• Re: Bug#1018110: RFS: hydrapaper/3.3.1-1 [RC] -- Utility that sets back

    From Tobias Frost@21:1/5 to Francisco M Neto on Sat Aug 27 15:10:01 2022
    On Sat, Aug 27, 2022 at 09:38:31AM -0300, Francisco M Neto wrote:
    Greetings!

    Thank you very much for taking the time to review this package! When it comes to packaging I clearly still have a long way to go.

    I have fixed all those issues, including #1018222. I have tested the package extensively and installed it on a fresh unstable installation running inside a virtual machine, and everything seems to be in working order.

    Before I re-submit this package for sponsorship, I do have one doubt, though: because it reached the NEW queue, I'm not sure if I should tag it 3.3.1-
    1 or 3.3.1-2. Which one should I use?

    Maybe I missed something, but there should be no NEW queue involved.
    However, said that, even if a NEW queue is involved, once accepted, this version
    number is used, so anyways you would need to increment.

    --
    tobi


    Thanks a lot!
    Francisco

    On Sat, 2022-08-27 at 12:00 +0200, Jeroen Ploemen wrote:
    Control: tags -1 moreinfo

    On Thu, 25 Aug 2022 15:49:14 -0300
    Francisco M Neto <[email protected]> wrote:

    I am looking for a sponsor for my package hydrapaper:

    hi Francisco,

    took a look but this package doesn't appear ready for uploading:
    * changelog: is that bug really fixed just by switching to gtk4?
    � There's still no dependency on python3-pil while the program is
    � directly importing from that module!
    * copyright: missing entry for the appdata xml file (cc0).
    * patches: forwarded upstream but the related merge request was
    � closed by yourself; why? is the patch still needed?
    * watch: multiple empty lines at EOF
    * control:
    �+ short and long description could use an update (upstream describes
    �� the program as a "Wallpaper manager with multi monitor support";
    �� mention additional supported desktop environments, etc.)
    �+ unused build-dep on python3-willow?
    �+ the build-dep on libwnck-3-dev appears to server no other purpose
    �� than pulling in the dbus-1 pkgconfig file from libdbus-1-dev; if
    �� so, you should depend on the latter directly
    �+ libhandy-1-0 is a hard dependency of gir1.2-handy-1 but not
    �� imported or linked directly in hydrapaper, so no need to duplicate
    �� that here
    �+ gir1.2-handy-1 itself looks isn't used at all in the new upstream
    �� release so that should go too
    �+ ${shlibs:Depends} is pointless for an arch:all Python package �������������������������
    Program fails to start (missing dep on something to ensure gi gtk4 is present, installing gir1.2-gtk-4.0 seems to fix that):
    Traceback (most recent call last):
    � File "/usr/bin/hydrapaper", line 60, in <module>
    ��� gi.require_version('Gtk', '4.0')
    � File "/usr/lib/python3/dist-packages/gi/__init__.py", line 129, in require_version
    ��� raise ValueError('Namespace %s not available for version %s' % ValueError: Namespace Gtk not available for version 4.0

    Same for adw:
    Traceback (most recent call last):
    � File "/usr/bin/hydrapaper", line 62, in <module>
    ��� gi.require_version('Adw', '1')
    � File "/usr/lib/python3/dist-packages/gi/__init__.py", line 126, in require_version
    ��� raise ValueError('Namespace %s not available' % namespace)
    ValueError: Namespace Adw not available

    Probably missing a dependency on python3-dbus too (imported by hydrapaperd)? And python3-pil as mentioned earlier.

    Please at least take a cursory look at upstream code when packaging
    major version bumps, and test your packages on a reasonably clean testing/unstable install before asking for sponsorship.

    Consider adding some basic automated testing, as even a trivial
    autopkgtest that just calls `hydrapaper --help' would have failed
    with errors similar to the ones listed above.


    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Francisco M Neto@21:1/5 to Jeroen Ploemen on Sat Aug 27 14:40:01 2022
    Greetings!

    Thank you very much for taking the time to review this package! When it comes to packaging I clearly still have a long way to go.

    I have fixed all those issues, including #1018222. I have tested the package extensively and installed it on a fresh unstable installation running inside a virtual machine, and everything seems to be in working order.

    Before I re-submit this package for sponsorship, I do have one doubt, though: because it reached the NEW queue, I'm not sure if I should tag it 3.3.1-
    1 or 3.3.1-2. Which one should I use?

