driver to refresh the screen after current bitmap changed. Note that moHidd_Gfx_ShowViewPorts is very different and moHidd_BitMap_UpdateRect will not be called if it succeeded!
obj - an object whose image to refresh x, y - A top-left edge of the rectangle to refresh width - Width of the rectangle to refresh height - Height of the rectangle to refresh
None.
This method is called also on offscreen bitmaps. You should track visible state of your bitmap and ignore these calls if it's not currently visible on the screen.
When working with graphics drivers this is the first object you get. It allows you to create BitMap and GC (graphics context) object. The class' methods must be overidden by hardware-specific subclasses where documented to do so.
Each display driver object internally stores a database of supported display mode IDs. This database is normally managed by base class, the driver does not need to reimplement respective methods. A display mode ID in AROS is a 32-bit integer value, the same as on AmigaOS(tm). However mode ID layout introduced by Commodore does not fit well for RTG systems. In order to overcome its limitations, display ID on AROS may have two forms: 1. A chipset mode ID. These are standard IDs defined by Commodore. You may find their definitions in graphics/modeid.h. 2. AROS RTG mode ID. An RTG mode ID is composed of three parts in the form: nnnn xx yy nnnn - monitor ID. This number is maintained by system libraries. IDs are assigned in the order in which drivers are loaded and display hardware is found. Drivers do not have to care about this part, and should normally mask it out if they for some reason look at mode ID. In order to distinguish between chipset mode IDs and RTG mode IDs, order number starts not from zero, reserving some space for C= chipset mode IDs (which appear to have order numbers from 0x0000 to 0x000A). Currently RTG monitor IDs start from 0x0010, however with time this value may change. So don't rely on some particular values in RTG IDs. Use cybergraphics.library/IsCyberModeID() function if you want to know for sure if the given mode ID belongs to an RTG driver. xx - A sync object index in driver's mode database. yy - A pixelformat object in driver's mode database. Normally the driver does not have to care about mode ID decoding. The mode database is maintained by base class. The only useful things for the driver are sync and pixelformat objects, from which it's possible to get different information about the mode. They can be obtained from the base class using HIDD_Gfx_GetMode(). Note that the driver object by itself does not know its monitor ID. Different displays are served by different objects, any of which may belong to any class. So all driver methods which return mode IDs will set monitor ID to zero. All methods that take mode ID as argument are expected to ignore the monitor ID part and do not make any assumptions about its value.
[.S.], void (*)(APTR userdata, OOP_Object *bitmap)
Set display activation interrupt handler. This handler needs to be called by hosted display driver, if host OS windowing system is used for the display and mouse input is handled by the host OS. This way the driver can tell AROS when a display window has been activated so that AROS will be able to switch current display correctly when working in a multi-display configuration. The function uses C calling convention and needs to be declared as follows: void ActivationHandler(APTR userdata, OOP_Object *bitmap); Parameters of this function will be: userdata - Whatever is specified by aoHidd_Gfx_ActiveCallBackData attribute. bitmap - Currently reserved. Drivers need to set it to NULL. The function can be called from within an interrupt, so usual restrictions apply to it. Set this attribute to NULL in order to disable activation handling.
When setting the activation callback function, be sure that you set correct userdata before you actually set the callback pointer. Otherwise your callback can be called with wrong data pointer. Only one activation handler can be installed. Installing a new handler replaces the previous one. Native displays do not need to implement this attribute because there can be no external activation events.
[.S.], APTR
Set user-defined data pointer for display activation handler.
[..G], OOP_Object *
Get a pointer to shared default GC object.
The returned GC is preset to the following: DrawMode = Copy FG = 0 BG = 0 LinePat = ~0 ColMask = ~0 You must not alter these settings even temporarily, because this GC is shared between bitmaps and between different tasks which may perform the rendering into different regions of the same bitmap (two windows on one screen, for example). This GC is intended to be used for internal copying operations.
[ISG], HIDDT_DPMSLevel
Gets or sets current DPMS level for driver's display. A value can be one of: vHidd_Gfx_DPMSLevel_On, vHidd_Gfx_DPMSLevel_Standby, vHidd_Gfx_DPMSLevel_Suspend, vHidd_Gfx_DPMSLevel_Off If the driver does not support some state, it's up to the driver what to do. Usually it is expected to ignore the request. Getting this attribute should return real current state.
[..G], STRPTR
Query CyberGraphX driver name. It is the same name which can be given to cybergraphics.library/BestCModeIDTagList() as CYBRBIDTG_BoardName value.
By default base class returns class name as value of this attribute. However this can (and must for some drivers listed in BestCModeIDTagList() documentation) be overriden.
[I.G], UBYTE
Specifies fixed framebuffer type used by the driver. The value can be one of the following: vHidd_FrameBuffer_None - the driver does not use framebuffer. vHidd_FrameBuffer_Direct - the driver uses framefuffer which can be accessed directly for both reads and writes. vHidd_FrameBuffer_Mirrored - the driver uses write-only framebuffer. This attribute has to be specified during driver object creation. If this is not done, the OS will use value of old aoHidd_Gfx_NoFrameBuffer attribute in order to distinguish between vHidd_FrameBuffer_Direct (for FALSE) and vHidd_FrameBuffer_None (for TRUE).
