AltME groups: search

Showing a face may have the effect of showing the parent-face. If 
you show each sub-face, you may be asking the View system to at least 
check if the parent-face needs showing, so, repeating work unnecessarily.

I have a view that works like simple Mapserver client with zoom box, 
etc. I made new version that supports panning, by dragging with mouse.

All worked fine (zooming, panning) until I changed simple loading 
map image from load to read-thru/to.

Zoom box worked, but while mapfile loads it slowly follows cursor 
instead that by script logic (that I did not change) it should stop. 
And my map pane no more returning to zero offset after map image 
loading.
How those view faces are connected with read-thru  ?
 
loadMapImg: func [
	mapUrl
	/local rezImg
][
	if exists? %mapImg.gif [ delete %mapImg.gif ]
	either attempt [
		rezImg: read-thru/to mapUrl %mapImg.gif
;		rezImg: load mapUrl	
	][
		return load %mapImg.gif
;		return rezImg
	][
		return none
	]
]]

When using a text-list, how do I reset the slider to the top if the 
data changes? For example: The slider is at a lower position, now 
I change the content to some smaler list and want to redrag the slider 
to the top and have the text-list show the first entry at the top.

And is there a simple way to add mouse-wheel support to a text-list?

I have done some work on it for the word-browser. Just look the code 
(desktop\rebol.com\tools\word-browser) :

- There is a patch to the resize function (maybe you can use it to 
refresh the list to the same size?) line 65

- There is a scroll-text-list func (line 414) that is called by the 
window/feel (line 381) for the mouse-wheel

Have also a look at line 259 : it's what is done to refresh the words 
text-list when you select a different category.

Thanks, I take a look.

Why FEEL actions (down, up, avay events) behaviour changes when I 
changed LOAD  for a face image to READ-THRU or LOAD-THRU?

Microsoft's Sparkle: Is It a Flash *cough* view *cough* Killer?  
http://www.eweek.com/article2/0,1895,1914903,00.asp

to clear the wait-list

 - This works AFAIK because event-port is a normal port and thus in 
 the wait-list. Everything there is processed on wait. 


So you could remove only this port temporary. Other tricks with that 
behavior: with read-thru rebservices/beer would still be active while 
downloading. maybe a bit less responsive, because a /no-wait introduces 
little delays?   

And you can caugh events before everything else, like this:

It works for me! Thanks!
Is it correct?

I looked into system/ports/wait-list - there is a lot variables and 
awake function.
For what else wait-list is good?

So I mean a bit detailed docs about wait-list system ...

yes, a block. with ports in it.
 echo on probe system/ports/wait-list
then look in %rebol-echo.txt with an editor.

i guess a search for "awake rebol" would turn something up.

You cannot see the text attribute, unfortunately, because it is covered 
by a subface, the blue progress amount.

If we set the font color to white and move the subface a little bit 
we can see the text.

There was some work on a progress in "RebGUI" group arround 24-nov-2005.


For those who don't like setting Altme message limit too high (sync 
bug) there is :
http://membres.lycos.fr/didec/rebol/altme-chat-reader.r
Save it to your %Altme\ folder for easier use.

well, exactly ... that is a mess however ...

PATH-THRU defines where the public cache is:
>> ?? path-thru
path-thru: func [
    "Return a path relative to the disk cache."
    url /local purl
][
    if file? url [return url]
    if not all [purl: decode-url url purl/host] [return none]

    rejoin [view-root/public slash purl/host slash any [purl/path ""] 
    any [purl/target ""]]
]

This means that the OS should provide standard ways of installing 
and uninstalling. But a good installer/uninstaller hasn't worked 
its way to be the standard yet.

now I understand Jaime's frustration .... IOS, as a top commercial 
product, was left completly unsupported ... I will neither use nor 
sell dead-end product ...

well, it sits somewher on feedback, somewhere in rambo, - how many 
times I should submit it to not sound as a broken machine? (which 
I am, repeating my complaints anyway :-)

but well, then we heard, IOS 2.0, somehow joined in effort with Coop, 
Altme2.0, whatever they are, will address some things, but so far 
there is no single pointer into what actually will or will not exist 
as a product. AltmMe moved to Qtask imo, Coop I don't know what it 
was supposed to be about, IOS 2.0? :-) Who knows ...

yes, of course, what else ... but that was a way of how to burry 
a product .... without me using it, I can't become knowledgable about 
it, I can't promote it, I can't make them another sales ...........

