[Boost-users] Simple telnet client demonstration with boost asio asynchronous I/O

[Jeff Gray]

I've been teaching myself how to use the asynchronous I/O capabilities
in Boost asio. I took the chat client example code that came with Boost
and modified it to act as a more general telnet client.

Hopefully, this may be of use to others trying to get to grips with the
basics of this way of working. More experienced developers can feel free
to suggest better ways of doing things here.

This is just a demonstration - it does not do all the niceties of a full
application, but hopefully covers the basics.

I wrote this for a Linux system, but the only system specific thing is
the code that enables single keypress input for cin.get(). I'm sure you
can easily find alternatives for your system to achieve the same result.

Enjoy,
Jeff

[Roland Bock]

Hi Jeff,

thanks, I am going to dive into Boost::Asio myself, soon, and will come
back to your example then :-)

Just wondering (while being aware, that this is not a fully fledged
telnet client):
I tried to connect to localhost, port 23. There is no service running,
so I would have expected the program to terminate (it terminates when I
specify a host which does not exist). I even sent 800MB of XML data into
the program. It did not complain and wrote it to Nirvana, I guess :-)

Regards,

Roland

> ------------------------------------------------------------------------
>
> _______________________________________________
> Boost-users mailing list
> Boost...@lists.boost.org
> http://lists.boost.org/mailman/listinfo.cgi/boost-users

_______________________________________________
Boost-users mailing list
Boost...@lists.boost.org
http://lists.boost.org/mailman/listinfo.cgi/boost-users

[Jeff Gray]

Roland Bock wrote:
> thanks, I am going to dive into Boost::Asio myself, soon, and will come
> back to your example then :-)
>
> Just wondering (while being aware, that this is not a fully fledged
> telnet client):
> I tried to connect to localhost, port 23. There is no service running,
> so I would have expected the program to terminate (it terminates when I
> specify a host which does not exist). I even sent 800MB of XML data into
> the program. It did not complain and wrote it to Nirvana, I guess :-)

You're quite right - I haven't tested this code for failure like this.
My initial thought is that it probably silently calls the close function
and exits the communication thread without exiting the application.

I'll look at adding this resubmitting the code - it should be a
straightforward change.

Jeff

______________________________________________________________________
This email has been scanned by the MessageLabs Email Security System.
For more information please visit http://www.messagelabs.com/email
______________________________________________________________________

[Andrew Maclean]

Here is a CMakeLists.txt file that will let you use CMake
(http://www.cmake.org) to create either a visual studio solution or a
kdevelop/unix makefile solution. Your code builds with no errors using
Visual Studio 2008 and (of course) Linux (Debian), I haven't tried a
mac.

Andrerw

--
___________________________________________
Andrew J. P. Maclean
Centre for Autonomous Systems
The Rose Street Building J04
The University of Sydney 2006 NSW
AUSTRALIA
Ph: +61 2 9351 3283
Fax: +61 2 9351 7474
URL: http://www.acfr.usyd.edu.au/
___________________________________________

[Jeff Gray]

Andrew Maclean wrote:
> Here is a CMakeLists.txt file that will let you use CMake
> (http://www.cmake.org) to create either a visual studio solution or a
> kdevelop/unix makefile solution. Your code builds with no errors using
> Visual Studio 2008 and (of course) Linux (Debian), I haven't tried a
> mac.

Thanks for that Andrew. Nice to see an easy portable build system.

I've worked on Roland Bock's suggestion of timeouts to handle failure to
connect & attached the working code.

When an error occurs in the class, in my example code, it sets a
variable which is checked by the main thread. I'm sure there are better
built in ways of dealing with this for Boost asio or threads.

Presumably, when the socket is closed, this means the io_service.run()
function returns and the thread is terminated. If anyone has any
suggestions how to detect any of this from the main thread to determine
when to exit, I'd like to hear them.

Regards,
Jeff

[Andrew Maclean]

It works nicely, this is a good example of using ASIO.

The CMakeLists.txt file could have been made more minimal but I hope
people find the extra bits useful as templates.

Regards
Andrew