A fixed framebuffer is a special bitmap in a fixed area of video RAM. If the framebuffer is used, the driver is expected to copy a new bitmap into it in HIDD_Gfx_Show() and optionally copy old bitmap back. A framebuffer is needed if the hardware does not have enough VRAM to store many bitmaps or does not have capabilities to switch the display between various VRAM regions. Some hardware suffers from slow VRAM reading. In this case you should use mirrored mode. If you use it, the system will hold a bitmap in the memory buffer, and update VRAM on demand (hence the name). An example of driver using a framebuffer is hosted SDL driver. By design SDL works only with single display window, which is considered a framebuffer.
[..G], BOOL
Return hardware sprite image types supported by the driver. The returned value is a combination of the following bit flags: vHidd_SpriteType_3Plus1 - color 0 is transparent, 1-3 visible (Amiga(tm) chipset sprite format) vHidd_SpriteType_2Plus1 - color 0 is transparent, color 1 is undefined (can be whatever, for example clear or inverse), colors 2-3 visible. vHidd_SpriteType_DirectColor - Hi- or truecolor image, or LUT image with own palette, perhaps with alpha channel
This attribute should return 0 if the driver does not support hardware mouse sprite at all. Software sprite emulation is done by graphics.library. Default implementation in the base class is based on aoHidd_Gfx_SupportsHWCursor value. This is done for backwards compatibility.
[..G], BOOL
Tells if the display driver is using hosted display in host OS' window, and mouse input is handled by host OS. Windowed displays may send activation events to AROS. This is needed in order to correctly handle display switch in a multi-display configuration (which means that the user has multiple windows on host OS desktop and can freely switch between them).
Even in fullscreen mode drivers should still return TRUE if the host OS manages mouse input (for example, X11 driver). If mouse input is not managed by the host OS (for example, with Linux framebuffer driver), return FALSE.
[..G], ULONG
Query video card's memory clock in Hz. 0 is a valid value meaning 'unknown'.
[..G], ULONG
Query total size of video card memory in bytes.
[I..], struct TagItem *
Specify a pointer to a taglist which contains description of display modes supported by the driver. This attribute is usually appended in moRoot_New method of the display driver class. This attribute is mandatory for the base class, otherwise driver object creation fails. Mode description taglist may contain the following tags: - Any sync attributes - these attributes will specify values common for all sync modes - Any pixelformat attributes - these attributes will specify values common for all pixelformat modes - aoHidd_Gfx_SyncTags - specifies a pointer to another separate taglist containing attributes for one sync (display) mode. If this tag is not supplied at all, a set of default modes will be generated for the driver. - aoHidd_Gfx_PixFmtTags - specifies a pointer to another separate taglist containing attributes for one pixelformat. This tag must be supplied at least once, otherwise driver object will fail to create. aoHidd_Gfx_SyncTags and aoHidd_Gfx_PixFmtTags can be specified multiple times in order to associate more than one display mode with the driver. Note that common values for sync and pixelformat objects need to be placed in the taglist before aoHidd_Gfx_SyncTags and aoHidd_Gfx_PixFmtTags. You may specify them again between these tags in order to alter common values.
Partial example code of display driver supporting a truecolor display with three resolutions: // Our pixelformat (24-bit 0BGR) struct TagItem pftags[] = { { aHidd_PixFmt_RedShift , 24 }, { aHidd_PixFmt_GreenShift , 16 }, { aHidd_PixFmt_BlueShift , 8 }, { aHidd_PixFmt_AlphaShift , 0 }, { aHidd_PixFmt_RedMask , 0x000000FF }, { aHidd_PixFmt_GreenMask , 0x0000FF00 }, { aHidd_PixFmt_BlueMask , 0x00FF0000 }, { aHidd_PixFmt_AlphaMask , 0x00000000 }, { aHidd_PixFmt_ColorModel , vHidd_ColorModel_TrueColor }, { aHidd_PixFmt_Depth , 24 }, { aHidd_PixFmt_BytesPerPixel, 4 }, { aHidd_PixFmt_BitsPerPixel , 24 }, { aHidd_PixFmt_StdPixFmt , vHidd_StdPixFmt_Native }, { aHidd_PixFmt_BitMapType , vHidd_BitMapType_Chunky }, { TAG_DONE , 0UL } }; // 640x480 resolution struct TagItem tags_800_600[] = { { aHidd_Sync_HDisp , 640 }, { aHidd_Sync_VDisp , 480 }, { TAG_DONE , 0UL } }; // 800x600 resolution struct TagItem tags_800_600[] = { { aHidd_Sync_HDisp , 800 }, { aHidd_Sync_VDisp , 600 }, { TAG_DONE , 0UL } }; // 1024x768 resolution struct TagItem tags_1024_768[] = { { aHidd_Sync_HDisp , 1024 }, { aHidd_Sync_VDisp , 768 }, { TAG_DONE , 0UL } }; // Mode description taglist itself struct TagItem mode_tags[] = { // Our driver supports a single pixelformat { aHidd_Gfx_PixFmtTags , (IPTR)pftags }, // Here go sync values common for all sync modes { aHidd_Sync_HMin , 112 }, { aHidd_Sync_VMin , 112 }, { aHidd_Sync_HMax , 16384 }, { aHidd_Sync_VMax , 16384 }, { aHidd_Sync_Description, (IPTR)"Example: %hx%v" }, // First resolution { aHidd_Gfx_SyncTags , (IPTR)tags_800_600 }, // Next two syncs will have HMax = 32768, as an example { aHidd_Sync_HMax , 32768 }, // Two more resolutions { aHidd_Gfx_SyncTags , (IPTR)tags_800_600 }, { aHidd_Gfx_SyncTags , (IPTR)tags_1024_768 }, { TAG_DONE , 0UL } }; // This is the attribute list which is given to New method // of the base class struct TagItem mytags[] = { { aHidd_Gfx_ModeTags , (IPTR)mode_tags }, { TAG_DONE , NULL } };
[..G], BOOL
Tells whether the driver does not need a framebuffer. Since v1.2 this attribute is obsolete. Please use aoHidd_Gfx_FrameBufferType in new code.
