World: r3wp

well, as for RebGUI, it is now nearly feature complete (in regards 
to R2 capabilities/limitations), so I don't want to open whole lot 
of possible new issues. As for R3, I propose graphicians contest 
to bring mockupus of possible R3 UI design. E.g. today I read that 
SkyOS got its design from one famous company ...

There is something that people that are making frameworks on the 
web are doing that I actually really like.


They hide all the UI until your mouse is over it.  What I like about 
this is that it reduces the noise of buttons you know are there, 
like [Close], [Minimize], [Resize].  If all windows conform, and 
you can truly trust them all to allow you to do everything, then 
I think this will become the preferred state.

But it only works if the UI is 100% consistent.

We are going to offer this as an option in Quilt.

http://www.hmkdesign.dk/rebol/tab-view/docs/example.png

Something I did a while ago

the button bitmaps are available in the parent directory: http://www.hmkdesign.dk/rebol/tab-view/tab-images.r

Feel free to grab them.

Cool.  Nicely done.

Perhaps buttons need a "touch" flag so that the side that touches 
another button takes of the rounded edges.

of = off.

yes, or simply ability to define what each edge to edge/corner to 
be round or not.  Automatic would be great though.

Yeah, automatic should prob be default.

I think we should nail the functionality down before worrying too 
much about the cosmetics.  Stack overflows and other issues are far 
more important to users.

For me, function is beauty

And I gather Windows Vista is going to bring a whole new heap of 
problems trying to match their new GUI

I think that can't be good for people with vision problems. I say 
stick to the simple old ways.

of rotating guis?

I was referring to Reichart's "hide-until-mouse-over" UI suggestion.

Oh...

Graham - you have still some problems with RebGUI non-stability?

Yes :(

but only since I incorporated Cyphre's radio-button

I seen stack overflow crashes due to infinite recursion

happens once a day now

ok, and when you remove it? goes it stable once again?

never saw it before... so haven't removed it yet.

had to make lots of changes in the source to accomodate the new way 
of getting data viz rg/picked and rg/selected

instead of get in rg 'data

what was wrong with get in rg data way? well, never mind, whatever 
.... should I ask Cyphre to look into it?

can you isolate the crash in some shorter example?

%radio-group.r ?

does it crash in all usage cases? or do you have any short code example?

IIRC Graham said on tabbing out, but hard to reproduce.

And his error-message said in "into-widget". Thats in %rebgui-edit.r 
, looks for a tabbable face and is recursive.

it's unpredictable.  Other errors in rebgui I've been able to reproduce 
with code examples.  Not this one.

A little bug is in the foreach. A pane-block can still conatin functions 
(for lists), and that is not checked.

But that should be more predictable IMHO.

and radio-button would need to usethat trick.

the smallest screen I've seen it on is with one radio-group and 2 
fields and 3 buttons

Would have to use what trick Volker?

put a function in a block in a pane.

for an iterated face.

usually the function is put in face/pane. but "face/pane: reduce[:iter-func]" 
works too AFAIK., that would crash 'into-widget because "tabbed? 
child" would now call it.

nothing fancy in my screen

but thats a theoretical possibility. the code does not look like 
it.

that's not Ashley's code ..

that was added by someone else to allow tabbing out of group-boxes

is the error always in 'into-widget ? Or could it be indirect recursion 
and we have to look somewhere else too?

its not the cvs-version?

svn..

Isn't it?