AltME groups: search

Another way of achieving the same effect is to simply populate %locale.dat 
with the required language/domain specific substitutions and let 
layout do its thing. e.g.:

	words: [
		"AAA" "My-AAA"
		"BBB" "My-BBB"
	]

then:

	drop-list data ["AAA" "BBB"]


I just don't see how changing radio-group and drop-list [only] solves 
anything. What about text-list, table, drop-list or any other widget 
that display's a value or values?

Ashley, for tables I have always a hidden column with a record ID. 
This ID is used through all the program.


For drop-lists I "store" the text of drop-list/picked at the moment. 
Hence if you have several languages, at one time you store the DE 
version at an other one the EN version. So, one solution is, I have 
to translate to the "master" language that's used in the database.


But what if syntex changes? Than I have to update all old entries. 
Using IDs frees me from this but adds an other level of indirection. 
The common way would be to normalize my database model and use other 
tables to get IDs, and than store those etc. IMO that's the way to 
make apps complex and slow.

It's two things: position and text independend (which solve I18N 
as well) storage.

I am not e.g. sure I would use native RebGUI localisation, because 
it goes from messages, UI elements, to of course data. And as for 
data, you have to establish the concept yourself.

but above would mean, that every translation for all tables would 
be in one 'trans table - not sure it is practical. Another model 
- simply have langs table, as in above example:

langs: [cz "Cesky" us "English" de "German"]
fruits: [
1 uk "apples"
1 cz "apples"
2 uk "bananas"
2 cz "banany"
3 uk "cherries"
3 cz "tresne"
]

select * from fruits where language = 'cz'


That is more practical - simply use one more key - that of your default 
or selected (preferred) language ...


In some sense, you don't need IDs, but I prefer to have one- it depends 
what kind of databases you use. But - it is surely nearly INSANE 
to use text, which can eventually change, as a key value (e.g. because 
of typo and later need to change that text)

want to get some text / labels in 
Arial Narrow""

	text "Some Narrow Text" font [name: "Arial Narrow"]

assuming you're on WIndows and that font is available of course.

I18N: We have written an own I18N system, that can handle all UI 
aspects dynamically. Table columns, drop-list data etc. Works very 
well. And it can be used not only inside GUI stuff but through the 
whole code.

For example you just specify the "master-language" drop-list data, 
provide a translation and it will change on the fly.

There are screenshots of a RebGUI and RebDB app posted at
http://peoplecards.ca/rebol/vidpics

Thanks.  This is posted more to support Ashley, and show off RebGUI. 
 I'm a big fan of
RebGUI.  :)

with SDK code? or rebgui and other public includes?

Thanks for the link, I've updated the RebGUI home page with a new 
"SItes using RebGUI" section (and removed the long-dead link to this 
AltME group as we don't seem to be broadcasting anymore ;(  )

We do need a directory selector that can access directories across 
the network.  At present i have to resort to using request-file,and 
selecting a file in the target directory.

directory selector that can access directories across the network

 We're limited to what "read %/" can pick up. Real solution is to 
 have a native request-dir func (and request-color, request-font).

background colour to an info field
 ... how about making it the same color as tooltips?

	field ctx-rebgui/colors/tooltip-fill

Any ideas
 ... looks like a RebGUI bug as I can reproduce it with:

	display "" [
		box tip "Some text"
		button [alert "Text"]
	]


Hmm, also noticed that tooltip and tooltip-time are dynamically escaping 
to the global context. Will look at this later today.

Build#94 uploaded to SVN. Contains minor tooltip, drop-list, request-menu, 
on-click and on-alt-click fixes.

Dave, easiest way to grab the latest SVN build (without using any 
client s/w) is via a simple REBOL script like this:


 write %rebgui.r read http://trac.geekisp.com/rebgui/browser/rebgui.r?format=raw


%rebgui.r (the complete merged distribution sans %tour.r) is now 
regenerated and uploaded every build.

Hmm, looks like the caret problem is confined to password, and the 
"current word" problem seems to have been fixed. This makes it a 
*lot* easier to track down now.

Is there a way to make the scroll automatically go down when adding 
text to an area?  I have an area I'm using as a status terminal window 
and I can't see the feedback until the task is done because the scroll 
stays at the top.

Scrolling is achieved by modifying face/para/scroll, where FACE is 
your AREA face.