Provides FALSE if not implemented in the driver.
[..G], ULONG
Gets total number of sync objects in the internal display mode database.
[..G], UBYTE
Specifies if the driver supports gamma correction tables. Default implementation in base class returns FALSE.
[..G], BOOL
Tells whether the driver supports hardware mouse pointer sprite. If the driver provides TRUE value for this attribute, it is expected to implement HIDD_Gfx_SetCursorPos(), HIDD_Gfx_SetCursorShape() and HIDD_Gfx_SetCursorVisible() methods. Mouse pointer counts for one hardware sprite, so if the driver implements also HIDD_Gfx_ModeProperties(), it should set NumHWSprites to 1 in order to provide valid information about display modes. The driver must implement this attribute if it implements HIDD_Gfx_ModeProperties(). Otherwise it will provide false information in graphics.library/GetDisplayInfoData(). Base class can determine NumHWSprites based on this attribute value but not vice versa.
Default implementation in the base class returns FALSE. This causes the system to use software sprite emulation. This attribute is obsolete and is used only by AROS graphics.library up to v41.2. In new drivers consider implementing aoHidd_Gfx_HWSpriteTypes attribute.
BOOL OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_CheckMode *msg); BOOL HIDD_Gfx_CheckMode(OOP_Object *gfxHidd, HIDDT_ModeID modeID, OOP_Object *sync, OOP_Object *pixFmt);
Check if given display mode is supported by the driver. Normally any resolution (sync) can be used together with any pixelformat. However on some hardware there may be exceptions from this rule. In such a case this method should be implemented, and check should be performed. The information provided by this method is used in order to exclude unsupported modes from the database Default implementation in the base class just returns TRUE for all supplied values. Note that this method can not be used in order to chech that the given mode is really present in the database and it really refers to the given sync and pixelformat objects. Use HIDD_Gfx_GetMode() for mode ID validation.
gfxHidd - A display driver object modeID - A display mode ID sync - A pointer to a sync object associated with this mode pixFmt - A pointer to a pixelformat object associated with this mode
TRUE if this mode is supported and FALSE if it's not.
Currently base class does not call this method after driver object creation. This needs to be fixed.
ULONG OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_CleanViewPort *msg); ULONG HIDD_Gfx_CleanViewPort(OOP_Object *gfxHidd, struct HIDD_ViewPortData *data)
Performs driver-specific cleanup on a given ViewPort.
gfxHidd - A display driver object. data - a pointer to a HIDD_ViewPortDats structure.
The same code as used as return value for graphics.library/MakeVPort().
When graphics.library calls this method, the ViewPort is already unlinked from its view, and the bitmap can already be deallocated. This means that both data->Next and data->Bitmap pointers can contain invalid values.
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_CopyBox *msg); VOID HIDD_Gfx_CopyBox(OOP_Object *gfxHidd, OOP_Object *src, WORD srcX, WORD srcY, OOP_Object *dest, WORD destX, WORD destY, UWORD width, UWORD height, OOP_Object *gc);
Perform rectangle copy (blit) operation from one bitmap to another. Given bitmaps may belong to different display drivers. The driver may attempt to use hardware for acceleration (if available), and if it's impossible, pass the operation on to the base class. Always check class of the supplied bitmap before attempting to look at its private data. A GC is used in order to specify raster operation performed between the source and destination according to its aHidd_GC_DrawMode attribute value.
gfxHidd - a display driver object that you are going to use for copying src - a pointer to source bitmap object srcX - an X coordinate of the source rectangle srcY - a Y coordinate of the source rectangle dest - a pointer to destination bitmap object destX - an X coordinate of the destination rectangle destY - a Y coordinate of the destination rectangle width - width of the rectangle to copy height - height of the rectangle to copy gc - graphics context holding draw mode on the destination
None.
