AltME groups: search

Is there a word similar to 'focus ?

after a couple of clears, the text in the field is no longer visible

Luc: dragger size is controlled by ratio, so just do a "sld/ratio: 
a / b show sld" or similar in your code.

Not with a pre-1.3 SDK. %rebgui-widgets.r uses some 1.3 specific 
words such as case, unless, etc. Once it starts making greater use 
of AGG-specific draw commands there will be no going back.

Pity, as a 1.3 SDK timeline has not been announced yet

I think the default behaviour for a tab within an edit-list should 
be to leave the list, and not insert a tab.

If the memory is allocated (in local variables) then you can't. Cos 
it is still in use.
Otherwise, try recycle/torture


But first, try to work out what stats is saying (cos I don't understand 
it) --- look at this:
  stats

  longstring: form now/precise loop 20 [append longstring random longstring]
  stats

Longstring is some 25meg long and not very compressible , but the 
2nd stats returns (for me) a slighly larger number than the 1st time.


Stats has some refinements now -- try help stats -- they might be 
more useful for tracking memory.

And maybe switch to core group for a general discussion.

i guess you had a lot of garbage already? this is from a fresh console.

Graham: If you want to reduce RebGUI's memory footprint just rename 
%dictionary/American.dat ... that alone accounts for more than half 
the memory footprint. ;)

Interestingly enough, adding:

	button "Clear" [clear system/locale/dict]

or

	button "Clear" [system/locale/dict: none]


to %tour.r have lesser and differing effects (from a stats / recycle 
perspective).

Gabriele -- no I was working from a fresh console.

Suggests the new GC is a bit odd, or platform dependant in some ways.

graham the bug with many password fields is undertable that's because 
it's designed in a field + password field window (like login requestor 
....)

As long as it can be fixed.  I have a screen I want the user to enter 
the password twice :)

Ashley, you might want to change this page http://www.dobeash.com/it/rebgui/
and remove the reference to it working with the sdk ( until RT releases 
a 1.3 compatible SDK ).

I'm currently developing a load-svg-icon function that will take 
an SVG file name and size, and return a scaled draw block. Some questions 
need addressing first though (feel free to post the GPL answers to 
the licensing group):


1) Can I distribute a set of GPL SVG icons with the RebGUI zip bundle? 
If so, under what terms / conditions?

2) If someone creates a RebGUI binary application (with the to be 
released SDK) can they include GPLed SVG files, and if so, under 
what terms / conditions?

3) Can GPLed SVG files be distributed under a different file name(s)? 
(e.g. rename gnome-gtk-folder-icon3.svg to open.svg)

4) What if I want to take a set of GPLed SVG files and distribute 
them as pre-converted draw-blocks (so RebGUI apps don't need to convert 
icons from SVG to Draw)?

Now some design questions:


1) Should the result of the proposed load-svg-icon be a draw block 
or an image?

2) Should RebGUI have a stock image cache or a stock draw-cmds cache? 
(i.e. is it better to cache draw cmds or images)

I hope so could be topo a starter point for a more sophisticate and 
polished library of widget to reploace default VID one

Hum GPL things must comes with GPL  other things warning authors 
is a good idea too (but some times they are so many that contacting 
eatch of them is hard ;))

for 2) it dependes mainly of the licence type of the binary ... GPL 
 is free to use/distribute modify until you stay in the GPL licence 
area ... And you keep original file names and original author name 
and you warn author that you retake there work and include it into 
an other work they can see to a gived location (it's most a way for 
main author to ensure that there rights and licence term is not derrived 
to an auther licence enclosement)

thre is nothing worst than spending time on a software put it to 
free offer and see an big company commercialise it under a changed 
name and claiming the author and the enteriority of there work ... 
We have to keep in mind that when Richard Mary Stallman makes the 
GPL  he was yet previsionning the fact that the informatic industry 
will lead later or sooner to the concentric monopol and to the software 
patent. So for him the only way to save the will, rights and work 
of a free sofware author is to impose a strict licence to enclose 
the area...

