World: r3wp

My first workaround checks if face is focused and does nothing then. 
but when you put something in face, this code should be activated 
on show.

Is it worth working on that, or can you live with current behavior?

hmm.. ive tried to use this rebgui stuff for the 1st time for a real 
problem, but it look buggy. a table cant be focused by tab (in spite 
of append ctx-rebgui/edit/tabbed 'table), a row in a table cant b 
focused by the program in an obvious way. the dropdown list doesnt 
rolls up after dropping down and it doesnt drops down if i click 
on its text.

am i just unlucky?

Graham: fixed it in another way after understanding more code.
Now at end of engage

  face/text-y: second size-text face ; only needed on real change, 
  but..

Which files have you altered?

rebgui-widgets.r and also rebgui-edit.r ?

Yes. But needed is only the line in rebgui-edit.r

I'm not clear on this.

using the original files from rebgui, I need to only make one change 
in rebgui-edit.r ?

Hmm, maybe i changed something else too. IIRC there was some problem 
with offset-to-caret somewhere. The fix to my fix is that one line, 
and defixing my change in %rebgui-widgets.r

Yes, another change in edit too, the
  0x2 + ;? heuristic patch

gee... I'm confused.  I put that line in the engage function after 
the switch and nothing happens.

do you have the complete files available with your fixes?

http://polly.rebol.it/test/test/rebgui/index.r

and the download.r there.

I ran download.r and it started up tour.r    ...was it supposed to 
do that?

Yes.

download.r is no longer the right name.

as my downloads now start a demo.

I note that editing is still broken though :(  if you enter lots 
of CRs in front of the sample text, the slider doesn't change to 
reflect this.

And it does not scroll when caret moves out.

Seems i commented something out?

I don't think it ever scrolled when the caret moved below the area 
margin.

Crossed fingers, but it seems to be largely working now.

I have a version here which does that. but do not know which version.

would be nice somehow to have version control with the editors we 
use.

There where to lines outcommented in %rebgui-edit.r . keeps caret 
visible now. (do not know if that has sideeffects.)

I wonder why Ashley couldn't keep the editor from VID.

He did and extended it, says the header. And made some wrong changes 
maybe. Maybe he got made because of the offset-to-caret-bug.

How do you get a face to come to the forefront? After I've "displayed" 
it and say, go and click on some other window, how do I get the initial 
face to pop back up. In my case, I have a table of data and when 
the user clicks it, I want to have a more detailed view of that line 
of data.

window/changes: 'iactivate ;IIRC

I'll try it. thanks.

Thanks Volker. That, plus a "show" worked.

Display is missing an options refinement.  I added one to 'display, 
and altered the view to this

			view/new/options view-face opts
		][
			view/new view-face
		]

either options [
			view/new/options view-face opts
		][
			view/new view-face
		]

didn't work though - when I tried using it with [ no-border ]

How would one get the layout for a window?  In VID, the layout can 
be assigned to a variable, but in Rebgui, that information is lost.

Would be the parent face of another face. Or patch display with a 
/set-option?

I'm sure it is, but that defeats the purpose of rebgui which is to 
build a complete gui toolkit without having to know the internals.

Had a quick look, the display function creates a local variable view-face 
which stores the layout.

So, if we add a return view-face to the end of the function, we can 
get 'display to now return the layout.

Cool... that works.  I can now get a png of the layout and dump to 
the filing system.

Is that Ashley I see online ??

Yep, just catching up with 5 weeks of snail mail and REBOL3 AltME 
messages ... should take another day.

Just to get you up to speed, I think the major problem to fix is 
the editor in the area.  To fix : cursor keys, and pageup/down behaviour 
especially at boundary, and also cr behaviour.  Highlighting is also 
problematic.

Also, need accessors to set radio buttons, and to move the cursor 
in a table, including selecting mulitple rows (oneTom)

Oh, and request-file needs the /keep refinement restored.

Ashley, have a look at this : http://www.rebol.net/cgi-bin/rambo.r?id=3801&

Seems RT have now fixed this bug.  Hopefully a new SDK will come 
out with the fix as well.

Another problem - RT's request-dir traps when you try to access a 
directory that is not available, but RebGui's crashes.  And it can't 
be trapped with a try block at the request-dir level.