World: r3wp

Shouldn't this be catched by the GC?

IMO Looks more like a View(text related) bug.

well, the face that SHOW is trying to show remains in memory, possibly 
with a reference to a parent-face that no longer exists. what exactly 
makes SHOW go boom, I'm not sure.

fix is done by unfocusing before destroying the pane. interesting 
bug.

There is a strange auto-complete (tab key) problem on View 2.7.7 
console,
connection: context [a: 1 b: 2]
test: context [a: 1 b: 2]
when you type conn then press tab key it writes just "connected?"
but its ok for test. no auto complete for "usage" function also.

I've found another tab-complete problem where an error is produced. 
Only seen it when running large apps though.

After loading RebGUI tab-complete gives error for undefined values 
also.

Endo, I see the error in View 2.7.6.4.2 as well.
>> connection: 1
then I type "connect" and press tab --> "connected?"

there's a built-in function "connected?" which lexicographically 
sorts before "connection", so it's completed first

That's the bug. It's supposed to stop the completion after the last 
letter common to all the possibilities.

And Rebol does that - try it with "con" and hit tab twice.  The bug 
is omitting user-defined words..

possibly only does defined while in the console

nope, it seems only *some* words starting with 'CON  are omited. 
  'CONTENT is also ignored in auto-complete... very strange

Clean console?
>> ? content
No information on content (word has no value)

possibly related to binding order on app startup... the words 'connect 
and 'Content are already defined in system/words when you boot up 
REBOL, but they are unset!

the auto-completion doesn't bring up values which are unset!


but its strange that defining them doesn't tell the auto-completion 
to use their "current" value, especially since we are operating in 
the global context.

It's not the fact that it's a user-defined word. It must be something 
else. It looks to me like it's because the word 'Connection (with 
capital 'C') already exists in system/words on startup. Defining 
'connection in the global context doesn't change the capitalisation 
on the existing symbol. So I think the completion function is case-sensitive 
Try in the console:
>> Abacus: 1
>> abalone: 2
>> aba
and press Tab Tab.

hey I think you got it.

How do I set the font of an area to Courier New ???

area with [font [ "Courier New" font-name font-courier]]  ;something 
like this?

area font [name: "Courier New"]

thanks!

Was the 'vectorial argument for text in the draw dialect removed, 
or broken?

It works fine in 2.7.6, but in 2.7.7 it causes rebol to segfault. 
 (linux)

Also, I thought there was a Draw group, but no longer see it.  If 
it was made private, can I be re-added?

There is no private grouop for Draw (unless it has an obscure name 
that I do not recognise).
Perhaps time to start a Draw group?

file a bug report for linux

Um, where?

it works on windows for me.

R2 bugs are still filed on rebol.net

i think that 2.7.7 crashes for most DRAW commands on Linux. at this 
point we have enough reports, it's worth making a ticket.

is there a programmtically solution to 're-center' the toplevel-face 
on the os-desktop after a face reseize?

you can add an event function to the window and trigger it on event/type 
= 'resize and then move the window according to its current size 
versus system/view/screen-size (or some such)

yes, nice idea henrik. this way I don't have to take care for every 
resize. thx... (so the path 'system/view/screen-size' looks don't 
exists, will check)

just found it -> system/view/screen-face/size

couldn't remember, so was betting you could find it yourself :-)

thx henrik ;-)

remember to use remove-event-func with the function again, when the 
window closes, otherwise amusing (depending on your mood) things 
happen.

There is an AGG group, but there may be cause for a new "Draw dialect" 
group.

how to detect  - which face - is envoking an event within a global 
insert-event-func?

do you want the face or the window?

window

so I have to filter the correct event (resize) and object (the window/face) 
to calculate the new pos on the desktop.

something like this...

insert-event-func 	[
    switch event/type [
        key         	[]  
        time        	[]
        close       	[]  
        offset      	[]  
        resize      	[
		switch event/face [
			X [recalculate pos for face x]
			...
			...
		]
	]  
             scroll-line 	[]  
             scroll-page 	[]  
	up 		[]  
	move 		[] 
	down 		[]  
	active 		[]  
	inactive 	[]  
	alt-down	[]  
	alt-up		[]   
	minimize	[]  
	maximize	[]  
    ]
    event
]

the window face name usually works, if you differentiate the names 
of the windows.

??

try probing face/text and see if it matches the window name

resize events only occur for windows... no?

don't know sure but is there no event/object message?

event/face perhaps

windows are passed in the event/face, that is sure... so if resize 
events occur only for windows, you should be pretty safe in assuming 
that in this case the event/face is always a window.

one thing though.  I've discovered that at low-level (view port wake 
event) resizing generates one event per mouse move.   


it sends all of them AFTER you finished resizing (pretty dumb) so 
you may end up with up to a 100 resizing events which are all, basically 
useless except for the last one.


I do not know if they are filtered out within the do-events (and 
thus within the event-func) but you should print out something to 
see if this is the case.