FC4t2 no good without LILO
Guy Fraser
guy at incentre.net
Wed Apr 13 22:12:17 UTC 2005
On Tue, 2005-12-04 at 22:23 -0400, Peter Jones wrote:
> On Tue, 2005-04-12 at 17:23 -0700, Mike Bird wrote:
> > I just checked the Redhat and Fedora distros which I still have around.
> > Here are the Lilo versions:
> [list of 8 different versions clipped for brevity]
> > Most of the version bumps were for new Redhat releases and involved no
> > source changes. There was only one patch to Redhat's Lilo in the last
> > two years: a single #define.
>
> People stopped finding so many bugs once we stopped using lilo by
> default. That's been explained to you several times now.
>
> You also need to look at all the other packages involved in setting up
> and maintaining the bootloader to. They're pretty complicated, and a
> lot of work.
>
> > Lilo version 21.4.4 actually dates back to 2000.
>
> And it needs plenty of work done, if we were going to consider using it
> again. More than grub currently needs, I suspect.
>
> > So we have a package that is essential in some circumstances,
>
> So far, you won't mention which circumstances those are. So I think
> most of us aren't considering this to be a fact.
>
> > that costs next to nothing to maintain
>
> That's simply not true -- there are several more patches to lilo that
> have been applied for RHEL, or that may unfortunately need to be
> applied. There's also very difficult logic in booty, anaconda, up2date,
> and likely several other packages that has to handle it.
>
> And there's also a *huge* support cost to everybody involved with
> helping other users, which you're obviously content to completely
> ignore. It isn't nearly as easy for most users as grub is.
>
> > , and that takes up <0.1% of a CD's space.
>
> I already told you, this really didn't affect things.
>
> > The experience of many people in the real world is that even a five
> > year old version of Lilo is more reliable than today's Grub.
>
> The plural of anecdote is not data. Right now, the number of people we
> can document who admit to having such experience is hovering menacingly
> around 5. And that's five really bad sources -- none of them will
> provide any detail on what's gone wrong at all!
Bull crap.
Everyone else probably gave up or found some other way around
the problem.
Why would I or anyone else enter a new bug report for a bug
that is already duplicated a bunch of times. I am sure there
are dozens of people who have the same complications that don't
post duplicate bug reports.
I am willing to guess that like me many people have opted for
booting off a drive attached to slower onboard PATA device,
because that was the only way they could get there machine to
run. Each time I installed a new drive or replace a drive with
a larger one, grub would fail and it only reports a stupid
*error number*. What good is that! Lookup the error number
and all it meant was that it could not find a requested file,
but gave no indication what file or device it was looking for
the file on. Are you seriously going to tell me that only 5
people have got the arcane error number after changing a non
boot drive?
I would have preferred to leave the ATA66 ports for the CD/DVD
drives. And use the faster SATA and PCI ATA133 controlers for
the hard drives.
It works but is not optimal.
>
> > Certainly a five year old version of Lilo handles software RAID better than
> > today's Grub.
>
> This is just plain insulting. You haven't even *tried* the software
> RAID support in today's grub. You've already said as much today!
>
> > And so the cardinals of the Redhat vatican issue a bull banning Lilo.
>
> Nobody's stopping you from using it, so quit pretending that's what
> we're doing. You have just as much choice as I do; you're choosing to
> whine on a mailing list about problems you won't even name, much less
> provide technical facts to support. I'm choosing to continue not
> resurrecting lilo, since none of the group of you who want it back are
> willing to provide any technical reasons whatsoever.
>
> Complaining more isn't going to change my mind. Continuing to refuse to
> provide any technical details on whatever your problem is won't make me
> any more likely to, either. If you really, really want lilo back,
> you're going to have to actually *convince* me that it's a good idea.
> You don't have to convince me that you've had a really bad experience,
> that's obvious. But if you can't tell me why, I can't help make sure
> anybody's experience is any better.
>
> --
> Peter
More information about the fedora-test-list
mailing list