Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.os.linux.networking > #1186 > unrolled thread

ETHO interfacing linux IP:stack?

Back to article view | Back to comp.os.linux.networking

Contents

  ETHO interfacing linux IP:stack? no.top.post@gmail.com - 2012-03-22 19:28 +0000
    Re: ETHO interfacing linux IP:stack? Tauno Voipio <tauno.voipio@notused.fi.invalid> - 2012-03-22 21:39 +0200
      Re: ETHO interfacing linux IP:stack? J G Miller <miller@yoyo.ORG> - 2012-03-22 20:07 +0000
        Re: ETHO interfacing linux IP:stack? Lew Pitcher <lpitcher@teksavvy.com> - 2012-03-22 16:27 -0400
          Re: ETHO interfacing linux IP:stack? Lew Pitcher <lpitcher@teksavvy.com> - 2012-03-22 16:44 -0400
          Re: ETHO interfacing linux IP:stack? Allodoxaphobia <knock_yourself_out@example.net> - 2012-03-23 02:43 +0000
        Re: ETHO interfacing linux IP:stack? Roger Blake <rogblake@iname.invalid> - 2012-03-22 21:12 +0000
      Re: ETHO interfacing linux IP:stack? "Chris Burrows" <cfbsoftware@hotmail.com> - 2012-03-23 19:10 +1030
      Re (2): ETHO interfacing linux IP:stack? no.top.post@gmail.com - 2012-03-23 23:21 +0000
      Re (2+): ETHO interfacing linux IP:stack? no.top.post@gmail.com - 2012-03-24 05:58 +0000
        Re: Re (2+): ETHO interfacing linux IP:stack? Tauno Voipio <tauno.voipio@notused.fi.invalid> - 2012-03-24 12:12 +0200
          Re (3) ETHO interfacing linux IP:stack? no.top.post@gmail.com - 2012-03-24 18:24 +0000
      Re (2+++): ETHO interfacing linux IP:stack? no.top.post@gmail.com - 2012-03-24 17:22 +0000
        Re: Re (2+++): ETHO interfacing linux IP:stack? Tauno Voipio <tauno.voipio@notused.fi.invalid> - 2012-03-25 15:17 +0300
          Re: Re (2+++): ETHO interfacing linux IP:stack? Jorgen Grahn <grahn+nntp@snipabacken.se> - 2012-03-25 12:31 +0000
          Re (2+++): ETHO interfacing linux IP:stack? no.top.post@gmail.com - 2012-03-25 14:26 +0000

#1186 — ETHO interfacing linux IP:stack?

I'm writing this while I'm thinking it through, so I'll have a record.
Duke mentioned the problem of email for LNO.
I don't like LNO much, compared to LEO, which I use all the time.
But LNO can run on a netbook using a framebuffer, without the
full X-window system.

So let's think this problem through.
And correct my mistaken assumptions and conclusions:-

N-O was originaly designed for RS232-modem,
and the full <IP/TCP stack> was in N-O [of course].

So for  <send string1 to IP/port>,
all stages/levels, from the app's passing the string,
down to the serial port driving the modem,
was handled by N-O [of course].

Under LEO, the underlying linux does the internet handling.

Without anlysing the source code, I want to know how the
N-O part of LEO interfaces with linux.

Does the fact that "everything in *nix is a file" make this
easy?

Since linux, which previously drove a modem via a RS232,
can easily be 'reconfigured' to drive a radio-modem via a
USB; and passing the same byte-stream in both cases, I'm
guessing that just the appropriate file-driver are needed.

So the byte-stream is just fed to any character-type
file-device, and the appropriate driver handles it.
So the input-feeder need not be concerned what is
receiving the input.

If that is the case, and although the netbook has no actual
RS232, if I create a /dev/tty0 [pseudo]file and [symbolicly]
link it to the /dev/<existingUSB> , which is connected to
the remote IP:port via linux, could I communicate with the
remote IP:port via LNO? Would LNO treat the /devtty0/
as if it was a PC's "COM" port?
I guess not, since during boot, linux would see that there
is NO RS232-port and .....?

