F9-Beta install from local harddisk - broken?

[Michael Schwendt]

On Fri, 04 Apr 2008 14:38:45 -0400, Jesse Keating wrote:

> If your idea is that software has to be done before you can test it,

Your words, your words.

> I don't see how you can ever test it.

By bringing it into a form where it can be used daily.
To have the Beta fail during installation where the Alpha had no
problems is weird.

> As for slow, we have all sorts of debugging turned on all over the place
> so that when things to crash we have enough breadcrumbs to figure out
> what went wrong and fix it.  Those will be removed over the next week or
> so, which will make it harder to get anything useful out of any bugs you
> find afterward.

Still, when I asked about this slowness (and even asked about kernel
level debugging, too), nobody wanted to confirm or deny. So, in that
old thread I went and booted Rawhide with F8`s kernel, it didn`t feel
snappier either.
 
> And we're sorry you had a grub bug, which may be a real bug, but it's by
> far not the common case and thus it didn't get any testing until you
> tried it.  It likely would still be broken in the final release if you
> didn't try it (not that I'm claiming it will be fixed, this may be one
> of those "don't do that" issues, but at least we'll know about it and
> can document it due to your testing).

The traceback in Anaconda related to GRUB device options is from a
real-world install scenario. It's very common to install GRUB not into
the MBR. The other bug in GRUB shell certainly is no
"don't do that" issue, because the GRUB shell is made for things
like that.