AltME groups: search

In the coming versions, I will remove the original demo and rely 
more on the docs. I might provide an updated demo later when the 
list view is in a more reliable state.

What I mean by this is that list could request a model to give the 
data to display for the specific row. Instead of enforcing a particular 
model for the data representation.

hmm... could you give an example? my brain is a bit fried right now 
:-)

Right now all *data model* is a block. But you may want to be able 
to access a file of bytes and some particular fields in that stream.

Nope. One of my problems with VID and RebGUI is that the gfx are 
to tightly couple with the data representation. So a change in data 
representation implies a change in the gfx code.

what it can do right now is append, change, remove and insert rows. 
so this would be kind of an ACQUIRE function that calls a controller 
function that returns a value that can be inserted into DATA with 
insert-row or append-row?

with the point being, that you call your data source and insert the 
data into the list using a single function

I'll take a look at it, maybe tomorrow. close to bedtime now :-)

Henrik, just looking at your list-view, and noticed a little thing 
that can be optimized:
b: []  insert tail insert tail b 1 2
b ;== [1 2]
b: []  insert insert tail b 1 2
b ;== [1 2]

I see you are successfully basing on FACE. Just a point about placement 
of custom facets; I would put all the facets that already exist in 
face at the top, then follow with all your custom ones. That way, 
someone who reads your code can see which is which. Furthermore, 
if by some chance the FACE definition ever changes (adding or removing 
a facet), it will be easier to see where the fault lies. Eg:
	list-view: face with [
		size: 500x300
		edge: make edge [...]
		
		; my custom facets

  colors: [...]  ; <--- this is the same name and I suppose the same 
  concept as used in VID 
				; (but it's still custom from list-view's point of view)
		select-modes: [single multi row multi-row column]
		; etc...
	]

What I meant to say about the face definition changing - say for 
instance a COLORS facet was added to FACE. Then, you would be unintentionally 
overriding the View system's one and probably introducing bugs, because 
the View system would expect it to be used a particular way and might 
render the face incorrectly given the data you've put into it. It 
might also change that data when you expect it to remain, etc etc.

anton:

First one: Alright. :-)

Second one: All that code will be thrown out and there will be a 
different demo later

Third one: You have a point. I'll look at the issue of sorting standard 
vs. custom facets.

Henrik, Cyphre once did a list-view for me. It has some really nice 
things in it. Take a look at http://www.robertmuench.de/projects/data-form
for some docs on it.

Let me know, if you want to take a look at the source code.

thanks! I'll take a look at it.

version 0.0.14 uploaded

Changes:
      Fix: Documentation updates
      Fix: Multiple lists accidentally shared the LIST-SIZE value
      New: Custom layout block can now have multiple rows
        at the cost of horizontal resizability
      New: Now using ROW-FACE to store the custom layout block
      New: If HDR-COLS is set to a single word in a block,
        it will take the width of the list view.
      Fix: Code size optimization of navigation functions

      Info: The version between 0.0.12 and 0.0.14 is mysteriously missing... 
      :-)


The demo has been changed a bit to test custom layouts and multiple 
lists.

http://www.hmkdesign.dk/rebol/list-view.r
http://www.hmkdesign.dk/rebol/list-view.html

please note that now when using custom layout, it's not possible 
to have a horizontallly resizable list, because it's hard to determine 
how the elements should be resized without making the layout code 
complex (if no one has any suggestions, that is...)

that's a good idea! I think I'll implement that soon.

That's an old version now.  As you've discovered, 'case isn't defined 
in the plugin.  I've used Ladislav's 'case for the later versions 
.. 
Don't know why you're seeing this now .. unless it was cached.

A later version if not the latest is http://www.compkarori.com/reb/pluginchat5.r

is something supposed to happen when you select a user from the list? 
It reacts very slowly and sometime require multiple clicks before 
it responds

So, looks like a bug in that version of the list-view.

I found a bug that might be related to that, but it only shows up 
if your row height is different than 20 pixels. It was fixed in 0.0.14.

btw - lock-lst should be preserved for real locking, I would maybe 
introduce - no-edit, or no-select flags ... just a naming convention 
...

