World: r3wp

http://shadwolf.free.fr/RebGUI-038-shad.zip

http://shadwolf.free.fr/RebGUI-038-shad.zip

REBGUI 0.38 Shadwolf version !!!


- FULL  retail files no need to merge them back  with previous REBUI 
version
- Didec-requestdir mini app 
- Listview as base object  
- Demo-listview.r shows how to use the listview widget.

lol nice... I plan on trying RebGUI for my next little project, but 
never downloaded it before...

yea rebgui is very very neat  it's growing in size but it's still 
pretty  cool  widget library . I recommand it to you as simple as 
Vid but mutch more sophisticated  ^^

field on-focus [system/view/caret: face/text true] on-unfocus [face/text 
= "Exit"]

fixes the exit-field-bug in action-demo

Somehow i dont like the names 'on-focus, 'on-unfocus. They feel loong 
somehow. how about

 field enter [system/view/caret: face/text true] exit [face/text = 
 "Exit"]

Not sure how on-focus works, ut returns a flag? Could it return the 
value for caret?
 field on-focus [face/text] on-unfocus [face/text = "Exit"]

What is the diference between on-unfocus and the usual action? can 
they be joined? WHat do i overlook?

The idea with the on-* naming is to clearly indicate an event. I 
toyed with using just focus / unfocus but these verbs could be confused 
with the functions of the same name. Likewise, 'enter could be confused 
with the 'return keyword and 'exit has the same issue with regards 
to both 'exit and 'quit. 'on-enter and 'on-exit are possibilities 
though.


on-focus and on-unfocus return true or false so as you can handle 
failure. This scenario is more likely with on-unfocus where you would 
want to trap and handle a field validation failure for example.


The difference between on-unfocus and the usual action (both may 
be triggered by pressing enter for instance) is that on-unfocus enables 
you to interrupt the normal work-flow by aborting the unfocus and 
subsequent focus operation - something you could not [easily] do 
normally. Also, for many widgets the usual action and on-* actions 
may be different (e.g. pressing enter to initiate the default action 
may be quite different to pressing tab and leaving the field).


For those familiar with Oracle*Forms, and similar products, these 
new actions are like the pre-field and post-field triggers (with 
the app-level ones mapping to pre-form and post-form).

Good questions all, keep them coming.

http://www.lexicon.net/antonr/rebol/rebgui/add-remove-tab-panels.r

Anton very nice script i have a firend that need this feature for 
one of his school project ^____^

firend = friend

Anton i like to see franch strings in your code ^___^

on-something is good, ashledy, go ahead with it ....

yes it's yet seen by most of w32 QT wXwindows GTK TCL/TK  coders 
 ^^^

DocKimbel used it for Uniserve too ...

focus/unfocus from Ahsley: Sounds very good and flexible to me. I 
like it a lot and yes, it meets the features I have in mind. I'll 
give it a try.

>> change-dir %../
== %/D/rebol/rebgui/

>> do http://www.lexicon.net/antonr/rebol/rebgui/add-remove-tab-panels.r
connecting to: www.lexicon.net
Script: "add/remove tab-panels" (8-Dec-2005)

** Access Error: Cannot open /D/rebol/rebgui/programs/rebgui/rebgui.r
** Near: do view-root/../programs/rebgui/rebgui.r
relayout: has [fenetre-cannibalisee]

table: Moving the last slider changes the space between draging-bar 
and text "Char" dynamically. Looks like a relativ calculation thing.

action demo: very cool. I'm now testing it.

Table: the labels assume the same alignment as their contents (i.e. 
the "Char" column is centered as is its label)

I've been looking at alternate UI / icon resources recently and stumbled 
across Tango - the standards they propose might just be what RebGUI 
needs. Take a look at:

	http://tango-project.org/Tango_Icon_Library
	http://www.tigert.com/archives/2005/11/16/tango-ui-for-web/


and let me know what you think about making RebGUI tango complient.

That first link should have been: http://tango-project.org/Tango_Desktop_Project

Shadwolf, that is not a coincidence, I think !

Your friend emailed me personally, so the script is for him, but 
also maybe for others here. :)

Tango: Very interesting stuff, and looks like these guys thought 
a lot about doing it right. I vote for using it for RebGUI. Frees 
us to do invent the wheel again.

I think Tango is nice, but I really much would appreciate or like 
to sea some standard way to add a textual description to icons, like 
often used nowadays. I know almost nobody will want to give up icons 
or graphical objects, but in order to make RebGUI as good as possible 
I think to add such a textual description by design would be very 
very good.

I think of something like another parameter to a button in which 
the label is placed. The UI could then offer a simple way to activate 
or deactivate the labels, the icons or have both, side by side or 
on top or below. 

This way we get the best of two worlds (only some space needed) - 
people can choose for themself and with the text it will be almost 
always clear what a button means.

Of course the could be done externally by adding a second row or 
column to every button, but this would separate what belongs together.
What do you guys - especially Ashley think ?

Anton you have to set do %rebgui.r instead of  do  /D/rebol/rebgui/programs/rebgui/rebgui.r

Anton yea but i wasn't aare of that  ^^

aare = 	aware

now  i will rework on rkini my irc client and base it on RebGUI

Shadwolf, that makes the bug simpler, but does not fix it, either 
:)

I can't locate the user's installation of rebgui, unfortunately.

(if it was in the public cache, I would have a chance).

Is there a way of selecting a specific tab using shortcut keys ?

Anton, most of us assume that rebgui.r is in the current directory 
as in Ashley's demos.

ok, I've fixed it.

Works well ...

<devil's advocate>With a project like RebGUI I would expect to be 
able to find a URL that I could use to execute the source within 
my application.  Preferably the link would include a COMPATIBILITY 
number so that any updates to RebGUI would be executed seemlessly 
using 'load-thru but if at any point an update is known to break 
compatibility with a previous version then it's downloaded from a 
different URL and hence wouldn't screw with my application.  We have 
to be able to leverage the infrastructure of the internet in our 
applications.  Having to load the application on the user's computer 
means that I can't use RebGUI for anything other than an encapped 
application because I don't trust my users to install the RebGUI 
scripts on their system correctly and sorry, but I don't want to 
have to include so many files in my application and I'm not going 
to take the time to copy them all into my application either.</devil's 
advocate>

Remember that RebGUI is still alpha and we probably want to incorporate 
Reb/Services at some point ... but apart from that an excellent suggestion. 
If you, or someone else, writes the necessary code (or pseudo-code) 
post it here and I'll include it in the next build. ;)

Does the on-focus and on-unfocus work for tab-panels as well? Didn't 
tested it.

One file - should be easy. Use prebol. Like encap: you get a single 
file. but that file is text.

That file you can do in your application. rebservice-lib seems to 
use that, + compression.

You can also use "-t source.r" in encap, to get source as by-product)

default empty value: How about a way to specify a default empty-value 
for a field? For example: I would like to set  0.0 if a user deleted 
all text in the field. So being able to specify a default value, 
that get's set if the field is left without anything entered would 
help a lot.

Ammon, that's not playing devil's advocate. That's simply proposing 
new functionality. Devil's advocate is just taking an opposite, usually 
negative, view, and trying to sink someone's idea. If you mean, however, 
just that you're advocating something, but you're being a devil by 
not providing any code, then ok :)

But I think the idea is a good one.

... but if at any point an update is known to break compatibility 
with a previous version...

   <-- this is difficult to know, to track accurately etc... it becomes 
   a new problem to be solved. I'm searching for a solution...

Maybe this is what  is wanted: