World: r3wp

cardinality of ANY is  0,N
cardinality of SOME is 1,N

I changed my mind, i uses the default caret and highlight behavior, 
much simple.

Ammon some doesn't work for SOME reason   ^^ .... probably  ^^

Most widgets are rectangular in shape ... any reason for this?  Are 
other shapes feasible?

So far, there is not much widgets, which could have different shape, 
no?

Lack of imagination?

edge / hgit testing is the main reason

easier computationally?

much

you have to use alternative methods like colour testing for non-regular 
shapes

well, what can you do with quad core cpus ??

I've watched those multitouch demos ... seem to use irregular shapes

ill dig out my radial menu thingy and update it for r3 and test it 
for speed.. have a old version at http://rebol.mustard.co.nz/readial.tar.gz

http://rebol.mustard.co.nz/radial.tar.gz.. oops

Ok, that is cute

meh - ive done it in assembly before - its rather sad in rebol :(

though I suspect it's not working ... I only see shapes and no text

its 2003 - back before the text in agg changed

ahh...

I'm surprised you were able to find it so quickly!

i downloaded my website today to look for some old sudoku code :P

Silverlight is starting to use radial menus too

My Daylight Map used color to detect irregularly shaped hit spots...

yeah colour testing is fast and maps nicely to functions

Are you still working on your own graphics system for R3?

http://ross-gill.com/r/daylight/map.jpg
http://ross-gill.com/r/daylight/map-zones.png

im waiting for Carl/Cypher to fix some of the view holes.. and also 
for modules and timers and threads :P

:(

view is currently swiss cheese and the underlying Core has some large 
gaps that really need addressing before GUI sees the light of day.. 
IMHO..

the presence of holes suggests some underlying substance

the thing that gets frustrating is when something works fine in earlier 
builds then mysteriously dies in later builds :P

thats a great colour example CHris :)

Pretty, is it not?  : )

its one example where the benefits of colour testing over poly testing 
are clear to see.

Doesn't scale well though...

of course.. zooming and SVG night be another method :)

With AA turned off...

with AA on you can do approximation mapping with tolerances

Sounds like it could get complex - it certainly is a useful pattern 
for many situations, especially when you can use graphics programs 
to develop the maps.

Hmm, isn't jus alpha transparency enough? Why to do any shape detection? 
We just need the ability to define level, which will let events to 
go to underlying faces, just like Amiga DE did it. And if you look 
at Cyphre's irregular window shape, you will see, that it simply 
is not rectangular, or is it?

do http://www.rebol.cz/~cyphre/trans-gui.r

Graham - there are no holes in R3. Most of the stuff is just non 
finalised design. Look at latest changes - absolutly necessary Core 
changes to get load, transcode, modules working. Last two days we 
saw two releases - A38, A39 with bunch of fixes. We are slowly getting 
there.

Cyphre can't do nothing yet, as Carl did some changes to the code, 
and Cyphre's version of R3/View is not compatible anymore. Of course, 
rebin isolation should help, and in month or two, we get there - 
modules are bing worked on right now, plug-ins and rebin are next 
one. Then some source-codes get released finally. In the meantime, 
Carl also updates docs - very important ...

Three releases - we're up to A40 now :)

Do we have a deeper text compositing support in AGG ? (woops...)

by deeper and since the number of font is limited the basic thing 
i need would be to get a something like "get-char-size char font-name 
font-size"

this what in text-draw.r (VID3 sources)
font-char-size?: funct [fstyle] [
	gob: make gob! [offset: 0x0 size: 300x100]
	gob/text: reduce [
		'font fstyle/font

  'para make fstyle/para [wrap?: off] ; avoid infinite loop font bug
		'anti-alias fstyle/anti-alias
		"M"
	]
	size-text gob
]

But i don't use this. Useless memory overhead (by creating a new 
gob) each time.

the gob size is important in the wrapping calculations

So what ?

scan-fontset: func [fname fsize /style fstyle /local tmpfnt font-metrics 
charlist n][
	font-metrics: copy []

 tmpfnt: make system/standard/font compose [name: fname size: fsize 
 style: (either style [fstyle][none])]
	charlist: copy ""
	for n 32 126 1 [append charlist to-char n]
	forall charlist [
		append font-metrics size-text make gob! compose [
			offset: 0x0
			size: (as-pair fsize * 2 fsize * 2)
			text: [font tmpfnt size fsize (to-string first charlist)]
		]
	]
	append font-metrics charlist
	return font-metrics
]

probe scan-fontset "Lucida Sans" 24