If not, I suspect that facilities exist to easily connect into
the IP:stack at various levels, and connecting at the highest
level would be most convenient, via an Oberon <-> C
interface ?

Thanks for any ideas,

== Chris Glur.

[toc] | [next] | [standalone]

#1187

On 22.3.12 9:28 , no.top.post@gmail.com wrote:
> I'm writing this while I'm thinking it through, so I'll have a record.
> Duke mentioned the problem of email for LNO.
> I don't like LNO much, compared to LEO, which I use all the time.
> But LNO can run on a netbook using a framebuffer, without the
> full X-window system.
>
> So let's think this problem through.
> And correct my mistaken assumptions and conclusions:-
>
> N-O was originaly designed for RS232-modem,
> and the full<IP/TCP stack>  was in N-O [of course].
>
> So for<send string1 to IP/port>,
> all stages/levels, from the app's passing the string,
> down to the serial port driving the modem,
> was handled by N-O [of course].
>
> Under LEO, the underlying linux does the internet handling.
>
> Without anlysing the source code, I want to know how the
> N-O part of LEO interfaces with linux.
>
> Does the fact that "everything in *nix is a file" make this
> easy?
>
> Since linux, which previously drove a modem via a RS232,
> can easily be 'reconfigured' to drive a radio-modem via a
> USB; and passing the same byte-stream in both cases, I'm
> guessing that just the appropriate file-driver are needed.
>
> So the byte-stream is just fed to any character-type
> file-device, and the appropriate driver handles it.
> So the input-feeder need not be concerned what is
> receiving the input.
>
> If that is the case, and although the netbook has no actual
> RS232, if I create a /dev/tty0 [pseudo]file and [symbolicly]
> link it to the /dev/<existingUSB>  , which is connected to
> the remote IP:port via linux, could I communicate with the
> remote IP:port via LNO? Would LNO treat the /devtty0/
> as if it was a PC's "COM" port?
> I guess not, since during boot, linux would see that there
> is NO RS232-port and .....?
>
> If not, I suspect that facilities exist to easily connect into
> the IP:stack at various levels, and connecting at the highest
> level would be most convenient, via an Oberon<->  C
> interface ?
>
> Thanks for any ideas,
>
> == Chris Glur.
>


Would you please make a glossary of the TLA's you used?

TLA = three letter abbreviation

Please get a basic text on TCP/IP networking, and read it first.
It will be very difficult to answer the above questions, as
it doen not make sense with the way SMTP mail is working.

-- 

Tauno Voipio

[toc] | [prev] | [next] | [standalone]

#1188

On Thursday, March 22nd, 2012, at 21:39:18h +0200, Tauno Voipio demanded:

> Would you please make a glossary of the TLA's you used?

        LEO == Low Earth Orbit

[toc] | [prev] | [next] | [standalone]

#1189

On Thursday 22 March 2012 16:07, in comp.os.linux.networking,
miller@yoyo.ORG wrote:

> On Thursday, March 22nd, 2012, at 21:39:18h +0200, Tauno Voipio demanded:
> 
>> Would you please make a glossary of the TLA's you used?
> 
>         LEO == Low Earth Orbit

Also
          LEO == LSI-Embedded realtime Operating system

http://cnu1.springnote.com/pages/3985325/attachments/1973837

But I doubt that this is the LEO that the OP refers to.

I can't find any Google references to LNO or N-O wrt Linux networking,
though.


-- 
Lew Pitcher

[toc] | [prev] | [next] | [standalone]

#1190

On Thursday 22 March 2012 16:27, in comp.os.linux.networking,
lpitcher@teksavvy.com wrote:

> On Thursday 22 March 2012 16:07, in comp.os.linux.networking,
> miller@yoyo.ORG wrote:
> 
>> On Thursday, March 22nd, 2012, at 21:39:18h +0200, Tauno Voipio demanded:
>> 
>>> Would you please make a glossary of the TLA's you used?
>> 
>>         LEO == Low Earth Orbit
> 
> Also
>           LEO == LSI-Embedded realtime Operating system

But, it apparently stands for

  Linux Eth Oberon
and
  LNO stands for Linux Native Oberon
and is apparently the runtime and support for the Oberon programming
language.

