AltME groups: search

Google browser extensions are written in Javascript - if you want 
to use REBOL, you'd either need to compile to Javascript or be a 
plugin (which is supported by Chrome). The closest thing in R3 to 
Google's browser extensions is the module!.

Have you looked to the LUA model? I posted you a link privately - 
seems similar ...

As for the JIT, I could write the compiler in REBOL and generate 
the intermediate code of the JIT, then pass that intermediate code 
to the JIT with a command. The JIT would then generate a function, 
add it to its list, and return the list index as an integer. That 
integer can be  used to create a new command!, which RX_Call can 
dispatch to the internal JITed function.

Since extensions can't be statically linked with R3, I can wrap a 
LGPL JIT like libjit. It should work great. I'll be stuck with the 
RX data model, but that cold be plenty for the types of functions 
you wold write just for speed.

a 64-bit issue?

Not a 64bit issue, a Vista+ issue.

Gabriele, I probably worded my last comment a bit strongly. (I don't 
want to enflame the situation..) Of course I highly value your judgement, 
opinions and suggestions. I should say, though, that I read your 
previous comment as somewhat dismissive of RobertS concerns / problem.

With your #[{ ... }]# suggestion, I think it looks pretty good, and 
could be extended with specification of an end delimiter. With so 
many computer languages of all different syntaxes, having a way to 
specify the delimiter allows it to flex to any situation. 
We could add a second string to the block, eg.
The end delimiter is END:
	#[{END} {... END}]#
or a simple newline char:
	#[{^/} {... ^/}]#
or a secure hash:
	#[{92838929838} {...92838929838}]#

The R3 docs for extensions display an example for reversing a string. 
I wonder if it's the same code as used internally for the built-in 
REVERSE function?

My PHP page has no escaped characters becuase I can use the so-called 
"here document syntax".  That is critcal for my templating needs. 
 If your templating will only be done by rebol developers, it is 
of course a moot point: Rebol developers do not need heredocs/  But 
are all users of rebol going to be rebol developers?  That is the 
dialect issue.  And in the case of PHP (jsut an example) many PHP 
users are not PHP developers.  The wiki page has examples from python 
or ruby if the mention of PHP offends purists .. ;-)

not to vex, I hope, but I just saw that PHP 5.3.0 added so-called 
"now docs" which are what I call literal strings or verbaibm strings 
in that no substitiutions occur so no escaping is done.  I believe 
this how at least comment {should behave in R3 if I use a { as in 
just print {this} { to get the effect you want }  which comment is 
not the same which has to explain that you do whis without the escapes 
- which is fine if thosei are the only and very  escapes in question 
...

The PHP restriction on HEREDOC and NOWDOC is that the closing identifier 
( a label, if you like) must appear after a nl and there must be 
ONLY a nl immediatly after its ; terminator  - a restriction  which 
seems not at all onerous in practice now that I am using them

In Firebird's sql language, we can specify the delimiter and then 
restore it  after dealing with a bunch of code/text.

No need for me to press enter.
Pekr, is your error true for all types of files?
Is capitalization a problem here? (try capital "C:".)

Are you on Vista, isn't the "windows" directory a protected system 
directory - perhaps that's causing an issue.

it does not seem to know Dir - maybe a path lookup problem?

That's a rebol error message, obviously, not a cmd one.

I don't know if you were in a cmd shell or a rebol one (!)

(Start with a fresh rebol console.)

Anton: my comment may be dismissive, but the point I'm making is 
that that is just my opinion, I'm not sure why I should be forced 
to "like" a heredoc syntax :) I understand that it is very handy 
on occasions, but the price is a ugly script, which I'd rather not 
pay if I can avoid it easily enough (for example, if I were generating 
a lot of PHP or JS, I would just build a dialect to do that, instead 
of using strings). Again, though it's funny for me to have to repeat 
it again and again, this is just IMHO, Robert should be free to do 
things his own way, and if REBOL can provide a syntax for heredocs 
without creating other problems I'm fine - I'll just avoid it (and 
I'll hope not many people start using it ;)

Gabriele, I had a look at a few of the heredoc formats and some of 
them do look quite ugly. I admit that I actually haven't used heredocs 
myself yet, only come across a few snippets here and there recently, 
so I guess I don't have much experience with them yet, and can't 
tell what problems they may cause. My initial thought was that they 
would be a good idea, since you could just paste foreign code straight 
into your script.