I am sure you can pick the parts you need from the above code, to 
create a nice little function which only appends text.
Note also the definition of LINEY, from further up in EDIT-TEXT: 
        textinfo face line-info 0 
        liney: line-info/size/y 
This depends on LINE-INFO, which is in system/view/line-info

So you will need to either bind your code to system/view and ctx-text, 
or refer to line-info and unlight-text using paths, eg:
	system/view/line-info
	ctx-text/unlight-text

btw - there is some area style, which scrolls when you hilite-by-mouse. 
It is for VID only and comes from french community, at least that 
is what Bobik said to me ...

I'm still having area issues.  If I reach the bottom of an area widget 
and hit return at the end of a line, the caret will often jump to 
the beginning of that line instead of making a new line.

on vid i attacgh a scroller and set that to 1.0 . maybe that works 
with rebgui too?

Hmm.. I've gone over all the GUIs and I've come to an observation 
(not a criticism)... 
There's no graphic designers here.

After looking at all of the various UI for Rebol that's currently 
available, and although they are all great in their own right, I 
can't help but think there's something missing.  I've found the rebol 
mashups I've done using Javascript, HTML, CSS and Flash.. to be so 
much more flexible, and of a higher quality... perhaps not in things 
like speed etc., but when it comes to eye-candy, View, and all of 
it's GUI libs come in third at best.


If you're running windows.. take a look at RASH.. a simple blue button 
that floats on top of all your other windows.. click it, and a panel 
pops up.. 

It uses Flash as the UI.. and it's really quite easy to tie Rebol 
into Flash..  and you can do it with core. Why there isn't more Flash 
/ Rebol mashups is weird

when it comes to eye-candy, View, and all of it's GUI libs come in 
third at best

 ... I'd exclude View from that statement. With AGG support built-in 
 it enables you to be as creative as you want; the fact that no-one 
 has built a nice Konfabulator-like dialect that uses View to its 
 full potential is another issue.

Ashley have you been able to reproduce the boundary issues with the 
area widget .. basically when trying to type into the widget at the 
bottom, and either have the caret fold back to the start of the same 
line, back into the word you're typing.

able to reproduce the boundary issues

 No, and I use area everywhere in my apps. Never come across this 
 problem before. What's the sequence of actions? i.e. it's gotta be 
 more complex than, "just type until it happens". I suspect you've 
 hit a nasty interplay between arrow/slider scrolling and caret positioning. 
 Fixable if I know how to reproduce it.

 the fact that Konfabulator-like dialect that uses View to its full 
 potential is another issue.


did you look at (short on time, cause everyone else was late ;-) 
 elixir devcon session ?


that just what it is...  100% AGG gui engine... with full AGG element 
based interaction and per stroke feel.

Ashley, I can't reliably reproduce this boundary issue.  How about 
putting in a check for when the caret moves to the next line after 
<enter>, to ensure it is not on the same line, and then writing this 
to the rebgui.log ?

graham... darn... should be putting time on elixir and liquid next 
week *I think*  but with all the things going on in my life right 
now I can't be sure.

effects/radius: 0 for group-boxes?
 ... effects/radius is global. It is currently used by:

	group-box
	panel
	tab-panel
	tooltip

and can be set as follows:

	ctx-rebgui/effects/radius: 0

Is there a way to make a horizontal scroll bar on a table?

 No, nor with text-list or area. Same issue with tab-panel. This is 
 quite deliberate and you may wish to refer to my comments from 10th 
 March this year headed, "A word on my design philosophy, ...", to 
 see why. ;)

and links active?

I'm of the "don't tell me about it unless I can do it camp" here. 
Nothing worse than hunting through a complex menu to work out what 
I can and can't do. 80% of what I want to do or see should be no 
more than one click away IMHO.

maybe we could hack in disable/enable functions? I did some, and 
Anton (or Volker) fixed it to work even with keyboard. The solution 
was to put transparent face with some effect into front of "disabled" 
element ...

table: We have extended TABLE to allow horizontal scrolling. You 
specify the number of visible columns and the rest can be scrolled.

how about just adding a disable flag to the button and put a disable 
check in the engage event?

perhaps better to remove them both visually and functionally?

Well, I am going to try it both ways and see what the users say

in lists and tables, scroll bars appear dynamicallly

New error for me ..


