Hi Niels,
Niels Thykier wrote:
> The options are confusingly similar due to an unfortunate choice many
> years ago.
Yeah, looks historically grown.
> It does not help that the text differs between dh_installinit and
> dh_installsystemd despite the options intend to have the same
> meaning.
Yep, was kinda irritate by this, but since systemd has to do
everything differently, I expected subtle differences in the semantics
and hence also that the option names had to be different for similar
reasons.
> > With a lot of experimenting (e.g. trying the alias --no-stop-on-upgrade
> > for --no-restart-on-upgrade), I figured out that this works as I want
> > it:
> >
> > ------------------------------------------------------------------------
> > override_dh_installinit:
> > dh_installinit --noscripts
> >
> > override_dh_installsystemd:
> > dh_installsystemd --no-restart-on-upgrade --no-restart-after-upgrade
> > ------------------------------------------------------------------------
> >
>
> Can I please have you use "--no-stop-on-upgrade" instead of
> "--no-restart-on-upgrade".
Can do, but won't do a separate upload just for this.
> The "on" vs. "after" upgrade is not kind to
> the eyes when they have different meanings
Well, depends, I have definitely more issues understanding the
semantic difference.
> and I intend to remove the "on" variant for exactly this reason.
But please not for dh v13, only for v14 upwards.
> The "no-stop-on-upgrade" is not entirely accurate either though,
Exsactly. That's what this bug report is mainly complaining about. :-)
> but at least it is less likely to be misread as "restart".
Well, in this case, I prefer the restart over the stop, because it's
mainly the restart which is the issue here. (Yes, it technically
includes a stop, but it doesn't need to.)
> > "Do not stop service on upgrade." is clearly not true unless additional
> > options are given.
> >
> > "Undo a previous --restart-after-upgrade (or the default of compat 10)."
> > is not true either there was no previous --restart-after-upgrade to
> > undo.
> >
> > I suspect that either
> >
> > * the logic which adds the snippets to postinst is flawed,
> > * the snippets added to postinst are flawed, or
> > * it's a pure documentation issue.
[...]
> It is the logic,
Ah, ok.
> although the documentation was not great either.
*nod*
Regards, Axel
--
,''`. | Axel Beckert <a...@debian.org>, https://people.debian.org/~abe/
: :' : | Debian Developer, ftp.ch.debian.org Admin
`. `' | 4096R: 2517 B724 C5F6 CA99 5329 6E61 2FF9 CD59 6126 16B5
`- | 1024D: F067 EA27 26B9 C3FC 1486 202E C09E 1D89 9593 0EDE