You must specify valid coordinates (non-negative and inside the actual bitmap area), no checks are done. It is valid to specify two overlapped areas of the same bitmap as source and destination.
IPTR OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_CopyBoxMasked *msg); IPTR HIDD_Gfx_CopyBoxMasked(OOP_Object *gfxHidd, OOP_Object *src, WORD srcX, WORD srcY, OOP_Object *dest, WORD destX, WORD destY, UWORD width, UWORD height, PLANEPTR mask, OOP_Object *gc);
Perform rectangle copy (blit) operation from one bitmap to another, using a cookie cutter mask. Given bitmaps must be on the same display driver. A GC is used in order to specify raster operation performed between the source and destination according to its aHidd_GC_DrawMode attribute value.
gfxHidd - a display driver object that you are going to use for copying src - a pointer to source bitmap object srcX - an X coordinate of the source rectangle srcY - a Y coordinate of the source rectangle dest - a pointer to destination bitmap object destX - an X coordinate of the destination rectangle destY - a Y coordinate of the destination rectangle width - width of the rectangle to copy height - height of the rectangle to copy mask - single bitplane mask gc - graphics context holding draw mode on the destination
TRUE is the operation succeeded and FALSE in case of some error, for example if the system was too low on memory.
You must specify valid coordinates (non-negative and inside the actual bitmap area), no checks are done. It is valid to specify two overlapped areas of the same bitmap as source and destination. Mask size must correspond to full source bitmap size (including alignment).
OOP_Object *OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_CreateObject *msg); OOP_Object *HIDD_Gfx_CreateObject(OOP_Object *gfxHidd, OOP_Class *cl, struct TagItem *tagList);
Create a driver specific Gfx Object of the type "classID"
gfxHidd - The graphics driver used to create the object. cl - The base OOP_Class of the object to be created. tagList - Object specific attributes.
pointer to the newly created OOP_Object, or NULL on failure.
Drivers should query the gfx.hidd, or support class for the base class Ptr that the driver objects should use. The gfx hidd itself defines the following -: GC A GC object is just used for data storage. It is possible to subclass, however it is not recommended since it may not be future-proof due to the fact GC subclasses can not be interchanged between different drivers. Avoid using custom GCs. BitMap Each graphics driver exposes at least one displayable bitmap class. More may be exposed at the drivers discretion to represent nondisplayable bitmaps or other driver specific bitmap types. Generally bitmap objects are never created directly. Instead they are created using the HIDD_Gfx_CreateObject() call. An implementation of this method in the driver should examine bitmap attributes supplied and make a decision if the bitmap should be created using the driver's own class or one of the system classes. A typical implementation should pay attention to the following bitmap attributes: aHIDD_BitMap_ModeID - If this attribute is supplied, the bitmap needs to be either displayable by this driver, or be a friend of a displayable bitmap. A friend bitmap usually repeats the internal layout of its friend so that the driver may perform blitting operations quickly. aHIDD_BitMap_Displayable - If this attribute is supplied, the bitmap NEEDS to be displayable by this driver. Usually this means that the bitmap object will contain video hardware state information. This attribute will always be accompanied by aHIDD_BitMap_ModeID. aHIDD_BitMap_FrameBuffer - The bitmap needs to be a framebuffer bitmap. A framebuffer bitmap is necessary for some kinds of hardware which have a small fixed amount of video RAM which can hold only one screen at a time. Setting this attribute requires that a valid ModeID be also set. aHIDD_BitMap_Friend - If there's no ModeID supplied, you may wish to check class of friend bitmap. This can be useful if your driver uses different classes for displayable and non-displayable bitmaps. By default base class will pick up friend's class and use it for new bitmap if nothing is specified, here you may override this behavior. If a driver wants to specify a custom class for the bitmap being created, it should pass the aoHidd_BitMap_ClassPtr attribute to the base class. Bitmap objects should not be directly created, otherwise necessary information provided by the base class will be missing. This method must be implemented by the subclass. aHIDD_BitMap_ClassPtr or aHIDD_BitMap_ClassID must be provided to the base class for a displayable bitmap.
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_DisposeOverlay *msg); VOID HIDD_Gfx_DisposeOverlay(OOP_Object *gfxHidd, OOP_Object *Overlay)
Deletes an overlay previously created by moHidd_Gfx_NewOverlay. Subclasses do not have to override this method unless they allocate anything additional to an overlay object in their HIDD_Gfx_NewOverlay() implementation.
gfxHidd - A driver object which was used for creating a GC. Overlay - Pointer to an overlay object to delete.
None.
BOOL OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_Gamma *msg); BOOL HIDD_Gfx_GetGamma(OOP_Object *gfxHidd, UBYTE *Red, UBYTE *Green, UBYTE *Blue);
Get current gamma table for the display. This method was neither ever implemented nor used. Currently obsolete and considered reserved.