    Thanks a lot!
    Francisco

    On Sat, 2022-08-27 at 12:00 +0200, Jeroen Ploemen wrote:
    Control: tags -1 moreinfo

    On Thu, 25 Aug 2022 15:49:14 -0300
    Francisco M Neto <[email protected]> wrote:

    I am looking for a sponsor for my package hydrapaper:

    hi Francisco,

    took a look but this package doesn't appear ready for uploading:
    * changelog: is that bug really fixed just by switching to gtk4?
      There's still no dependency on python3-pil while the program is
      directly importing from that module!
    * copyright: missing entry for the appdata xml file (cc0).
    * patches: forwarded upstream but the related merge request was
      closed by yourself; why? is the patch still needed?
    * watch: multiple empty lines at EOF
    * control:
     + short and long description could use an update (upstream describes
       the program as a "Wallpaper manager with multi monitor support";
       mention additional supported desktop environments, etc.)
     + unused build-dep on python3-willow?
     + the build-dep on libwnck-3-dev appears to server no other purpose
       than pulling in the dbus-1 pkgconfig file from libdbus-1-dev; if
       so, you should depend on the latter directly
     + libhandy-1-0 is a hard dependency of gir1.2-handy-1 but not
       imported or linked directly in hydrapaper, so no need to duplicate
       that here
     + gir1.2-handy-1 itself looks isn't used at all in the new upstream
       release so that should go too
     + ${shlibs:Depends} is pointless for an arch:all Python package                          
    Program fails to start (missing dep on something to ensure gi gtk4 is present, installing gir1.2-gtk-4.0 seems to fix that):
    Traceback (most recent call last):
      File "/usr/bin/hydrapaper", line 60, in <module>
        gi.require_version('Gtk', '4.0')
      File "/usr/lib/python3/dist-packages/gi/__init__.py", line 129, in require_version
        raise ValueError('Namespace %s not available for version %s' % ValueError: Namespace Gtk not available for version 4.0

    Same for adw:
    Traceback (most recent call last):
      File "/usr/bin/hydrapaper", line 62, in <module>
        gi.require_version('Adw', '1')
      File "/usr/lib/python3/dist-packages/gi/__init__.py", line 126, in require_version
        raise ValueError('Namespace %s not available' % namespace)
    ValueError: Namespace Adw not available

    Probably missing a dependency on python3-dbus too (imported by
    hydrapaperd)? And python3-pil as mentioned earlier.

    Please at least take a cursory look at upstream code when packaging
    major version bumps, and test your packages on a reasonably clean testing/unstable install before asking for sponsorship.

    Consider adding some basic automated testing, as even a trivial
    autopkgtest that just calls `hydrapaper --help' would have failed
    with errors similar to the ones listed above.


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

    iQIzBAABCgAdFiEEPlgWVZo9XXifkM/x0wsWlNaS+/AFAmMKEEcACgkQ0wsWlNaS +/DSHw//TLQI8Fb1W9Ks1DLK5e3d5cIVYoCEM4Sp7gZDrmBPSha+yczg83DFKSnE ai5PMpe81SlZY0xKwg4bNMmq1wAfyvas61FzHlyROakCGBccBK9ySILVJ3EXs/Jb JaRt9L8xtnZHFOVjeT/Ax+7bVUiJwzu0R9XZsojAR6QSIfgJcZZVYkd8TO8RbmPl honwIAjKlarUqwmhSrW5QtHlHx7ibbMg7+E6aYSwlTKaqsJWcma87A5P3MtrB9SN gbJmOChykxOVtYBXoSThaAplf6Lr6t6liZvKFtO/vys/uQyt0iqcnl1NUN//r5aL v0N9mccxf/jggKul4oAwbeTPZVWwu2nm2yNdGaA+Gx2v61Wr3VyT6+vOD5XxS4zH m2eIp52jG6DGXTzZW9ebIMyHgrzMNIgArjviToiiL4gsJX7tcUrV5sstzB7pE+oi /dktiakh19mGTMMngImF2Iv05AMzPshaAzDCPTc7Jw7L6aiQmCgDVs0A3tC99/hC wTNePoeHKp7ywZTyxZKiD/OBOvKfkwCNxsOfbp9ebpSEOmMcnB4ViUWg1ZK1rk4p CxOJHlymEbNA8Lz2wzdFdTsAV46gFafKjhUeS+uz41cjIpSbqwBBXaLiUMUrWjdG 6y8GUvAqLdQTHn+zStSMNUm6/g+GUAnMm/8XLRUeex1K1D5E31I=
    =wQrp
    -----END PGP SIGNATURE-----

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Francisco M Neto@21:1/5 to Tobias Frost on Sat Aug 27 16:00:01 2022
    Hello tobi,

    On Sat, 2022-08-27 at 14:51 +0200, Tobias Frost wrote:
    Maybe I missed something, but there should be no NEW queue involved.
    However, said that, even if a NEW queue is involved, once accepted, this version number is used, so anyways you would need to increment.

    I see. I'm working on fixing all the issues you and jcfp encountered, and adding a couple of superficial autopkgtests.

