Re: dual boot Hurd/Linux?

[Samuel Thibault <sthibault@debian.org>]

Martin-Éric Racine, le mar. 29 avril 2025 12:30:14 +0300, a ecrit:
> ma 28.4.2025 klo 13.47 Samuel Thibault (sthibault@debian.org) kirjoitti:
> >
> > Martin-Éric Racine, le lun. 28 avril 2025 13:19:44 +0300, a ecrit:
> > > ma 28.4.2025 klo 10.59 Samuel Thibault (sthibault@debian.org) kirjoitti:
> > > >
> > > > Martin-Éric Racine, le lun. 28 avril 2025 10:02:19 +0300, a ecrit:
> > > > > su 27.4.2025 klo 13.11 Samuel Thibault (sthibault@debian.org) kirjoitti:
> > > > > > Martin-Éric Racine, le dim. 27 avril 2025 11:29:21 +0300, a ecrit:
> > > > > > > The key point is that without keyboard-configuration, the console will
> > > > > > > fail to load on reboot,
> > > > > >
> > > > > > That shouldn't be so, it should at least be able to start in english
> > > > > > mode.
> > > > >
> > > > > Heck, even after keyboard-configuration and locales are configured,
> > > > > upon reboot, Hurd still tells me that console launch failed because it
> > > > > cannot find a keymap for locale C.
> > > >
> > > > Please paste the exact error message, so we can check exactly what this
> > > > is about.
> > >
> > > Message gone after a few reboots.  It actually was about no compose
> > > map for locale C.
> >
> > Is that really an error, and not just a warning?
> 
> Error. The console exited with an error due to no Compose map for locale C.

That should be turned into a warning then. Missing compose is really not
a severe problem.

> > > Returning to the failed ACPI shutdowns, the console now shows the
> > > following at the end of the process initiated by 'exec sudo poweroff:
> > >
> > > hwsleep-0079 hw_legacy_sleep  : Entering sleep state [S5]
> > >
> > > Except that it doesn't go to sleep. It merely reaches the end of the
> > > halt sequence. Anyhow, S5 is supposed to mean poweroff, not sleep.
> > >
> > > The obvious question:  where do I find the defaults used for the
> > > shutdown sequence?  Something apparently decided to use sleep instead
> > > of poweroff.
> >
> > That goes through the shutdown translator, in hurd/shutdown/shutdown.c,
> > that calls the acpi_sleep() RPC, implemented in hurd/acpi/acpi-ops.c
> > S_acpi_sleep
> 
> Noted.
> 
> I take it that Hurd doesn't use bin:powermgmt-base or anything similar
> to communicate with ACPI?

Nobody contribute anything beyond the acpi translator, so far.

Samuel