AltME groups: search

A sandbox is a long-time wish of mine:
   http://www.rebol.org/ml-display-thread.r?m=rmlNVBC

REBOL3's security policies start to make it possible:
   http://www.rebol.com/r3/docs/functions/secure.html

yay! now it's a wish of two :)

Anyway to clear the screen in the console short of starting a new 
session?

We have to wait a bit for Carl to come-up from his cave ... hopefully 
he is cooking something here or there :-)

Just to be picky: you can't actually infer from `actor: make native! 
[[end!]]` how much a scheme does

rebol3 is not in a process of production but design + production 
so I would not be too restrictive on carl jumping from one thing 
to another. he is also only one, so he can't just do r3 for 2 years 
straight

Next up is http://www.rebol.net/r3blogs/0274.htmlcompressed modules/scripts. 
I have some interesting tricks to make them seamless to use and make, 
but I have to get a little sleep before I can really dig in to them.

Not a bad idea - what uses it?

This doesn't look good ... http://www.curecode.org/rebol3/index.rsp


new issues are being created, but there's a sharp decline in closed 
issues

We were focusing on other things for a while. We're bug fixing now 
though.

Is there a way to specifiy that you want a timestamp vs date ?

NOW/time returns the time. It's not a timestamp though.

I want to specify that a datestamp is the required parameter .. and 
that a date is not enough

date with a time ...

So, some DBs have a date type, and a datestamp type

A SQL timestamp is different than a datetime. There's no real timestamp 
type in REBOL, the closest we have is STATS/timer.

Thanks. The closest we have that is still a datetime is NOW/precise/utc.

Let's say that the function you just wrote requires that the user 
enter the date with a time ....

If you are getting the data from a string and can afford to be strict, 
use TO-DATE. Like this:
>> d: to-date "24-Jan-2010/0:00"
== 24-Jan-2010/0:00
>> d/time
== 0:00
>> d: to-date "24-Jan-2010"
== 24-Jan-2010
>> d/time
== none

Sometimes you want to import a module more than once, sometimes you 
don't. If you want to redo it every time then don't set a name in 
the spec - that turns it into a mixin. And named modules can be upgraded 
at runtime.

Use the version to trigger an upgrade of a named module.

Time ago I have written about porting REBOL3 to the .NET platform 
and Java VM.

The reason ? There is a big universe of libraries and frameworks 
immediately usable that would bring many developers to the REBOL 
platform.

Considering .NET and JAVA VM as alternate OS to develop REBOL3 for 
would be a big step forward to me.

In life we must first accept a comprimise to bring our ideology to 
the world.

I like the idea of a REBOL executable whose size is less than 1MB 
but I want to use it as soon as possible and in many scenarios as 
possible.

Would it be so difficul for REBOL Tech. to understand that this starting 
compromise would be instead the best possible way to pubblicize REBOL 
?

If you agree with me, what about writing a small letter to Carl explaining 
our desire ? If many developers sign this letter we could try to 
convince him to make this step once R3 will go to beta.

Then, with the time, the standalone version of REBOL3 will surely 
gain additional libraries and componet but until then we will have 
a  great ecosystem already available !

... even back at that time, some ppl objected, that coding REBOL 
in JAVA or .NET would make it really slow. Instead of that, what 
was proposed was the integration work. Now with R3 Host Kit and Extensions 
infrastructure, it is really doable. R3 core already runs on many 
systems, and the porting was done in a week or two? I think that 
even ARM version is in the works ....

Getting REBOL to JAVA VM might be more interesting, as there is probably 
many HW platforms JAVA runs on, but REBOL does not. But OTOH - those 
porting efforts are going to be mainly in community hands anyway. 
Only the language interpreter itself stays closed-source, the rest 
is open-sourced. As for me - I prefer getting R3 in a raw state to 
many platforms, instead of slowing it down and porting it to JAVA 
or .NET directly.