Imagine tomorrow REBOL  becomes open source (GPL  doesn't means that 
RT will abandon their author rights far of that ...) the Sun retakes 
the code of rebol  (Hum they are not in this mood I know I say Sun 
to not have to say microsoft or APPLE) change the names and the copyrights 
/auythor reference to their own and publish a RAVA (clone of rebol) 
and most of all they patent it forbiding to RT to exploit any more 
their beloved REBOL ... RT to claim his rights and make iprohibitate 
the RAVA will have to make a trail and that 's exactly the weak point 
where the big company thinks they are propotent ... But as far as 
you can prouve the anteriority of your creation and the code retake 
you are ensured to win the trail ... But then it depends how many 
justice action they will intent If the bg company thinks the  technologie 
is prior for his futur incoms they can go very far ...

So to resume GPL is an enclosement for the free software domain creation 
and prevent people to claim the paternity of a freee software that 
not belongs to them ;)

th only person that can split the licence for a project is teh author 
it self and one import thing is to publish the work ... that way 
your GPL  work will gain public notority the date of publication 
into a tiers site like sourceforge or rebol.org can be recieved by 
the court as proof of the anteriority of your work upon the derivation 
;)

4) What if I want to take a set of GPLed SVG files and distribute 
them as pre-converted draw-blocks (so RebGUI apps don't need to convert 
icons from SVG to Draw)?  Normally the LGPL force you to keep the 
 original names and author reference but you can warn the author 
and ask them a special permission to not repect this term... But 
then you obviously have to mention it with a text like " Thank to 
the authors of GTK for allowing us to exploit their SVG icon set 
and allow us to repack and interger it to the rebGUI project"

Actually, in this case the SVG is the source and the Draw blocks 
are the "binaries". You can change the file names of the "binaries" 
and distribute the source SVG files in a seperate archive with their 
original file names.

That's why i put "binaries" in quotes before when referring to the 
preconverted draw blocks you mentioned before. Since SVG is text, 
even human-readable (with difficulty at times) the SVG files can 
be considered the source and the draw converter can be considered 
the compiler. Still, I agree that it would be best to make your own 
if you can, since the GPL is really poorly suited as a picture content 
license - they should have tried one of the Creative Commons licenses 
instead.

From the AGG Group: "Having our own SVG icons is a good idea as we 
don't have to worry about licence / distribution issues and we can 
tweak the generation to make it as AGG compatible as possible. To 
do that we need to settle on *one* tool so as we generate consistent 
SVG code, and we should generate SVG that uses a single unit of measure, 
preferably pixels, as it's a real pain trying to handle multiple 
units of measure within the same file."

shadwolf: Once you've settled on the "best" SVG generation tool to 
use (with feedback from folks here), then create a simple SVG icon 
(a shaded "next" arrow for instance) and post a link to the file. 
I'll have a look at the underlying SVG code it generates and we'll 
see if it's clean enough to use easily. We can then start creating 
more complex icons and improving the SVG renderer as we go. My hope 
is that we will end up with a nice set of REBOL / RebGUI SVG icons 
that are free of any licence / distribution issues.

Reasons:

	1) Solid authoring tools
	2) Access to a large number of existing SVG icons (if needed)
	3) The standard for vector graphics
	4) Interoperability

 5) Publicity ("we support SVG icons", as opposed to, "our icons are 
 just REBOL blocks")


If we generate our own content then there is nothing stopping us 
having icons in source form (SVG) *and* pre-compiled form (draw block).

There was a version of paint.r that output to SVG. I think Jeff did 
it

Don't know inkscape, But I'm on a Draw editor script since yesterday 
;-), but, hey, chhhh...

see screen shots entry in the menu of the inkscape website to have 
a view on what this tool can do for us ;)

I am a bit surprised not to see any DRAW SVG tool yet.... even just 
a small one.

