(tests guix-download) fails when DNS is bogus

  • Done
  • quality assurance status badge
Details
4 participants
  • Leo Famulari
  • Ludovic Courtès
  • Maxim Cournoyer
  • ng0
Owner
unassigned
Submitted by
Leo Famulari
Severity
normal
L
L
Leo Famulari wrote on 23 May 2017 18:24
(address . bug-guix@gnu.org)
20170523162431.GB15379@jasmine
The guix-download test includes this snippet:

------
# Make sure it fails here.
if guix download http://does.not/exist
then false; else true; fi
------

Unfortunately, many ISPs (such as T-Mobile) return bogus results for
otherwise unclaimed domain names, causing this test to fail.

Does anyone know if there is some domain that is designed to fail as a
"standard", as http://example.com is intended to be used for examples
of good domains?
-----BEGIN PGP SIGNATURE-----
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=H/x/
-----END PGP SIGNATURE-----


M
M
Maxim Cournoyer wrote on 23 May 2017 18:47
BC77E34E-40B0-402A-9C85-4223FAB5A8CB@gmail.com
On May 23, 2017 9:24:31 AM PDT, Leo Famulari <leo@famulari.name> wrote:
Toggle quote (15 lines)
>The guix-download test includes this snippet:
>
>------
># Make sure it fails here.
>if guix download http://does.not/exist
>then false; else true; fi
>------
>
>Unfortunately, many ISPs (such as T-Mobile) return bogus results for
>otherwise unclaimed domain names, causing this test to fail.
>
>Does anyone know if there is some domain that is designed to fail as a
>"standard", as <http://example.com> is intended to be used for examples
>of good domains?

Could we simply mock the expected response?

Maxim
N
(name . 27039)(address . 27039@debbugs.gnu.org)
E1dDDHH-0002gK-UV@rmmprod05.runbox
On Tue, 23 May 2017 12:24:31 -0400, Leo Famulari <leo@famulari.name> wrote:

Toggle quote (15 lines)
> The guix-download test includes this snippet:
>
> ------
> # Make sure it fails here.
> if guix download http://does.not/exist
> then false; else true; fi
> ------
>
> Unfortunately, many ISPs (such as T-Mobile) return bogus results for
> otherwise unclaimed domain names, causing this test to fail.
>
> Does anyone know if there is some domain that is designed to fail as a
> "standard", as <http://example.com> is intended to be used for examples
> of good domains?

http://fail.0

fails to resolve with "ping: bad address 'fail.0'. This was run on an OpenNIC connected computer.
L
L
Ludovic Courtès wrote on 23 May 2017 23:03
(name . Leo Famulari)(address . leo@famulari.name)(address . 27039@debbugs.gnu.org)
87shjv70by.fsf@gnu.org
Leo Famulari <leo@famulari.name> skribis:

Toggle quote (11 lines)
> The guix-download test includes this snippet:
>
> ------
> # Make sure it fails here.
> if guix download http://does.not/exist
> then false; else true; fi
> ------
>
> Unfortunately, many ISPs (such as T-Mobile) return bogus results for
> otherwise unclaimed domain names, causing this test to fail.

Yes, this has been reported several times in the past and marked as
“wontfix”. :-)

Toggle quote (4 lines)
> Does anyone know if there is some domain that is designed to fail as a
> "standard", as <http://example.com> is intended to be used for examples
> of good domains?

There’s an RFC defining example.org et al. These domain names are
guaranteed to exist, so we’d be testing something different; also, I
don’t know if the RFC defines pages guaranteed to be 404, for instance.

Ludo’.
L
L
Ludovic Courtès wrote on 4 Jun 2017 23:04
control message for bug #27039
(address . control@debbugs.gnu.org)
87vaobh3bt.fsf@gnu.org
tags 27039 wontfix
close 27039
?