At a command prompt, call cmd /? to get help on cmd.exe and its commands.

Sounds like Pekr is trying to capture the output of a directory listing 
in R3, is that correct?

anyway Pekr, just use:

call "cmd dir > somefile.txt"

then you can just read the file back into a string.

Paul - I know that is an option ... but what is it good then to get 
'call output into console, if it can't be even easily captured into 
a string?

Isn't that a limitation of R3?

Is it me, or the example to read a dns address is broken ???
>> print read dns://www.rebol.net

Geez...

>> first wait read dns://www.rebol.net
read
== 216.240.11.203

>> first wait read dns://216.240.11.203
read
== "www.rebol.net"

A little bit obfuscated, but it works...

Nicolas, VIEW in its built in state doesn't work properly, so do 
something like this:

>> load-gui
>> a: load http://arkandis.com/images/blog/google.gif
>> b: make gob! [image: a]
>> view b

What is the point of /any on dir?

As far as I can see, as soon as you add a wildcard char, dir?/any 
always returns true

    dir?/any %ddddff.file
    == false                              ;; no wild ==> false
    dir?/any %ddd?dff.file
    == true                                ;; wild ==> true
    dir?/any %ddd*dff.file
    == true                               ;; wild ==> true

    dir?/any http://www.test.com/file.html
    == false                             ;; no wild ==> false
    dir?/any http://www.test.com/*.html
    == true                              ;; wild: true

looks like a bug

definitely a bug. please add it to curecode, thanks.

It's already there - add your code as a comment on the existing DIR? 
/any ticket.

I have heard at least one other opinion, that name "command" might 
be too worthy to waste on simple and signle thing as wrapping merely 
a funciton call. Of course even Devices have commands, but those 
are not probably rebol level related and influence nothing ...

Carl said that devices would be a variant on extensions.

No, I said, that device commands probably don't influence rebol level 
code, so that we are eventually free to rename extension command 
to extension! term. That was just a thought, in case we would have 
some better usage for 'command term later. Nothing really important 
imo. Extension commands are not much advanced - they are just descriptor 
for particular meaning.

and to just describe meaning of wrapping a call, I though external!, 
extern!, ext!, wrap! could be used instead. Probably too late to 
try to convince anyone :-)

Those are all modifier words. We need a noun.

btw - "external", english-wise, can't be thought as a noun?

I just read it, and was about to reply to your counter-post, just 
to one point.
External

 is an adjective. When used as a noun, its subject is implied. Bad 
 form.

as for Chat - is there any shortcut to know, where I am posting by 
default? I mean - the only commands I use is "n" for new posts. The 
I want to add a reply. But adding a reply is different then simply 
posting "p". But - when adding a reply "r", I specify the message 
number, so I can be sure, where it goes. But where does simple "p" 
post it?

The "original post" you are replying to is displayed by n or nn with 
a > on the left of it. I usually enter the number for the post I 
want to reply to first, which displays in and sets the >. Then I 
use r. I wish there was an rp command...

It's why I always go somewhere on purpose, either by typing  in the 
header number manually or reading a message on the subject.

I've had a week of dealing with personal/family stuff. Some of us 
don't have blogs/twitter :-/

Is this a bug?
blk: [a: 23 b: 34]
to-object blk
** error - invalid argument
make object! blk
== make object! [a: 23 b: 34]

What's the difference between the make and to functions?

As for bug#706, it's a file permissions thing. You can set the permissions 
of the directory that REBOL is installed in, so that only authorized 
users can write to it (generally, only the administrator). This means 
that %rebol.r can't be written by untrusted scripts. If you let %rebol.r 
be loaded from just any diirectory, regardless of who has permissions 
to write to the directory, then you have enabled %rebol.r to be used 
as a malware installer.

I will have to think about it, but if you think about script's current 
dir = user has already right to place script there, why should placing 
user.r there be a security violation? Well, truth is, that most of 
the time I am local admin, so my knowledge is limited here ...

The %rebol.r file is only meant to be used by the person who installs 
REBOL, for the purposes of adding custom stuff that is specific to 
a particular computer. Loading standard libraries will be handled 
by the module system instead. This is the same reason why we are 
planning to replace %user.r with a preferences file.

