AltME groups: search

@Graham: you are using web sockets for a wierd job. The shortest 
explanaiton for web socket is: it is the way to trigger an event 
on the client side from the server side via already open tcp (http) 
connection. ofcourse that connection is full duplex, client and server 
both can send a utf8 encoded (currently no binary) string messages 
anytime.

Sure .. just wondering if it's easier to do thru web ports and if 
the server can notify me thru an open port when it succeeds or fails 
rather than me poll the server

Notification thru websocket port: sure, you can have that, but you'll 
need: a ws client for REBOL, a websocket app and a RSP script for 
that (the websocket app will need a timer to trigger queries on a 
RSP script to get the email status or hack in local MTA data to get 
the info to avoid querying a RSP script).

Since the File API is asynchronous, files could be loaded and processed 
in the background while the user does other things. This could also 
be useful for creating web applications that work offline.

And it's supposed to be 40x faster too, since the processing is done 
client-side.

well remark allows you to do this on ther server... and you could 
cache it so its just processed once.

Section 9 "Connectionless push and other features" in the W3 doc 
is interesting.

heh .. I was looking and looking and trying to find new glasses ...!

no, it works, and the pick function is (in pop)
pick: func [
	"Read the Nth message from the POP port" 
	port
][
	read-message port 1
]

I confused because, first p simply calls pick, my pick function inside 
an attempt block, so it cannot return an error.

but "first port" returns an error, if no data. its not a big deal. 
but spent some hours on it, then I saw that "pick port 1" returns 
none instead of error (as expected)

So I don't know what is different between "first port" and "pick 
port 1"

and pick p 1000000 is the same?

ws protocol is event based in its specification, so made it as no-wait. 
then and event based protocol can be built on this.

Endo, for the moment, use 4 spaces for tabs, and you're formatting 
will stay intact.  I'll change the script to replace tabs with spaces.

At present I have to manually edit the httpd.cfg and restart ( maybe 
reloading works ) Cheyenne when I add a new user.

I'm wonding if I can just create a new httpd.cfg from a template, 
make the changes to support the port I want to provision, create 
a sym link in the new directory and then run from the sym link...

Graham: domain or maybe subdomain? if subdomain and maybe domain 
also you could probably put nginx in front of cheyenne and use rewrite 
rules to do this (depending on the concrete situation)

And that way I wouldn't have all the issues of mismatched ssl certificates

for subdomains it's simple (and much cheaper than domains, since 
you don't have to buy each one) .. you use rewrite rules and modify 
domain to a get parameter that cheyenne can act on

And is it easy to redirect subdomains to other ip addresses?

what would be the best way to resize and crop an image after it's 
uploaded on cheyenne .. using commandline imgmagick, I think oldes 
was making an magemagick binding once?, are there more light ways, 
can pure rebol do it?

Very cool Oldes! Thanks a lot! (and cool galery on that page)

Godaddy allows you to create up to 99 subdomains for each domain, 
and if I choose forwarding with masking, then the browser redirects 
but the address bar remains unchanged

First of all you must add DNS A records to point to your IP adress. 
And then setup your server (Cheyenne, Apache, Nginx..) to resolve 
the full subdomain.domain names.

I think you want multiple subdomains, one for each user, and you 
can simplify that by defining one catch-all IP number in DNS. Then 
define all the virtual hosts in Cheyenne. You do have to add every 
subdomain to Cheyenne, but this way you don't have to maintain DNS 
for each

if u want something nicer, multicast dns, and a superfast and well 
thought out control panel, check out dnsmadeeasy.com

I used encmd now ... but it still appears in the system tray.  I 
right click on the system tray icon, and use "run as a service" and 
then it stops running

and disappears from the running processes

not tested, but if you want to start something on windows start up, 
just add link into this folder> C:\Documents and Settings\<%your-user-name%>\Start 
Menu\Programs\Startup   (if you have english windows) The folder 
names are localized so it will look different in other languages)

ok, so get it running as a service you need to run as administrator 
first, and then use the "run as service" menu

ie. the cgi, rsp and rsp web app are not

I wonder how slow it might be to use a coding style where each page 
is represented by a Rebol object, and to display the page back to 
the browser client, you have a render function that iteractively 
calls the render function of each element of the object.  the object 
initializes by reading the session and request objects when it is 
first instantiated

I've in the process of rewriting a site I built 3 years ago in YUI 
and it is so tedious and repetitive.

Endo: I will need a rebol code obfuscator for making one of cheyenne 
webapps also run locally. I don't know how encryptor would work since 
local app would have to decrypt it to run it so others can see that 
and decrypt it also in same way.

@janko: I use Zend Encoder to encode my PHP scripts before give them 
to my customers, so they cannot see/change my scripts. Zend Optimizer 
is a free tool which installed on IIS or Apache, it simply reads 
the encoded php files decode them and execute. It really difficult 
to crack it, not impossible ofcourse but it gives enough security 
to protect sources.