I get the corner piece at a different size from the scroller.

when you click on a name in the list, ip and port information is 
given.

the background slows everything down so as per Allen's suggestion, 
I have set saved-area: true for all the buttons.  I did the same 
for the main message area which fixes up the slow scrolling issue, 
but causes a resize error sometimes.  Not sure how to fix this as 
resetting the saved-area to true after a resize doesn't fix it.

0.0.15 uploaded

Changes:
      New: Updated documentation with images
      New: DATA can now also be a single block of values
      Fix: IN-COLS is no longer mandatory
      Fix: MAIN-COL is no longer mandatory
      New: Default WIDTHS now a fraction value.
      New: Fractional widths of the list width as decimals
      Fix: List size calculation optimizations

      Fix: Scroller width is now always equal to the corner reset button 
      width
      New: SCR-WIDTH lets you set the scroller width
      Fix: AGG is no longer a requirement
      New: CLEAR to quickly clear the list

The files have moved again:
    http://www.hmkdesign.dk/rebol/list-view/list-view.r

    Docs are available in makedoc2 format at:
    http://www.hmkdesign.dk/rebol/list-view/list-view.txtand
    http://www.hmkdesign.dk/rebol/list-view/list-view.html

Seems to be a missing graphic for the docs

yes, some graphics are still missing. there is a bug in the width 
calculation code, which I want to fix for 0.0.16, so I thought that 
would have a higher priority

but I can tell you what our 300 users are used to, working with 3 
grids we introduced within last 10 years, what they valued most - 


- column resizing was not an issue at all - proper grid can calculate 
size to set it automatically - we had two modes, auto column-width 
adjust, or static, set in config

- drag and drop was rarely used for columns. You can do it othe way 
around - let users to choose which columns they want to see, in what 
order. Let them save the config!!! Without it, forget about your 
grid being powerfull - what ppl are looking for is comfortable view 
of their data. Each person is used to slightly different view

- horizontal scrolling is a must. How cbut I can tell you what our 
300 users are used to, working with 3 grids we introduced within 
last 10 years, what they valued most - 


- column resizing was not an issue at all - proper grid can calculate 
size to set it automatically - we had two modes, auto column-width 
adjust, or static, set in config

- drag and drop was rarely used for columns. You can do it othe way 
around - let users to choose which columns they want to see, in what 
order. Let them save the config!!! Without it, forget about your 
grid being powerfull - what ppl are looking for is comfortable view 
of their data. Each person is used to slightly different view

- horizontal scrolling is a must. How cbut I can tell you what our 
300 users are used to, working with 3 grids we introduced within 
last 10 years, what they valued most - 


- column resizing was not an issue at all - proper grid can calculate 
size to set it automatically - we had two modes, auto column-width 
adjust, or static, set in config

- drag and drop was rarely used for columns. You can do it othe way 
around - let users to choose which columns they want to see, in what 
order. Let them save the config!!! Without it, forget about your 
grid being powerfull - what ppl are looking for is comfortable view 
of their data. Each person is used to slightly different view

- horizontal scrolling is a must. How cbut I can tell you what our 
300 users are used to, working with 3 grids we introduced within 
last 10 years, what they valued most - 


- column resizing was not an issue at all - proper grid can calculate 
size to set it automatically - we had two modes, auto column-width 
adjust, or static, set in config

- drag and drop was rarely used for columns. You can do it othe way 
around - let users to choose which columns they want to see, in what 
order. Let them save the config!!! Without it, forget about your 
grid being powerfull - what ppl are looking for is comfortable view 
of their data. Each person is used to slightly different view

- horizontal scrolling is a must. How cbut I can tell you what our 
300 users are used to, working with 3 grids we introduced within 
last 10 years, what they valued most - 


- column resizing was not an issue at all - proper grid can calculate 
size to set it automatically - we had two modes, auto column-width 
adjust, or static, set in config

- drag and drop was rarely used for columns. You can do it othe way 
around - let users to choose which columns they want to see, in what 
order. Let them save the config!!! Without it, forget about your 
grid being powerfull - what ppl are looking for is comfortable view 
of their data. Each person is used to slightly different view

- horizontal scrolling is a must. It is so basic feature, I wonder 
why noone is requesting it? grids based upon list show its main flaw 
- non ability to scroll horizontally. Real life apps do consist of 
more fields usually then you can fit into your screen ;-)

- sometimes users requested split-view, so that few left columns 
got locked and did not scroll horizontally ....

