On Mon, 05 Dec 2022 at 01:21:19 +0100, Samuel Thibault wrote:
> Simon McVittie, le lun. 05 déc. 2022 00:15:01 +0000, a ecrit:
> > Upstream gdm3.service (or possibly gdm.service, I think we do some renaming
> > for historical reasons) has
> >
> > [Install]
> > Alias=display-manager.service
> >
> > so that `systemctl enable gdm3.service` will create a symlink
> > /etc/systemd/system/display-manager.service ->
> > /lib/systemd/system/gdm3.service.
> > This assumes that only one display manager can be enabled at a time, and
> > preferably only one *installed* at a time: if both gdm3 and lightdm try
> > to do that, if I understand correctly, the result will be unpredictable.
> >
> > In Debian, instead of gdm3.service having that, the debconf question
> > creates that symlink programmatically if and only if gdm3 is the selected
> > display manager (in particular, the symlink isn't created if you have
> > installed both gdm3 and lightdm, and chosen to use lightdm instead).
> >
> > If I understand correctly, any mechanism that makes the symlink exist
> > should be enough for "Before=display-manager.service" to do what you want
> > it to do. Please try it?
>
> Ah, indeed, that works, sorry for not having actually tried, because
> lightdm does have these two lines. But then, with this reasoning,
> lightdm is bogus by having these two lines?
I honestly don't know. Maybe this is wrong in lightdm (and possibly
other *dm implementations), or maybe removing those lines in gdm3 is
unnecessary delta and we don't really need that part of the patch?
We need a small amount of delta in the source file gdm.service.in *anyway*,
so if this is an unnecessary patch, it's not a particularly expensive one.
Note that both gdm3 and lightdm override dh_installsystemd to do nothing,
so probably it doesn't matter very much in practice unless a sysadmin
explicitly runs `systemctl enable gdm3.service` or
`systemctl enable lightdm.service` themselves.
smcv