My son, who helped with View 1.3 is using Fireworks to create the 
graphics, then a small REBOL script to convert to DRAW format.

Pekr: I know that Cyphre has put a few large city maps into DRAW, 
and they looked pretty good. We just need to get him to tell us more 
about it.

unless there's a non compete clause.

sure the ideal thing would be to have a rebol clone of inkscape ... 
But this will take lot of time to write ...

These are things to be added... but by a team of people.  Should 
move this discussion out of this group.

Most of us just want a decent cross platform development toolkit.

do %rebgui.r
display "RebGuiTest" compose/deep/only [
        label "Query" t: text 40 "" return
		a: area 100x50 "In the beginning, "
        button "Insert" [
				insert a/text "Title{}"
				; edit-text a enter

    insert tail a/text "And from a barely documented VID, a need for 
    something bare but easy to document: RebGui.  And the user saw it 
    was good."
				; edit-text a enter
				show a
        ]

do [at face/pane/3/text 10 insert tail face/pane/3/text "there was 
View 1.3.  "]
] ; End Display
do-events

How to move the cursor to inside the brackets of Title{->Here<-}.

How to enter a newline, and more generally use the function defined 
in rebgui-edit.r into the area?
** Script Error: edit-text has no value
** Where: action
** Near: edit-text a enter

More generally, how to enter a string and move the cursor after the 
insertion of it.  Thanks for an answer to this 'II am New' question 
pertaining to RebGui.

This should get you going:

display "Test" [
	a: area 100x50 "Title {}"
	do [
		rebfocus a
		system/view/caret: back system/view/caret
		show a
	]
]


The RebGUI edit-text function (and indeed the majority of the edit 
context) is designed to be used from within RebGUI widgets ... it's 
not "user-space" functionality. Manipulating text and / or focus 
is done via the standard view mechanisms (system/view/caret, focal-face, 
highlight-start & highlight-end) ... RebGUI just provides an encapsulated 
front-end to these. Depending on what sort of operations need to 
be performed (e.g. move the cursor to the Nth character), we can 
certainly add additional user-space functions (much like rebfocus), 
we just need to define what these functions are and how common their 
use would be.

Looks good under Opera. Has the SVG renderer you and Vincent have 
been working on got to the stage where it can display these? If so, 
post a link and I'll integrate that in. The entry point will be a 
function named 'load-svg in a separate file so we don't have to worry 
about "widget clutter", and I've already added a new 'icon widget 
and updated toolbar widget which use the 'load-svg function ... so 
we should be ready to go.

actually not because we have several problems on recursion group 
interpretation on gradient effect (fill and for border line fill) 
so now we are trying to make a better engine that can be use more 
efficient ly rebol object! properties we are basing our work on xml-to-object 
function ...

SVG format is so blury that you can't match it to a predeterminate 
struture you have to do a structure analyst for each leaf of the 
the object tree then get the needed information from leaf and nodes 
then convert them to rebol then apply the draw

I think that's because technologie issue (graphics libs that would 
be used to normally render SVG files to screen are different from 
AGG ) then you have plenty of possibilities SVG can be seen as a 
VID  widget tree where youonly get relevent informations but the 
structure have some recusing consept

for example the path tag that describes a shape (AGG/draw word) have 
obligatory a d field but then to handle the rendering options you 
have lot of ways to handle it

you can see the use of a style field in the path tag that encap all 
the graphical need (internal color / gradient effect, border size 
color /gradient etc...) or this information can be exploded in as 
many field that you have description

so you have an XML  struture we transforme it to a REBOL Object! 
tree then this tree must be glanced and translated to have the AGG 
draw block ;)

hum I don't know if we can be so drammatical and have a so drastic 
point of vue ....

