perlgssapi-developer

[Perlgssapi-developer] Re: [rt.cpan.org #19524]

[Achim G. <ac...@gr...>]

On Saturday 27 May 2006 16:15, Florian Ragwitz via RT wrote:

> Sat May 27 10:15:39 2006: Request 19524 was acted upon.
> Ticket <URL: http://rt.cpan.org/Ticket/Display.html?id=19524 >

Thank you for sending the bugreport!

> The submitter complains that the modules tests will fail while being
> offline, as some of them require network access.

I've checked this, he is right.
I get the same fails after shutting down the networkinterface.

(I think it depends on the underlying Kerberosimplemention what happens,
I'm using MIT 1.3).

> I'd suggest to either skip all tests that require networking unless some
> condition is met, for example if $ENV{TEST_ONLINE} is set, or to be
> smarter and try to detect if a network connection is available when
> running the test suite.

Skipping tests on some conditions is no problem,
I can add that.

But I dont't know *what* condition to test.
Sure I can test if $ENV{TEST_ONLINE} is set, but who or what software
sets that System-variable?

Can you help me and answer what's the condition to test to?

Achim

[Perlgssapi-developer] Re: [rt.cpan.org #19524]

[Florian R. <ra...@de...>]

On Sat, May 27, 2006 at 05:12:12PM +0200, Achim Grolms wrote:
> On Saturday 27 May 2006 16:15, Florian Ragwitz via RT wrote:
> > I'd suggest to either skip all tests that require networking unless some
> > condition is met, for example if $ENV{TEST_ONLINE} is set, or to be
> > smarter and try to detect if a network connection is available when
> > running the test suite.
>=20
> Skipping tests on some conditions is no problem,
> I can add that.
>=20
> But I dont't know *what* condition to test.
> Sure I can test if $ENV{TEST_ONLINE} is set, but who or what software
> sets that System-variable?

The user. Just like those pod and podcoverage test lots of people use
nowadays.

Those tests usually get skiped unless $ENV{TEST_POD} is set and the SKIP
message tells the user that, if he wants pod testing, he needs to set
that variable.


-Flo

--=20
BOFH excuse #392:
It's union rules. There's nothing we can do about it. Sorry.

[Perlgssapi-developer] Re: [rt.cpan.org #19524]

[Achim G. <ac...@gr...>]

On Saturday 27 May 2006 17:19, Florian Ragwitz via RT wrote:
>        Queue: GSSAPI
>  Ticket <URL: http://rt.cpan.org/Ticket/Display.html?id=3D19524 >

> Those tests usually get skiped unless $ENV{TEST_POD} is set and the SKIP
> message tells the user that, if he wants pod testing, he needs to set
> that variable.

I've had a look into tho t/test.t file:
The problem is the usage of gss_nt_service_name
in import() what makes the test always using the DNS.
(What fails if DNS it not available).

I have changed that (not to use DNS),=20
tested with MIT and Heimdal,=20
commitet changes to SVN Rep.

https://svn.sourceforge.net/svnroot/perlgssapi/GSSAPI/trunk

Please let me know if that changes work for you,
I will release GSSAPI 0.22 to CPAN if you think it works.

Thank you,
Achim

[Perlgssapi-developer] Re: [rt.cpan.org #19524]

[Florian R. <ra...@de...>]

On Sat, May 27, 2006 at 06:14:22PM +0200, Achim Grolms wrote:
> On Saturday 27 May 2006 17:19, Florian Ragwitz via RT wrote:
> >        Queue: GSSAPI
> >  Ticket <URL: http://rt.cpan.org/Ticket/Display.html?id=3D19524 >
>=20
> > Those tests usually get skiped unless $ENV{TEST_POD} is set and the SKIP
> > message tells the user that, if he wants pod testing, he needs to set
> > that variable.
>=20
> I've had a look into tho t/test.t file:
> The problem is the usage of gss_nt_service_name
> in import() what makes the test always using the DNS.
> (What fails if DNS it not available).
>=20
> I have changed that (not to use DNS),=20
> tested with MIT and Heimdal,=20
> commitet changes to SVN Rep.
>=20
> https://svn.sourceforge.net/svnroot/perlgssapi/GSSAPI/trunk
>=20
> Please let me know if that changes work for you,
> I will release GSSAPI 0.22 to CPAN if you think it works.

I verified that your changes fix the problem.


Thanks,
Flo

--=20
BOFH excuse #103:
operators on strike due to broken coffee machine