AltME groups: search

This is my first pass at writing a smtp server ...  works from telnet 
but not from a mail client so far.

REBOL [
	Title: "SMTP daemon"
	Author: "Graham Chiu"
	Version: 0.0.1
	Date: 5/3/2005
]

install-service [
	name: 'smtp
	port-id: 25

	multi-line-end: rejoin [crlf #"." crlf]
	stop-at: crlf
	
	server: make object! [ user-data: none ]
	
	maildir: %mail/ ; store mail here
	
	save-mail: func [ data /local mailbox ][

  if not dir? mailbox: rejoin [ maildir server/user-data/email "/" 
  ] [
			if not exists? join maildir %misc/ [
				mailbox: make-dir join maildir %misc/
			]					
		]
		if not exists? join mailbox "mail.txt" [
			write join mailbox "mail.txt" "" 
		]
		write/append join mailbox "mail.txt" join crlf to-string data	
		print dehex data
	]
	
	on-new-client: has [su] [
			su: server/user-data: context [
			state: copy "command"
			email: none
	 	]
		stop-at: crlf
		write-client join "220 mail.compkarori.co.nz SMTP" crlf
	]

	smtp-rule: [
		"HELO" thru newline
			( write-client join "250 mail.compkarori.co.nz SMTP" crlf ) |
		"EHLO" thru newline
			( write-client join "500 not implemented" crlf ) |
		"MAIL" thru newline
			( write-client join "250 OK MAIL FROM" crlf) |
		"QUIT" thru newline
			( write-client join "221 Good Bye" crlf close-client) |
		"RSET" thru newline

   ( write-client join "250 OK RESET" crlf server/user-data/state: copy 
   "command" ) |
		"NOOP" thru newline
			( write-client join "250 OK NOOP" crlf ) |
		"VRFY" thru newline

   ( write-client join "252 send some mail, i'll try my best" crlf ) 
   |
		"EXPN" thru newline
			( write-client join "500 not implemented" crlf ) |
		"RCPT" [ thru "<" | thru ": " ] copy name to "@" thru newline 

   ( server/user-data/email: form name write-client join "250 OK RCPT 
   TO" crlf ) |
		"DATA" thru newline

   ( stop-at: multi-line-end server/user-data/state: copy "body" ) 
	]
	
	on-received: func [data /local su] [
		su: server/user-data
		print join "Data: " data
		switch su/state [
			"command" [
				if not parse data smtp-rule	[ 
					write-client join "500 command not understood" crlf
				]		
			]
			"body" [
				; reject if we don't have a RCPT command first
				if none? su/email [
					write-client join "500 no email address received" crlf
					stop-at: crlf
					su/state: copy "command"
					return
				]
				
				; write the body of the message somewhere

    save-mail rejoin [ "Received: from somewhere at " to-idate now newline 
    dehex data newline newline ]
				stop-at: crlf
				write-client join "250 OK MAIL received" crlf
				su/state: "command"
				su/email: none
			]	
		]
	]
]

create a directory called "mail" and each subdirectory holds a "mailbox" 
for the incoming mail.  If a mailbox does not exist, it gets routed 
to mailbox "misc".

ah... found a problem.

REBOL [
	Title: "SMTP daemon"
	Author: "Graham Chiu"
	Version: 0.0.2
	Date: 5/3/2005
]

install-service [
	name: 'smtp
	port-id: 25

	multi-line-end: rejoin [crlf #"." crlf]
	stop-at: crlf
	
	server: make object! [ user-data: none ]

 clear-server: server/user-data [ state: "command" email: computer: 
 none ]
	
	maildir: %mail/ ; store mail here
	domains: [ "@compkarori.co.nz" ] ; list of accepted domains
	
	save-mail: func [ data /local mailbox ][

  if not dir? mailbox: rejoin [ maildir server/user-data/email "/" 
  ] [
			if not exists? mailbox: join maildir %misc/ [
				mailbox: make-dir join maildir %misc/
			]					
		]
		if not exists? join mailbox "mail.txt" [
			write join mailbox "mail.txt" "" 
		]
		write/append join mailbox "mail.txt" join crlf to-string data	
		; print dehex data
	]
	