- proper keyboard handling is a must - enter=edit, ins=insert record, 
del=delete record, move by arrows, pgdn, pgup, end home, ctrl plus 
pgdn, pgup, home, end, left, righ ....

Eliza is a hotty!!!

ah damned, I should request new notebook, sorry, my keyboard is so 
broken - few keys don't work and I have to keep "a" in clipboard, 
making wrong presses with ctrl plus c which is near ctrl plus v

pekr, config saving will come, but I'm not sure yet how to do it. 
maybe an object or something.


Horizontal scrolling, I need to figure out. It will be an additional 
mode that can be activated by setting FIT-WIDTH to FALSE. By then, 
a scroller will appear along the bottom and different policies regarding 
column drawing need to be made. A little complex to figure out, but 
it will come.

No. There is text-list/update. Partially broken. I have a fix somewhere, 
but not used for some month.

LIST-VIEW version 0.0.16 uploaded.

Changes:

      Fix: LAST-CNT crashed on empty SORT-INDEX after filtering

      New: FILL flag to set whether or not to paint all rows in the list 
      view

      New: COLORS/4 contains the color of the background behind the rows.

      New: When resizing OUT-COLS in runtime, WIDTHS is reset to default 
      values

      Fix: Columns can now be appended or removed to OUT-COLS in runtime
      Fix: Changed behaviour so that DATA is always treated as
        a block of blocks if manipulated by LIST-VIEW itself.

        Only external input of a single block value can change this.
      Fix: INSERT-ROW-HERE, REMOVE-ROW-HERE, CHANGE-ROW-HERE
        failed on empty DATA
      New: Stricter typechecks on manipulation functions
      Fix: Values are now copied into DATA
      Fix: Sorting did not work, when using only one DATA column

      New: Allows using keyed blocks as input for manipulating functions.

      Fix: Header size now calculated with OUT-COLS also, which removes 
      a crash
      New: Created a test suite in file list-demo.r
      New: Moved demo code out of this file

    Main file is available at:
    http://www.hmkdesign.dk/rebol/list-view/list-view.r
    Demo and testcases available at:
    http://www.hmkdesign.dk/rebol/list-view/list-demo.r
    Docs are available in makedoc2 format at:
    http://www.hmkdesign.dk/rebol/list-view/list-view.txtand
    http://www.hmkdesign.dk/rebol/list-view/list-view.html

LIST-VIEW 0.0.17 uploaded.

Changes:

      Fix: FILTER was broken. Seems to be a bug in clearing bitsets.

      New: EDT-FLD-ACT and EDT-LST-ACT to let you perform functions when 
      tabbing out of an inline editing field

      Fix: Column selecting marked columns beyond the end of DATA.

      Fix: Fractional widths are now properly resized, when resizing the 
      list

      New: Inline Editing! Can be activated with the EDITABLE? flag (default 
      on)

pekr, well, I find the names to be logical, but if we can agree on 
better names, it's only a matter of search/replace

if you don't want any auto resizability, then MAIN-COL should be 
set to NONE, but this isn't tested thoroughly yet. The point is that 
in my lists, I mostly have some data columns that need to be short 
and then one long description column. It fits very well to have only 
that column to be resizable, plus the rows always neatly fits horizontally 
to the list view size.