I've been thinking of how to separate code from display and have 
come up with the following ideas http://rebol.wik.is/Cheyenne/Page_Title/HTML_dialect


Basically the idea is to create all the html, js, css beforehand, 
and then emit in a template.  When the parser runs through the layout 
block, it builds all the CSS and JS links.  Once it reaches the body, 
it recursively parses the div tags building and closing them.  In 
the same way html, head and body tags are built and closed.

Hopefully this will result in more maintainable code .. as the whole 
page is now Rebol.  And like classical programming, we "define" all 
our variables first .. and then use them later on.

actually it's more high level. you get loops and mapping of blocks 
and objects to tables, etc.

and I use it also on the R2 and R3 pages for GUI images, that I serve 
locally, as you have seen.

I can see if I can pull a few out. The reason the HTML dialect is 
not published yet, is because the docs contain examples about redirecting, 
and coincidentally, redirects to an empty page. :-)

I think what I am suggesting is a MVC approach, and so clearly the 
template at the bottom is my View, and the code at the top is the 
Model and Controller.  I imagine that it would be quite simple to 
change the View with this approach

sample template and actual ouput http://rebol.wik.is/Cheyenne/Page_Title/HTML_dialect_demo

This uses the Yahoo YUI grid t1 example

A html dialect that builds tables and such would be used outside 
of the template.. which only does a simple substitution

I've been thinking of how to separate code from display and have 
come up with the following ideas


I've seen much benefit when exchanging one syntax for another, especially 
when the 'other' moves away from a standard.

Ok, removed the 'div and fixed a few typos

Terry, I'm trying to get away from mixing html and rebol inside the 
view part of the page ...  this type of stuff   Hello <%= select 
session/content 'username %> ... looks hard to maintain.

I've looked at some php pages with js , sql etc and it just looks 
like a nightmare to maintain

That's why web devs love it so much. Job security and status as magicians

I think the term is alchemist, and not magician!

looks hard to maintain.


In PHP, I process/generate all the hypertext, and avoid the 'dynamic' 
aspect of web page creation altogether, so my page templates look 
like so;

<html>
<head>
<?= $cssincludes ?> // CSS
<?= $headincludes?> // JQuery, specific javascript functions  
</head>
<body>
<?= $body ?>
</body>
</html>

As long as you have just simple pages, it's fine. You can use this 
with RSP easily as well.. but if you want to look at "hard to maintain" 
PHP code.. and have good nerves..just download Magento's sources 
and try to figure out what's going on.

This is a little odd .... I created a PDF and sent it back to the 
browser using the response, and application pdf, and the browser's 
security prevents it from downloading ....on some of the user's PC's 
( at a hospital ), but if I write the pdf to the disk and then redirect 
the browser to the file in the www directory, there's no complaint!

Use wireshark and check for differencies

The difference is between a GET and POST ... I wonder if the hospital 
has configure their security this way

And cannot you store the generated PDF in temp and redirect to it's 
location using the GET request?

Graham: Terry, I'm trying to get away from mixing html and rebol 
inside the view part of the page ...  this type of stuff   Hello 
<%= select session/content 'username %> ... looks hard to maintain.

Wicket and Click are great example of how to get away from using 
scripting inside template files or even tags.

The question is the remark you make to Terry: I'm trying to get away 
from mixing html and rebol inside the view part of the page ...  
this type of stuff   Hello <%= select session/content 'username %> 
... looks hard to maintain.

So chyenne loads page.html and does what?

And is it inside a webapp?

and most html is dynamically generated for me ...

I'm assuming that the JS and HTML are part of your template

well, the current rsp script name may be accessible ..and if so, 
you can just load in the html page

so your "rsp" page could be .html and the template .htm

And so I did. Then I used the app-init.r to test the results, and 
the before and after page intercept did not work.

and does rsp execute at all?

and it does when you use a rsp extension

The re-usability that you are looking for (i.e. the JS, etc) could 
be met with a simple configuration file and page fragements.

but .. show me the pudding and I'll eat it :)

:) We need to write some code then. See ya. And thanks for the time.

I don't find much use of rebol view. I'm interested in browser based 
stuff, and that's the only way I'd spend time with rebol. I'd love 
to take a look at Temple.