	on-new-client: has [su] [
			su: server/user-data: context [
			state: copy "command"
			email: computer: none
	 	]
		stop-at: crlf
		write-client join "220 mail.compkarori.co.nz SMTP" crlf
	]

	smtp-rule: [
		"HELO" copy name thru newline
			( write-client join "250 mail.compkarori.co.nz SMTP" crlf 
				if not none? name [
					trim/head/tail name
				]
				server/user-data/computer: form name
			) |
		"EHLO" thru newline
			( write-client join "500 not implemented" crlf ) |
		"MAIL" thru newline
			( write-client join "250 OK MAIL FROM" crlf) |
		"QUIT" thru newline
			( write-client join "221 Good Bye" crlf close-client) |
		"RSET" thru newline
			( write-client join "250 OK RESET" crlf clear-server ) |
		"NOOP" thru newline
			( write-client join "250 OK NOOP" crlf ) |
		"VRFY" thru newline

   ( write-client join "252 send some mail, i'll try my best" crlf ) 
   |
		"EXPN" thru newline
			( write-client join "500 not implemented" crlf ) |

  "RCPT" [ thru "<" | thru ": " ] copy name to "@" copy domain to ">" 
  thru newline 
			( 
				either find domains domain [

     server/user-data/email: form name write-client join "250 OK RCPT 
     TO" crlf 
				][

     write-client join "553 sorry, that domain is not in my list of allowed 
     rcpthosts" crlf
					server/user-data/email: none					
				]
			) |
		"DATA" thru newline

   ( 	stop-at: multi-line-end server/user-data/state: copy "body" 
				write-client join "354 start mail input" crlf
			) 
	]
	
	on-received: func [data /local su] [
		su: server/user-data
		; print join "Data: " data
		switch su/state [
			"command" [
				if not parse data smtp-rule	[ 
					write-client join "500 command not understood" crlf
				]		
			]
			"body" [
				; reject if we don't have a RCPT command first
				if none? su/email [
					write-client join "500 no email address received" crlf
					stop-at: crlf
					su/state: copy "command"
					return
				]
				
				; write the body of the message somewhere

    save-mail rejoin [ "Received: from " su/computer " ( " su/computer 
    " [ " client/remote-ip " ]) " to-idate now newline dehex data newline 
    newline ]
				stop-at: crlf
				write-client join "250 OK MAIL received" crlf
				su/state: "command"
				su/email: none
			]	
		]
	]
]

I have mine running as a test on 192.168.1.202

from another pc, or from local host  ... set-net [ [gchiu-:-nowhere-:-com] 
192.168.1.202 ]

send [gchiu-:-compkarori-:-co-:-nz] "testing ..."


and the email is stored in the mail/gchiu directory appended to mail.txt

A couple of questions:
1. how to timeout the client after a period of inactivity?

2. how to process multiple clients at the same time, or to refuse 
a client connection while an existing connection exists?

I've got a prototype version of my mail server now running.

set-net [ youremail-address 203.79.110.37 ]
send [gchiu-:-compkarori-:-com] {your test message ... }

Anyone want to try it out for me ?

I've posted a new version 0.0.4 : http://www.compkarori.com/vanilla/display/Smtpd.r

After repeated testing, I have bumped up the version to 0.1.0 at 
http://www.compkarori.com/vanilla/display/Smtpd.r


As explained in the URI, a form of teergrubbing ( anti-spam ) is 
implemented.

And that is with a block period of 10 seconds.  The greylisting paper 
suggests using 1 hour and I'll try that now.

As he says, it's very easy to write a new service for it.

Last I checked uniserve uses a directory structure to handle stuff. 
So I wonder if the encaped app needs to replicate this structure.

In short.

Protocols and services (or whatever else) are encapped as data in 
a block! (a cache sort-of).

instead of 'do/'load, you have to use 'do-cache/'load-cache in your 
script.


The cache functions are initialize differently either you are encapping 
or doing the script.

if you do cgi, then you have to launch a new version of the encapped 
script for each cgi ?