BOOL OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_GetMaxSpriteSize *msg); BOOL HIDD_Gfx_GetMaxSpriteSize(OOP_Object *gfxHidd, ULONG Type, UWORD *Width, UWORD *Height);
Query maximum allowed size for the given sprite type.
gfxHidd - A display driver object Type - Type of the sprite image (one of vHidd_SpriteType_... values) Width - A pointer to UWORD where width will be placed. Height - A pointer to UWORD where height will be placed.
FALSE is the given sprite type is not supported, otherwise TRUE.
Default implementation in the base class just return some small values which it hopes can be supported by every driver if the driver supports given sprite type. It is strongly suggested to reimplement this method in the display driver. Width and Height are considered undefined if the method returns FALSE.
BOOL OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_GetMode *msg); BOOL HIDD_Gfx_GetMode(OOP_Object *gfxHidd, HIDDT_ModeID modeID, OOP_Object **syncPtr, OOP_Object **pixFmtPtr);
Get sync and pixelformat objects for a particular display ModeID.
gfxHidd - pointer to a driver object which this ModeID belongs to syncPtr - pointer to a storage where sync object pointer will be placed pixFmtPtr - pointer to a storage where pixelformat object pointer will be placed
TRUE upon success, FALSE in case of failure (e.g. given mode does not exist in driver's internal database). If the function returns FALSE, sync and pixelformat pointers will be set to NULL.
Every display mode is associated with some sync and pixelformat object. If the method returns TRUE, object pointers are guaranteed to be valid.
OOP_Object *OOP_DoMethod(OOP_Object *o, struct pHidd_Gfx_GetPixFmt *msg); OOP_Object *HIDD_Gfx_GetPixFmt(OOP_Object *gfxHidd, HIDDT_StdPixFmt pixFmt);
Get a standard pixelformat descriptor from internal pixelformats database.
gfxHidd - A display driver object pixFmt - An index of pixelformat (one of vHIDD_StdPixFmt_... values)
A pointer to a pixelformat object or NULL if lookup failed
Pixelformat objects are stored in a global system-wide database. They are not linked with a particular driver in any way and completely sharable between all drivers.
OOP_Object *OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_GetSync *msg); OOP_Object *HIDD_Gfx_GetSync(OOP_Object *gfxHidd, ULONG num);
Get a sync object from internal display mode database by index
gfxHidd - A display driver object to query num - An index of sync object starting from 0
A pointer to a sync object or NULL if there's no sync with such index
ULONG OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_MakeViewPort *msg); ULONG HIDD_Gfx_MakeViewPort(OOP_Object *gfxHidd, struct HIDD_ViewPortData *data)
Performs driver-specific setup on a given ViewPort.
gfxHidd - A display driver object. data - a pointer to a HIDD_ViewPortData structure.
The same code as used as return value for graphics.library/MakeVPort().
When graphics.library calls this method, a complete view is not built yet. This means that data->Next pointer contains invalid data and needs to be ignored. It is valid to keep private data pointer in data->UserData accross calls. Newly created HIDD_ViewPortData is guraranteed to have NULL there.
ULONG OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_ModeProperties *msg); ULONG HIDD_Gfx_ModeProperties(OOP_Object *gfxHidd, HIDDT_ModeID modeID, struct HIDD_ModeProperties *props, ULONG propsLen);
Obtain an information about the video mode. Video mode description structure may grow in future, so be careful and always check propsLen parameter value. A system may ask you for less data than you can provide. Always return an actual value. Do not just zero out fields you don't know about, this is not expected to be backwards compatible.
gfxHidd - a pointer to a display driver object whose display mode you want to query modeID - a mode ID to query props - a pointer to a storage area where HIDD_ModeProperties structure will be put propsLen - length of the supplied buffer in bytes.
Actual length of obtained structure
Returned data must reflect only real hardware capabilities. For example, do not count emulated sprites. The system takes care about emulated features itself.
OOP_Object *OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_NewOverlay *msg); OOP_Object *HIDD_Gfx_NewOverlay(OOP_Object *gfxHidd, struct TagItem *tagList);
Create a video overlay object.
gfxHidd - A graphics driver object on whose display you want to create an overlay. tagList - A list of overlay attributes. See overlay class documentation for their description.
Pointer to the newly created overlay object or NULL in case of failure.
Default implementation in the base class always sets VOERR_INVSCRMODE error and returns NULL meaning that hardware overlays are not supported. There's no sense in software implementation because the software is supposed to handle software rendering itself.
HIDDT_ModeID OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_NextModeID *msg); HIDDT_ModeID HIDD_Gfx_NextModeID(OOP_Object *gfxHidd, HIDDT_ModeID modeID, OOP_Object **syncPtr, OOP_Object **pixFmtPtr);
Iterate driver's internal display mode database.
gfxHidd - A driver object to query modeID - A previous mode ID or vHidd_ModeID_Invalid for start of the iteration syncPtr - A pointer to a storage where pointer to sync object will be placed pixFmtPtr - A pointer to a storage where pointer to pixelformat object will be placed
Next available mode ID or vHidd_ModeID_Invalid if there are no more display modes. If the function returns vHidd_ModeID_Invalid, sync and pixelformat pointers will be set to NULL.