So we at least know a bit, what is planned ;-)

Sounds like a lot of work still ahead...for R3 beta

Petr, see it as a smart marketing action. Make it hidden, mighty 
and people will take a look. IMO Apple knows this to perfection. 
Creating a huge hype & buzz upfront new products.

No doubt everyone's list will be a little different

I wonder if Carl isn't just following a GTD scheme, which basically 
means working on whatever motivates him the most right now, because 
that overall gives the highest level of productivity.

and sometimes leaving a problem alone for a while allows a much better 
solution present itself.

well, it seems to me the community is plenty involved, but it could 
be a bit better.

I agree with Graham - in fact it is highly demotivating and frustrating 
situation. Some points of my analysis are ignored, and exactly predictable. 
When website redo and 2.x upgrade was announced, I knew exactly what 
happens and what it means for R3. Not a one week halt, but it is 
stalled for more than one month already.

Sorry, I'm playing devil's advocate here. But, really, there is still 
a ton of things to do that doesn't involve Carl. The function reference 
manual is only 30% done and that took me a week to do. So why don't 
a couple of people here continue the work? I won't have time for 
the next few months. Carl has asked for this several times, and this 
is something that really needs to be done.

Henrik ... do we even have a license?

A number of us said that R3 chat was a waste of resources ... and 
... now, barely anyone uses it.

Pekr, Graham knows more about schemes and protocols now than he did 
before. I'd say there is a win there.

Who says that Carl isn't using it? I see some private messages trickling 
in once in a while. And Graham, you didn't respond to my question 
that I posted there 3 hours ago to help resolve the document editing 
situation on rebol.com.

That's a manual page, so that looks OK. Can you edit a page like 
this one:

http://www.rebol.com/rebol3/

To edit a page, you need a "edit" menu on the page

ok? I get: "Restricted action. You have attempted to access a section 
of this website that requires administrative access."

http://www.rebol.com/r3/docs/functions/read.html

see the word "edit" top right ... I think that's the only time you 
can edit a page

also I can see three more items checked on the priorities page from 
the last time I checked it a few hours ago, so docs conversion is 
progressing.

posted 4 and 5 days ago. a couple more earlier in the month.

While I agree that the website overhaul has a very bad timing, I 
also think that the new website is an incredible improvement

There is an unheard of (for RT) transparence of work done on the 
website (http://www.rebol.com/recent.html). A cynic might think, 
though, that it can surely be only a matter of time until that page 
is locked down and protected and whatnot.

A lot of us never visit the website ...

Still, I don't think that all this shiny website freshness does justify 
the total lack of R3-related activity and/or communication. Just 
as one could sense at least a tiny bit of R3 momentum ...

but we need first a good R3 GUI

a R3 ODBC scheme and others (mysql etc...)

a Cheyenne with R3 would be great too

it would be a good showcase for the possibilities of R3

Paul, the R2 console is Windows only, so Carl doesn't want to re-use 
that console, but it could probably be used as a temporary measure 
(that would be another diversion). Use something like this for now:

xyz: read clipboard://

for all the worried, Carl twitted "Wrapping up website changes. Was 
a huge, intense project, but had to be done. The site is much more 
manageable as a result."

graham, yes i would like to contribute to do a pestore like in R3

but i am not a guru in Rebol. and for me it is important to try to 
do this application to learn & to show the best practicies in R3, 
with the help of some guru ;-)

Paul, what I'm saying is that the R2 console is very Windows dependent. 
It doesn't exist on any other platform. This situation must change. 
Meanwhile the functionality would also likely improve, so we could 
get a graphical console.

Until someone finds a better console.

>> ? frame!
frame! is a datatype
It is defined as an internal context frame
It is of the general type internal

No values defined for frame!

Ashley, a revision to your post:
Changed to typesets, not removed:
	any-block!
	any-function!
	any-string!
	any-type!
	any-word!
	number!
	series!


Non-functional legacy, not newly added, likely to be removed or replaced 
by user-defined types:
	rebcode!
	struct!

frame! is used in extensions as well. There It's a linked/cascaded 
data-structure to get access to call parameters and construct return 
values like block!

