World: r3wp

ref-a: none
val-a: "A"
ref-b: true
val-b: "B"

f/opts [a: ref-a if ref-a[va: val-a]  b: ref-b if ref-b[vb: val-b]]

but then it needs a composeto avoid name-clashes. maybe not the best 
idea.

you lost me :-) but it's ok, I'll try something with build-path...

You can do this:

do to-path reduce ['do-a-function either condition ['ref][none]] 
<argument>
but if it's less clumpsy, you deside!

; Try this way, no reduce or to-path...

do either condition ['do-a-function/ref] [:do-a-function] <argument>

; Or this
do pick [do-a-function/ref do-a-function] condition <argument>

Graham:

>> word: to lit-word! 'literal
== 'literal
>> parse [literal] [word]
== true

lots of different solutions to my problem... thanks everyone :-)

a comment on ATTEMPT. I think it's a wonderful function, but don't 
overuse it because it can make your code hard to debug when nothing 
happens where there should be an error.

true

i only use it for single calls.

Steve, when it comes to SIXTH, I wouldn't worry too much.

Is there a way to set and read windows environmental variables ?

'get-env ?

thanks.

Does it work ?

works with "path" ..

env

'scuse me...

Perhaps I need to reboot for rebol to pick up user env variables 
?

>> get-env "APPDATA"

== {C:\Documents and Settings\Gabriele Santilli\Dati applicazioni}

I had to reboot for this to work.

You mean, you set a new user env variable, and rebol wasn't able 
to pick it up straight away ?

logout and login probably sufficient

didn't try that ...

i'm running into a strange issue with the prerebolizer... it is taking 
a part of one line from my main rebol file and then a few blocks 
of another rebol file and overwriting another seemingly random rebol 
file with this data... has anyone run into this problem or have any 
idea what could be causing the problem?  i am using Preprossor 2.0.0

I haven't seen that Robert. Can you narrow it down to a very small 
example?

>> o: make object! [time: does [now]]
>> third o
== [time: func [][now]]
>> reduce third o
== [7-Dec-2005/3:23:23+1:00]
>> third o
== [time: 7-Dec-2005/3:23:23+1:00]

>> o: make object! [time: does [now]]
>> third o
== [time: func [][now]]
>> reduce copy/deep third o
== [7-Dec-2005/3:25+1:00]
>> third o
== [time: 7-Dec-2005/3:25+1:00]

Why is the block not copied?

ah, using GET instead of THIRD allows me to copy it

I have one question regarding security - can actually rebol words 
be secured/protected? I noticed, when looking at get-net-info and 
trying to execute it, that get-reg is probably being unset internally 
after get-net-info gets executed. Isn't it sign rebol's security 
model is not sufficient?

I am not trying to say that we need modules, but I would like to 
know your opinion? Can some rebol word/value be really protected? 
We have 'unprotect available, so actually what is 'protect good for?

can I e.g. create word in context, which would be visible only in 
terms of such context? Being it value, function, whatever, prevented 
to be outputted by 'get, 'probe, 'print, 'source and other functions?

Henrik: what happens there is that the word TIME is set to the result 
of the function.

AFAIK not. 'protect is for debugging AFAIK.

But i can work secure if you use 'secure. The way is, as long as 
i dont touch user-data, i am secure. Thats why rebol has get-reg 
on startup, no user-code loaded, no harm possible. Before i touch 
user-code (or data), i tighten security as much as possible. unset 
dangerous words (if i am paranoid i disable "make struct!". I use 
secure to restrict file-access only where needed, disable shell and 
library, maybe network-access. A script/bug can not undo this things, 
and IMHO that is quite secure.

Petr: you're not talking about security, you're talking about encapsulation, 
i.e. making parts of your code unaccessible from other code.

Henrik, copy/deep does not copy functions. and you have a function 
here, try mold/all instead of probe.

volker: it's just that TIME is bound to the object, and he's changing 
its value to that date...

so even if copy/deep copied functions, he'd still get that result 
above.

Got it. was to hastily. good riddle :)

volker, I don't get what serialization with mold/all does?

i misunderstood, you can ignore that. mold molds functions like their 
sourcecode. mold/all molds functions specially, so that you see what 
a function and what only source is. Thought that would give a hint 
to the problem.

I don't think it's useful to solve my problem, but very interesting 
anyway, so I asked :-)

I'm actually looking for a way to get input data as a block:

[a: 1 b: 2 c: 3]

made into an object which is:

make object! [a: b: none c: does [something to make the c value]]

so I would get an object like this:

make object! [a: 1 b: 2 c: <return value from object function>]

Gabriele - encapsulation is kind of security. And in fact, if some 
code is encapsulated, or not being able to be called from outside 
of some context, it is kind of secure

yes, but that is another level of security, where you trust some 
modules of a script but not others. i guess that will come with REBOL 
3

Interesting question on ml:

join 'a 'b
== "ab"

type? join 'a 'b
== string!

Is that correct? Should not be a word instead?

My guess is that as word! values are not series! then join doesn't 
produce them...

In the same way that, eg.
>> join 1 2
== "12"

>> join true false
== "truefalse"