Here is an example of a starter that can be used for doing the script 
and to encap it :

This is a strip down of the script I use for encaping lecture-forum.

Got a working pop server now for uniserve .. connected to a firebird 
database.

Now to get the smtp service to save files to the database as well.

I've further improved my spam busting mail service.  If I get an 
email from a new user, and that email passes all my spam tests, I 
now will send an smtp challenge to the mail server that should have 
sent that mail.  I know some mail servers will respond that every 
user combination exists, but others will not, and that will all help.

I've got to the point where I can receive mail from the smtp service, 
store it in a database, and retrieve it using pop3.

even better still, I had over a 1000 smtp connections over night, 
and I caught all the spam :)

It will help me determine how effective this is.

I need people who get a fair bit of spam, and don't mind their mail 
sitting on my server.

Got DideC.. who used Rebol to send it.  We should really have a message-id 
generator as part of the send command.

Haven't so far seen Ammon's alternate mail.  And Sunanda's suffered 
a write failure .. in my save-email routine :(

the addres stuff you see at the top of an email has about as much 
to do with delivery as the picture on a postcard

I must have allowed it thru as it was a RCPT TO:<[gchiu-:-compkarori-:-com]>

I don't see it either. Dual booting is a clumsy solution. It would 
be a lot more fun to just let them run inside OSX like OS9 apps run 
under OSX currently.

pekr, it's fake. it's just a joke image

http://www.sixapart.com/...everyperson a blog ;-)

with a lot of zeros:-))

uf, so the only advantage of mozilla mail is now gone :-) they go 
ms outlook route - one big file, binary, what a route ....

About the Email.. I think they dont understand where email is going.. 
If they would have build an email client that would support encryption 
or packaging from text to grafics they would be on the route to the 
future.. For..re-inventing the wheel with lots of bells and advertising.. 
(yes im a little anti java and .net internet applications currently 
;-) Ever stranger...NET is dead and also is Java..still many use 
it.. i dont get it..

Isnt it more that the Java community is moving other ways because 
of the onsupport by SUN? And I have read from a .NET developer from 
Microsoft that the complete architecture of .NET was wrong wrong 
the start and that they where running into problems and because of 
that they wuold not survive time?

I can't agree to .Net, but maybe that is because we are so much of 
a MS independent. We run mainframe, unix servers etc., it has tradition 
within our company, MS has NOTHING to offer to us in that regard 
....

Oke SAP and Oracle is a different story... ;-)

I wished that java and .NET started like this -> 

What would we put inside a binary if we only would have 600 Kb to 
be filled..

yes that a pitty.. I know back in 1997 where Lotus had to fight with 
Microsoft about the email..Lotus was already party slow on the interface 
part..The webconfiguration was not even to handle that..that slow.. 
But i happy to read its still alive because it had then some nice 
touches..

I agree..but its a contradiction ...the idea of company's floating 
on opensource and sell their applications for big money... that happens 
just too often...If you develop with opensource be that flexible 
and sell it for a low street price.. In this case SUN is taking the 
honour but not the money...pitty...

What is the big change in Firefox 2 that made them bump the major 
version number? (It looks more like a 1.7 from the tiny list of changes 
they have there ;)

http://minimsft.blogspot.com/2006/03/vista-2007-fire-leadership-now.html
<--- seems there are a few management problems in Microsoft Land

http://10000th.com/

a flash site to showcase Tag Heur.

try out how you can flip the pages and even tear it out.

a little dated already but i like the idea behind the opensource 
movie creation with blender.. here the website: http://orange.blender.org/media-gallery

I cant realy place this product -> http://www.runrev.com/but Its 
somehow nice but also somehow very ugly product... Anyone any experies 
with it? Its commercial though it like the GUI part of Rebol in a 
very Big box.. (in size that is..)

Evolution Robotics today announced a strategic alliance with WowWee 
Robotics(TM) to integrate Evolution's technologies for vision and 
navigation into the next generation of WowWee products. http://robotgossip.blogspot.com/2006/03/wowwee-partners-with-evolution.html

Actualy we should have a rebol version of thisone !!! ;-) --> http://tryruby.hobix.com/

IE 7 is one of the buggiest pieces of garbage I've ever come across.. 
here's one problem I had.. (a quoted solution.)