An error has occurred.  If this occurred with an upgrade, please 
revert to the older version for the meantime, and report the error 
message as documented below.

make object! [
    code: 501
    type: 'access
    id: 'not-open
    arg1: "Port"
    arg2: none
    arg3: none
    near: [repeat i p/cols [
            line/pane/:i/offset/x: col-offset 
            line/pane/:i/size/x: p/widths/:i - sizes/cell 
            all [
                p/pane/2/show? 
                i = p/cols 

                line/pane/:i/size/x: line/pane/:i/size/x + (p/size/x - p/pane/2/size/x 
                - (line/pane/:i/offset/x + line/pane/:i/size/x))
            ] 

            line/pane/:i/text: replace/all form pick p/data index - 1 * cols 
            + i "^/" "¶" 

            line/pane/:i/font/color: either find p/options 'no-action [
                black
            ] [
                either find picked index [white] [black]
            ] 
            col-offset: col-offset + pick widths i
        ]]
    where: 'confirm
]

Well, that looks like the face-iterator logic (used by table and 
text-list) ... any idea what was being done at the time (i.e. table 
just about to be displayed, row double-clicked on, etc)?

Build#95 uploaded to SVN. Contains new 'on keyword and request-char 
function. 'on is used to specify multiple action handlers at once, 
as in:

	display "" [
		field
		field on [
			scroll [print scroll/y]
			focus [print "Got here!"]
		]
	]

%tour.r updated to demonstrate new request-char function.

I was looking at (and don't grasp all the implications of) the parsing 
 set-word!  occurances and how a set-path! may be inserted.  But...it 
looked too simple, so it had to be wrong.  :)  And again, I don't 
get all the implications, so I'm completely accepting of Ashley's 
nice short no.

so I can add traps around them and then log them to the rebgui.log

Are these errors you're generating or your users are generating? 
I've got a dozen or so users who've been pounding away at a large 
RebGUI app for almost 2 years now, and the rebgui.log files on their 
PC's are pretty much empty. Might just be a case of YMMV ;)

1.3.61 is actually older than View 1.3.2
I've renamed mine to:
rebview1.3.061.3.1  and  rebview1.3.2.3.1
so that they sort properly, by release.

Unfortunatelly, before it was fully finished, RebGUI got some api 
changes, and I am not sure how long would it take to adapt.

and you can find it in the widgets folder. It is just not included 
in the demo.

and assign them a width of .001 which does the trick.

I found grid.r on trac.geekisp. How do I download the script from 
here? I've tried to cut and paste but hte line numbers come along 
with it. And how do I use it with rebgui?

download - we all use TortoiseSVN, which you can install as an Explorer 
extension to Windows. Then via right mouse button context menu you 
choose -update, and point it to svn archive - you can find link somewhere 
on rebgui site.

and as I said - grid.r will not work imo. There were some changes 
in rebgui and someone would have to adapt it.

I will check if I have such one, and send you email eventually. I 
switched jobs inbetween, so dunno if I have it here.

I'm running he test app with grid.r and rebgui in same directory. 
Do  I need to "Do" grid.r?

Regarding my boundary issues with rebgui and the area widget, this 
is what happens


If I am typing at the bottom of the screen and the scroller bar is 
shorter than the scroller height, and the bar is placed up against 
the top, then entering a return places the caret at the head of the 
current line instead of at a new line.

If I move the scroll bar to the bottom of the scroller widget, <return> 
then it does move to a new line.

There are no problems if the scroll bar occupies the whole of the 
scroller widget.

re: boundary issue. Is this the same "issue" as:

	display " [
		area "a^/b^/c^/d^/e^/f^/g^/h^/i^/j^/k"
	]


Then click on the first line to position the caret next to the "a". 
Now drag the scroller down and observe how the caret position changes 
to remain visible. Is this behaviour correct? (BTW, an answer of 
"No" raises a whole bunch of other issues).

I can only reproduce the behaviour you describe (boundary scrolling) 
with the test case posted above ... and in that case it's a "feature" 
not a bug. But the situation you describe is different in one important 
respect ... it doesn't insert a newline. Do you have a simple test 
case which demonstrates this? (It might be a combination of a certain 
area size, font-size and number of lines).

Now paste the sentence you wrote above 7 x into the area box ...the 
caret stays between the n and a in combination.  The scroll is not 
at the bottom.  Move the caret down one by using the arrow keys, 
and hit enter.  The caret moves to the beginning of the line instead 
of creating a new line.

After I arrow down and hit enter I get a newline as follows:

demonstrates ...
|area ...
area ...


RebGUI build#95 with default ctx-rebgui/behaviors (I deleted %ui.dat 
first to be sure).

tried again with build #95 - and problem still is present.

If you couldn't see it, the server got reassigned another IP address 
and so port forwarding failed :(

I downloaded it (476Kb) but under both Quicktime and VLC it appears 
as a large gray rectangular area for about 11 seconds!?

No dice. Anyone else able to download and view it?

and Brian, are you able to reproduce the bug demonstrated?

