Re: [PATCH v2 0/6] maintenance: use packaged systemd units

> Hi Max
> 
> On 22/03/2024 22:11, Max Gautier wrote:
> > * Distribute the systemd timers used by the `git maintenance start` with
> >    the systemd scheduler as part of git, rather than writing them in
> >    $XDG_CONFIG_HOME.
> > 
> > This allows users to override the units if they wish, and is more
> > in-line with the usual practices of distribution for systemd units.
> 
> Thanks for suggesting this, I think this is a useful change, but the
> implementation could be improved.
> 
> > We also move away from using the random minute, and instead rely on
> > systemd features to achieve the same goal (see patch 2). This allows us
> > to go back to using unit templating for the timers. This is also a
> > prerequisite to have static unit files.
> > 
> > Note that even if we really need more specific OnCalendar= settings for
> > each timer, we should still do it that way, but instead distribute
> > override alongside the template, for instance for weekly:
> > 
> > /usr/lib/systemd-user/git-maintenance@daily.timer.d/override.conf:
> > [Timer]
> > OnCalendar=<daily specific calendar spec>
> 
> We should definitely do that. Using systemd's random delay does not prevent
> the different maintenance jobs from running concurrently as one job may be
> started before a previous job has finished. It is important to only have one
> job running at a time because the first thing "git maintenance run" does is
> to try and acquire a lock file so if the hourly job is running when the
> daily jobs tries to start the daily job will not be run.

> 
> > The cleanup code for the units written in $XDG_CONFIG_HOME is adapted,
> > and takes care of not removing legitimate user overrides, by checking
> > the file start.
> 
> This points to an alternate strategy for supporting user overrides - don't
> overwrite the unit files if the user has edited them. I think that there is
> still a benefit to moving to system wide unit files though as it means that
> if we improve the unit files in the future systemd will pick up these
> improvements automatically. That is an improvement over the status quo where
> the users' unit files are written once and never updated.
> 
> I think it would help to reorder the changes in this series as follows:
> 
> 1 - simplify the invocation of "systemctl --user"
>   This would be the current patch 3 without adding "--force" or
>   moving "--now" combined with the relevant test changes from patch 6.
>   This would make it clear that those changes are a simple clean up that
>   is independent of the other changes made in this series.
> 
> 2 - don't delete user edited unit files
>   This would be based on the current patch 4 and would show that we can
>   avoid deleting unit files that the user has edited without the other
>   changes in this series. This change should have an associated test.
> 
> 3 - start using systemd's random delay function
>   This would be the current patch 1 without the template changes and the
>   commit message should explain that it is in preparation for disturbing
>   system-wide unit files.
> 
> 4 - install system-wide systemd unit files
>   This would be based on the current patch 2 with the addition of
>   overrides to prevent more than one job running per hour. The unit
>   files should be installed under $XDG_DATA_HOME when $(prefix) starts
>   with $(HOME), not just when they are equal. The associated test
>   changes from patch 6 should be moved here as well as the "--force"
>   change from patch 3.
> 
> 5 - documentation updates
>   I'm on the fence about having these in a separate commit like the
>   current patch 5 or updating the documentation when the code is
>   changed.