This is just so 
Microsoft" it should almost be expected.


I had recently installed IE7 in an unsupported way using the instructions 
found here.  The nice thing about this, is that it lets you run IE7 
side by side with IE6.  As a developer, there's no way to just let 
IE7 install itself over IE6, so I thought this would be a good solution.


Fired up IE7 for the first time and it took about 2 minutes for me 
to realize there is just no possible reason why anybody would find 
this useful at all.  Not for end users... not for developers... it 
just doesn't work right.  The new toolbars are not that special either, 
IMO.


So, that was it.  At least for Beta 2.  Fast-forward one week and 
I'm doing some serious testing of one of my new apps.  Of course, 
I'm testing on the fly in Firefox but testing in both browsers after 
finishing all pieces of major functionality.  Enter my URL into IE, 
press enter... and bang... up comes Firefox with the page I loaded!?!?! 
 Uh... what?


Google to the rescue.  A search for "IE Launches Firefox" returned 
only 2 results... but luckily, one of them had the solution.  It 
seems that a registry key installed when IE7 is run causes this situation. 
 Just brilliant.  From the IE Blog... locate this registry key and 
remove it: HKEY_CLASSES_ROOT\CLSID\{c90250f3-4d7d-4991-9b69-a5c5bc1c2ae6}

As stated... it fixed my problem.  Thanks Microsoft...

The windows registry is probably the most "crappy" (native english 
speaker, please replace this word by the word that is in my head 
and I can't find) invention of all the computer history.


If you have a problem and it's not hardware, then it's 99% chance 
it's a registry problem. The 1% rest is for files and DLL problem.

My boring all day experience.

Bill already own the high score in "Bird names" (direct french translation 
for  a non offensive manner to represents offensive words ;-) given 
by all the world computer guys who tried to make Windows X Y working.

I think he is rising over Adolph H. in this area (no, ok, bad joke 
here).

hmm, hmm - http://www.eweek.com/article2/0,1895,1944044,00.asp - 
Ruby on rails takes on Java ... interesting how Carl is right - ppl 
don't buy technologies, they buy solutions ... and Rails is solution, 
which makes Ruby a king ... although it all is surrounded by nice 
hype .....

RoR like most other framework, sells the ease to get things done. 

What I think makes them standout, is the way they SHOW it.


1. Ruby itself is always a 'simpler' approach to Java. That is one 
EASE


2. They SHOW you on their webcast...and they said it themselves. 
'SHOW, Don't Tell'
http://rubyonrails.org/screencasts


3. And they also SHOW it by using tools other programmers use, eg 
TextMate, drawing a targetted group
potential user.


The success of RoR puts a lot of big companies to shame. (given the 
resources they have)

Being symbolic, a query can be as little as 3 bytes.

Whoever heard of a 3 byte database query?

Digital is only saturated Analog, so a bit can be split to various 
levels, but then logic become fuzzy.

in 2 weeks time  a cooperation of micorsoft  together with Apple 
can build very nice applications that bring money ;-)

who would think of this.. actualy its a nice idea...-> http://ilps.science.uva.nl/MoodViews/

A Year in the Life of a BSD Guru

 Josh Berkus of PostgreSQL http://blogs.ittoolbox.com/unix/bsd/archives/008710.asp

All those Google applications. Wouldn't it be a lot better user experience, 
if REBOL clients were made, instead of having the applications inside 
a browser? Might be a good plan:

1) Get people away from MS products. It's not the best solution having 
everything inside a browser, but it lets them know again, that they 
have a choice.

2) Now people have choices, they can choose the best solution, which 
is REBOL reblets.

OTTAWA — California-based linguist Paul Payack expects the English 
language to gain its one-millionth word this autumn. The language 
has come a long way indeed, as the English would say, in 400 years. 
In 1582, the English grammarian Richard Mulcaster could say that 
the language was "of small reach, stretching no further than this 
island of ours, nay not there over all.

http://www.languagemonitor.com/

Ah the real problem with it is a single point of view, and that point 
of view needs a projector next to it. Neat trick though.

