Re: Breakage from grub-mkconfig changes for grub-file

[Andrey Borzenkov]

В Вт, 24/12/2013 в 00:34 +0000, Colin Watson пишет:
> On Mon, Dec 23, 2013 at 11:21:38PM +0100, Vladimir 'φ-coder/phcoder' 
> Serbinenko wrote:
> > On 23.12.2013 23:01, Colin Watson wrote:
> > >   This should be redesigned so that there is some way to declare in a
> > >   grub.d script that it requires multi-platform support and should be
> > >   run multiple times.  (It *must* be this way round so that upgrades
> > >   work properly.)
> > 
> > The idea was that platform-independent scripts were still runnable,
> > they'll just produce the same output N times and this list is just an
> > optimisations, specially to avoid running os-prober N times.
> 
> Granted, but in some cases those scripts might not be idempotent:
> consider a user-written "42_custom" (or whatever) script that adds a
> menu entry, for instance.
> 
> > The alternative will be to have something along the lines of different
> > hashbang or implementing this functionality as sh functions.
> 
> How about this simpler option: any script that needs to be run for each
> platform could have a magic comment that we grep for in grub-mkconfig.
> 

I have a feeling that we are doing it backwards. What we actually want
is to know file type. So what about making grub-file print it instead of
having ever growing list of conditions? I.e.

grub-file --print {cpu|os|bits|...} 

and simply using it in every script to generate run-time condition like
in (simplified)

cpu=$(grub-file --print cpu)

cat << EOF
if [ \$grub_platform = $cpu ] ;  then
menuentry ...
EOF

And this could be wrapped in grub_mkconfig_platform_condition function.


> > >   The platform names used in grub-mkconfig (x86 i386-xen-pae x86_64-xen
> > >   mips mipsel sparc64 powerpc ia64 arm arm64) are not the same as the
> > >   platform names used in the GRUB build system, but yet they're exported
> > >   across the interface to /etc/grub.d/ as GRUB_PLATFORM.  This is messy
> > >   and confusing, and it's not clear what promises we make about future
> > >   changes here.
> > > 
> > >   We should rationalise this before issuing anything as part of a stable
> > >   release, perhaps by adopting the same target_cpu/platform terminology
> > >   used in the build system.  Furthermore, if we made the namespaces
> > >   match up then it would be fairly straightforward to only run grub.d
> > >   scripts for platforms for which we have installed GRUB modules, which
> > >   seems as though it would be sensible.
> > 
> > GRUB platform names don't match with the OS compatibility. On x86 other
> > than xen you can use the same kernel on all the platforms. On ARM, for
> > what is arm-uboot platform for us may require different kernels for
> > different hardware.
> 
> OK, but if it is a different concept then it should have a different
> name, not "platform" - otherwise it just seems confusing.
>