Ah, got it! I tried it again but without inserting a return after 
each copy and I see the problem (finally). It's actually inserting 
a return at the tail of the text then repositioning the caret to 
the start of the next line. The 2nd half of that operation is correct. 
Will debug this soon.

I used nVidia cards exclusively, hating ATI. So, I had defined accelerator 
keys for moving windows from one monitor to other, typicall ctrl 
alt num1 (or num2). That worked with AltME and other View based apps. 
I have also some Win32 wrappers to do so - resizing, window z-axis 
positioning, setting window position, etc.

Uploaded area caret positioning fix to SVN. "Fix" was to comment 
out Anton's code in the area widget that kept the caret inside the 
visible part of the area. This changes the behavior of area and may 
have other unforeseen consequences (i.e. fixing one issue may have 
created another).

Feasible to be able to double click on a spell choice , have that 
replace the current highlighted word and then automatically move 
on to the next word?

I had a try but it momentarily highlights the next word and then 
the highlight vanishes

Yes, just add a "show face/parent-face" after the set-values and 
it'll work fine.

and there are no actions  associated with the edit-list.

Looks like its been triggered by a double click. I don't think edit/drop 
list handled those... but they usually don't need to as the first 
click of a double click sequence is enough to select an item and 
close the popup. Odd.

Question: How hard would it be to write a LAYOUT function that will 
transform the RebGUI layout into a HTML page?


Of course for all the action code etc. we need to find a nice way 
(but won't be priority 1 for me). The idea is to have a way to get 
my RebGUI apps to the internet.


Of course if we would have a Rebol plug-in this would solve it a 
bit but I don't expect this within the next 6 months. And, plug-in 
will not fit all user needs.

Can you transform to Curl 5.0 instead ( http://www.curl.com) ?  
Or write a Curl macro to start  the task ?  I have a few curl notes 
at eclectic-pencil and eclectic-pencil.blogspot.com  ( I work in 
Curl at a s/w co.  servicing web browser needs of F500 + int'l ) 
    The macros  {example    and   {x-example-ref    are the coolest 
    Curl is kinda link Scheme+ Tcl too ... if you have looked at 
Groovy HTML-builder then Curl will look familiar.  The REGEXP is 
Perl-like    Curl is a web-content language in the functional-mold 
rather like a JavaScript with its own RTE.   There is also TiScript 
 ( www.terrainformatica.com  ) and of course XUL for Mozilla ....

Curl: Do you mean Curl as basic system or transforming to Curl? IIRC 
there system is quite fat and requires special servers etc. They 
have a complicated licensing scheme etc.

So, I agree with Petr, why use Curl if we have Rebol?

It's just an HTML alternative ( HTML+CSS+JavaScript )

I was once an OS.2 Zealot and a Prolog Zealot.  I love Rebol, but 
it will not make me either a Rebol bigot or a Rebol zealot.  Even 
a good advocate should not come across as either ...  Curl is not 
server-side; it is just client-side.  If what you have is a 'layout', 
why transform to HTML just to get a page into a browser while waiting 
(6+ mo)  for a Rebol plug-in ?  Curl is a web content language.  
It's good for that.  And it has great sliding panes and all just 
like scriptaculous but a s simple and consistent as good Rebol code. 
 Think of it as Scheme/Dylan for web pages instesad of HTML.

RobertS - just don't try to suggest we are REBOL zealots. Maybe I 
was Amiga zealot, but that time is long gone. OTOH I strictly refuse 
to be shy of REBOL anymore? What does Curl gives me? It is just another 
technology. I was open about Curl, or was it other technology? And 
it was Robert who suggested that the thing is slow, and that his 
fried had to restart the project because of that.

Curl and client side? Wait, is that a plug-in, or what? I don't want 
another plugin. I accept already existant plugins, or the technologies 
I like, and it is REBOL. So, ROBERT is looking for the REBOL to browser 
converter. And it is news to me the world of web browsers supports 
Curl out of the box. It does not. And that is exactly the reason 
why Curl is absolutly zero solution here ...

And I guarantee you,that REBOL plug-in, if done right, is a killer 
product. Let's show that html monkeys trying to do stupid grids slow 
as mollasses that REBOL can be light years faster. Web should never 
get past static page displays :-)

Then there are the slow-as-molasses industries that replaced 3270 
with PC's just in time to discover 'thin client' as the 'new terminal'.

Diskless-Network-PC? Give me 6 months of high fashion and the 'new 
black' any day.  Client app? Pass the X-server wlll ya ....  ;-)

How hard would it be to write a LAYOUT function that will transform 
the RebGUI layout into a HTML page?

