World: r3wp

new GC is inserted into 102 version but I francly don't see any différence 
in tour.r with RebGUI 0.22  we start a 6800 Ko and the memory is 
cleaned when 8500 Ko is reached and cleaned  to 7400 Ko ...

(regarding de/allocation) I don't know if it would change something 
in our case: in /View, event objects are generated by user/system. 
The event object carrying mouse 'move information may not be used, 
but one can only know and discard it at the end of the event processing 
chain (a global event-func or a face feel) - could be tricky to do

new GC: it must be various corrections to crashes and speedups. I 
don't know if anything has changed on conditions for garbage collection 
triggering.

I think the memory leak problem with VID  is closely related with 
way data are alocated in general

for example the mouse pointer event is stored into the same var but 
betwin two allocations you have no memory clear so glob vars are 
the same the content is not the same and the only thing that change 
is the memory adress related to the new data allocated

this make increase the memory use until it reach a certain level 
then it's trash maybe it could be bettre to make on new data allocation 
the systematical deallocation of the previous content instead of 
having them surviving into the memory ...

each event is a new object, you can't reuse the same memory place. 
example:

t: copy [] a: insert-event-func [append t event event]
view layout [button "unview" [unview]]
remove-event-func :a
print length? t

foreach e t [probe e/1]

in this case why the previous obj is not trashed those object have 
a short life time this muts be taked in count I think

In all cases we can solve this by patching the show but this force 
the intensive use a recycle witch was  not designed to do ths ...

Maybe recycle could take an argument, which specifies a word to take 
particular note of.

yes this would be a first step  ;) instead of scanning comparing 
the whole memory

but the point is that we don't know exactly how recycle works

does it scan the entire memory ? does it make the comparaso betwin 
the whole memory content table and the actual rebol word content 
?

If rebol keeps track of all previous existing data doesn't it simpliest 
to trash those datas when a next allocation is done ?

Can we move this discussion to the View or Performance group? ;)

sure :)

Hi Ashley and Shadwolf  sorry if i didn't answer before i was out 
of town ( mother's day =P ) let me explained to you why the auto-fill 
on a drop-list ...

If you have a big drop-list ( non editable ) with 60 salesman ( big 
store ) and you shadwolf because of your name are close to the 50th. 
position the way to select your name frome that drop list would be 
with the mouse

wouldn't be easier if you typed sh on the drop-list to select your 
name instead ?

and no i am not talking about an edit-list. And is not an auto-fill 
per se routine. More like a keystroke capture routine so that it 
would do a search on the list  = )

Luisc: usually, drop-lists don't work well with big lists - it's 
quite hard on long lists to select something without mistakely closing 
the list. A classic text-list is far more suited for anything longer 
than 10 elements. A 60 items drop-list will raise a lot of problems: 
if the widget is too low on screen, most of the choices will be hidden.

Unless your droplist widget adds a scrollbar when there are more 
elements than you specify space in the widget for, like in Windows.

the combo-box? yes but it's one of the more complicate, buggy and 
error-prone interface in Windows. it's like a popup textlist associated 
with an editable field. (had painful moments with it in Visual Basic)

I'd have to agree with you if that error-prone complaint is about 
the usability of that kind of control once it's implemented. Still, 
is there a better way of doing what it does? The Mac way perhaps?

Yes about usability. It's a all-in-one widget: one can use it to 
replace menu, lists, simple requesters (like the fonts and colors 
selection in MSWord), toolbars, auto-complete fields, etc. So it's 
mostly a shortcut to put more widgets (or compact them) in a window 
and can be confusing for users: what will happen if I select something, 
write in the field, click outside the list, click inside on the left, 
on the right, hit enter? It's a full requester in one widget: you 
can do the same with a request-list -like requester with ok/cancel 
buttons.

Vincent i understand what you are saying and i totally agree with 
you. It is not a big issue all i am trying to convey is that if you 
have a non-editable field with different choices it would be easier 
for some of us  to select those choices with the keyboard instead 
of the mouse.

when you select the language for windows installation, and the timezone 
etc, they use these very large unwieldly drop lists :(

Oh, I hate that. Even worse, most of the time you can speed up the 
process of selecting from a dropdown by typing the first letter of 
your selection, but all of the time zone selections start with ( 
so that speedup doesn't work with them.

Luisc we have 2 item in fact edit-list to handle large data with 
some useful stuf like list vertical scroller and auto-fill field 
This is what you want. And the drop -list widget that is for handle 
tinny list of data :)

Luisc so what you need is yet planned and it's called editor-list 
not drop-list ;)

ah! thank you shadwolf   = )

your welcome ;)

lastest version of my work on Listview widget is available here http://shadwolf.free.fr/rebgui-list49.r

as shown is the example programmer hav only to give to list-viewdata: 
field the new list content then make a show listview to see the chages 
made on the listview.

I'm going to work some days on MDPWiki Project so during this time 
I want you to make me lot of comments and suggestion on listview 
widget. Once  I'm done with MDPwiki I will retake all the comments 
and suggestions and add them to listview ;)

sugestion can be new widget types, bugs reports etc...

Can we prefix all files required for a valid distribution of RebGUI 
with regbui-* please? Makes adding it to existing frameworks etc. 
much simpler.

Robert normally rebgui files are disigned to be placed into a sub 
sub folder

in your main script you put a do %sub-folder/gui.r and that's all 
;)

sometimes its easier to put all in one folder. you see all at once, 
no need for tools to recurse, and ios uploads more easily.

i would like something short, rg- or such.

How about a merge of all RebGUI files into a single file called %rebgui.r?

for me the packagin way is insignificant ;) must be done has the 
people perfer having several files is more easy to see what is integrated

packaging: I vote for it. What's up with the requestors? Are these 
always needed? If possible we should find a way to pack those things 
together that are always needed and let the rest be optional.

Prefix rg- is OK with me. But remember, people don't know what rg- 
means, and rebgui is only four letters more but hundreds of questions 
less. Never shorten things just to shorten them... semantic is the 
key.

rebgui needs to be easy to use: Just copy, load and use. :-)) That's 
what I like about it.

Now that makes for a really nice slogan, Robert: "only four letters 
more but hundreds of questions less". It's as true as that of the 
well-known picture, which says more than thousand words. Comes down 
to: One word says more than thousand letters ;-)

Regarding scrollers and sliders in RebGUI: I examine a strange behaviour 
here. If you grab a knob near it's edge (left/right for horizontal, 
top/bottom for vertical scrollers) and drag it, you'll notice how 
the mouse pointer's offset to the knob changes relative to the amount 
that the knob changes it's offset from the scrollers boundaries. 
I haven't seen something like that in any of the (not so many) UIs 
I've came across, so I'd suggest that this gets fixed to conform 
to what people are used to.