What do you mean by " IOS clients "? it does not allow multiple exes, 
but it uses a subfolder for each server. so you can have multiple 
ios-acounts.

for usb i once provided a solution. an encap which changed the registry-entry 
to point somewhere else, eg the place it was called from. if you 
call that from usb, usb is the link-root for new starts.

well i think those codes are good for research of the best and cheaper 
solution but it  would  be better to add them one day to rebGUI  
to get a centralized way for user to get all optimised kind of widgets 
and not to have to endless browse internet to search the needed script 
in the better version

shadwolf: I have no experience with RebGUI but I think it can make 
use of VIDPatch. It's just a model how to expand VID. This can be 
implemented into RebGUI natively.

Kru: nice!

After looking quickly at the code, I see you patch each face feel, 
and text editing function...

Does using insert-event-feel is not enough to handle the tooltips 
stuff (Also I know that it's hard to know witch is the face that 
is the final destination of the event) ?


Also, there is a 'help facet in VID-face and VID styles that is use 
nowhere. I though it could be used for handling tooltips strings. 
Then just have something in the global event handling that check 
the 'help of the face under the mouse and flash tooltips if any.

DiceC: Hm I have to try it. Most of patching is due to drag'n'drop 
I think. Haven't seen the code in a while...

Mmm.. I don't know if it would be so soon. That change would break 
a lot of code.

Shock, horror ... you mean scripts *rely* on a bug?! ;) Better to 
fix it sooner than later then.

I prefer to fix code writen to patch a Rebol error, than to let the 
bug in Rebol like this.

There's actually an old patch in there already, it just hasn't made 
it into a release yet. Thanks.

there's a filter isn't there for sorting out code from application 
problems?

I don't know of a filter just for that. You could probably use the 
"Other" category for those kinds of entries, to break them out a 
bit.

in windows...  just cause its making me mad... how does REBOL/view 
1.3.2 go about sourcing the damned user.r and rebol.r files... cause 
right now, I once again have some icons which source them and some 
which don't.    :-(


basically if you double click an *.r file it does not work, but if 
you launch rebol.exe or a shortcut to it... it works.

ARRGH ! and I though  v1.3 was about keeping install simple   :-(


I have set both HKEY registry entries to the same dir... but why 
are there different reactions depending on the way you launch rebol?

try doing a clean install. note that uninstalling may leave some 
registry keys (in particular, it will leave them if they have been 
customized).

after about an hour of fidling around...  I realised that the user.r 
is now ALWAYS sourced from the actual windows's user dir.  This is 
a good step forwards.

the issue is that the window's user "home" path is just freakish 
and very innaccessible from a dumb user's point ot vue...

I'm not addressing "you" and M$ does not care about Linux.  But what 
they did do was make a system that allows multiple people to use 
the same computer and keep their date apart.  It is wrong, and badly 
done, but the "correct" way to do it for thier system.



As to Carl allowing other ways to do it, that is the real question 
I assume you are asking.

even if I include a rebol.r file at that path... it is ignored...

Has anyone written a script/function to display pie-charts?

the cause is imo unbound "", 'copy "" solves it so that it only blinks 
and then is correctly replaced by clear field = no artifacts .... 
those unbound "" shared values are imo a REAL EVIL in rebol, as well 
as shared subobjects etc. I can't seen ANYONE, who would not come 
to problems ....

RRDTool is a pretty cool logging and graphing tool. Very flexible, 
but I think it's a bit difficult to use. a REBOL equivalent that 
would be easier would be cool.

ah, it's not a Rebol tool.