OOP_Object *OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_NominalDimensions *msg); OOP_Object *HIDD_Gfx_NominalDimensions(OOP_Object *gfxHidd, UWORD *width, UWORD *height, UBYTE *depth);
gfxHidd - The graphics driver used to create the object.
ULONG OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_PrepareViewPorts *msg); ULONG HIDD_Gfx_PrepareViewPorts(OOP_Object *gfxHidd, struct HIDD_ViewPortData *data, struct View *view)
Performs driver-specific setup on a given view.
gfxHidd - A display driver object. data - a pointer to a chain of HIDD_ViewPortData structures. view - A pointer to graphics.library View structure being prepared.
MCOP_OK if there was no error or MCOP_NO_MEM otherwise. MCOP_NOP is not allowed as a return value of this method.
graphics.library calls this method in MrgCop() after the complete view is built. data->Next pointer contains valid data. This function can be repeatedly called several times, and there is no cleanup counterpart for it. This should be taken into account in method implementation.
BOOL OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_QueryHardware3D *msg); BOOL HIDD_Gfx_QueryHardware3D(OOP_Object *gfxHidd, OOP_Object *pixFmt);
Query if the driver supports hardware-accelerated 3D graphics for the given pixelformat.
gfxHidd - A display driver object pixFmt - A pointer to a pixelformat descriptor object
TRUE if the driver supports hardware-accelerated 3D for the given pixelformat, FALSE otherwise.
HIDDT_ModeID *OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_QueryModeIDs *msg); HIDDT_ModeID *HIDD_Gfx_QueryModeIDs(OOP_Object *gfxHidd, struct TagItem *queryTags);
Obtain a table of all supported display mode IDs The returned address points to an array of HIDDT_ModeID containing all ModeIDs supported by this driver. The array is terminated with vHidd_ModeID_Invalid.
gfxHidd - A driver object which to query. querytags - An optional taglist containing query options. Can be NULL. The following tags are supported: tHidd_GfxMode_MinWidth (ULONG) - A minimum width of modes you are interested in tHidd_GfxMode_MaxWidth (ULONG) - A maximum width of modes you are interested in tHidd_GfxMode_MinHeight (ULONG) - A minimum height of modes you are interested in tHidd_GfxMode_MaxHeight (ULONG) - A maximum height of modes you are interested in tHidd_GfxMode_PixFmts (HIDDT_StdPifXmt *) - A pointer to an array of standard pixelformat indexes. If supplied, only mode IDs whose pixelformat numbers match any of given ones will be returned.
A pointer to an array of ModeIDs or NULL in case of failure
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_ReleaseModeIDs *msg); VOID HIDD_Gfx_ReleaseModeIDs(OOP_Object *gfxHidd, HIDDT_ModeID *modeIDs);
Free array of display mode IDs returned by HIDD_Gfx_QueryModeIDs()
gfxHidd - A driver object used to obtain the array modeIDs - A pointer to an array
None.
BOOL OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_SetCursorPos *msg); BOOL HIDD_Gfx_SetCursorPos(OOP_Object *gfxHidd, WORD x, WORD y);
Set current mouse pointer position. This is a real position on top-left image corner relative to top-left corner of the physical display. Neither logical screen origin nor hotspot are taken into account here. The default implementation in the base class does nothing and just returns TRUE. If a software pointer emulation is used, this method will never be called.
gfxHidd - a display driver object, on whose display you wish to position the pointer x - An x coordinate of the pointer (relative to the physical screen origin) y - A y coordinate of the pointer (relative to the physical screen origin)
Always TRUE. Reserved for future, do not use it.
This method is called by graphics.library/MoveSprite() which has no return value. However, for historical reasons, this method has a return value. Drivers should always return TRUE in order to ensure future compatibility.
BOOL OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_SetCursorShape *msg); BOOL HIDD_Gfx_SetCursorShape(OOP_Object *gfxHidd, OOP_Object *shape, WORD xoffset, WORD yoffset);
Set mouse pointer shape. A pointer image is contained in the specified bitmap object. The bitmap object may contain a colormap if the system wants to specify own colors for the pointer. The supplied colormap will also contain alpha channel values. It is up to driver what to do if, for example, alpha channel is not supported by the hardware. Or if given bitmap type is not supported (for example truecolor bitmap on LUT-only hardware). It is expected that the driver converts bitmap data to a more appropriate form in such a case. A hotspot is given as an offset from the actual hotspot to the top-left corner of the pointer image. It is generally needed only for hosted display drivers which utilize host's support for mouse pointer. The default implementation in the base class just does nothing. A software mouse pointer is implemented in a special layer called fakegfx.hidd inside graphics.library. If a software pointer emulation is used, this method will never be called.
gfxHidd - a display driver object, for whose display you wish to change the pointer shape - a pointer to a bitmap object, containing pointer bitmap xoffset - a horizontal hotspot offset yoffset - a vertical hotspot offset
TRUE on success, FALSE on failure
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_SetCursorVisible *msg); VOID HIDD_Gfx_SetCursorVisible(OOP_Object *gfxHidd, BOOL visible);
Control mouse pointer visiblity. The default implementation in the base class does nothing. If a software pointer emulation is used, this method will never be called.
gfxHidd - a display driver object, on whose display you wish to turn pointer on or off visible - TRUE to enable pointer display, FALSE to disable it
None.
