World: r4wp

OK.  Thanks!  I'll give that a try.

Are you trying it on RaspberryPi? I'll be interested too to know 
if it can work on the RPi already.

That was my plan, to see if it would work there yet.

Bo: this should help you: http://stackoverflow.com/questions/3122448/where-does-opencv-install-its-libs-in-ubuntu

I already found it.  Thanks.

Unfortunately, I am getting the same error trying to run it on Windows 
and Linux-ARM:

# chmod 755 opencv-cam-test
.# /opencv-cam-test

./opencv-cam-test: symbol lookup error: ./opencv-cam-test: undefined 
symbol: isNaN

On Windows I get:


The procedure entry point isNaN could not be located in the dynamic 
link library MSVCRT.DLL.

Gregg, you can DO READ %file in console-pro, or DO READ "url"

You can also give console-pro a file or URL to DO on the operating 
system command line

Thanks Kaj!

Ah, then need to get libcurl...

And sqlite3...

Do I need everything in %MSDOS/Red/?

Pretty much, but it's all there, so you can unpack it in one go

Try VIEW ["Hello" button "OK" [quit]]

The zip doesn't want to download for me. Must be a sign I should 
run lean for the moment, and bug Doc for a few features. :-)

I'm really impressed with all you've done Kaj. And since I'm going 
to start on a DLL interface shortly myself, I will no doubt bug you 
for advice.

I will be mostly offline this weekend as I'm making a trip by the 
coast.

Have a nice break Doc.

I've hacked Ladislav's INCLUDE to let me build .red scripts, and 
I built a console with my mezzanines. When you get back, I'll chat 
with you about it, and send it to test. It compiles fine, but some 
funcs fail if compiled in. If I then paste them into the console, 
they work fine. e.g.

negate: func [n [number!]] [n * -1]

zero?: func [n [number!]] [n = 0]
divisible?: func [a b] [0 = remainder a b]
even?: func [n [number!]] [zero? remainder n 2]
even?_a: func [n [number!]] [0 = remainder n 2]
even?_b: func [n [number!]] [divisible? n 2]
remainder: func [
	a [number!]
	d [number!] "Divisor"
][
	a - (d * (a / d))
]

zero? 0 == true
zero? 1 == false
remainder 10 3 == crash!
divisible? 10 3 == false
divisible? 10 2 == false  INCORRECT
even? 10 crash
even?_a 10 == false  INCORRECT
even?_b 10 == false  INCORRECT

Thanks, Gregg. Did you wait for the zip to be prepared? It takes 
the server around two minutes to create your archive of the latest 
versions

You can also use Fossil to download and update it comfortably:

http://web.syllable.org/news/2012-11-18-20-47-Red-high-level-programming-language-first-alpha.html

ZERO? is already implemented as a fast routine in my common deposit:

http://red.esperconsultancy.nl/Red-common/dir?ci=tip

Ah, I didn't konw it built on demand. I was impatient.

I will steal your ZERO? :-)

That's fine. There are several things in my repositories that may 
not be suitable for inclusion in core Red. Originally, Doc wanted 
natives, but he has also included some mezzanines by now. My Red-common 
repository is self-sufficient, but the others are bindings that core 
Red may not want to depend on, except perhaps functions from the 
C library, because core Red already links to that

Yeah, I'm posting mezz stuff here so it can be evaluated and discussed, 
along with how to organize things, without polluting the core yet.

Funny, it looks like you introduced a crash! type next to error! 
above :-)

:-)

For performance in the interpreter, better write remainder as a - 
(a / d * d)

However, all these simple number handling functions are better written 
as routines for performance

Sure.Since 100'000 calls still seem instant, I'm OK with it for now. 
:-)

Baby steps.

It sure is nice to develop in the interpreter and port to Red/System 
with minimal changes

Agreed.

Also looking at what minimal behavior we need in the kernel to build 
on.

You mean the operating system kernel?

No, the Red/System kernel, sorry.

Just checking. It's a relative notion

Could just say "in Red/System" :-)

In %boot.red, FOREACH says it only accepts a word, but it also seems 
to take a block of words just fine. Should block! be added to the 
word arg spec?

Block!: yes.

Seems I broke something in the commit for the issue #455. All words 
are preset to unset instead of their value.

value/header: TYPE_UNSET  ? ;-)

That shouldn't be a problem because, it's on the symbol creation, 
before it gets a reference to a value...

The issue is the value slots in the global contexts now always return 
TYPE_UNSET...will track it in a few minutes.

It seems the issue is only when using the console.