It's a hard job to handle SVG structure and make our decoder able 
to decode all needed  information that's all ;)

for lineargradient SVG gives us a transform information (can be matrix, 
rotation,scale , skew, etc...) then we have an offset  x1 X2 y1 y2 
then we have  color informations spread method

a bit off topic: iTunes has a widget for rating of songs with stars 
one to five. Basically it's a widget that has 5 different states, 
depending on where you click on it. Would that be on the todo list 
for RebGUI? I'm planning to have such a thing in my next rewrite 
of my listview, I showed you some time ago.

but nothing prohibes you to make a sutch widget for rebgui ;)

This topic has been quiet for a while now .. anything happening??

Is there a tool to enumerate all the path to the text fields in a 
RebGui application.   Rebol is great to draw interfaces, but boy 
is it owerwhelming to find the path of those fields.  I did not find 
the way (if there is one) with Anamonitor003.   To have an idea of 
what I talk see my never finished pet at www.biblex.org.

Graham: End of fin year and Q2 GST.

Normand: Try the following;

display "Test" [
	a: text "Name"
	b: field
	do [c: face]
]
do-events
halt

ashley: dragging the cursor (mouse left/right while left button down) 
on top of a password field with hidden text, makes the cursor position 
not correspond accurately to the position of the mouse. very very 
small thing, I know, but if you want it perfect. :-)

I notice that there is no cursor in a text field now when the field 
is empty.

Yes, the list works to show stuff, but I can't select an entry and 
get a feedback into my code. Or is it possible?

you have a built in var in it to retrieve information

so the programmer that use the listview widget have only to make 
a test on listview/picked content if none -> no line is picked (so 
for example prompt an alert to the user either make processing of 
the listview/picked content

Robert, I'm waiting for a few things in the REBOL world to stabilize 
a bit before proceeding. Specifically:

1) The new REBOL/View system documentation
2) The new OS X version of REBOL/View
3) New SDK builds
4) Couple of AGG fixes

The two big design issues I'm grappling with at the moment are:


1) Look & feel, especially in light of the availability of View on 
OS X

2) Whether to revert to View 1.2 compatibility (and not use AGG) 
given the lack of a 1.3 SDK

2) IMO RT can provide them quite fast. So we just need to bug Carl 
about it. Shouldn't be a show-stopper.

OTOH i had a really interesting look at my floor, suddenly sitting 
there, after reading there is a running OS/X, in such a short time 
after 1.3 and Carl saying "now we really do that". OK, he said that 
the n-th time, but this time he did, in short time. so..

I know ... I just received email from old time reboller who's name 
I will not disclose, but he is right - why we are still pretending 
we have some secret weapon? The rest of the world really does not 
care:


  I don't want to know anything that the public won't be told. Why 
  do I
need a special handshake just to find out what REBOL intends 
  to do
someday? I just want to be a satisfied customer, I have no 
  desire to
be part of a secret club. Aren't plans and roadmaps what 
  you share
with customers? I have no need for secret organizations, 
  closed door
policies, or special preferential relationships.

Good 
  ideas need to withstand the harsh light of day, which is a
by-product 
  of openness and transparency. MS and other companies do not
care 
  one bit if REBOL is open-source-- they are so far beyond caring
about 
  that. RT has no first-mover advantage or special secret that
makes 
  REBOL a dangerous disruptive technology, and if it did,
open-source 
  or not, they could recreate it within a year (maybe not
compact, 
  but no one will care). Remember MS employs some of the top
scientists 
  from functional programming in their labs.

... all good things here - as working cooperation of dev community 
with RT should be used as marketing advantage and should be described 
on web-site as a good, working model .... that should happen on REBOL.com 
- but creating of new site needs real strategy, especially marketing 
wise ...

This is fixed by adding a reduce to the button data (rebgui-widgets.r), 
as in

		data:	reduce [
			#[image! 7x14 64#{
				...

inverse logic

 triggered a thought: what about a 'share-keyword and clone by default? 
 for performance [  text share[para] ]

If para was converted to a block, then make would copy it for us 
automatically.

Actually, yes, the default face has a para object, even though a 
lot of styles don't necessarily use para features (or even text at 
all), so sharing is worthy.