BOOL OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_Gamma *msg); BOOL HIDD_Gfx_SetGamma(OOP_Object *gfxHidd, UBYTE *Red, UBYTE *Green, UBYTE *Blue);
Set current gamma table for the display. A gamma table consists of three 256-byte tables: one for red component, one for green and one for blue.
gfxHidd - A display driver object Red - A pointer to a 256-byte array for red component Green - A pointer to a 256-byte array for green component Blue - A pointer to a 256-byte array for blue component
FALSE if the driver doesn't support gamma correction, otherwise TRUE
BOOL OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_SetMode *msg); BOOL HIDD_Gfx_SetMode(OOP_Object *gfxHidd, OOP_Object *sync);
Update display mode according to changed sync object
gfxHidd - A display driver to operate on sync - A modified sync object pointer
TRUE if everything went OK and FALSE in case of some error
This method is used to inform the driver that some external program has changed sync data and wants to update the display if needed. It's up to the implementation to check that current display is really using this sync (frontmost screen uses this mode).
OOP_Object *OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_Show *msg); OOP_Object *HIDD_Gfx_Show(OOP_Object *gfxHidd, OOP_Object *bitMap, ULONG flags);
Change currently displayed bitmap on the screen. The bitmap object supplied must have been created with aHidd_BitMap_Displayable attribute set to TRUE. The function's behavior differs a lot depending on whether the driver uses a framebuffer or video hardware is able to switch screens itself. If the driver uses a framebuffer bitmap, it is supposed to copy the supplied bitmap into the framebuffer and return a framebuffer pointer. It also can be asked to copy back old framebuffer contents into previous bitmap object. It is driver's job to keep track of which bitmap object was displayed last time. This is what default implementation does. Note that it is very basic, and even does not support changing display resolution. It's not recommended to rely on it in production drivers (unless your video hardware supports only one mode). If the driver does not use a framebuffer, it is supposed to reprogram the hardware here to display an appropriate region of video RAM. Do not call the base class in this case, its implementation relies on framebuffer existance and will always return NULL which indicates an error. It is valid to get NULL value in bitMap parameter. This means that there is nothing to display and the screen needs to be blanked out. It is valid for non-framebuffer-based driver to return NULL as a reply then. In all other cases NULL return value means an error. Please avoid returning errors at all. graphics.library/LoadView() has no error indication. An error during showing a bitmap would leave the display in unpredictable state. If the driver does not use a framebuffer, consider using HIDD_Gfx_ShowViewPorts(). It's more straightforward, flexible and offers support for screen composition.
gfxHidd - a display driver object, whose display you wish to change. bitMap - a pointer to a bitmap object which needs to be shown or NULL. flags - currently only one flag is defined: fHidd_Gfx_Show_CopyBack - Copy back the image data from framebuffer bitmap to old displayed bitmap. Used only if the driver needs a framebuffer.
A pointer to a currently displayed bitmap object or NULL (read FUNCTION paragraph for detailed description)
Drivers which use mirrored video data buffer do not have to update the display immediately in this method. moHidd_BitMap_UpdateRect will be sent to the returned bitmap if it's not NULL. Of course display blanking (if NULL bitmap was received) needs to be performed immediately.
VOID OOP_DoMethod(OOP_Object *obj, OOP_Msg msg);
Indicate upcoming machine reset. Obsolete. Since graphics.library v41.4 this method is not used any more. Considered reserved. Do not use it in any way.
None.
None.
ULONG OOP_DoMethod(OOP_Object *obj, struct pHidd_Gfx_ShowViewPorts *msg); ULONG HIDD_Gfx_ShowViewPorts(OOP_Object *gfxHidd, struct HIDD_ViewPortData *data, struct View *view);
Show one or more bitmaps on the screen. It is completely up to the driver how to implement this function. The driver may or may not support hardware-assisted screens composition. Bitmaps are sorted in the list in descending z-order. The driver is expected to put at least frontmost bitmap on display. It is valid to get NULL pointer as data parameter. This means that there's nothing to show and the screen should go blank. Bitmaps display offsets are stored in their aHidd_BitMap_LeftEdge and aHidd_BitMap_TopEdge attributes. This function is not expected to modify their values somehow. They are assumed to be preserved between calls unless changed explicitly by the system. If you implement this method, you don't have to implement HIDD_Gfx_Show() because it will never be called. Note that there is no more error indication - the driver is expected to be error-free here.
gfxHidd - a display driver object, whose display you wish to change. data - a singly linked list of bitmap objects to show
TRUE if this method is supported by the driver, FALSE otherwise
--background_planarbm-- | aoHidd_PlanarBM_AllocPlanes | aoHidd_PlanarBM_BitMap |
This is a class representing a planar Amiga(tm) bitmap in AROS graphics subsystem. When you create an object of this class, an associated planar bitmap will be created. However, it's possible to use this class with pre-existing bitmaps, making them available to the graphics HIDD subsystem.