Cyphre is working on a pie chart library for Qtask. Ask him.

;Robert, should be easy enough to do a script based around these 
examples from the draw docs

Does anyone have a clue when rebol 'draw will support fonts in Linux?

I recall there was a script to upload pics to website. I think it 
was used for the Devcon images. Anyone know if that is available?

I lost a little the track regarding 'ASYNC.. I understand that 1.3.x 
will not have kernel 'async support in /view but Carl is talking 
in his Blog about /async handler..Is that a special /core beta? or 
is it from Gabriele's async handler? who can help me out here..

aaa oke a flaw... Right thanks for the info

It was not so much that there was flaw. But that it introduced a 
new set of issues on how to create programs that behave correctly. 
Because the async mode could interrupt any excuting sequence.

And because there were so many other bugs in the core. The decision 
was to fix to bugs, and have a stable version that have nail as many 
as possible before introducing the async event model.

Is there anything I can do about the "no 'wait in view handlers" 

I try to use an sqlite call in a button, but sqlite uses wait internally 
...

a few nested waits are usually no problem. maybe
sqliting?: false
..
button [

 either sqliting?[alert "sqliting already"][sqliting?: true sqlite 
 sqliting?: false]
]
;if your user clicks very fast.

My script looks like this

view layout [
   title "Address"
   across
   name-list: list 300x400 [
      text bold 400 [

         person-info: first sql rejoin [{select * from person where guid like 
         "%} face/user-data {%";} ]

         ;person-info: ["1" "ingo" "hohmann" "static test" "" "" ""]
         vals: get person-disp

         repeat i length? person-info [insert clear vals/:i person-info/:i]
         show w
      ]
   ]
   supply [
      count: count + cnt
      face/color: ivory
      face/text: none
      if even? count [face/color: ivory - 50.50.50]
      if none? v: pick list-data count [exit]
      face/text: reform [pick v 4 "-" pick v 2 pick v 3]
      face/user-data: pick v 1
   ]
   sld: slider 16x400 [
      c: max 0 to-integer (length? list-data) * value
      if c <> cnt [cnt: c show name-list]
   ]
   w: box 400x400 with [pane: layout address/make-person-lay]
]


If I click on a list entry, it's like all other entries are removed 
from the list. Wherever I click I always get the same entry, if I 
move the mouse over the list, all other lines are emptied when the 
mouse moves over them.
Using the static test-data everything warks fine

I don't think this is an sql problem, but a list problem.

(and also not a problem of WAIT)

I suspect your INSERT CLEAR may be causing a problem.

I now don't think that's where the problem is occurring. Will make 
a stripped down list.

I just tried adding a wait in there, and it still works. So you are 
right, it's not the wait ...

It's a string copying issue somewhere.

O'k I found it - and of course it was not a problem in Rebol.

Somehow I had changed setting Language for non-Unicode programs in 
Regional and Language options Advanced tab.

Is there a default uptodate MS-Windows /VID somewhere for view? I 
want to skin /view to be MSWindows.. I know there was once a windows 
skin but was it Win95?

a bug? view layout [ch: choice 100 "a" "b" "c"  btn "zmen" [set-face 
ch "c" show ch]]

I know - it is just that View 1.3.2 is regarded being a stable release 
.... but the thing is, you really can't depend upon proper style 
behavior ... it is really tiresome to still "debug" VID ....

there's a million things wrong with VID, but right now there isn't 
much to do but work around it

the thing is: what can we do about it right now? not much, other 
than either wait, or come up with sneaky solutions on our own. some 
did a solution and made RebGUI. I made LIST-VIEW, because I needed 
it and that's really the only way forward. Carl can't work on 50 
things simultaneously (or can he?) and designing a new VID, I don't 
think Carl can do that in an afternoon.

go to RAMBO and check for my name and you will see that once I find 
a bug, I report it ... not to mention tonnes of my reports in earlier 
days ....

you should be carefull to your statements, as some ppl simply use 
REBOL as a tool, nothing more. Not like us. And to them your claims 
will become arrogant ...

petr, the ticket is there and won't be forgotten. there's no problem 
with that. but, to fix the red icon problem you need more than that 
- you also need to release a new version of ios.

