1. Forum
  2. Usenet
  3. LINUX.DEBIAN.BUGS.DIST

  • Bug#1104275: pique: autopkgtest failure, possibly not allowed in main

    From =?utf-8?Q?=C3=89tienne?= Mollier@21:1/5 to All on Mon Apr 28 23:20:01 2025
    Control: tags -1 + confirmed upstream
    Control: forwarded -1 https://github.com/tony-travis/PIQUE/issues/4

    Hi Chris,

    Thank you for your report, I have informed upstream on their
    issue tracker.

    About the point of policy violation, the package has d/t/control
    restriction "needs-internet" as described below from autopkgtest
    documentation README.package-tests.rst:

    | needs-internet
    | The test needs unrestricted internet access, e.g. to download test data
    | that's not shipped as a package, or to test a protocol implementation
    | against a test server. Please also see the note about Network access later
    | in this document.

    I believe that if the autopkgtest should not attempt to reach
    the internet from our infrastructure, then maybe such tests
    should be skipped, since they are clearly identified as such in
    their d/t/control information. That being said, I agree I am
    not fan of tests which require connectivity, because as seen
    once again, they are fragile and break as soon as the remote
    resource is taken down, corrupted, etc.

    Hopefully this is fixable by shipping new test data, but in the
    meantime, the only way out seems to disable autopkgtest.

    Have a nice day, :)
    --
    .''`. Étienne Mollier <emollier@debian.org>
    : :' : pgp: 8f91 b227 c7d6 f2b1 948c 8236 793c f67e 8f0d 11da
    `. `' sent from /dev/pts/2, please excuse my verbosity
    `- on air: Genesis - Home By The Sea

    -----BEGIN PGP SIGNATURE-----

    iQIzBAABCgAdFiEEj5GyJ8fW8rGUjII2eTz2fo8NEdoFAmgP7pcACgkQeTz2fo8N EdoOdg//RLzJL607EqqeRtml0oJ96gKrtzuLh4TZf5RnQksZbXcdLcZtRXs7dSvI rsJ0pD+e98orMYwxSf4XbxXdYt/Q4e41NoOxfc0TQks2QOCxC+/QxyzPrRexGwH8 2IT0Yx7Hrg75trpCo3NPGC5rQkme2vFz+PeM719wIO6uCUnWT8bHMFd7yt7TC1TH acfv5dyZZ+xJ2a9nSlX8SNHt67VuKvryPIcPWfDp45DndmYAFWFprtaDCYHNzcfg 5B0SMVk3hoHuMG8rR7dRW6V5VE8essnDmO5PBTD4lKBCwVF75t9Pag8iLlyRTKrW twitAWT76DED82eivWhro2Hav0IsfJPmVRmQ1/X3F0NRIU9PSWXgG30rtk6oCkuG Cn7QcQJ411ZywhvvzV8mm1XP95zs2MZ9PB3+B4X2TQwMghfRBY7IKfQSwqKt9cuv iNm4UZcAUy/nEKqoTPitVvpYmQijCPJzxR2DJzvKjSWwGdg9OM7hJwIDj2ivJRdZ 0RF54k3PPoBnE9ssMgiHnqKNdrv98Xxb+j4WX/uyk2wN+T5unyxnC+/159lkrmgC NTFpk2+MzHPo47gjO3htunzlYhX8JyV1BP9Dfx0/DbxMtsHCR5Yib/i1E+G2dfId yEFnjF08t5nT1v5NKXZvn46evE0oUkedwhqrT3rl
  • From Chris Hofstaedtler@21:1/5 to All on Tue Apr 29 01:50:01 2025
    * Étienne Mollier <emollier@debian.org> [250428 23:09]:
    About the point of policy violation, the package has d/t/control
    restriction "needs-internet" as described below from autopkgtest >documentation README.package-tests.rst:

    | needs-internet
    | The test needs unrestricted internet access, e.g. to download test data >| that's not shipped as a package, or to test a protocol implementation
    | against a test server. Please also see the note about Network access later
    | in this document.

    I believe that if the autopkgtest should not attempt to reach
    the internet from our infrastructure, then maybe such tests
    should be skipped, since they are clearly identified as such in
    their d/t/control information. That being said, I agree I am
    not fan of tests which require connectivity, because as seen
    once again, they are fragile and break as soon as the remote
    resource is taken down, corrupted, etc.

    Having thought about this a minute more: the "needs-internet"
    description you are quoting is a technical description of an
    autopkgtest feature.

    ftp-masters doc is a policy statement.

    I *think* the policy statement will have to win over the feature
    description. (Obviously we technically could do lots of things, but
    then various policies prevent us from doing so in Debian main.)

    Chris

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)