    Thanks again,
    Francisco

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

    iQIzBAABCgAdFiEEPlgWVZo9XXifkM/x0wsWlNaS+/AFAmMKItwACgkQ0wsWlNaS +/DrFA/9EgtNq8zP9qgZQWtOORUylzuFevDVFTqi97R7T00UulR0sw//4vHkMnKZ 13p8vTtn9cFRstmH8JRJ7mjuocywy90sZSibKSjI4upXmNn9J0I038KQTvjCV6RT TViniwHkQMrCgoD5TmCufNRXPax7nf38Eu2EiZmsiB+1nSjSQJuqvKtL4pFOobtG QlD+MHm9J9L4SLtoG4T4+iO5y7vdnSCnXFB+Gq2WmZMn9/0BASABl6DItAt3O3nT inlZuKZzxhcUOV6TvC3sQnC7wNOvqExi+J6jLQSMOhcT1PiBS8XbgM8pQfL4Ub5G olVZ6nnbpJMm2e1Ok1JL7dsWBRQbMwJcEJ6VCQx1w5dD7j5eKiHMt/K0SgC2WdEL 22cMo+Tb9VB2Q5UgoRrPxpmHBKa65EDYUB3d1U6Sk4+wEFqwh2cIC3pNM0qrF2VY D0TH5idvunOyQH8W7+MtR8qpqw+r+W7oDh5VpF8whzuzi14kqWUpYnJLi1WD/osZ rPtmGtqYHcH6hz9JUbs809dyo1MAq2aP18TZLBm88swjMT3k86AkSsA/I+CzZove gOfWv4INftcZlqu9TjkXIvi+76Xu6HGnIrANohlEmorPxhXHCiUZUcxLH0cWn1pn ysu4EEwMdzfwFyrHVVCCrXCzB/T0dSqPVZntuT1t0E4xGfYEK7U=
    =Em/1
    -----END PGP SIGNATURE-----

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Francisco M Neto@21:1/5 to Francisco M Neto on Sat Aug 27 18:00:01 2022
    Hello again!

    I have fixed all issues. I even included a superficial test, as jcfp suggested. 

    I looked into adding a more meaningful test, but I need to figure out exactly how to run a test that involved a graphical application, so I'll leave it for a future release. I even added such a test to d/t/control, but even with it succeeding I realised it was a false success because the testbed did not have
    a graphical environment; so I left it out. Maybe I can set up such a test using xvfb?

    Anyway, it should pop up on mentors.d.n at any moment.

    Thanks again
    Francisco

    On Sat, 2022-08-27 at 10:57 -0300, Francisco M Neto wrote:
    Hello tobi,

    On Sat, 2022-08-27 at 14:51 +0200, Tobias Frost wrote:
    Maybe I missed something, but there should be no NEW queue involved. However, said that, even if a NEW queue is involved, once accepted, this version number is used, so anyways you would need to increment.

            I see. I'm working on fixing all the issues you and jcfp encountered,
    and adding a couple of superficial autopkgtests.

    Thanks again,
    Francisco


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

    iQIzBAABCgAdFiEEPlgWVZo9XXifkM/x0wsWlNaS+/AFAmMKPVcACgkQ0wsWlNaS +/DOdg//SKCTlvV1E2yYuPvRT9CqfHLk3/1VH48b3+7cJQqm4nwycqVu/YIbwx5w Fy6OGSt7d0KKVp3IVOu0Eclz2yW4WKofcfGPOwXBR7rm3lH0LGy5GaTDuYyStNgM VmLTtGpFhrfAnll1DFgrFt5lO3ENKfcOjxFus8m04X8xO38qtE4uAWVOdus9stbg /Zu4OwaafLjhB+J2Rmc/2Rgsbeg7r68vlviwL6FHAQqHBKtC+HREmGeDLGJI2/5n xnDQ2SeoVcvLglB9VN3OqBYg7unS3jRpbjBBTyBPeADeMeo8qypZfxGG3M2aW5Yl AtwzHCd4ctlRlUBBuvygfLk7/srcBqu5rJaFIawQZ6GOG/XPITY3B7rkqhumN80t D9PLqpnkrrwabVSuV8mg01uHFhLcBQb8Xd75PXeJnjJXkSHjLGsPZoA7JAVdsIQM I7qw3+vXijhCmSEbX4VXN59TvKV83ncTm+Ih467eK/mmJMFJNnYs6nwieQbhdR62 ixmBcmW6waWb3Ikv4fm6FJZcIjkwrphlBEyZQCVmHXyDTzFlj2bxFwd+GI7KEc/f uwI6T16mI88m2YhJfKiKeS7j9iQjbs9OdYd1eauv4Q7kq9wfzV6hpUAaB6cocaBn LD4DLKocMdggz1Cfp0maFTTYZWOI5DiAk7hx7v++Smi/dJn8SEk=
    =iHYx
    -----END PGP SIGNATURE-----

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