develooper Front page | perl.perl5.porters | Postings from October 2017

[perl #132363] Net::Ping 2.61 test failures

Thread Previous
From:
James E Keenan via RT
Date:
October 28, 2017 13:46
Subject:
[perl #132363] Net::Ping 2.61 test failures
Message ID:
rt-4.0.24-17266-1509198373-1493.132363-15-0@perl.org
On Fri, 27 Oct 2017 09:52:32 GMT, mhasch-cpanbugs@cozap.com wrote:
> 
> This is a bug report for perl from mhasch-cpanbugs@cozap.com,
> generated with the help of perlbug 1.40 running under perl 5.26.1.
> 
> 
> -----------------------------------------------------------------
> [Please describe your issue here]
> 
> As of version 2.61, the Net::Ping test suite makes assumptions
> about the IPv4 address 172.29.249.249.	This is an RFC 1918
> address that may or may not be reachable in any local area
> network.  Making tests fail based on its reachability is
> therefore non-portable and should be avoided.
> 

For the record, I note that Net::Ping tests using 172.29.249.249 were introduced long before 2.61.  They date at least to commit f569508e in October 2002.

> While replacing the RFC 1918 address by an RFC 5737 address
> might reduce the risk of it actually being connected there
> is no guarantee any address usage convention at all will be
> met within private networks.  Therefore, in the test suite,
> (non-)reachability of a particular address should always be
> tolerated.
> 
> Interestingly, some tests against "known" internet hosts are
> already skipped if an initial ping of "google.com" fails for
> some reason.  Giving up this particular trigger and making all
> connectivity-related checks report skip reasons rather than
> failures could be a way to cope with most private and firewalled
> network situations.
> 
> Even better though would be a test strategy aiming for code
> quality criteria independent of outside conditions.  This might
> be achieved using mock-up interfaces or at most localhost.
> 
> If nothing else, at least provide an opt-out configuration
> parameter for remote network-aware tests.  Currently, PERL_CORE
> in the environment will have that effect.  Unfortunately, this
> parameter usually is a MakeMaker constant not under control
> of the user (through distroprefs, say).  But not only core may
> need to skip those tests.
> 
> Regards,
> -Martin
> 
> [Please do not change anything below this line]
> -----------------------------------------------------------------
> ---
> Flags:
>     category=library
>     severity=high
>     module=Net::Ping
> ---
> Site configuration information for perl 5.26.1:
> 
> (stripped by reporter, considered irrelevant)


-- 
James E Keenan (jkeenan@cpan.org)

---
via perlbug:  queue: perl5 status: new
https://rt.perl.org/Ticket/Display.html?id=132363

Thread Previous


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About