Is it a product? If so, there should be a ! before the group name.

I wouldn't consider it a "product". it is a collection of modules.

We could as well use the group power-pack, although it was meant 
a bit differently - to create package of most usefull add-on stuff, 
so e.g. mysql protocol, rugby, uniserve, etc.

of note, you don't see that from the website, but the repository 
includes a test suite and benchmark suite as well. they are undocumented... 
but i guess they can be useful for other projects as well...

if i had the time, i would add everything to the power mezz. the 
detective stuff needs some work though before being added here. (convert 
to a module, change the format, add tests if possible, and so on)

I would suggest adding a charset in the header of source code pages, 
this one for example http://www.rebol.it/power-mezz/mezz/html-to-text.r
I get this line:

break-at: complement charset [#"0" - #"9" #"A" - #"Z" #"a" - #"z" 
{"'#$%&([^{@} #"" - #""]   <- last 2 chars are ?

you'll find: break-at: complement charset [#"0" - #"9" #"A" - #"Z" 
#"a" - #"z" {"'#$%&([^{@} #"^(80)" - #"^(FF)"]

uploaded (to sourceforge) a zip with just the modules (.r files). 
no tests, docs, etc. for the lazy :-)

it should not clash with any built in ones available in the future, 
because it's a module.

For that matter it works (sort-of) the same, so it's a swap-in replacement. 
Hence the "not bad" :)

Thanks ... I'll give it a go.  But it seems that it would require 
a bit of work to use this within an encapped script?

I can use decode-email-field to decode various encodings of subject. 
But I wasn't able to figure out how can I decode the content of an 
email which is in my case encoded with quoted-printable / utf8 . 


I found to-netmsg library on codeconscius.com code that loads the 
email text and parses it into structure. It doesn't decode the subject 
=?UTF-8?B...?= but it does the content. I could use that and power 
mezz to get the effect I want. If there is a way to encode content 
in power-mezz I would rather use it alone.

Is there any documentation on how to use these modules? I've had 
a couple of hours this beautiful weekend and all I did was try to 
find a way to load load-html.r with no results. Dependent modules 
change the current directory in the console and as a result one module 
cannot find another another. Yes, I am a poor soul on windows. Heck, 
I got to take the kids to the Fall Festival and am late already.

Well, well. I thought I could just type and things heppen. I actually 
need to read code. Bummer. This guy, BrianH, is a heartless man.

The package you mention has files with a different extension, .rlp. 
Apparently the package needs to be  cooked on a strange OS different 
than Windows. (Gabriele: It currently requires GNU Make, however, 
those crazy people  still using Windows could rewrite remake.r to 
do all the work).

http://gnuwin32.sourceforge.net/downlinks/make-bin-zip.php

has a single binary in the bin/ subfolder, "make.exe", that's all 
you need.

I am not an RLP advocate, for good reasons, but it has its fans and 
I respect that. It's a matter of taste.

Note, I mean that I am not a LP advocate. However, if you are a LP 
fan, I *definitely* recommend RLP, it's a great implementation of 
the LP concept and a great piece of code.

Florin, if you go here: http://sourceforge.net/projects/rebol-power-mez/files/


you'll find power-mezz-built-1.0.0.zip which contains pre-built .r 
scripts. Inside tools/ , there is a module-console.r which defines 
a IMPORT function so that you can just write "import %mezz/load-html.r" 
and have its exported words available in the console.

Otherwise, from your own code:

    do %<path to power-mezz>/mezz/module.r
    load-module/from %<path to power-mezz>

    module [
        Imports: [%mezz/load-html.r]
    ] [
        ; your code that uses load-html
    ]

of note, if you have the normal package, from windows you can go 
to tools/ and double click on wetan-test.r (assuming you have View 
associated with .r files), and it will just ask you for a file to 
process. select any of the .rlp file to process it and produce .html 
and .r files.


make is only required if you want everything to happen automatically. 
i'm not sure if it is obscure, as it is the tool used by basically 
every C programmer on this planet since 1980 or so...

yes, html string should be enough.