-- 
Lew Pitcher

[toc] | [prev] | [next] | [standalone]

#1192

On Thu, 22 Mar 2012, Lew Pitcher wrote:
> On Thursday 22 March 2012,  miller@yoyo.ORG wrote:
>>> On Thursday, March 22nd, 2012, Tauno Voipio demanded:
>> 
>>> Would you please make a glossary of the TLA's you used?
>> 
>>         LEO == Low Earth Orbit
>
> Also
>           LEO == LSI-Embedded realtime Operating system
>
> http://cnu1.springnote.com/pages/3985325/attachments/1973837
>
> But I doubt that this is the LEO that the OP refers to.
>
> I can't find any Google references to LNO or N-O wrt Linux networking,
> though.

LEO == Law Enforcement Official

Don'cha watch NCIS?  (Never Call It Silly)

Jonesy

[toc] | [prev] | [next] | [standalone]

#1191

On 2012-03-22, J G Miller <miller@yoyo.ORG> wrote:
>         LEO == Low Earth Orbit

LEO == Lyons Electronic Office

-- 
-----------------------------------------------------------------------------
  Roger Blake (Change "invalid" to "com" for email. Google Groups killfiled.)

  "Climate policy has almost nothing to do anymore with environmental
   protection... the next world climate summit in Cancun is actually
   an economy summit during which the distribution of the world's
   resources will be negotiated." -- Ottmar Edenhofer, IPCC
-----------------------------------------------------------------------------

[toc] | [prev] | [next] | [standalone]

#1193

"Tauno Voipio" <tauno.voipio@notused.fi.invalid> wrote in message 
news:jkfv56$8gb$1@dont-email.me...
>
> Would you please make a glossary of the TLA's you used?
>
> TLA = three letter abbreviation
>

LNO = Linux Native Oberon:

LEO = Linux ETH Oberon

USB = Universal Serial Bus

TCP = Transmission Control Protocol

Regards,
Chris Burrows

Astrobe: ARM Oberon Development System
http://www.astrobe.com

[toc] | [prev] | [next] | [standalone]

#1195 — Re (2): ETHO interfacing linux IP:stack?

In article <jkfv56$8gb$1@dont-email.me>, Tauno Voipio <tauno.voipio@notused.fi.invalid> wrote: 

> On 22.3.12 9:28 , no.top.post@gmail.com wrote:
> > I'm writing this while I'm thinking it through, so I'll have a record.
> > Duke mentioned the problem of email for LNO.
> > I don't like LNO much, compared to LEO, which I use all the time.
> > But LNO can run on a netbook using a framebuffer, without the
> > full X-window system.
> >
> > So let's think this problem through.
> > And correct my mistaken assumptions and conclusions:-
> >
> > N-O was originaly designed for RS232-modem,
> > and the full<IP/TCP stack>  was in N-O [of course].
> >
> > So for<send string1 to IP/port>,
> > all stages/levels, from the app's passing the string,
> > down to the serial port driving the modem,
> > was handled by N-O [of course].
> >
> > Under LEO, the underlying linux does the internet handling.
> >
> > Without anlysing the source code, I want to know how the
> > N-O part of LEO interfaces with linux.
> >
> > Does the fact that "everything in *nix is a file" make this
> > easy?
> >
> > Since linux, which previously drove a modem via a RS232,
> > can easily be 'reconfigured' to drive a radio-modem via a
> > USB; and passing the same byte-stream in both cases, I'm
> > guessing that just the appropriate file-driver are needed.
> >
> > So the byte-stream is just fed to any character-type
> > file-device, and the appropriate driver handles it.
> > So the input-feeder need not be concerned what is
> > receiving the input.
> >
> > If that is the case, and although the netbook has no actual
> > RS232, if I create a /dev/tty0 [pseudo]file and [symbolicly]
> > link it to the /dev/<existingUSB>  , which is connected to
> > the remote IP:port via linux, could I communicate with the
> > remote IP:port via LNO? Would LNO treat the /devtty0/
> > as if it was a PC's "COM" port?
> > I guess not, since during boot, linux would see that there
> > is NO RS232-port and .....?
> >
> > If not, I suspect that facilities exist to easily connect into
> > the IP:stack at various levels, and connecting at the highest
> > level would be most convenient, via an Oberon<->  C
> > interface ?
> >
> > Thanks for any ideas,
> >
> > == Chris Glur.
> >
> 
> Would you please make a glossary of the TLA's you used?
> 
> TLA = three letter abbreviation
> 
Some say 'foo'. Some say baa. Or being Friday night, is it 'bar'?

> Please get a basic text on TCP/IP networking, and read it first.
> It will be very difficult to answer the above questions, as
> it doen not make sense with the way SMTP mail is working.
> 
It's not about SMTP.  Surely once TCP/IP is 'hooked into',
SMTP should work if telnet can work ?
> -- 
> 
> Tauno Voipio
> 
Someone who KNOWS how linux implements TCP/IP, would know
how the system described below would/could be 'spliced in'.
-------------------
ReadMe: Linux Native Oberon System 3 Release 06.Jan.2002

Linux Native Oberon is an emulation of Native Oberon for Linux. The goal is 
binary compatibility with Native Oberon. For information about Native 
Oberon see the homepage at http://www.oberon.ethz.ch/native/
...
Linux Native Oberon doesn't change the interface of any module. All driver 
modules which are changed in LNO are statically linked in oberon.lno. If you 
first install Native Oberon and then LNO on an own partition you will be 
able to start oberon from both, PC Native or from Linux.
...

[toc] | [prev] | [next] | [standalone]

#1196 — Re (2+): ETHO interfacing linux IP:stack?

In article <jkfv56$8gb$1@dont-email.me>, Tauno Voipio <tauno.voipio@notused.fi.invalid> wrote: 

> On 22.3.12 9:28 , no.top.post@gmail.com wrote:
> > I'm writing this while I'm thinking it through, so I'll have a record.
> > Duke mentioned the problem of email for LNO.
> > I don't like LNO much, compared to LEO, which I use all the time.
> > But LNO can run on a netbook using a framebuffer, without the
> > full X-window system.
> >
> > So let's think this problem through.
> > And correct my mistaken assumptions and conclusions:-
> >
> > N-O was originaly designed for RS232-modem,
> > and the full<IP/TCP stack>  was in N-O [of course].
> >
> > So for<send string1 to IP/port>,
> > all stages/levels, from the app's passing the string,
> > down to the serial port driving the modem,
> > was handled by N-O [of course].
> >
> > Under LEO, the underlying linux does the internet handling.
> >
> > Without anlysing the source code, I want to know how the
> > N-O part of LEO interfaces with linux.
> >
> > Does the fact that "everything in *nix is a file" make this
> > easy?
> >
> > Since linux, which previously drove a modem via a RS232,
> > can easily be 'reconfigured' to drive a radio-modem via a
> > USB; and passing the same byte-stream in both cases, I'm
> > guessing that just the appropriate file-driver are needed.
> >
> > So the byte-stream is just fed to any character-type
> > file-device, and the appropriate driver handles it.
> > So the input-feeder need not be concerned what is
> > receiving the input.
> >
> > If that is the case, and although the netbook has no actual
> > RS232, if I create a /dev/tty0 [pseudo]file and [symbolicly]
> > link it to the /dev/<existingUSB>  , which is connected to
> > the remote IP:port via linux, could I communicate with the
> > remote IP:port via LNO? Would LNO treat the /devtty0/
> > as if it was a PC's "COM" port?
> > I guess not, since during boot, linux would see that there
> > is NO RS232-port and .....?
> >
> > If not, I suspect that facilities exist to easily connect into
> > the IP:stack at various levels, and connecting at the highest
> > level would be most convenient, via an Oberon<->  C
> > interface ?
> >
> > Thanks for any ideas,
> >
> > == Chris Glur.
> >
> 
> Would you please make a glossary of the TLA's you used?
> 
> TLA = three letter abbreviation
> 
The names of the arguments are arbitrary.
It's only the underlying concepts that matter.
I.e. HOW2 connect the program that 'rides on linux',
to the ppp layer.

The more I try to 'translate' from the original text,
the more I add distortion.

[toc] | [prev] | [next] | [standalone]

#1197 — Re: Re (2+): ETHO interfacing linux IP:stack?

On 24.3.12 7:58 , no.top.post@gmail.com wrote:
> In article<jkfv56$8gb$1@dont-email.me>, Tauno Voipio<tauno.voipio@notused.fi.invalid>  wrote:
>
>> On 22.3.12 9:28 , no.top.post@gmail.com wrote:
>>> I'm writing this while I'm thinking it through, so I'll have a record.
>>> Duke mentioned the problem of email for LNO.
>>> I don't like LNO much, compared to LEO, which I use all the time.
>>> But LNO can run on a netbook using a framebuffer, without the
>>> full X-window system.
>>>
>>> So let's think this problem through.
>>> And correct my mistaken assumptions and conclusions:-
>>>
>>> N-O was originaly designed for RS232-modem,
>>> and the full<IP/TCP stack>   was in N-O [of course].
>>>
>>> So for<send string1 to IP/port>,
>>> all stages/levels, from the app's passing the string,
>>> down to the serial port driving the modem,
>>> was handled by N-O [of course].
>>>
>>> Under LEO, the underlying linux does the internet handling.
>>>
>>> Without anlysing the source code, I want to know how the
>>> N-O part of LEO interfaces with linux.
>>>
>>> Does the fact that "everything in *nix is a file" make this
>>> easy?
>>>
>>> Since linux, which previously drove a modem via a RS232,
>>> can easily be 'reconfigured' to drive a radio-modem via a
>>> USB; and passing the same byte-stream in both cases, I'm
>>> guessing that just the appropriate file-driver are needed.
>>>
>>> So the byte-stream is just fed to any character-type
>>> file-device, and the appropriate driver handles it.
>>> So the input-feeder need not be concerned what is
>>> receiving the input.
>>>
>>> If that is the case, and although the netbook has no actual
>>> RS232, if I create a /dev/tty0 [pseudo]file and [symbolicly]
>>> link it to the /dev/<existingUSB>   , which is connected to
>>> the remote IP:port via linux, could I communicate with the
>>> remote IP:port via LNO? Would LNO treat the /devtty0/
>>> as if it was a PC's "COM" port?
>>> I guess not, since during boot, linux would see that there
>>> is NO RS232-port and .....?
>>>
>>> If not, I suspect that facilities exist to easily connect into
>>> the IP:stack at various levels, and connecting at the highest
>>> level would be most convenient, via an Oberon<->   C
>>> interface ?
>>>
>>> Thanks for any ideas,
>>>
>>> == Chris Glur.
>>>
>>
>> Would you please make a glossary of the TLA's you used?
>>
>> TLA = three letter abbreviation
>>
> The names of the arguments are arbitrary.
> It's only the underlying concepts that matter.
> I.e. HOW2 connect the program that 'rides on linux',
> to the ppp layer.
>
> The more I try to 'translate' from the original text,
> the more I add distortion.



I'm pretty sure that you do not want to connect directly
to the PPP (packet transport) layer from an application program.

The electronic mail, as we know it in the Internet, runs on
the SMTP (and POP or IMAP) protocols. All of these protocols
are running on TCP stream transport. TCP uses IP for moving
the segments as IP packets. IP can use several different
protocols for moving the packets. Some of them are SLIP,
PPP and Ethernet (with ARP).

The stack looks like this:

   SMTP   POP    IMAP
          TCP
          IP
   SLIP   PPP   Ethernet
     Modem      Net cable

Helper protocols left out.

In Linux, the way to access the protocol stack is via the
C library, libc, using socket connections. There are
different kinds of sockets for connecting to different
layers of the stack.

My understanding of Oberon is that it contains interface
modules from Oberon code to the C library, which is the
native connection in Linux.

There are special interfaces, tunnel and ethertap, for
creating non-standard network interfaces at the network
or physical layer.

The question is:

What are you really attempting to achieve?

-- 

Tauno Voipio

[toc] | [prev] | [next] | [standalone]

#1202 — Re (3) ETHO interfacing linux IP:stack?

In article <jkk6ms$jt$1@dont-email.me>, Tauno Voipio <tauno.voipio@notused.fi.invalid> wrote: 

> On 24.3.12 7:58 , no.top.post@gmail.com wrote:
> > In article<jkfv56$8gb$1@dont-email.me>, Tauno Voipio<tauno.voipio@notused.fi.invalid>  wrote:

> > I.e. HOW2 connect the program that 'rides on linux',
> > to the ppp layer.
> >
> 
> I'm pretty sure that you do not want to connect directly
> to the PPP (packet transport) layer from an application program.
> 
> The electronic mail, as we know it in the Internet, runs on
> the SMTP (and POP or IMAP) protocols. All of these protocols
> are running on TCP stream transport. TCP uses IP for moving
> the segments as IP packets. IP can use several different
> protocols for moving the packets. Some of them are SLIP,
> PPP and Ethernet (with ARP).
> 
> The stack looks like this:
> 
>    SMTP   POP    IMAP
>           TCP
>           IP
>    SLIP   PPP   Ethernet
>      Modem      Net cable
> 
> Helper protocols left out.
> 
> In Linux, the way to access the protocol stack is via the
> C library, libc, using socket connections. There are
> different kinds of sockets for connecting to different
> layers of the stack.
> 
---> That's the key:libc, using socket connections.
I'd like to connect to linux at the highest/application level.

> My understanding of Oberon is that it contains interface
> modules from Oberon code to the C library, which is the
> native connection in Linux.
> 
> There are special interfaces, tunnel and ethertap, for
> creating non-standard network interfaces at the network
> or physical layer.
> 
> The question is:
> 
> What are you really attempting to achieve?
> 
> -- 
> 
> Tauno Voipio
Currently and for the forseeable future good text processing is the
most important facility for general computing.

ETHO [family] is the best -- IMO.
ETHO dates back to the 90's with versions:
DOSbased, Win?, NativeX86
 and a later version LinuxNativeOberon,
 then LEO [a pasted the full name previously].
 
 The source code of all modules for the different versions,
 is identical, except for the hardware interfacing.
 
 Because I had problems of my ISP introducing TxAuthenticate,
 I had to look into the mailer's source code.  But I want to be able
 to do USEnet & http also- like I can onj THIS LEO.  I've been using 
 LEO for some years. And I open exta copies to handle problematic
 tasks, perhaps the way an emacs-guru would.
 
 I've installed non-X linux onto a CompactFlash on a netbook,
 which allows LNO to run [but not LEO].
 
 When LNO was installed on a PC with a serial-port, it could
 ppp via the modem.  But the netbook has no serial port, and
 I have already sweated-bullets to get the netbook under
 linux to communicate with a wireless-modem.
 
 So I want to be able to use the superior TextUserInterface of
 LNO, and connect to the internet via the linux - wireless modem.
 
 == Chris Glur.
 
 

[toc] | [prev] | [next] | [standalone]

#1201 — Re (2+++): ETHO interfacing linux IP:stack?

In article <jkfv56$8gb$1@dont-email.me>, Tauno Voipio <tauno.voipio@notused.fi.invalid> wrote: 

> On 22.3.12 9:28 , no.top.post@gmail.com wrote:
> > I'm writing this while I'm thinking it through, so I'll have a record.
> > Duke mentioned the problem of email for LNO.
> > I don't like LNO much, compared to LEO, which I use all the time.
> > But LNO can run on a netbook using a framebuffer, without the
> > full X-window system.
> >
> > So let's think this problem through.
> > And correct my mistaken assumptions and conclusions:-
> >
> > N-O was originaly designed for RS232-modem,
> > and the full<IP/TCP stack>  was in N-O [of course].
> >
> > So for<send string1 to IP/port>,
> > all stages/levels, from the app's passing the string,
> > down to the serial port driving the modem,
> > was handled by N-O [of course].
> >
> > Under LEO, the underlying linux does the internet handling.
> >
> > Without anlysing the source code, I want to know how the
> > N-O part of LEO interfaces with linux.
> >
> > Does the fact that "everything in *nix is a file" make this
> > easy?
> >
> > Since linux, which previously drove a modem via a RS232,
> > can easily be 'reconfigured' to drive a radio-modem via a
> > USB; and passing the same byte-stream in both cases, I'm
> > guessing that just the appropriate file-driver are needed.
> >
> > So the byte-stream is just fed to any character-type
> > file-device, and the appropriate driver handles it.
> > So the input-feeder need not be concerned what is
> > receiving the input.
> >
> > If that is the case, and although the netbook has no actual
> > RS232, if I create a /dev/tty0 [pseudo]file and [symbolicly]
> > link it to the /dev/<existingUSB>  , which is connected to
> > the remote IP:port via linux, could I communicate with the
> > remote IP:port via LNO? Would LNO treat the /devtty0/
> > as if it was a PC's "COM" port?
> > I guess not, since during boot, linux would see that there
> > is NO RS232-port and .....?
> >
> > If not, I suspect that facilities exist to easily connect into
> > the IP:stack at various levels, and connecting at the highest
> > level would be most convenient, via an Oberon<->  C
> > interface ?
> >
> > Thanks for any ideas,
> >
> > == Chris Glur.
> 
> Please get a basic text on TCP/IP networking, and read it first.
> It will be very difficult to answer the above questions, as
> it doen not make sense with the way SMTP mail is working.
 
I read the book - ten years ago.
This needs thinking beyond-the-book.
It's not restricted to SMTP.

Here's a paste of the interfacing code of 'a related/family'
which was also modified to 'ride on linux':-
<><><><>
cmdStr := "(/usr/lib/sendmail -oi -t < /tmp/Mailer.Send; rm /tmp/Mailer.Send) &";
Unix.System (SYSTEM.ADR(cmdStr), LEN(cmdStr));
(* Unix.System("/usr/lib/sendmail -oi -t < /tmp/Mailer.Send"); *)
(* Files.Delete("/tmp/Mailer.Send", res) *)
<><><><>
If he was able to interface with the linux screen & mouse and fetch
http, I'm wondering why he used such a naive hack for sendingMail,
given that the system must have had a mailer, in the original 
'nativeX86'/non-linux-based version.

Of course I may study the source code, but I want to try to
analyse it at a higher-level first.

== TIA.

[toc] | [prev] | [next] | [standalone]

#1213 — Re: Re (2+++): ETHO interfacing linux IP:stack?

On 24.3.12 7:22 , no.top.post@gmail.com wrote:

>
> Here's a paste of the interfacing code of 'a related/family'
> which was also modified to 'ride on linux':-
> <><><><>
> cmdStr := "(/usr/lib/sendmail -oi -t<  /tmp/Mailer.Send; rm /tmp/Mailer.Send)&";
> Unix.System (SYSTEM.ADR(cmdStr), LEN(cmdStr));
> (* Unix.System("/usr/lib/sendmail -oi -t<  /tmp/Mailer.Send"); *)
> (* Files.Delete("/tmp/Mailer.Send", res) *)
> <><><><>

Here, you're getting out of Oberon and running a
subordinate process instead of interfacing to the
mail system.

This is the same as a shell commands:

/usr/lib/sendmail -oi -t </tmp/Mailer.Send
rm /tmp/Mailer.Send

I sure hope that the command is not the original Sendmail.
It is notoriously difficult to configure correctly. Exim
is one better substitute.

-- 

Tauno Voipio

[toc] | [prev] | [next] | [standalone]

#1214 — Re: Re (2+++): ETHO interfacing linux IP:stack?

["Followup-To:" header set to comp.os.linux.networking.]