Interesting item from the past .. both Brett and Andrew have droppped 
out of the Rebol scene ... the latter not by choice though

Temple is just an engine, so it can run in any context. It was mostly 
an experiment, and the HTML parser was very basic. I have a much 
better parser now, adapting the functions from Temple should not 
be hard. (left as a nice exercise to the reader :-)

http://www.rebol.it/power-mezz/mezz/load-html.html

i have some old stuff scattered around. i'm trying to put most of 
the newest stuff in the Power Mezz package which is on Sourceforge 
and www.rebol.it

Doc says he will try and fix the 'alert issue ... though he's not 
sure it will fix the ability to run as a service

I think it's Gabriele's module system. Similar functionality is standard 
in REBOL 3. When you export explicitly, you can choose what to export 
and what to leave internal to the context, so you have minimum potential 
for conflicts with the global context

Kaj: well, it's not my module system, but it's basically the same 
thing as using SET 'WORD instead of WORD: , except that it is more 
readable (a newbie can guess what "export" means, while it's harder 
to guess that SET 'WORD is different from WORD:)


Florin: since R2 does not natively include the concept of "modules", 
all of us have been doing various things to avoid polluting the global 
context too much. I eventually got tired of all this and created 
this instead: http://www.rebol.it/power-mezz/mezz/module.html

Hi all, just to let you know that I'm upgrading Cheyenne websocket 
support to the last ws draft RFC, the handshaking part has changed 
and is not compatible with the previous version, causing all current 
Cheyenne ws apps to not work with latest browsers (like Chrome 6).

node.js is a v8 javascript interpreter that could be used server-side 
to make tcp server.

So, not only web server but server with persistent connection too, 
like web sockets.

Javascript on the server side isn't a new idea, but V8 is very fast 
and more light without the html dom complexity.
It's all event driven and seems very fun to play with it.

I distribute encapped versions only when making new releases. If 
you need a intermediary revision encapped, just let me know the OS 
and encap kernel and I'll make you one. Also, I think I should release 
a new version now, the current SVN code base seems stable enough 
now.

hum...I forget about the SDK licensing issue, you can't distribute 
encapped binaries that allow running arbitrary REBOL code (Cheyenne 
allows that through CGI and RSP scripts).

This is what the SDK license says (so related to any encapping kernel): 
[...]

REBOL API Restriction. If you choose to export the REBOL API (function 
and objects of REBOL) for users of programs created by REBOL/SDK 
(in other words, you allow users of your programs to write and execute 
REBOL scripts within your program), you must obtain a separate commercial 
distribution license from REBOL Technologies. If your program does 
not allow users to access the REBOL API, then no additional license 
is required.
[...]
http://www.rebol.com/docs/sdk/sdkintro.html

Pekr, the definition : "(in other words, you allow users of your 
programs to write and execute REBOL scripts within your program)" 
seems broad enough to me to include CGI. But as I'm not a native 
english speaker, I might understand it wrongly.

I discussed this with Carl before and it's okay.

Anyway it seems that restriction is now obsolete as ssl and odbc 
are now free.  And IOS is dead.  The license prevented you from creating 
a clone of IOS

I updated to the lastest svn and encapped using encmd.exe from 14 
March 2008

It says on hover, listening on port 8000 and cgi/rsp/html work

I click on the run as service and it changes to "Cheyenne is running" 
with no port now specifed.  Html/rsp and cgi all now fail.

I confirm this behaviour, but if you launch Cheyenne by using right-click 
on cheyenne.exe and select "Run as administrator", then once in service 
mode, it works ok.

Yes, though oddly that's is what I first tried and it didn't work. 
 Which is why I switched to running from the dos shell.  Maybe I 
didn't get the right sequence!

Because Cheyenne must start as root user to be able to open listen 
ports < 1024, then only, SU to a non-root user. Unfortunately, all 
the log files generated during the root phase of Cheyenne's boot 
process are owned by root, so can't be used anymore once su-ed to 
a non-root user. An easy way to fix that (but looks a bit ugly to 
me) could be to chown( ) / chgrp( ) all generated log files before 
giving up root rights...A cleaner way would involve rethinking the 
whole boot up process of Cheyenne and Uniserve with deep code architecture 
impacts (too much work). Any idea on the best way to solve these 
issues are welcome.

re Cheyenne PID: your OS ID returned by system/version/4 might not 
be the same as the Linux one. Open %misc/os.r and add a new entry 
in the SWITCH block with your system/version/4 value and [%misc/unix.r], 
then try again.

I see it only evaluates that when it finds the library interface, 
and that is still not available in R2/Core

Strangely, it just worked now...I've tried it several times a few 
minutes ago and was getting only some parts or nothing at all...

No, I complained about it on RebDev, but no response, and no promised 
2.7.8

Kaj: I'm applying your patch on mod-userdir.r. I'm changing the user 
/ group arguments datatype from word! to string!, this both reduces 
the code size in mod-userdir and makes the dialect less confusing 
IMO when you use named uid/gid. So, instead of :

user www
group users

we'll have now :

user "www"
group "users"

Do you agree with that change?

And, in Core and others, could be cmd line option.

Carl, did you see our above discussion on getting Cheyenne to shed 
root access, and my request in R3 Chat to have Library access enabled 
in R2/Core?

rebcmd has odbc, ssl and library access.

Making the REBOL side process a command-line option for /Core and 
build option for /Command would be a good improvement.

I deleted the app-init.r and still get the error