No install needed. The whole %rebol.r concept is meant for use by 
sysadmins who want to control what programs can be run on their system 
- there is even a Windows Policy that lets you limit which directories 
programs can run from. Regular users would never use %rebol.r at 
all.

In general, if I were a sysadmin, I would not load any functions 
in %rebol.r - I would just put a call to SECURE in there, to limit 
access to my resources. This would help me lock down my user's scripts. 
I wouldn't use %rebol.r at all for user's personal computers, just 
for public-use computers.

%rebol.r is very usefull when it works under R2 (which it doesn't 
most of the time).  you can put a few generalized functions you always 
want access to... setting lib pats, protecting/securing them, loading 
some libs by default, etc.

btw the extensions really are a god send... its cool you wrapped 
them into modules.  good work.

or a way to detect whether capslock is down or up?

If by alt-up/alt-down you mean alt-key modified up-key/down-key in 
opposite to plain up-key/down-key, with current R2 that's not possible 
to detect, afaik since not all OSes R2 has been designed for feature 
such a system specific key. Detecting caps lock isn't supported, 
too.

When investigating the creation of a MAP function in REBOL 2, I found 
that sending functions with refinement to map required some extra 
work (the need for a DO). The rules about get-words as arguments 
has changed in REBOL 3. Maybe I should talk to Carl about it, but 
I could discuss it with you guys first to not disturb Carl too much. 
First a REBOL 2 version of MAP, that can't cope with refinements:


>> map: func [:f l /local r] [r: clear [] foreach i l [append r f 
i] r]
>> map sine [0 30 90]
== [0.0 0.5 1.0]


f is the function, l the list and r the result. i is an item in the 
list. The critical part is

append r f i


The function f is evaluated taking the argument i. Easy to read and 
understand. But it can't cope with refinements, which are seen as 
the path! datatype. Example:

>> map sine/radians reduce [0 pi / 6 pi / 2]
== [sine radians sine radians sine radians]


This can be fixed by putting a DO before f. Now it works both with 
and without refinements:


>> map: func [:f l /local r] [r: clear [] foreach i l [append r do 
f i] r]
>> map sine [0 30 90]
== [0.0 0.5 1.0]
>> map sine/radians reduce [0 pi / 6 pi / 2]
== [0.0 0.5 1.0]

In REBOL 3, the function is not evaluated:


>> map: func [:f l /local r] [r: clear [] foreach i l [append r f 
i] r]
>> map sine [0 30 90]
== [sine sine sine]

Including DO just makes it worse:


>> map: func [:f l /local r] [r: clear [] foreach i l [append r do 
f i] r]
>> map sine [0 30 90]
== [make native! [[
        "Returns the trigonometric sine."
        value [number!] "In degrees by default"
        /radians "Value is specified in radians"
    ]] make native! [[
        "Returns the trigonometric sine."
        value [number!] "In degrees by default"
        /radians "Value is specified in radians"
    ]] make native! [[
        "Returns the trigonometric sine."
        value [number!] "In degrees by default"
        /radians "Value is specified in radians"
    ]]]

To make map behave correctly, I have to do something like:


>> map: func [:f l /local r] [r: clear [] foreach i l [append r do 
reduce [f i]] r]
>> map sine [0 30 90]
== [0.0 0.5 1.0]
>> map sine/radians reduce [0 pi / 6 pi / 2]
== [0.0 0.5 1.0]

Is this ok and accepted behaviour? Will it break many scripts?
(Sorry for the long post, but I wanted to be precise.)

BrianH: thinking about submitting 'Call function for fixes. I gave 
it some thoughts, and I think I am not satisfied with answer, that 
'call will be part of Open Host code - we are waiting for host code 
for 2 years, and there is no guarantee, that we will see it anytime 
soon. I think call should be fixed, or it is big let-down in comparison 
to R2. It messes with console in such a way, that it looks inconsistent, 
and its output can't be trapped easily. It might be a show-stopper 
for some ppl, in regards to R3 deployment. What do you think? Maybe 
it can be improved a bit? We are seening good changes to many natives, 
so why to wait with call for host code release?

Speculating about set- and get- datatypes in relation to:
http://www.rebol.net/cgi-bin/r3blog.r?view=0229#comments


In R2, we have get-word!, set-word! and set-path!. R3 brought us 
get-path! too. Is it a good idea to have things like get-paren! and 
maybe even get-block! and set-block! ?

