> Hi Lukasz,
>
> so, here are a few more answers to your questions:
>
> On Wed, 14 Oct 2020, Lukasz Zemczak wrote:
> > * How badly is calibre broken on focal right now? Is it really
> > unusable in its current state? Examples of how broken things are so
> > that we can understand the situation better
>
> Yes. I just installed a vm, updated to the latest version of the
> packages of focal, installed calibre, started calibre, and got
>
> norbert@ubuntu2004-vm:~$ calibre
> Traceback (most recent call last):
> File "/usr/bin/calibre", line 20, in <module>
> sys.exit(calibre())
> File "/usr/lib/calibre/calibre/gui_launch.py", line 73, in calibre
> main(args)
> File "/usr/lib/calibre/calibre/gui2/main.py", line 543, in main
> listener = create_listener()
> File "/usr/lib/calibre/calibre/gui2/main.py", line 514, in create_listener
> return Listener(address=gui_socket_address())
> File "/usr/lib/calibre/calibre/utils/ipc/server.py", line 110, in __init__
> self._listener._unlink.cancel()
> AttributeError: 'NoneType' object has no attribute 'cancel'
> norbert@ubuntu2004-vm:~$ vim
>
> So well, it is completely useless.
>
> This can also be seen by the list of upstream bugs reported
> https://bugs.launchpad.net/bugs/1899700
> https://bugs.launchpad.net/bugs/1899674
> https://bugs.launchpad.net/bugs/1899355
> https://bugs.launchpad.net/bugs/1899035
> https://bugs.launchpad.net/bugs/1899029
> https://bugs.launchpad.net/bugs/1898940
> https://bugs.launchpad.net/bugs/1898904
>
>
> > * How does the automated test coverage on calibre look like? Do all
> > new features come with unit testing? What about autopkgtests (I don't
> > think I see any?)?
>
> There are not autopkgtests, but there is an extensive test suite built
> into calibre.
>
> > * What would be the acceptance criteria for the new version? What
> > testing should be performed to make sure the new version works as
> > expected and doesn't regress any existing users (assuming calibre in
> > focal right now is at least usable to some extent)
>
> Since it does not even start, I guess there is no regression for
> focal users, only for those upgrading from a previous release.
>
>
> Together with YOKOTA Hiroshi (in Cc), who has done most of the work on
> recent packaging, I have prepared a version for focal (SIP4, debhelper
> 12), built it on my focal machine, and successfully run it. Source and
> amd64 packages are available here:
>
> deb http://www.preining.info/debian focal main
> deb-src http://www.preining.info/debian focal main
>
> (signed with my gpg key https://www.preining.info/rsa.asc)
>
> What are the next steps you are expecting from me?
>
> - prepare a package for groovy and separately for focal?
Yes.
> - what are the version numbers you want to see?
If it's based on e.g. 5.3.0+dfsg-1, you want
5.3.0+dfsg-1ubuntu0.20.04.1 for focal
5.3.0+dfsg-1ubuntu0.20.10.1 for groovy
or
5.3.0+dfsg-1~ubuntu<20.04.1/20.10.1> would work
too, but I think the first version is better.
> - how should we proceed?
Open - or repurpose an existing bug - with the SRU template:
https://wiki.ubuntu.com/StableReleaseUpdates#SRU_Bug_Template
e.g.
[Impact]
calibre does not start, is an unstable pre-release with a lots of bugs
[Test case]
calibre should start and be functional
[Regression potential]
It did not start before, so it can hardly regress
[Other info]
RT members agreed in <msg>... that we can update to the
latest release.
Attached are debdiffs against the version <ver> in Debian.
with a bit nicer wording and more details :D
And close it in the changelog with LP: #thatbugnumber.
Then attach a debdiff against the debian version to the bug
report for each focal and groovy, and subscribe ~ubuntu-sponsors.
--
debian developer - deb.li/jak | jak-linux.org - free software dev
ubuntu core developer i speak de, en
--
ubuntu-devel mailing list
ubuntu-devel@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel
