World: r3wp

join always returns a string unless the first value is of type? series.
>> join 1 2
== "12"
Try
   source join
to see the actual code.

JohanAR, use this instead:

attach: func [a b][to type? a join a b]

thanks!

On to the next question :) Why is '= valid but '< or '> aren't?

My guess is, that it's becuase < and > are part of tag type, like 
<tag>. So it's a restriction to make value evaluation easier internally.

>> blk: [=]
== [=]
>> blk: [<]
** Syntax Error: Invalid tag -- <
** Near: (line 1) blk: [<]
>> blk: [<tag>]
== [<tag>]

>> type? first [<]
** Syntax Error: Invalid tag -- <
** Near: (line 1) type? first [<]
>> type? first [ < ]
== word!

hehe, funny! :)

I guess it is special-cased in the loader.

>> attempt [[<]]
** Syntax Error: Invalid tag -- <
** Near: (line 1) attempt [[<]]
>> attempt [load "[<]"]
== none

This relates to the sandboxing discussion I was just having with 
Paul.

The first error isn't caught because the load happens before the 
attempt is called.

you can ofcourse use to-word "<" but it's not very good looking if 
you have to cover all of <, <=, >, >=, <> etc.. managed to avoid 
the case by using compose instead of reduce though

Is it possible to define a function, which takes another function 
as argument and that the argument func must take exactly two arguments 
in it's turn?

This is how you can pass a function to a function:

>> a: func [a [function!]] [ print a 1 2]
>> b: func [x [integer!] 
y [integer!]] [add x y]
>> a :b
3

Sorry about the formatting; let my try again:
>> a: func [a [function!]] [ print a 1 2]

>> b: func [x [integer!] y [integer!]] [add x y]

>> a :b

3

I think you will need to "manually" check that the supplied function 
uses the correct number of arguments

>> c: func [a [function!]] [if 2 = length? first :a [print a 1 2]]
 
>> c :b
3

>> d: func [][print "no args"]
>> c :d

== none

Mm, I ended up writing something similar to that. Found out that 
I also have to check if third first :a is /local, which would also 
be a valid function

getargs: function [
	:fun "Get arguments from this function"
][
	out pblock
][
	out: make block! 10
	parse third :fun [
		some [
			string!		; Strip comments
			|

   [ [word! | get-word! | lit-word!] (pblock: [any-type!]) opt [string!] 
   opt [set pblock block!] opt [string!] (append/only out pblock) ]
		]
	]
	out
]

a little messy, but it appears to work :)

You don't have to manually check to see if the function takes two 
arguments; this is good, because doing so is awkward. All you really 
need to do is call the function in parentheses ( ) or at the end 
of a code block, so that it can't take more than two arguments even 
if it tries. If it takes less than two arguments, who cares?

Be sure to assign the use or assign the result of the function call 
inside the parentheses, rather than outside them, in case the function 
takes less than the number of arguments you are expecting it to.

Do
    source replace
in REBOL 2.7.6 for an example of this - look for the parentheses.

The ARRAY function in 2.7.6 does the end-of-the-code-block method 
for function value arguments.

If the datatypes the function value is expecting don't match the 
data that your function is passing to it, no problem: A nice error 
message will be generated at runtime that the developer of the function 
value can use to fix their function value. :)

assign the use
 -> "use"

Carl had a nargs mezz somewhere (maybe he posted it on the ml?)

http://www.fm.tul.cz/~ladislav/rebol/argstake.html

; what is the rationale, I wonder of the difference
>> lit-word? first [ 'thang thing]  
== true
>> lit-word? 'thang
== false
>> type? first ['thang thing]
== lit-word!
>> type? 'thang
== word!

type? first ['thang]
'thang is not being evaluated

type? 'thang

'thang is being evaluated. The result of that evaluation is a word!

.....But with 'reduce, 'thang is being evaluated, so:
    type? first reduce ['thang]
    == word!

thang -ks

I can't believe I missed all those REBOL tutorials available on Youtube 
by Nick Antoniacco. Here's the complete list, in case newbies missed 
them too:

http://musiclessonz.com/rebol_video_links.html

Interesting guy http://rockfactory.us/Teachers.html

He is the top one "Nick"

Didn't RT link to those from REBOL.com at some point? If not, they 
really should.

looking at source of 'send, am I right, if there is an error executing 
send, than the port is not closed, this will leave a lot of open 
connections when batching?