Carl's set-word! in a block problem could be solved with:

user: [name: "Steve" age: 38]
user/:(age:)


About get-block! and set-block!, today we can set many values with:

set [a b c] [1 2 3]

Why not just write:

[a b c]: [1 2 3]

And a get-block! like:

:[a b c]

should return a block with values like

reduce [a b c]

Just thoughts.

user: [name: "Steve"  age: 38  country: France]
I just have a question. 
How Carl knows my Vital record...

[a b c]: [1 2 3] and :[a b c] , very nice, very rebolish 8)

About set-block and get-block.

If it's only to save the use of SET and REDUCE, i think it's a little 
luxurious.

It would have to work like set-word!s for contexts, to have more 
meaning. Could save a line of code.

...
it's not a good example ;-)
i do this, context [foo: fum: 3]

you can with R3
>> set c: construct [a: b:] [3 5]
== [3 5]

>> c
== make object! [
    a: 3
    b: 5
]

or more generally: it can occur often that you need to set multiple 
words sequentially. there should be a syntax for that.

set c: context [a: b: none] [3 5]

what happens when you say:

[a [b c]: d]: [1 [2 3] 4]

I'm not sure it is. :-) but many things don't seem very useful on 
the surface. I'm still thinking in terms of setting mulitple words 
with multiple values in one operation. I hate picking words out of 
a block, one at a time. It becomes more powerful when you replace 
the blocks with words. Then you can use the same program structure 
to set one-to-many, one-to-one, many-to-one and many-to-many words.

yep... like :

z: [b c] 
[a :z d]: [1 [2 3] 4]

if z s a word, IT gets assigned [1 2]  if z is a get-word, its content 
is assigned to the content [1 2], so in the above, b and c would 
be 1 and 2 respectively.

double colon makes no sense... what are you assigning to z?  a new 
datatype?

but Henrik's  :z:  has me puzzled here:

[a :z: d]: [1 [2 3] 4]

what would the above mean?  hehe

and what is your :z: ?
a new datatype too, much more confusing to my mind

no :z already exists.   if z where a function, it returns the function 
itself, not its evaluation.

that is a new datatype, a GET-SET

it does in path notation.  its also silently ignored in the command 
line as long as the get-set word actually holds a word within  :-)

in a path, it works though.

but in plain REBOL, it has to be a new datatype

no, the path evaluation scheme uses a string parser

a path is a string in memory

I think I see what you mean when you say a new datatype...  yes, 
it would be a get-set word, but the path notation already makes it 
clear how that should work... and I've used it often

so semantically its nothing new.  but in the context of the above 
multi-assign, I concur that the meaning becomes a bit obscure.

it has to be a new datatype

just as get-word and set-word are different datatypes.  this is just 
the combination of both actually, code wise, its probably a one hour 
affaire for Carl, maybe even less.

posted a note about it in R3 chat... just to see his reaction  :-)

a: [x y]
b: 'a
:b: [22 33]
== x: 22
== y: 33

probably a flame :)

A path is not a string in memory, btw. It's more like a block.

well, i only said that without knowing the true truth, only to gain 
a victory against Maxim

spoken like a true french would...

if you can't beat him... lie!    ;-p

so it's a mid-lie

half a truth is no truth AT ALL !!!!

assuming is just lying with a good case!

In a decade of REBOLing, Ive tried to use it exactly twice, and ended 
up having problems, so I created global words instead and forgot 
about it.

Actually, there is too much evolutions, i can't focus on a specific 
project

Haven't used R3 for anything serious yet.


but now that extensions are upon us, the OpenGL integration will 
allow me to use R3 for a *VERY* serious project.  still need Carl 
to fix the make/copy issue before I can convert most of my cool stuff 
over to R3 (like liquid and globs).

I hope extensions will be in the linux and mac ports shortly.

OpenGL integration is a too big project for me, i lost my motivation 
after some days

Its working on my system  :-)


when A80 is out, depending on what he will have added to the API, 
I will try to port a simple OpenGL game to REBOL  :-)

well, i have some fun coding a proxy for R3 since some days, but 
I reach my limit attention

porting relavance to REBOL is also another big project of mine, probably 
will start in under a month  :-)

Re ALIAS -- it has several security and other issues.

Enabling it via SECURE is a good way of limiting the inadvertent 
damage it can do....And of drawing attention to its dangers for any 
casual dabbler.