On Mon, Sep 05, 2011 at 06:50:34PM -0000, <email address hidden> wrote:
> I was just curious as to where the issue resided. Is the grub.cfg stanza
> syntax erroneous or is it correct and some downstream function at fault?
The syntax was wrong. It was produced because Ubuntu prefers to run
without /boot/grub/device.map, which is an upstream goal but not one
that's generally enabled by default as yet. We have a patch to improve
behaviour in that situation, but at some point when merging a new
upstream release I missed a new situation that needed to be handled.
So it's not really possible to say either that it was an upstream bug or
an Ubuntu-specific bug in this case; it was a bit of both.
Either way, it wasn't terribly worrisome because the following 'search'
command generally finds the correct device by UUID, and the 'set root='
line is only a fallback mechanism. Still, it was untidy and confused
people, so I wanted to clean it up anyway.
On Mon, Sep 05, 2011 at 06:50:34PM -0000, <email address hidden> wrote:
> I was just curious as to where the issue resided. Is the grub.cfg stanza
> syntax erroneous or is it correct and some downstream function at fault?
The syntax was wrong. It was produced because Ubuntu prefers to run device. map, which is an upstream goal but not one
without /boot/grub/
that's generally enabled by default as yet. We have a patch to improve
behaviour in that situation, but at some point when merging a new
upstream release I missed a new situation that needed to be handled.
So it's not really possible to say either that it was an upstream bug or
an Ubuntu-specific bug in this case; it was a bit of both.
Either way, it wasn't terribly worrisome because the following 'search'
command generally finds the correct device by UUID, and the 'set root='
line is only a fallback mechanism. Still, it was untidy and confused
people, so I wanted to clean it up anyway.