...before entering Gates' lodge-style, 66,000-square-foot home overlooking 
Lake Washington with a reported seven bedrooms, six kitchens, 24 
bathrooms, a .....

he should be careful with windows that can be flipped over and are 
hinged at the center of the frame. I have those in my livingroom. 
When opened at a certain angle, the reflection in the glass is  directed 
towards the roof. Sometimes a bird would get "caught" in the reflection 
while sitting on the roof and start attacking the window glass repeatedly. 
It looks really funny, but you have to clean up the mess afterwards: 
I suppose repeatedly banging their head against the window glass 
every 2 seconds for 10 minutes makes for a bad stomach...

because you loose a lot of the goodies in OSX if you don't develop 
for the OS. not even the integration put forward here could make 
up for that

surely, opening up the windows API to mac users... sounds like a 
case to code in OSX natively  hehehe

yeah making media player work inside a mac border, doesn't make it 
any better.

First, Cringley is a little off here. What he suggests is basically 
a description of Wine, but he says it is not like Wine because of 
a complete misunderstanding of what Wine is. Stupid.

http://news.com.com/2100-1028_3-6064016.html?part=rss&tag=6064016&subj=news
<--- US Congress to prepare a new and tougher DCMA

The problem in this case is that the software itself became illegal, 
which is similar to lock picking tools being illegal.


Then the transport of these tools (links on websites) were made akin 
to transport (trafficking).


The solve is simply to fracture all software into so many pieces 
that it is impossible to point in any one direction.


A paperclip here, a tin of graphite there, etc.  And in theory, it 
may be possible due to the web not to have any group of these items 
in your possession at the same time.


Another option is much simpler…send the files out (to some other 
country), have them return unlocked.  This would pose an interesting 
problem since clearly you are receiving a copy some IP.  So then 
one should receive a "protected" copy using a different protection 
system.  No that part has to be proven as well.


It is all very interesting…but in the end we must abide by the law 
until we can vote it to change.

I'm just a little baffled that it would have to go in the direction 
where words like "fighting terrorism" need to be used. I remember 
an ad that ran here a few years ago with footage of Columbian drug 
dealers, terrorists and animal smugglers that said if you bought 
pirated access cards for satellite TV viewing, you'd be supporting 
drug dealers, illegal trading and terrorists. (Geomol has probably 
heard of TV3 and Viasat) I couldn't believe a private company would 
play on that, but they really did. Of course it changed nothing, 
since people are not that stupid and out of touch with reality.


But doing the same kind of lobbying to a government could be very 
dangerous. Are the people in the government easier to affect? Probably 
if you are waving enough money in front of them.

To me, it is all very simple. I just install a program that protects 
all content on my computer before *anything* comes on my PC. Now, 
every piece of software that tries to get anything of my computer... 
of course the protection I offer is cumbersome, so most software 
will circumvent it... and I can sue them. Of cousre I can publish 
something to protect me, espscially from a non-US country.

Imagine that people would run such a piece of software when the sony 
rootkit came out.

Cool. It almost seems like there is a real person inside the suit. 
:-)

But you know what the first application of this technology is going 
to be - military - killing people.

The next application will be smaller  - robbing a convenience store.

actually in Japan it's expected that these robots will be used to 
aid older people in their homes. the number of young Japanese people 
is on the decline and there will be a shortage to help elders in 
a few decades, so Honda started a humanoid robotic development program 
20 years ago. Asimo is where they are now.

a look at Symbian & UIQ Motorola 1000 SmartPhone - http://www.osnews.com/story.php?news_id=14558

Rebol is a PERFECT match for Cellphone aps.

REBOL is a tiny download, but a very large memory footprint when 
running.
I remember that was one reason for it never being ported to PDAs.

Cellphones also have memory limitations -- both in maximum megs and 
underlying mapping models.
Has the hardware curve caught up with REBOL's needs yet?

H blogged a couple of times about winCE being complete pants on a 
cellphone.
And hinted that a port of REBOL to winCE is possible given time:
http://www.rebol.net/article/0217.html

The Nintendo Wii controller looks ingenius in action. I wonder if 
there will be a homebrew scene for Wii...

