On Sep 23, 2008, at 10:17 AM, Will Woods wrote:
On Mon, 2008-09-22 at 13:34 -1000, David Cantrell wrote:
Perhaps, but my personal take on it is there is the do-I-care issue
and the technical issue.
For me, I don't care what is on a user's HFS or DOS partition. Maybe
it's bootable, maybe it's not. But it's not our problem.
Except when we set up bootloader config for the other OSes on the
system, as we do in anaconda:
http://wwoods.fedorapeople.org/screenshots/f10b-boot-config.png
I wonder what happens if you try to install Fedora on a system that's
already got Mac OS X and Windows dual-booting? Two "Other" entries,
identical except for the partition number?
Well, that's true. I guess I mostly hate it how Linux is expected to
be friendly to neighboring operating systems while Windows and OS X
get to stomp all over the system. Why can't we stomp all over the
system?
The technical issue is we can't really make the assumption that all
bootable HFS partitions on ppc systems are MacOS X.
What? Why not? Unless I've seriously misunderstood this part of booty,
that field is just a human-readable string that identifies the system
type.
Change it to "Mac OS" and we'll be right easily 95% of the time, and
the
rest will be owned by hackers who are obviously competent enough to
know
what HFS-based, non-Mac OS system they're dealing with.
It is just a human-readable string that identifies what's going to
boot off that drive, but I was trying to say that you can't assume
what's on it based just on the filesystem type. To really know, you
have to examine the contents of the filesystem.
But yes, you're right, we'd be correct for most cases by just making
the assumption.
I mean, it's
probably a safe guess, but it's not guaranteed. The same with DOS
partitions. We can't assume it's Windows. The only way to know for
sure what's on those filesystems is to dig down in to them and see
what will boot from them, and then that brings in a whole set of
knowledge that isn't really part of anaconda.
..which is why this is a patch to booty, not anaconda.
Missed that, sorry.
If you want me to extend the patch to actually try to identify the OS
*for sure*, with version numbers and everything, that's fine. But that
seems like an awful lot of additional complexity to handle HFS-based
systems that *aren't* OS X or NTFS/FAT-based systems that *aren't*
Windows - especially when, once again, if we guess *wrong*, it's
just an
incorrect label that the the user can correct.
That would be a waste of time.
I really think this is one of those lucky situations where we can
guess
right 95% of the time, and when we guess wrong, the failure is
harmless
and trivial to correct. Refusing to guess because of that 5% chance of
failure seems like a cop-out.
Forget my original analysis. That's me trying to aim for the perfect
coding solution or whatever for utopia land. I guess the assumption
for the labels is probably fine for us.
--
David Cantrell <dcantrell@xxxxxxxxxx>
Red Hat / Honolulu, HI
_______________________________________________
Anaconda-devel-list mailing list
Anaconda-devel-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/anaconda-devel-list