[Coco] BASIC09 GPLOAD clarification?

[CoCo Mongrel]

Thanks Willard, that makes much sense.


On Sun, Jun 5, 2011 at 7:47 PM, Willard Goosey <goosey at virgo.sdc.org> wrote:

> On Sun, Jun 05, 2011 at 05:51:03PM -0500, CoCo Mongrel wrote:
> > I'm a little uncertain about how GPLOAD in GFX2 is supposed to work.
>  More
> > importantly, how does it actually work?  The FM says it's supposed to
> load a
> > Get/Put buffer with data from a disk file.  The parameters listed in the
> > documentation (aside from the "GPLOAD" command) are group, buffer,
> format,
> > xdim, ydim, and size.  But nothing is mentioned about a path list, or a
> > pointer to a disk buffer, or anything that looks likely to actually get
> data
> > into the G/P buffer.
>
> Most of the calls by GFX2 (and it's C version, CGFX) don't actually do
> anything but tell Windint (cowin in NitrOS) (1) to do stuff.  In this
> case, GPLOAD gets windint ready to read binary data into a get/put
> buffer.  Immediately after issuing a GPLOAD, you need to write the
> graphics data to the screen.
>
> A quick, context-less example (page 256, "The Complete Rainbow Guide to
> OS-9 Level II Volume 1: A Beginners Guide to Windows")
>
> DIM Buffer:BYTE
> [...]
> RUN GFX2("gpload",1,2,7,40,19,190)
> FOR i:=0 to 189
>   READ Buffer
>   PUT #1, Buffer
> NEXT i
>
> Technically, run GFX2("GPLOAD",<params>) writes the characters (hex) 1B 2B
> <params> to the window.   That's all it does.  Windint then expects <size>
> bytes of data to be written to that window.
>
> You can do the same thing from the shell with display ("display 1b 2b
> ...") or from BASIC09 with PUT or PRINT.  GFX2 provides a human-readable
> interface.  Much easier to read, understand, and change.
>
> The talk about disk files is misleading.  It comes up because its common
> for windint bitmaps to be disk files, and these can be loaded by
> merge'ing the file to a window.  MERGE /dd/sys/stdptrs > /w1 for instance.
> These files look like: 1B 2B <gpload params> <data>.
>
> The example above reads its graphics data from B09 DATA statements, but
> could just as easily be a disk file.  The ultimate source doesn't matter
> to windint.
>
> Windint, you see, is almost completely controlled by escape sequences
> like this.  It's a very "clean" design for a graphics system.
>
> >  Also, the example given at the bottom of the page
> > (9-98 of the BASIC09 Reference) uses the "DEFBUFF" command rather than
> > "GPLOAD".
>
> Typo.  The Tandy manuals are notoriously full of 'em. :-(  On the same
> page, the format code should refer to Table 9.6...
>
> >Am I just misunderstanding the purpose of GPLOAD, or is this
> > mal-documented?
>
> At best, the GFX2 manual entry is _very_ poorly phrased.  The Tandy CGFX
> manual says it a little better: "Prepares to load a get/put buffer...
> the system loads the next bytes written to that path into the specified
> get/put buffer."
>
> The docs for Mike Sweets' CGFX7 says "GPLoad will generate the necessary
> loading header to preload a get/put buffer. ... A write should directly
> follow this call to insure the buffer is not loaded with garbage..."
>
> Sometimes you see or hear of "clip", "clp", or "buff" graphics files.
> These are like merge-able gpload files such as stdfonts, but aren't fully
> specified.  They don't have the GPLOAD escape codes, the group number,
> or the buffer number.  They start with <type> <x-size> <y-size> <length>
> <data>...  To load one of these from the shell you would "display 1b 2b
> <group> <buffer> ; merge <file.clp>, or something similar in BASIC09 or
> C.
>
> Well, 4 different books later, I hope you understand this a little
> better.  I'm no LII graphics master, but I do have a grip on GPLoad. :-)
>
> (1)I don't know if grfint supports get/put buffers or not.  I never used
> grfint, certainly never tried to program for it.
>
> Willard
> --
> Willard Goosey  goosey at sdc.org
> Socorro, New Mexico, USA
> I search my heart and find Cimmeria, land of Darkness and the Night.
>  -- R.E. Howard
>
> --
> Coco mailing list
> Coco at maltedmedia.com
> http://five.pairlist.net/mailman/listinfo/coco
>