 Dynamic or static conversion? I think it's possible to map most VID/RebGUI 
 styles/widgets to CSS and HTML with Javascript required for a few 
 of the more complex widgets; so reproducing simple layout forms online 
 is trivial, more complex apps with tab-panels, etc a whole lot harder.


This reminds me of some of the R&D work I was doing with IBM back 
in the early 90's. Layouts were IODEFs (Input/Output definitions) 
where you hooked a data source (e.g. DB, Terminal, Webpage) to an 
output target (e.g. Screen, printer, Webpage) with zero application/logic 
changes. The entire app was stored in a code repository across a 
couple of simpe DB2 tables. Anyway, I digress.

HTML: Doing static layouts would be OK in the first step.


And, this declarative stuff like hooking input-streams, to output-devices 
is IMO a good aproach. A lot of technologies from the mid 80s / erlay 
90s are possible these days. IMO there exist a lot of forgotten things 
that are pretty cool.

Graham;  I'm being facetious; but who, other than you, writes apps 
that need that many tabs.  You're adding too many features.  :)  
More seriously.  I've hacked an overlay widget by pulling the tab 
pane out of  Ashley's tab-panel (build 95)  that uses the same layout 
model and his select-tab action.  Put up a couple of arrows and you 
can stack as many tabs in an overlay space as you'd like (they would 
not scroll one at a time; it'd function as banks of tabs; one tab-panel 
per overlay "tab" but it'll save a fair amount of screen real estate). 
 I could post the code somewhere if you'd like.

Multiple lines of tabs are sometimes solved with just one line, and 
"scrollable" tab panel style - you get arrows on the left and on 
the right side, so you can have as many tabs as you wish. Of course 
pushing users to scroll is not very much intuitive, but it solves 
the case ...

Code at http://peoplecards.ca/rebol/ld2-rebgui.rUse it just like 
a tab-panel.  I use the text that would go  on a tab for a comment. 
 Use show widget/select-tab n  to switch between overlays.  The code 
 conditionally loads rebgui and the widget; so if you are modifiying 
anything, you'll have to unset ctx-rebgui for each test load.

Yes, windows ... and a mouse.  Speed is set to just to the right 
of middle.

and yes, it's not easily reproducible

How does one reset a radio-group to it's unselected state?  I tried 
rg/init, and rg/select 0 ( which crashes )

No, the value of face/picked contains the current selection and is 
set by clicking a radio button or using the select-item function 
which simulates a click. By default it is not possible to simulate 
a null click.

display "" [ radio-group data [ "one" "two" ] [set-focus fld ] fld: 
field ] do-events

shows the focus and then loses it

That's not a trivial thing to do with the current design ... and 
then add the complexity of action types (single, double or alt-clicks; 
combined with various on-* triggers ... it gets messy very quickly).

I guess I should not put complicated code inside the action block 
... and just factor it all out.

what about refreshing the table data and keeping the existing sort 
in place?

Build#101 uploaded with fix to caret problem. All widgets defined 
in behaviors/caret now focus on event/offset and only default to 
tail if focus change was not caused by a mouse click (i.e. tab or 
a manual set-focus).

and have it restored when we clear the field

It's getting a bit awkward now .. I have to set the image at startup, 
restore it on an unfocus event, and then restore it again if I clear 
the field.

I think I need to store the image in the field itself .. perhaps 
as a new type of field widget

And a related question.. Is it possible to resize the window when 
my picture is changed, so that the new picture doesn't get resized 
to the previous one's size?

Well, RebGUI allows for widget design but that can seem complicated 
at first.  Another method is judicious use of the span attributes 
and surrounding widgets with box elements set up like elastic widgets 
that will take/give all the space during a resize.


Or wait till Ashley reads this and responds as he always has the 
best and simplest advice.  :)

Ashley, we talked about this before.  I  am looking for a way to 
spell check in a popup or static window attached to the area that 
is being checked, and to double click on the choice, have it replace 
the highlighted text, and then move on to the next text which is 
then highlighted.  I got it working except for the part of highlighting 
the next word that is suspect.  It highlights and then immediately 
loses the highlight .. perhaps something to do with the double click 
and focus system?

I remember looking at this and concluding that ctx-edit.r would have 
to be changed rather dramatically to allow this ... it has to do 
with the order in which set-focus and show are called. Several minor 
hacks I experimented with (mainly in set-focus) just didn't cut it. 
I put it in the "too hard" basket but may revisit this now that the 
long-running (for me at least) pick-list bug has been fixed; and 
I also have a few users asking about this as well.