Well, I consider it a bug if you alter the text in one field, and 
experience unwanted side effects in other fields.

So, when you create a text field, it needs it's own paragraph object.

I agree that this is an undesired feature, I also agree that sharing 
the para object is generally a good idea. One way around this would 
be to change the focus function such that the shared para object 
is "reset" upon change of focus.

Seems a reasonable thing to do.

display "" [ button "Print" ]


shows a button with text "Imprimer" , and one with "Save" shows "Enregistrer" 
...

Latest build available at: http://www.dobeash.com/files/RebGUI-034.zip


*** Unzip this file into your existing RebGUI 0.3.0 distribution. 
Requires View 1.3.1 ***

Highlights include:

	- request-dir requestor updated
	- alert and question requestors added

 - Tab focus expanded to include edit-list handling (see %focus-demo.r 
 for details)
	- Button image now resizes correctly

 - text-list widget totally rewritten to use [newly documented] View 
 1.3.1 iterated faces
	- text-list now supports Ctrl+click, Shift+click, Ctrl+A

 - text-list scrollbar now reacts to current display state (eg. disappears 
 if not required, alters dragger ratio as rows are added, etc) 
	- check and radio widget aesthetics improved

 - added options support to layout function and text-list and check 
 widgets

 - changed base-size from 4x5pixels to 5x4 pixels (better granularity)
	- reworked all widget sizes to account for above
	- fixed scaling problems with many widgets
	- updated focus-demo.r

 - greatly enhanced and expanded %tour.r to demonstrate more variations 
 and usage cases
	- added "live" unit-size and font-size drop-lists to %tour.r
	- numerous other minor bug fixes and coding improvements

In the works


 - table widget being rewritten along the same lines as the new improved 
 text-list widget
	- aiming for a 0.4.0 beta release prior to DevCon2005

With the text in an edit list highlited, click on the drop down arrow, 
click on a new item, then tab .. and the above occurs.

Still a beta release .. not a general release.

Mchean all you have to do is put a comment  ";" at the line and it 
should work =)

I guess this is more of a Rebol problem, but alerts don't stay infront 
of the parent window.

Image accepts a file for the image, but not an image type.

When a window containing a text-list is maximised, the scrollers 
disappear.

I thought I'd have a go at building an application with RebGUI, and 
I'm starting with an open source electronic medical record.
I'm just building the gui first off 

http://www.compkarori.com/emr/remr.exe
http://www.compkarori.com/emr/remr.r

Help appreciated :)

Good first cut. I''d standardize label / field widths with composed 
words; eg.

window-size: NxN
tab-size: window-size - NxN
label-width: N
field-width: N

display "Test" compose [
	...
	label (label-width) ...
	field (field-width) ...
	...
]


This way you can globally experiment with different sizing layouts 
without too much effort.


If you come across anything that would make writing an app even easier 
(be it a RebGUI usage hint or enhancement request) then be sure to 
mention it here. Thanks. ;)

Hi, just few notes .... 


1) why new versions are not released as complete ones? The download 
is small already. I tried to point out my friend to it, and he missed 
0.3 version or so, which is required ...


2) box definitely does not resize properly. Still, so far, Romano's 
system was the most complete and the least errorless version I saw. 
Try to move resizing window here or there and you will get even cases 
as three lines of color boxes, no spacing, mixed together ...


3) lists - are we ready to overcome rebol limitation here? It works 
better, but still not flawlessly and not in system friendly manner: 
a) when moving "too fast" with mouse, it stays open b) in above and 
and other cases, ESC shoudl close it c) I understand we have use 
some "excuses" and workaraound for now, but that is not the way to 
go in future d) when moving away, it should stay open, last hilited 
item should stay hilited, close on esc, close on click-away, should 
be foxusable, ability to be driven by keyboard  - that is the only 
system friendly way


