World: r3wp

Thanks.

In 'display, there is this code 

        foreach window view*/screen-face/pane [
 	                if title = window/text [return]
        ]

which should be changed to:

        foreach window view*/screen-face/pane [
 	                if title = window/text [return none]
        ]

so that if you have: 


win: display "title" [] ... an error is not produced as win does 
not have a value

Don't know if this is in the current table widget .. but if you click 
on the header of an empty table

make object! [
    code: 303
    type: 'script
    id: 'expect-arg
    arg1: 'index?
    arg2: 'series
    arg3: [series! port!]
    near: [index? find parent-face/data last-selected]
    where: 'action
]

Looks like this bug is still in current table widget.

Can't reproduce that with the following line:

	table options ["A" left .5 "B" left .5] data []

do you have a simple test case?

I wonder if it's related to the previous problem...with tables.

I have a table populated .. and remove each row.  Once it's empty, 
and I try and sort, I get the error.  But I have not updated to the 
latest table code yet.

Ok, problem is fixed in latest rebgui.

Build#42 (Trac only) has your "return none" fix plus a fix to the 
long-standing tool-bar tooltip bug (tooltip not hiding after tool-bar 
button clicked).

I would like to see an enhancement to tables so that /selected/block 
returns series of blocks.

the reason is that if you are have anonymous tables, the only way 
to work out the length of a row is to take all the selected and divide 
by the length of picked

How about just adding 'cols and 'rows attributes then?

What if rows have variable lengths ?

don't think allowed ...Anton

(Ok, sorry, that was not a well-considered comment of mine.)

Build#45 (Trac only). Includes the following enhancements:

- display now returns face

 - clear-widget uses radio-group/select-item and table/text-list select-row
 - table/rows attribute added to complement table/cols
 - text-list/rows attribute added
 - get-input and set-input funcs added


These later functions allow you to easily get and put values into 
a display or tab-panel, even one containing nested grouping widgets 
such as 'group-box or 'tab-panel. Handles the following input widgets: 
area check check-group drop-list edit-list field group-box password 
radio-group slider tab-panel table text-list.


The /type refinement of get-input is usefull in design/debug mode 
to get a formatted list of widget type/value pairs.

Example code to demonstrate the use of these two functions.

d: display "Test" [
	after 3
	area "area"
	check true
	check-group data ["check" true]
	drop-list data ["drop-list"]
	edit-list data ["edit-list"]
	field "field"
	group-box "group-box" data [
		field "field"
	]
	password "password"
	radio-group data [1 "radio-group"]
	slider data .5
	tab-panel data ["A" [field "Tab-panel"] "B" []]
	table options ["col" left 1.0] data ["row1" "row2"]
	text-list data ["line1" "line2"]
	button [a: get-input d b: get-input/type d halt]
	button [

  put-input d ["Text" true [false] "A" "B" "text" ["xx"] "" 1 1 ["Bob"] 
  [1] [2]]
		show d
	]
]

put-input?

or set-input ?

I'm still open on the names, and will probably add a 'clear-input 
type function to round out the set. I chose the word 'input as both 
'form and 'values implies a superset of input [values]. I also considered 
other action words such as read, write, load and save; but to my 
mind they are so well established in REBOL that they already have 
other meanings (e.g. load-form, like load-image, would imply loading 
a form from disk).

'set is OK except that the logical complement is 'unset.

load-fields?

the ? above is because you introduce it as saying you have implemented 
'set-input and then the example uses 'put-input ....

Ah, missed that. It should be get and put.


load-fields suffers from the dual problems of "load" (as described 
above) and begs the question, "what about other widgets that are 
not fields but except input". We'de have folks looking for load-area, 
load-text-list, etc

fill-fields

Ashley, please take a look at the xpeers RebGUI stuff. We have added 
a lot and should align the code lines.

Resizing: I have a question how to best implement the following:
- I have three tables from left to right

- I want to make a horizontal resizing in that the middle and right 
table are moved and narrowed.
- The left most tables only uses #X
- The moving should be right aligned to the left side table


I hope you understand what I mean. Using #WX screws up the layout 
as the gaps between the table get to big

any idea how I can get the desired resizing effect?

Not exactly sure what you want, something like this:

	display "Test" [
		table #W options ["" left 1.0] data []
		table #X options ["" left 1.0] data []
		table #X options ["" left 1.0] data []
	]

But than the tables 2 and 3 don't change their width, right? So what 
I want is that all three tables get narrowed and are still horizontally 
aligned with the distance given by SPACE 2x2 for example.

Current system does not support more than one resize per axis (i.e. 
you can have one #W per x axis and one #Y per y axis). What you are 
asking for is a form of proportional resizing.

Robert - what are your xpeers additions to RebGUI?

We enhanced a lot of the widgets etc. to much to list here. The problem 
is I don't find the time to document everything and publish it. So, 
if someone want to do it, let me know.

are results even mergable?

I will need to release grid too :-( I want to have it out 1.12 as 
is - just a bit of further documenting needed

It takes time but yes, the changes are mergeable.

Overall I think I have invested about 8000€ into RebGUI so far.

What are you building?

uh, so you use Rebol commercially?

btw - isn't area enhanced to the state, that hiliting text by mouse 
will cause it to scroll?

Robert - any screenshots? I do remember some of your older stuff, 
which looked pretty. I wonder if you use default RebGUI look, or 
you tweked it a bit, to something more pleasantly looking?

area remains a big weakness for me.

Graham, I have build a cost-analysing software. With this it's possible 
to identify cost optimizing potential within the complete supply-chain.

you should sell it to SAP :-)

Yes, it's commercial. The app is supported by hughe benchmark databases. 
I have about 400MB raw data, pressed into 66 MB final database.

Petr, I'm currently the interims purchasing manager for SAP AG region 
Central:-)) So, three days a week I'm in Walldorf.

Let's see... ;-)

:-)

I hope we meet once I am at IBM :-)

Styling: I kept it mostly as is, because of lack of time. It's OK 
but a bit more eye candy would be nice.