Ports are closed when the garbage collector gets them. If you are 
batch sending, the next time you call the SEND function it will assign 
none to the smtp-port local variable, rendering the previous reference 
assigned to that variable into garbage. The next time RECYCLE is 
run, any open ports in the garbage will be closed.

ok, but between recycles, I could reach the maximum connections limit 
of the smtp server?

also, 'send should probably return a value of type logic!

Recycle yourself if you're worried. Interesting idea on the return 
value.

Thanks BrianH, I'm trying to find a bug, I get random port none not 
open errors but they appear randomly everywhere, only thing sure 
is they come only after using 'send , but it may be word! conflict, 
variable rewrite, etc.. just maybe someone has a better idea where 
to look for.. 8/

looks like an asynchronous event error popping in the middle of execution 
of other code, possible??

like this error:
code: 501
description: access error, Port none not open
near:
[if none? i [return none] 
either
]
where: confirm

or

code: 501
description: access error, Port none not open
near:
[if empty? block: reduce block]
where: confirm


oly thing in common, is where: confirm and that I see only in net-utils, 
any idea ?

even had a recycle..
code: 501
description: access error, Port none not open
near:
[recycle 
wait 0.2
]
where: confirm

here is the net-log, a couple of this (no problem)::

Net-log: ["Opening" "tcp" "for" "esmtp"]
Net-log: [none "220"]
Net-log: "220 domain.com running core"
Net-log: [["EHLO" system/network/host] "250"]
Net-log: {250-domain.com hello mail.local (10.5.32.4)}
Net-log: "250-PIPELINING"
Net-log: "250-8BITMIME"
Net-log: "250-BINARYMIME"
Net-log: "250-CHUNKING"
Net-log: "250-ENHANCEDSTATUSCODES"
Net-log: "250-ETRN"
Net-log: "250-AUTH CRAM-MD5 DIGEST-MD5 NTLM PLAIN LOGIN"
Net-log: "250 SIZE 2147483647"
Net-log: ["Supported auth methods:" [cram plain login]]
Net-log: ["MAIL FROM: <[majordomo-:-domain-:-com]>" "250"]
Net-log: "250 2.1.5 sender OK"
Net-log: ["RCPT TO: <[user-:-domain-:-com]>" "250"]
Net-log: "250 2.1.0 forwarding to [user-:-domain-:-com]"
Net-log: ["DATA" "354"]
Net-log: "354 send your message"
Net-log: [none "250"]
Net-log: "250 2.6.0 message received OK"
Net-log: ["QUIT" "221"]
Net-log: "221 2.0.0 domain.com closing connection"

than suddenly this one:

Net-log: ["QUIT" "221"]
Net-log: ["Opening" "tcp" "for" "esmtp"]
Net-log: [none "220"]
Net-log: "220 domain.com running core"
Net-log: [["EHLO" system/network/host] "250"]
Net-log: {250-domain.com hello mail.local (10.5.32.4)}
Net-log: "250-PIPELINING"
Net-log: "250-8BITMIME"
Net-log: "250-BINARYMIME"
Net-log: "250-CHUNKING"
Net-log: "250-ENHANCEDSTATUSCODES"
Net-log: "250-ETRN"
Net-log: "250-AUTH CRAM-MD5 DIGEST-MD5 NTLM PLAIN LOGIN"
Net-log: "250 SIZE 2147483647"
Net-log: ["Supported auth methods:" [cram plain login]]
Net-log: ["MAIL FROM: <[admin-:-domain-:-com]>" "250"]
Net-log: "250 2.1.5 sender OK"
Net-log: ["RCPT TO: <[user-:-domain-:-com]>" "250"]
Net-log: "250 2.1.0 forwarding to [user-:-domain-:-com]"
Net-log: ["DATA" "354"]
Net-log: "354 send your message"
Net-log: [none "250"]
Net-log: "250 2.6.0 message received OK"
Net-log: ["QUIT" "221"]
Net-log: "221 2.0.0 domain.com closing connection"


error code: 501
description: access error, Port none not open
near: [value: either series? :value [copy value]]
where: confirm

note the difference that seams to trigger the problem is the first 
line, the 'quit, where does it come from?