World: r3wp

Henrik, could you test ticket 744? I don't have a Mac.

Re: #744. They no longer crash but the "launched" script uses the 
same console session as the original script so keyboard input and 
screen output gets mixed up. This is different from the behaviour 
under Windows where R3 opens a new console session for the "launched" 
script.

The following session shows that launch does not effectively run 
a script as a separate process and return immediately.

>> ans: "yes"      

== "yes"

 >> do [ while [ans <> "no"] [ans: ask "(l)aunch, yes or no ?" if 
 ans = "l" [launch %launched.r]]]
(l)aunch, yes or no ?l
Checking for rebol.r file in /Users/peter/Desktop/Rebol3/
Checking for user.r file in /Users/peter/Desktop/Rebol3/


Evaluating: /Users/peter/Desktop/Rebol3/launched.r

launched script
Do you want to continue running the launched script?yes

yes

Do you want to continue running the launched script?yes
yes
Do you want to continue running the launched script?yes
yes
Do you want to continue running the launched script?no
no
(l)aunch, yes or no ?y
(l)aunch, yes or no ?y
(l)aunch, yes or no ?no
== none

The source of %launched.r is:

REBOL []

print "launched script"

ans: ""
while [ans <> "no"] [

   print ans: ask "Do you want to continue running the launched script?"
]

Same problem with call:

>> call "pwd"

/Users/peter/Desktop/Rebol3

== none

I'll mark it as tested then, and we can post any tweaks to the behavior 
as new tickets.

Would you like me to open a new ticket for the above?

I checked call on Windows and it doesn't work either. I opened ticket 
757.

CALL works correctly on Windows. The behavior you describe in ticket 
757 is by design. See the comments for that ticket.

The Mac version should work like the Windows version does.

I am not assuming that it does though - I can't test it.

The mac version appears to be working. At least call "ls" gives me 
a directory listing.

and returns none

Could you post system/options from R3 on Mac? Just run R3 and type 
in system/options

I can't paste here, as copy/paste is broken. Will be a couple of 
minutes...

http://rebol.hmkdesign.dk/files/sysops.txt

I look forward to AltME 3.

I look forward to a VMWare with working copy/paste :-/

Does LAUNCH work for you?

It launches a script alright, but am not sure what proper behavior 
would be for returning.

NONE is returned, even though the script itself returns an integer. 
I suppose that's correct?

Yeah. CALL doesn't return a value yet. I'm waiting for the host source 
before I make proposals.

Or rather, CALL retrns none.

LAUNCH also returns none.

>> source launch
launch: make function! [[
    {Runs a script as a separate process; return immediately.}
    script [file! string! none!] "The name of the script"
    /args arg [string! block! none!] "Arguments to the script"
    /local exe
][
    if file? script [script: to-local-file clean-path script]
    exe: to-local-file system/options/boot
    args: to-string reduce [{"} exe {" "} script {" }]
    if arg [append args arg]
    call args
]]

oh

I see bugs already. LAUNCH could use a revamp like LOAD.

could you explain?

Not correct for script = none. Might be some other tweaks too.

Does R3 have a way to launch a Rebol console?

apart from calling the R3 binary.

LAUNCH, perhaps?

I don't get the question, sorry. Please explain what you want.

sometimes you want to start a Rebol console from an encapped application.

but you don't want to have to need to have rebol.exe locally.

That can be used to debug the application with access to all the 
variables etc.

Encapping is going to be different in R3, as is the console. It is 
intended that a graphical console will be developed, and callable 
from any graphical host. Hosts will be the new encapped apps.

a graphical console will be good.

We want our REPL :)

How is error handling handled?

I suppose that restricting console access for encapped apps would 
involve simply not including console code?

Similar to R2, though the errors start disarmed and need to be armed 
with DO. CAUSE-ERROR wraps this.

Henrik, sure :)

I have some radical ideas for a console. Carl will hate them. :-)

Carl doesn't have to use them - behold the power of modules.

That's good. Perhaps for R3/Plus

Ah, vector! fixed, great! :)

#460 not fixed

Now it's not possible to make signed/unsigned vectors using "make 
vector! [- 16 integer! 100]" ? Or is the syntax different? I cannot 
find it anywhere.

Ah, #460 is not fixed because maximum-of doesn't work at all. This 
should be first line in the 'maximum-of function: "if not skip [size: 
1]"

Rebolek - by testing in console - you have wrong order of arguments, 
try:

make vector! [integer! 16 100]

make vector! [unsigned integer! 16 100] ; there was change from hyphen 
to work "unsigned", dunno why - http://curecode.org/rebol3/ticket.rsp?id=350