Is there a way to have "invisible" text in a face, i.e. invisible 
text even if ALL [STRING? FACE/TEXT NOT EMPTY? FACE/TEXT] is true?

This seemed to be the natural approach; I can't come up with a solution 
to this question (driven by cure curiosity).

I don't think you can do it in a single face... but so I can understand 
the goal, why do you need it ?

Here's a way using DRAW dialect text:

Actually, my idea was to have a FIELD style with DRAWn text without 
having to write my on CARET-TO-OFFSET and OFFSET-TO-CARET as mezzanines. 
Maybe that's a bit over-the-top, though.

Put a feature request in Rambo (if it's not there already). I am 
sure I also wanted to use such a feature before for such things.

Hi Anton, about the list bug with sqlite. We were both right. It 
_is_ a series copy bug. And it _is_ an sql problem ... sql reuses 
the series it has given as a return value on the next call :-(

Volker, of course you're right, I always forget that. So here's a 
proof-of-concept of an AGG field style:

is it possible to save a layout as image, e.g. png?

Is it possible use fonts which are not installed? (under windows, 
at the moment) 

I have a dir with lots of font files, and would like to write a font 
preview utility for it, but as these fonts are not installed, I would 
have to reference the font-files, not the name.

There was some discussion and a couple of scripts on the ML some 
time back about this: http://www.rebol.org/cgi-bin/cgiwrap/rebol/ml-display-thread.r?m=rmlYDRQ

Has anybody every tried to create a fish-bone-diagram dialect?

No.

But I used to be a fan of allclear -- they devised a very simple 
text-based dialect from which they can generate numerous diagrams.
http://www.allclearonline.com
It could probably be re-engineered in REBOL in a few weeks :-)

This is probably sooo stupid.. because i work always with a make 
face I now need to add a simple face to a layout.. ;-) But i simply 
dont get it running... ie.

Sunanda, thanks for the link. Never found a tool where I could use 
a dialect to create a flow-chart.

aaa wait... its the 'make face that needs to go inside a layout ofcourse...

yes yes.. i think thats it ;-) I forgot to do a make face of the 
bolck..

1) copying it all into a block and append
2) using make face layout inside a layout
3) using a layout with a make face with
??

Hi Rebolinth, there's a compromise between the low-level MAKE FACE 
and the LAYOUT [] approach for single faces called MAKE-FACE. It's 
basically a way to create single faces using the VID style name and 
a spec block for of specifying facets. Of course, MAKE FACE should 
have the lowest overhead, whereas - just as it says - LAYOUT is more 
for describing whole layouts. You may find MAKE-FACE slightly easier 
to create VID-compatble faces without getting into too much detail.

With is nice i.e. to set the background color of a text style (where 
colors used in the dialect are assigned to font) :

I'm testing a tab button group with BTN look. What do you think? 
http://www.hmkdesign.dk/tabview.png

the idea is for them to be toggles in a group, where only one of 
them can be selected. then there needs to be an easy method to perform 
some kind of action, such as selecting a mode or changing a layout 
in a pane

the point is to have a button group that actually looks like a button 
group

I like the BTN style, but it's underused in VID for other applications, 
such as a tab button group

second, I use pane switching a lot and it's a hassle to write your 
own routines for that all the time. it should be much simpler.

Just write a couple of functions, face-to-front and face-to-back 
and you can reuse them many times.

The point is that it should be easy for everyone to do this. VID 
does not offer a standardized way to do this and you need to study 
how to switch layouts in panes, etc. It should work out of the box

Default size of the tab-panel is calculated as the maximum of the 
three layout pane sizes. The three texts are available in face/texts 
so can be processed by init easily to create the three buttons. PANES 
would be a dialect word that simply sets face/panes, absolving the 
need for WITH in the most often used cases.