[Prev][Next][Index][Thread]

Re: A92: PROPOSAL: TI-89/92 Screen Library


Do you think it would slow it down too much if we somehow made a zoom function for
it? So if someone ran a 92 program on the 89 it would just have a zoomed out look
to it, and vice versa. If we did it in grayscale it might be semi-readable.

Bryan Christopher Turner wrote:

> // Structure for describing any type of screen
> typedef struct GraphStruct
>     {
>     char height,width;
>     GraphStruct *next;          // Next graph struct on stack.
>     short numBitPlanes;         // The number of structures to follow
>     struct
>        {
>        short planeAddrDiv8;
>        char *planeAddr;
>        } thePlane [];           // This is an array
>     } GraphStruct;
>
> GraphStruct *GraphInit()
>     It figgures out what calc it's on and returns a set of flags for the
>     programmer to use.  These include the current LCD buffer address, height,
>     width, and wether grayscale is running, and it's depth.
>
> GraphStruct *GraphNew( short numBitplanes )
> // 1 bitplane  = B&W
> // 2 bitplanes = 4 grayscale
> // 3 bitplanes = 8 grayscale
> // etc..
>     Creates a new LCD buffer of the correct size and returns a pointer to it's
> GraphStruct.
>
> void GraphFree( GraphStruct *theGraph )
>     Frees an allocated buffer.  Correctly frees all bitplanes, and also calls
>     GraphRestore() if the bitplane given is the current bitplane.
>
> void GraphSwitch( GraphStruct *theGraph )
>     Switches the LCD to point to a new LCD buffer.  This function would place
>     the old buffer address on a stack for later retrieval.  Also, it would take
>     care of installing or uninstalling the grayscale interrupts for the buffer.
>     As long as all programs use this function to switch the LCD buffer, there
>     will never be a problem!
>
> void GraphRestore()
>     Pops off the old GraphStruct and makes it the current one.  This removes
>     interrupts, etc.. and also switches the LCD base address.
>
>    Anyone have other ideas/features?  Remember, this has to be able to work on
> calcs with different sized screens, and different methods to switch the LCD
> buffer address.  I included the baseAddrDiv8 field for TI-92 type interrupts,
> although this may change after we find out how the '89 works.
>
> --Bryan
> bcturner@eos.ncsu.edu
References: