Re: RFE: ncat should call shutdown on EOF

[Henri Doreau <henri.doreau () gmail com>]

2012/11/7 David Fifield <david () bamsoftware com>
> Henri, I bisected and identified this revision as the one causing some
> existing ncat-test.pl errors. Would you check it?
>
> FAIL Messages are logged to output file (tcp)
>      Server logged "abc\n" at ./ncat-test.pl line 670.
> PASS Messages are logged to output file (udp)
> FAIL Messages are logged to output file (sctp)
>      Server logged "abc\n" at ./ncat-test.pl line 670.
> FAIL Messages are logged to output file (tcp ssl)
>      Server logged "abc\n" at ./ncat-test.pl line 670.
> FAIL Messages are logged to output file (sctp ssl)
>      Server logged "abc\n" at ./ncat-test.pl line 670.
> FAIL Debug messages go to stderr (tcp)
>      Read timeout at ./ncat-test.pl line 690.
> FAIL Debug messages go to stderr (sctp)
>      Read timeout at ./ncat-test.pl line 690.
> FAIL Debug messages go to stderr (tcp ssl)
>      Read timeout at ./ncat-test.pl line 690.
> FAIL Debug messages go to stderr (sctp ssl)
>      Read timeout at ./ncat-test.pl line 690.
>
> David

Hi,

isn't this the expected behavior? If I understand the test correctly it
writes things on the client side, closes client stdin (EOF), then writes
on server side. The test fails because only the first part gets logged
(i.e.: what was written on client side before EOF).

Unless I'm missing something, this behavior is consistent: EOF triggers
the added shutdown() statement and communication ceases between client
and server.

I can mark the tests as XFAIL and maybe try to write new ones (perl
doesn't look that terrible at second glance). If I'm mistaken I would
need some quick explanations on what is the actually desired behavior.

Regards

--
Henri
_______________________________________________
Sent through the nmap-dev mailing list
http://cgi.insecure.org/mailman/listinfo/nmap-dev
Archived at http://seclists.org/nmap-dev/