[I..], BOOL
Set this attribute to FALSE if you want to create an empty bitmap object containing no bitmap data. Useful if you want to create an empty object to be associated with existing bitmap later.
This attribute is obsolete. It's equal to supplying aoHidd_PlanarBM_BitMap attribute with a NULL value.
[ISG], struct BitMap *
Allows to specify or retrieve a raw planar bitmap structure associated with the object. Useful for direct access to the bitmap within subclasses, as well as for associating an object with already existing BitMap structure. It is valid to pass this attribute with a NULL value. In this case the object becomes empty and contains no actual bitmap.
If the object was created with own bitmap data (with no aoHidd_PlanarBM_BitMap specified during creation), this data will be deallocated when you set this attribute. It's up to you to deallocate own bitmaps, set using this attribute. Even if the object is disposed, it won't deallocate user-supplied bitmap.
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_GC_SetClipRect *msg); VOID HIDD_GC_SetClipRect(OOP_Object *obj, LONG x1, LONG y1, LONG x2, LONG y2);
Install a clipping rectangle on a GC.
obj - a GC object x1, y1 - top-left coordinate of the clipping rectangle x2, y2 - bottom-right coordinate of the clipping rectangle
None
Since the GC is just a data container, installing clipping rectangle doesn't magically applies it to all operations. Graphics driver method which uses the GC needs to support it explicitly. Currently clipping is supported only by Draw and DrawEllipse methods. Use this method if and only if the GC object was created by you. graphics.library internally operates on temporary GC objects, which are allocated only partially. They don't have storage space for clipping rectangle data, and attempt to use this method on such a GC will result in memory trashing.
VOID OOP_DoMethod(OOP_Object *obj, struct pHidd_GC_UnsetClipRect *msg); VOID HIDD_GC_UnsetClipRect(OOP_Object *obj);
Uninstalls the clipping rectangle (whatever it is) from the GC.
obj - a GC object
None
--background_overlay-- | aoHidd_Overlay_Error | aoHidd_Overlay_SrcFormat | aoHidd_Overlay_SrcHeight |
aoHidd_Overlay_SrcWidth |
Objects of overlay class represent hardware video overlays. Current hardware supports only one video overlay per screen, however in future the situation may change. hidd.graphics.overlay is an interface name. There's no such public ID since there's actually no base class for the overlay. The whole implementation is hardware-dependant and needs to be done separately for every driver. Overlay classes do not need to be public. It's up to display drivers to manage them. A moHidd_Gfx_NewOverlay method of graphics driver class is used to create overlay objects.
[I..], ULONG *
Specifies a pointer to ULONG location where error code will be written. This attribute can be used for overlay creation in order to be able to get an information about the actual failure reason. Resulting error code can be one of VOERR_... values defined in cybergraphx/cgxvideo.h: VOERR_OK - there was no error VOERR_INVSCRMODE - no (more) hardware overlays are supported on this card VOERR_NOOVLMEMORY - there is not enough VRAM to hold overlay data VOERR_INVSRCFMT - given source pixel format is not supported by the card VOERR_NOMEMORY - there is not enough system RAM for internal driver needs
[I..], ULONG
Specifies source data pixel format. The value should be one of SRCFMT_... constants defined in cybergraphx/cgxvideo.h: SRCFMT_YUV16 - 16-but YUV SRCFMT_YCbCr16 - 16-bit YCbCr SRCFMT_RGB15PC - R5G5B5, little-endian SRCFMT_RGB16PC - R5G6B5, little-endian
Not all formats can be supported by all drivers. Use aoHidd_Overlay_Error attribute in order to get an explanation why overlay creation fails.
aoHidd_PixFmt_CgxPixFmt |
[..G], ULONG
Returns pixelformat number according to CyberGraphX standard or -1 if the pixelformat has no correct representation in CGX (for example, planar pixelformats).
aoHidd_ColorMap_NumEntries | moHidd_ColorMap_GetColor | moHidd_ColorMap_GetPixel | moHidd_ColorMap_SetColors |
BOOL OOP_DoMethod(OOP_Object *o, struct pHidd_ColorMap_GetColor *msg); BOOL HIDD_CM_GetColor(OOP_Object *obj, ULONG colorNo, HIDDT_Color *colorReturn);
obj - colorNo - colorReturn -
HIDDT_Pixel OOP_DoMethod(OOP_Object *obj, struct pHidd_ColorMap_GetPixel *msg); HIDDT_Pixel HIDD_CM_GetPixel(OOP_Object *obj, ULONG pixelNo);
obj - pixelNo -
BOOL OOP_DoMethod(OOP_Object *obj, struct pHidd_ColorMap_SetColors *msg); BOOL HIDD_CM_SetColors(OOP_Object *obj, HIDDT_Color *colors, ULONG firstColor, ULONG numColors, OOP_Object *pixFmt);
obj - colors - firstColor - numColors - pixFmt -