On Sun, 2012-03-25, Tauno Voipio wrote:
> On 24.3.12 7:22 , no.top.post@gmail.com wrote:
>
...
>> cmdStr := "(/usr/lib/sendmail -oi -t<  /tmp/Mailer.Send; rm /tmp/Mailer.Send)&";
...
> Here, you're getting out of Oberon and running a
> subordinate process instead of interfacing to the
> mail system.
...
> I sure hope that the command is not the original Sendmail.

All Unix mail systems have a /usr/lib/sendmail, so this name doesn't
indicate which one is running.

/Jorgen

-- 
  // Jorgen Grahn <grahn@  Oo  o.   .     .
\X/     snipabacken.se>   O  o   .

[toc] | [prev] | [next] | [standalone]

#1215 — Re (2+++): ETHO interfacing linux IP:stack?

In article <jkn2c2$u3a$1@dont-email.me>, Tauno Voipio <tauno.voipio@notused.fi.invalid> wrote: 

> On 24.3.12 7:22 , no.top.post@gmail.com wrote:
> >
> > Here's a paste of the interfacing code of 'a related/family'
> > which was also modified to 'ride on linux':-
> > <><><><>
> > cmdStr := "(/usr/lib/sendmail -oi -t<  /tmp/Mailer.Send; rm /tmp/Mailer.Send)&";
> > Unix.System (SYSTEM.ADR(cmdStr), LEN(cmdStr));
> > (* Unix.System("/usr/lib/sendmail -oi -t<  /tmp/Mailer.Send"); *)
> > (* Files.Delete("/tmp/Mailer.Send", res) *)
> > <><><><>
> 
> Here, you're getting out of Oberon and running a
> subordinate process instead of interfacing to the
> mail system.
> 
> This is the same as a shell commands:
> 
> /usr/lib/sendmail -oi -t </tmp/Mailer.Send
> rm /tmp/Mailer.Send
> 
Yes I just put that in there, for you to feel familiar and comfortable.
It's rather trivial to write/modify the complete pop/smtp in ETHO.
Here's some more code snippets, which I had to look at to add TxAuthenticate:
From the RFC where you have to give: UserID
	Connect(S, host, POP, res);
		IF res = 0 THEN NS.ReadString(S, s);
			IF s[0] = "+" THEN SendCmd(S, "user", user); 
------------> which goes to:
-> SendCmd(S: NS.Connection; cmd, arg: ARRAY OF CHAR);
->  NS.WriteBytes(S, 0, i+2, buf)
WriteBytes (C: Connection; pos, len: LONGINT; VAR buf: NetBase.Data);
->	TYPE
		IPAdr = NetBase.IPAdr;
		Connection = POINTER TO ConnectionDesc;  <-- C: Connection
		ConnectionDesc = RECORD
			res: INTEGER;
		END;
		
	TYPE
		IPAdr = ARRAY 4 OF CHAR;    <-- NetBase.IPAdr;
		Data = ARRAY OF SYSTEM.BYTE;
		Sockaddrin = RECORD
			sinfamily: INTEGER;
			sinport: INTEGER;
->			sinaddr: IPAdr;
			sinzero: ARRAY 8 OF CHAR
		END;
============================
Now lets get back to the real problem:
I'll consider the 'Mailer' because I'm most familiar with it
compared to News & http-fetch.

LNO is running on linux. `pstree` shows that it's a descendant of `bash`

A=>LNO: Mailer < ........<ser-port < modem

B=> Linux < pppd chat script<...<tttyUSB0/1 <wireless-modem

The final target hardware is a netbook, which has NO ser-port,
but I can test on a PC which has USB and ser-port.

I connect to the-net via the linux-chat-script;
and then I tried to 'talk via LNO' by having /dev/tty1 linked to
/dev/ttyUSB0. It fails and LNO hasn't got `ping`to help test.

Firstly I linked BOTH /dev/modem for Linux and
 /dev/tty1 for LNO  /dev/ttyUSBO, but that
 seems strange.
 
 This is really a problem for kernel-programmers?
 

[toc] | [prev] | [standalone]

Back to top | Article view | comp.os.linux.networking

csiph-web