4) text-list multi mode - ctrl works, shift too, ctrl A too, but 
not in a system friendly way once again. Maybe I should check first, 
but IIRC, it should work following way - ctrl selects particular 
items. BUT - it should also deselect them - try ctrl A and then, 
holding Ctrl, press some item - it does not deselects them - that 
is imo wrong. Also - shift should mark all items between point of 
last press and active mouse position, deselecting all the rest, even 
if previsously selected ...

scaling - cool! But resizing works a bit strange sometimes, especially 
if you lower window size under some acceptable limit .... dunno if 
that one could be automatically calculated or not, but sometimes 
you get destructed (non-functional) gui

Luisc: its ok im a newbie too

Ashley, I've also come across the need to use width keywords.  I'm 
still not sure what the best method may be, but I'm leaning toward 
devising a generic UI grid.  At its simplest, it'd be like a 2-cell 
(four guideline) html table with labels on the left, controls on 
the right.  But this could also broken and repeated across the width 
of a form, with keywords/parameters used to position elements across 
the guidelines.  This is a thought in progress, unfleshed.

Somewhat in the mold of a Wordprocessor, but with form elements instead 
of paragraphs...

How to clear a field ?  This doesn't work without screwing up the 
text cursor...

 display "" [ f: field 30 [ show-text f copy "" ]] do-events

Graham,

	display "" [field [show-text face "" rebfocus face]]


but wrapping this in a 'clear-face func is the way to go. Thanks, 
added to the list.

Shouldn't that be part of show-text ?  If you're going to preset 
a number of fields eg. to reset some defaults, you don't really want 
to rebfocus face on each one as well as do a show-text ?

Depends. Take the following code:

display "" [
	a: field "A"
	b: field "B"
	c: field "C"
	button "Clear" [
		show-text a ""
		show-text b ""
		show-text c ""
	]
]


Clicking "Clear" first behaves as we expect, but clicking in a field 
then "Clear" is problematic. The solution, as in VID, is to unfocus 
whenever soemthing else is clicked on. Added to the list.

Pekr, the RebGUI resizing model is simple and works well. The "problems" 
you are now seeing are caused by the fact that the latest build prevents 
widgets being assigned negative size values (which prevents one class 
of errors (sporadic crashes) but can cause anomalies when windows 
are resized below a certain point). RebGUI assumes that if you allow 
widgets to be resized you will *also* specify a window min-size that 
can comfortably hold all widgets. The next build will include an 
updated %tour.r and documentation that addresses this.

Could rebgui calculate a minimal size for a layout?

Go here for a list of all other issues: http://www.dobeash.com/it/rebgui/issues.html

Certainly, and indeed that was how it was handled before I "discovered" 
min-size ... but handling this non-natively (i.e. not via min-size) 
looks and feels a bit crude as well as being inconsistent with how 
other [non-REBOL] apps handle resize.

i would use a combination, window with min-size, but if min-size 
is smaller than min-layout-size, dont shrink layout further. keeps 
at least layout intact and user can resize back.

since users (or Pekr;) see a destroyed layout the current way.

But this allows you to resize a window below an invisible threshold 
only to have your resize [partially] ignored when the layout jumps 
back to it's minimum size. I still maintain that setting a realistic 
min-size is not only subjective but the role of the GUI designer 
... if RebGUI tries to infer this setting it will invariably get 
it wrong ("why did it let my title field shrink to less than three 
characters ... everyone knows a one-char field is useless" type of 
issues).

No, this allows the layout to jump back to a bigger size if the designer 
made a mistake with min-size.

Correct, the code in question prior to latest build was:

	if find face/span #"W" [face/size/x: face/size/x + delta/x]
	if find face/span #"H" [face/size/y: face/size/y + delta/y]


which works well except that negative sizes seem to cause downstream 
problems. Suffice it to say that I'm now leaning towards a layout-min-size 
model to at least prevent negative sizes in the first place! ;) Good 
discussion.

And in your example, the right hand text list demonstrates this point 
for me losing it's sliders as well as another problem - a tiny square 
widget.