Later I'll implement your much wanted system for horizontal scrolling 
and in that mode, MAIN-COL won't be needed as the list will take 
up the amount of width that it needs. Then it'll work more like a 
standard list.

aligning column headers with list data doesn't work yet when using 
a custom row face. there are apparently still issues with MAIN-COL 
and it really wants to autoadjust the column width. otherwise the 
solution would be to just provide a set of pixel widths. I'll see 
what I can do about this.

I have a large view script that work fine with the previous verison 
of view, but that gives the following error witht he new version 
of view:

Misplaced item: "imanXw"
Misplaced item: "Quit"
Misplaced item: [

    (write %session-time.txt (session-time + (either now/time >= time-start 
    [session-seconds: now/time - time-start] [se
ssion-seconds: ((now/time + 24:00) - time-start)])))
    quit
]

http://www.compkarori.com/reb/pluginchat23.r

which has support for RMB on the users .. but a visual glitch in 
the scrollers.
Doesn't work under plugin.

well, i think RT have to release a signed version .. but as carl 
says, it's difficult.

MS had to disable the NS-compatible plugins by court order - some 
jerk (EOLAS) patented the idea of a plugin in a browser, and the 
overloaded US patent office didn't notice the prior art, much of 
which predated the Internet. Still pending, but MS preemptively dropped 
the patented method (the EMBED tag). EOLAS is waiting to see how 
this turns out before suing the rest of the world.

what's the best/most "friendly" and brief way to get a basic font 
object?

There's a tiny explanation here:

http://www.rebol.org/cgi-bin/cgiwrap/rebol/ml-display-thread.r?m=rmlZVTK

So, Henrik, seemns to me you need a news section on your website, 
and then you can simply post one line of updates : )
This is looking really good.

yeah looks like I'll need that. I'm also making a Rebsite now. I 
think that's a little easier than posting a large list of file names 
here

See, now this is why we need a cetralized website where you can post 
your news : )  everyone is having to set up the same thing, seems 
a little silly.

could also be a news item viewer for Viewtop? I think it's a bit 
of a shame to stuff everything in little dull text files.

I think of it the other way around, "send" the material in XML, RSS, 
ATOM, what ever, an let each system deal with showing it.  But no 
matter what, we need a website where all this stuff comes together.

should follow the width, but not working for me.  Is there a place 
in the source I can start the edge to a width I want ?

I use the docs as kind of a testing ground too. If you see something 
in the docs which you are trying to do, it should work, because the 
images in there are always generated from the latest version

made a rebsite under Public/HMK Design. should be a little easier 
to run the demo now

I have a question about the matrix functions in Draw.  Is it true 
they can only be applied to an existing draw sequence?  that is ,I 
am creating novel points to insert into a polyline that is yet to 
be drawn, it seems I cant  use the matrix functions to generate these 
points.

TomC, I think I understand your question. I had a similar problem. 
I wanted to calculate where some points in the middle of a curve 
would be, if that curve were drawn.

This is a problem I am having setting up a view on a linux virtual 
server. My cgi look like...

There probably is no X server, so this might explain the error. Is 
there a work around, or do I really need X?

I dont need view, its just easier to do using it, since the project 
involves layering gif images. My expectations were that View would 
work as long as I didnt try to show a window. Luckily I have a a 
backup plan in place, just use core, so I set that up core last night.

I'd like to be able to use Core as a CGI processor for its image 
creation features......But I've never had a happy time testing it.

Anyone know how to popup a windows *without* grabbing focus ?

I pictured this being a 3 col listview, with
first column - left aligned, bold
second column - centered, not bold
third column - right aligned, not bold

Graham, if you mean a REBOL window, I think the only option is to 
change the active window after it comes up. If launching another 
app, you can tell it not to activate the window.

Yes, but there's not a 'dont-activate option for VIEW AFAIK.

It should be possible. It's a parameter to the ShellExecute and SetWindowPos 
API functions.

If you have a working example for windows, that would be great ...:)

Just a way to bring up a new window without it grabbing focus.

default as in when using FONTS, it seems to use a different font 
object than when not using FONTS

Don't have that handy for a VIEW call Graham.

I'd like a way of adding a value to the list without changing the 
row focus

that way you can have a row with a status column which for example 
reads 'offline or 'online. Then you could say [status = 'offline] 
would result in some kind of output in that cell, either a change 
of color or font or an image. basically anything that you can do 
with a face

there is a little bit about it in the docs

Henrik.  What you are doing is really amazing.  I really enjoy watching 
projects advance.  Graham, Ashley,  John Niclasen and now you have 
let us watch an immature app/style/control grow into a more mature 
being.  I thank you all for this.

I was hoping Cyphre's style creation talk would be much on this exact 
topic of the steps needed to create a new style and walking through 
the creation of one.  Mainly to see if this was something that was 
rudimentary (even newbies could do) or very involved (only for the 
seasoned programmer).

I use Cyphres list style and asked him at one point what was involved 
in making the familiar style/control that allows selection of items 
in one list and moving them over to a second list either by double 
clicking on it or pressing buttons between the two lists.  I have 
no idea what this style is called, but it sure would be interesting 
seeing how you would implement this with your list style.

Do you plan to use the 'words facet to allow a VIDish way to specify 
(i.e) data.
So you can write :

	view layout [list-view data ["one" "two" "three"]]

Is backeffect a new facet in VID ?

brock, it's been a long and tough ride to get this far on my list 
view. This is actually somewhere around my 11th or 12th attempt at 
a list view where 3 of them were almost as functional as this one! 
So it has been because of this knowledge that it was possible to 
build LIST-VIEW this fast.

Have a look to the current list style. Use Anamonitor and look at 
the 'words facet. It's pretty easy to do.

didec, I'm thinking about it, but it'll break a whole lot of stuff, 
and I hate long variable names :-)

alright, I'll have a look at it

graham, I'll provide a translation table :-)

it's just a search and replace.

When I read 'change-row-here, I just wonder "Where?".

When I read 'change-selected-row, I know where. It also implied that 
it can happen that there is NO selected row.

If there is always a current row (but that is not always selected) 
then use 'change-current-row instead.

there is a problem with GET-SELECTED-COL, because it could be misinterpreted 
as you want to return the entire column, and not the single value 
at a specific position in a row. I've not adopted the term "cell", 
but would that be appropriate? I normally only associate it with 
spreadsheets.

it is a table cell though

ryan, that's a possibility

yeah, however there might be a slight confusion: Currently the -HERE 
functions take in account filtering and sorting, so that you always 
edit or retrieve the right row regardless of these settings. the 
other functions get and put directly in DATA. The behaviour would 
either need to change or I'd have to add another refinement

there could be a /direct refinement.

suggestion:

insert-row - would add a row at the selected position


insert-row/at - would add a row at the position that can be seen 
in the list after filtering and sorting


insert-row/at/direct - would add a row directly in DATA at the position

insert-row/direct - would be ignored? (I don't like that)

I dont know, good luck with that guys. Seems to have a whole lot 
of innuendo's. BTW Henrik, this is a very needed control, I commend 
your efforts!

wait a minute... 'at is an existing function. will that work?

a less elegant option would be to use separate functions: INSERT-ROW, 
INSERT-ROW-AT, INSERT-ROW-RAW

doing it this way, will require a minimum of changes to the code

Since here is a refienement, it assumes you the specify. selected 
behavior by default.

well, thinking about it, /here is a pretty good suggestion. Then 
we'd have INSERT-ROW, INSERT-ROW/HERE, INSERT-ROW/HERE/RAW

Anybody got a rebol routine for saving gifs? Perhaps the one I wrote 
years ago and happen to lose?

I sometimes miss a "soft" SHOW, which only shows faces that have 
the SHOW? flag already to be true. Is there an easy way around that?

and if I have a large set of faces in a pane?

It does get complex, and possibly too slow, but you may be able to 
patch show usefully for your application.

Check out a patched SHOW near the bottom of http://www.lexicon.net/antonr/rebol/gui/mimic-do-event.r

It uses a TREE-FACE function.to recurse to all subfaces of the face 
you are SHOWing.