World: r3wp

sunanda, that is a hell of a nice cheat!  never thought of that one.

pity it can't be made into a function  :-)

that is not the current context, rather, the current "function name". 
(and Max... I think I posted that a long time ago ;) I was using 
it in my profiler)


Awi, the thing is that functions don't have names. Also, by "context" 
in REBOL we mean something closer to what other languages call "frame", 
however REBOL has no concept of a "current frame" or of a "parent 
frame". R3 though has (had?) a STACK function that will probably 
give you want you want.

You can get the name of the word that the function was called through 
in the current call stack. This isn't *the* name of the function, 
but it's *a* name of the function. Note: "pity it can't be made into 
a function  :-)" This is because if you put that trick in a function, 
then *that* function is the function that will be referenced.

R3's STACK returns information about the entire call stack, though 
the 'debug security setting would put some limits on it. SECURE is 
disabled in recent builds so we don't know what the limits are; we 
could try older versions to find out what those limits used to be. 
Also, see http://issue.cc/r3/885for a bug in STACK.

Thank you guys, I can never learn enough of Rebol.

>> a: "1.  "  ;here's  a space
== "1.  "   
>> layout [text a]
>> a 
== "1."

That looks more like a quirk of the function, than as something one 
should explain to the beginners.

The function looks like trying to be "cleverer" than the user.

I feel  that 'a should not be modified?

probe select system/view/vid/vid-styles 'text
...
if all [not flag-face? self as-is string? text] [trim/lines text]
...

Just don't know how one should set the flag as-is... I'm not using 
view many years

Googled....
>> a: "1. " layout [text as-is a] a
== "1. "

Ok

[trim/lines text]   or [trim/lines copy text]  ?

the latter one, if you don't want the original string modified.

0MQ binding.. that is awesome! Does/will this work for R2 also maybe? 
:)

ups .. wrong chanel

VID question (R2): Is it possible to get the cursor position in the 
scroll-line event? I wanted to use the scroll to zoom in (like in 
google maps), and to zoom in to the right area, I needed the cursor 
position. Thanks for the help.

You can get cursor position using CARET-TO-OFFSET

Unluckily the face I am using is a plain panel, so no text there. 
I just tried PRINT CARET-TO-OFFSET pnl-map "", and it returns none.

That's the mouse pointer, not the cursor (sorry, terminology isn't 
very portable to REBOL).

You probably need to remember the offset from the last 'move event, 
since the scroll-line uses the event offset parameter itself.

yes, you need to hack the event engine a little bit.   


As gregg says, you need to have a memory of the last move event to 
get its position  and store it (you can do this with an event-handler). 
 glayout and GLASS do this for handling scrollwheel events.


what I also do is find the top-level face which is under the mouse-cursor 
and fire off my own events from the scroll-wheel instead of relying 
on a text field.

again, you can trap the scroll-wheel events in the event handler.


if you want to have a ready-made solution, download glayout.r from 
rebol.org  and look at the hacked WAKE-EVENT function.


it already does all of this and wraps it up by adding new function 
you can add to your face/feel  object in order to handle scroll-wheels.

the wake-event function needs a few other functions which are all 
in the glayout module, but it should be easy to keep just what you 
need and run that before the rest of your script.

Thanks for the help, I am still digging around, will let you know 
the result. I will have to translate all that to RebGui though.

Is there a better way than this to convert a block to string, then 
back to block?
blk-to-send: reduce ['my-function 1 2 3]
str-zmq: mold blk-to-send
blk-received: first to-block str-zmq
Thanks for the help!

use LOAD instead of FIRST TO-BLOCK

Thanks Rebolek!

Is there a way to start the word browser directly without going thru 
the viewtop?

do http://www.rebol.com/view/tools/word-browser.r


You can <alt>-click (or right-click) icons in the viewtop to see, 
where the script come from.

Awesome

Is there a way to to get the time in a datetime value without using 
refinement?
>> d: now
>> d/time
== 12:09:58

Is there something like select d 'time OR pick d 'time ? Thanks.

>> fourth now

== 13:23:31

Works in both REBOL 2 and 3.

This works in R3:
   >> pick now 'date
    == 30-May-2011

This is cool! Thanks!

I encountered another problem:
>> 1.48297457491612E-2 + 0.985170254250839
== 1.0
>> arccosine/radians 1.48297457491612E-2 + 0.985170254250839
** Math Error: Math or number overflow

** Near: arccosine/radians 1.48297457491612E-2 + 0.985170254250839

>> arccosine/radians probe (1.48297457491612E-2 + 0.985170254250839)
1.0
** Math Error: Math or number overflow

** Near: arccosine/radians probe (1.48297457491612E-2 + 0.985170254250839)
>> arccosine/radians 1.0
== 0.0

That's a rounding problem. You can check such numbers in R3 by:

>> to binary! 1.48297457491612E-2 + 0.985170254250839
== #{3FF0000000000001}
>> to binary! 1.0                                     
== #{3FF0000000000000}

I don't understand it, from 'probe', it seems like it already returned 
1.0, why arcosine/radians still get > 1.0

>> system/options/decimal-digits: 17
>> to decimal! #{3FF0000000000001} 
== 1.0000000000000002

>> (1.48297457491612E-2 + 0.985170254250839) = 1.0
== true

decimal-digits tell, how many digits, you wanna see. But when you're 
out there at 15-17 digits, you can't count on the last one, so it's 
choosen to be not seen.

= (or equal?) is not exact. Use == (strict-equal?)

>> (1.48297457491612E-2 + 0.985170254250839) == 1.0
== false

Makes sense?

Ok, I got it now. Thanks Geomol!

welcome

Is there a best practice for such situation? Always rounding to 0.0000000000000001 
?

well, I just tried (R2):
>> (1.48297457491612E-2 + 0.985170254250839) == 1.0
== true

In R2, integers are 32 bit, in R3 64 bit.