World: r3wp

rebol.org?

Graham: on close you can just close the port (so that it is removed 
from the wait list). apparently not doing that causes an error (not 
sure if this is new, or i just never catched it before; anyway i 
think rebol should not be calling the awake after the port has been 
closed)

Anton - "MOLD can produce a string with braces instead {}"  - That's 
true, but only if the string contains new line or " char. You cannot 
use such a chars in file names, so I think it's pretty safe to use 
mold for the above example.

It's possible, although certainly rare.  I've accidentally made files 
with newlines in the filename.  And I just checked - the quote mark 
is also valid (on linux)

Also, length is an issue:
>> mold [Long line?]
== "[Long line?]"

>> mold [Long line? lets make it really really really really long 
now]

== {[Long line? lets make it really really really really long now]}

(Rebol 2.7.6 on Ubuntu)

in R2 ... isn't this wrong?

>> a: first do "['dd]"
== 'dd
>> type? a
== word!
>> a: to-lit-word a
== 'dd
>> a
== dd
>> type? a
== word!

the way I see it, the above means that words always aggressively 
evaluate lit-words, 


shouldn't a value returned as a lit-word stay that way until an eval 
of your choice is perfomed on it?

This may help, but I don't play with lit-words very often:
>> a: first do "['dd]"
== 'dd
>> type? a
== word!
>> type? :a
== lit-word!

yeah, I guess you're right... I didn't think about it this way.


its the purpose of a word to evaluate its content, and get-words 
to return them un-evaluated.


still, this specific case isn't very obvious.  guess I put my "newbie 
hat" for a few minutes there  ;-)

in a dialect, via block-parsing, this is a very tricky detail!

I suggested quite a few times to Carl to specifically suppress this, 
but my proposal was not accepted in this case, maybe if you wrote 
it as a CureCode wish, Carl would find out I am not the only one 
suggesting it?

As Iskata underlined it, i don't see anything specific.
a -> is evalatued, :a is not.

The VM always tries to evaluates/reduces the words, not the get-words.

*evaluated, *Izkata

but lit words are a datatype.   when such a lit word is encountered 
in a do block and evaluated, it is reduced to a word.  that is ok.


but when it is *stored* AS as a lit-word, it should not be evaluated. 
 remember that words may contain words, which will be evaluated.


so why should lit-words be evaluated too?  the basic word containing 
a word already does that.

Hmm... let me check that...

the specific is that 

a: 'z      and     a: to-lit-word 'z

are not equal expressions.

so why should evaluating a also evaluate z in the second form.

since it will in the first form.

its the same as saying that blocks would ALWAYS be reduce everytime 
you evaluate a word which holds a block.

to me a lit-word is not a reference, its a token.  nothing more. 
 so evaluating one should just return itself, like strings & blocks.

sorry... evaluating a *word* which holds a lit-word, should just 
return the original lit-word.

Ok I see your point now.

But I have to admit I've never been bored by this cascade evaluation.scheme 
:)

@Gab .. since the sub-port is closed, it can't receive events to 
pass on to the main port so yes, it shouldn't receive any more events.


Just wondering why putting a 'true in the handler at the end of the 
'close event handler also causes REBOL to exit the View event handler 
as well ...

it quits the wait []  call.

the same happens if you return true from the view wake-event.

I guess it must.

In the async-http ( or any other protocol ), when exactly is the 
'init function called?

Something I'd forgotten ... but you can set custom headers using 
the header keyword.  Not documented though.

read/custom url [ header [ cookie: "blah blah" ]]

The header keyword is not needed if you do a post though.

Anyway, my async-fetch-s3object is now working :)

Oldes - "only if the string contains new line or " char" - that's 
wrong. It depends on the string length. also, mold escapes characters 
the REBOL way which you DON'T want.

>> "a long string"
== "a long string"

>> "a very long string with no newlines or quote characters, that 
still uses braces instead of quotes because it's long"

== {a very long string with no newlines or quote characters, that 
still uses braces instead of quotes because it's long}

the init function is called when the port is created from a url! 
or block! spec. eg. p: make port! ahttp://...

calling OPEN on a url! or block! spec also creates the port! first, 
so init is called.

So, after the port is created ..

i think the native code does some very basic initialization of the 
port! then it calls init

ok.  thanks.

Is there a good method to conditionally remove a char in a string 
without having to manage the string? That is: Find char X at location 
Y and remove it if it's there, otherwise just return the string as 
is.

good method = fast method

the first occurence only ?

yes, I was thinking of the case of removing a conditional comma at 
the tail of an assembled string as quickly as possible.

first occurrence = just one occurrence

something like...
head any [remove find string char string]

hmm.. yes, that might work, thanks

Parse  may be faster, though

especially within R3

hmm, the string is required to be present twice, but I guess that's 
ok.

Probably a little faster with R3...
also string remove find string char

Regarding Romano's atcp protocol, http://www.rebol.it/romano/atcp-protocol.r
 there is a little bug at the bottom of the page I think

He has 
		if find system/components 'ssl [
			net-utils/net-install 'assl self 0
			net-utils/net-install 'atsl self 0
		]

but I think he meant

		if find system/components 'ssl [
			net-utils/net-install 'assl self 0
			net-utils/net-install 'atls self 0
		]

Good catch, Graham

I emailed him.... hope he has time to spare from his Java development 
to fix it :)

Gab's async protocol, the awake event takes two parameters
Romano's atcp protocol, the awake event takes one parameter ....