著者: Werner Mahr 日付: To: dmo-discussion 題目: Re: MythTV 31 in experimental
Am Montag, 2. März 2020, 08:38:46 CET schrieb Christian Marillat: > On 01 mars 2020 17:45, Werner Mahr <werner@???> wrote: > > First thing: I would suggest that python3-mythtv is either renamed to the
> > old, generic python-mythtv or gets a conflict for this.
>
> Done.
Jep, works. I would have preferred to have it named python-mythtv, as this
allows just a normal apt-upgrade instead of a dist-upgrade, but I guess it's a
policy thing here.
> > Second thing: the user in the unit-file is hardcoded. In the old
> > /etc/default/ mythtv the was an other user set, which leads the backend
> > to fail starting with an not meaning error 130.
>
> I'm not sure here, but my package doesn't provide a service file or
> /etc/default/mythtv file.
The v30 mythtv-backend shipped that file, I saw you changed the EnvironmentFile
in the mythtv-backend.service and played a bit with that.
The bad news: Changing the user to as isn't possible with this, so users like
me, that run the backend as their normal user to easier access mythtv-setup
have to edit the unit manually.
The good thing, placing a symlink to the old-file or just moving it there
works, as the old one isn't removed while upgrading.
Maybe you can consider to ship a file for that, with all the params predefined
there. In my cas I have to edit the unit to remove the --syslog... stuff, as
mythtv has now a --systemd-journal param to directly log there. I read
somewhere that rsyslog will be replaced by systemd-journald in the future (or
at least everthing is prepared that it could be replaced).
And btw. the old defaults calls it EXTRA_ARGS, while the unit uses
ADDITIONAL_ARGS.
> > Beside these two things, upgrade worked flawless.
>
> Good to hear that.
>
> Thanks for the feedback.
And just as an additional info, changing the dh_sequence_python3 to the older
dh --with-python3 and freetype-dev to freetyp6-dev makes it work on stable,
too.