Script: {Allows "importing" modules at the console} (none)
Script: "Modules for REBOL 2" (none)
>> import %mezz/load-html.r

>> result: load-html "<p>This is a paragraph<p>This is another one"

== [root none [] [html [...] [] [head [...] [] [title [...] []]] 
[body [...] [] [p [...] [] [text [...] [value "This is a paragraph...

>> import %mezz/trees.r
>> print mold-tree result

[root [] [html [] [head [] [title []]] [body [] [p [] [text [value 
"This is a paragraph"]]] [p [] [text [value "This is another one"]]]]]]

modules were designed to help with building larger applications. 
i understand it can get in your way if you are doing something very 
small. You can still just DO the scripts but then you need to resolve 
the dependencies yourself, ie. do all the needed scripts in order. 
(there is a function inside tests/run.r that returns a block with 
all the modules a module depends on.)

Is there a way to navigate this tree? Something similar to, let's 
say xslt path?

If I want to find a tag with a given ID, how is it done?

you can navigate with GET-NODE, I haven't needed something like XSLT 
Paths yet (I had something closer to that in Temple).

root: load-html ...
p: get-node root/childs/html/childs/body/childs/p ; for eg.


most of the logic to do what you want is already in trees.r actually, 
because of the rewrite-tree function (that I don't use anymore). 
Anyway, a simple way would be:

get-node-with-id: func [root id] [
    if id = get-node root/prop/id [return root]
    foreach child get-node root/childs [
        if get-node-with-id child id [return child]
    ]
    none
]

I've got a meeting to run to -- will check back in couple of hours 
 :-)

Patrick, e:/Projects/... is not a valid rebol file path.

try with something like:


    power-mezz-path: %/E/Projects/PT/Rebol/power-mezz-built-1.0.0/
    do power-mezz-path/mezz/module.r
    load-modules/from power-mezz-path
    ; etc.

The particular script I am writing is called GET ADDRESS.  This script 
takes a CSV file called contacts which has first and last name, city 
and state of all of my friends that I'd like to get addresses for 
Christmas cards, but have forgotten or misplaced.


So far, the script takes each entry and sends it to SUPERPAGES.com 
where the HTML sent back contains the information.  Right now, I'm 
simply saving the HTML as a file for each entry in my CSV.


What I would like to do is somehow parse the HTML from it and extract 
out the address lines, zip code, phone number etc.  But I admit that 
parsing through HTML is daunting to me.  So after looking around 
on the internet, I discovered HTML-TO-TEXT in your Power Mezz.  


That is where I am now, trying to figure it out and see how it works. 
 I've read some of your documentation, but I admit, I am still in 
the dark as to how it works -- at least for my application.  Any 
advice you have is welcome.

Thanks in advance.

I don't think that's a good function to use for that. It seems to 
me it's meant for making readable text, not processable text

Use "5.10 Parse HTML text into a tree" instead

To be honest, if you just like to parse some HTML page to get some 
parts of it, you don't need to use Power Mezz at all.. I'm using 
Rebol more than 10 years and still consider PM as a too complex staff 
for me. If you are REBOL newbie, better to start reading REBOL doc. 
In your case something about parsing.

There is a lot of pages about 'parse' on net... for example this 
one: http://www.rebol.com/docs/core23/rebolcore-15.html

But when the HTML and the data become more complex, there are so 
many exceptions you have to program, that a real HTML parser becomes 
more convenient

would a real HTML parser convert the data to a REBOL object?

a good one would be to convert R3 rich text to HTML and vice versa.

Kaj, I think it's the other way around! I found when the HTML and 
the data become more complex, then a simpler "hack" parse job is 
more likely to survive changes to the source. This happened to me 
several times with a weather forecast and television guide scraper 
etc. that I made (and remade, and remade..).