Heard its a far descendant of APL, and fast. Found this: http://www.kx.com/news/in-the-news/pr-041228-vector.php

Cyphre did find some link to beta Ruby/AGG release, which one note 
of author, stating something like "Download beta release, but this 
is unexpectadly slow :-(" ... which could mean there is still a long 
way to go for them, and we can have REBOL 4.0 by that time ...

graham - my reference was not to Ruby/AGG, which apparently can be 
a good product, but to statements like "rails .... even faster than 
rebol" - whish sounds, from someone who is interested in languages 
so much, like a plain nonsense, comparing language to framework ...

The most curious thing I remember Carl saying was that he said he 
would do it now if he had a million $. That was around the 2005 devcon.

not ... as a long term amigan here, not really sure for modules ... 
modules were there long enough, no? :-) It was components ... big, 
colorfull announcemend, yada, yada ... removed later, when R# went 
to sleep :-))

the problem is that there is soo much to do. Rebol covers so much 
ground, there would be enough work for 30-40 top-skilled software 
engineers over the next 2-3 years working full time to cover everything. 
I would for example like to see just a bit of focus on the Word Browser, 
make it complete so you could add comments. It's hard because there 
are a 100 other things to do. It's like playing tennis and someone 
throws 50 balls at you. You'd need a really big racket.

It is better to be a polyglot than succumb on the I don't understand 
that, so I will not use it, and by the way I will bash any other 
language that I don't  get.

On the other hand, if Pekr wants to be a Rebol wizard and specialize 
in that, it's his privilege to do so..

I do believe that the competition is great, and by the way for a 
much as I love Rebol. I don't see the like the ultimate end of computer 
science,, as matter of fact I see some drastic limitations like Rebol 
can't bootstrap Rebol, because rebol is not compilable.

Yes, Pekr should become a Rebol wizards and demostrate so. Instead 
of talking with property about other languages.

Ruby and Rails both are making a lot money for some people. How many 
Pro-Applications makin $$$$$ does rebol has. And I mean public knowledge 
not embedded internal projects?

My knowledge of this is IOS and QTask. I don't know how much IOS 
produce, but I believe Qtask or more specifically Quilts will make 
a lot of money.

Yeah, I can do a lot of  this with Rebol, but I will spend more than 
10 minutes to have basic CRUD (Create, Read, Update, Destroy) code, 
and much more it I want all of the rest.

Jaime - you never so my code, so don't try to pretend I am not able 
to code. That I don't produce any code does not mean I am not able 
to. I have different orientation at my current job, that is all. 
And of course, I am really not a good coder to provide code of Cyphre 
level quality.

I did not bashed Ruby, but exactly the same nonsense attitude some 
guys are showing over the net. Noone would bark at Ruby, if Rails 
would not exist. Now quite opposite - I can see some of my friends 
talking Ruby as a cool language, because it has Rails, while asking 
them for some language details, they are NOT able to provide any.

Talking of rails .. I've got a dvd player (Sony) where the sled has 
jumped the rails :(  Anyone have a link on how to fix these things?

Pekr, I am not being elitis, I just think your comments are off place. 
If you can not learn a language doesn't mean the language is bad. 
Ruby is a great language and the fact that it has rails, gems, rake, 
and some many other modules makes it great. By great I mean that 
it enhance productivity.

How many times has Rebol send you in loopholes, because a feature 
doesn't exist like certificates, one that you request every so often. 
That is a productivity problem, one that a large community addresses 
quickly.

And Rebol is very good a creating dialects.

Volker compilation of rebol is due to context free grammars. If you 
take this restriction you could probably have a bootstrap of rebol. 
But you can not compile just any rebol program. So this is a problem.

Pekr, I guess my reaction was due to this comment "Good old Jaime 
adheres to hype :-)", this is simply  not true. I recommended Rails 
because I liked the productivity boost that I got in a recent project, 
not because of its popularity.  I  have read about different Programming 
Languages and found that  each one  has its strengths and its weaknesses. 
But independent of this you can always learn a new technique that 
will expand your horizons..

Can we build a rebol framework similar to rails?