Apple Boot Camp turns Intel Macs into PCs-only

[YKhan]

They install Windows XP, and boot into it, Boot Camp works fine. They
then try to boot back into OSX afterwards, they can no longer get back.


PCWorld.com - Users Find Flaw in Boot Camp
http://www.pcworld.com/news/article/0,aid,125393,tk,dn041306X,00.asp

 

[Tony Hill]

They install Windows XP, and boot into it, Boot Camp works fine. They
then try to boot back into OSX afterwards, they can no longer get back.


PCWorld.com - Users Find Flaw in Boot Camp
http://www.pcworld.com/news/article/0,aid,125393,tk,dn041306X,00.asp

I love how people are whining and complaining about how a piece of
freely distributed *BETA* software has bugs in it! I feel absolutely
no sympathy for any of these idiots!

To quote directly from the Boot Camp download page that all these
people went through:

"Warning: Boot Camp Beta is preview software licensed for use on a
trial basis for a limited time. Do not use Boot Camp Beta in a
commercial operating environment or with important data. You should
back up all of your data before installing this software and regularly
back up data while using the software."

 

[Keith]

I love how people are whining and complaining about how a piece of
freely distributed *BETA* software has bugs in it! I feel absolutely
no sympathy for any of these idiots!

This seems to be a pretty obvious and nontrivial flaw. To get to
an open *beta* with this sort of fault is not right, particularly
with a closed system.

To quote directly from the Boot Camp download page that all these
people went through:

"Warning: Boot Camp Beta is preview software licensed for use on a
trial basis for a limited time. Do not use Boot Camp Beta in a
commercial operating environment or with important data. You should
back up all of your data before installing this software and regularly
back up data while using the software."

Sure, but this is not some sort of strange problem. This should
have been caught before it was released into the wild.

 

[Tony Hill]

This seems to be a pretty obvious and nontrivial flaw. To get to
an open *beta* with this sort of fault is not right, particularly
with a closed system.

I don't know how obvious the flaw really is. Clearly it's only
happening to a fairly small percentage of users, and there aren't all
that many details as to what's going on. As best as I can tell the
issue is that WinXP is over-writing part of the partition tables.
This is not at all abnormal, Windows always re-writes your MBR and
sets the partition that it's being installed to as the active
partition. I'm assuming that Boot Camp has some sort of method to get
around this but under certain situations it fails.

FWIW I've run into the same sort of issue a dozen times over trying to
do multiple OSes with Linux and Windows. If you ever install Windows
AFTER Linux than it will hose the Linux booting system (Lilo or Grub).
It's easy enough to fix as long as you have a bootable CD that lets
you re-create everything.

 

[Gnu.Raiz]

Didn't MS do a release of server 2003, or something straight from beta,
instead of letting it cook for a while?

After all if it's good for MS, then it's good for us right? Something
on the lines of the old saying if its good for GM, <insert company
here> then it's good for the USA!

That is why I never allow Windows on my MBR, it's grub or freebsd boot
loader for me. Or an extra Windows dedicated drive, after all drive
space is cheap. Might as well give Windows what it wants a dedicated
MBR.

The real question should be is how does OSX handle NTFS file systems, I
bet they have the same problems as Gnu/Linux has, which is read only
for safety.
I only see this more of a proof of concept, and for gamers who might
want to play games while in Windows mode.

This is how I see it, if you support BootCamp, then you really need to
support NTFS, which mean you need to be able to write to a Windows
partition?

Why, well for the same reason you support BootCamp, it's only a matter
of time that people will be asking for this. Then you must not forget
the DRM, even if it's the same computer but different os, if you DRM a
file, but want access to it in a different OS will you be able to? Or
what I am saying will your hardware be smart enough to tell the
difference, without any major performance hit?

Gnu_Raiz

 

[Keith]

I don't know how obvious the flaw really is. Clearly it's only
happening to a fairly small percentage of users, and there aren't all
that many details as to what's going on. As best as I can tell the
issue is that WinXP is over-writing part of the partition tables.
This is not at all abnormal, Windows always re-writes your MBR and
sets the partition that it's being installed to as the active
partition. I'm assuming that Boot Camp has some sort of method to get
around this but under certain situations it fails.

"Small percentage?" How many users are there? Come on, it's a *closed*
system. It's not even like a release from Redmond that had a problem with
some esoteric RAID card.

FWIW I've run into the same sort of issue a dozen times over trying to
do multiple OSes with Linux and Windows. If you ever install Windows
AFTER Linux than it will hose the Linux booting system (Lilo or Grub).
It's easy enough to fix as long as you have a bootable CD that lets you
re-create everything.

Sure, but that's on *PCs*, not a closed sysetem that I would *ASSUME*
Appple has tested.

 

[George Macdonald]

I don't know how obvious the flaw really is. Clearly it's only
happening to a fairly small percentage of users, and there aren't all
that many details as to what's going on. As best as I can tell the
issue is that WinXP is over-writing part of the partition tables.
This is not at all abnormal, Windows always re-writes your MBR and
sets the partition that it's being installed to as the active
partition. I'm assuming that Boot Camp has some sort of method to get
around this but under certain situations it fails.

FWIW I've run into the same sort of issue a dozen times over trying to
do multiple OSes with Linux and Windows. If you ever install Windows
AFTER Linux than it will hose the Linux booting system (Lilo or Grub).
It's easy enough to fix as long as you have a bootable CD that lets
you re-create everything.

You just jogged my memory about an experience I had with WinNT 3.51(?) on a
Digital Alpha PC years ago. I had two separate disks, one for each OS and
WinNT noticed the 2nd disk on its 1st boot after install. It asked if I
wanted to "initialize" it - obviously I chose no but it still clobbered the
Unix OS on that other disk. Plus ça change..........

 

[Tony Hill]

"Small percentage?" How many users are there? Come on, it's a *closed*
system. It's not even like a release from Redmond that had a problem with
some esoteric RAID card.


Sure, but that's on *PCs*, not a closed sysetem that I would *ASSUME*
Appple has tested.

The new Apple's are pretty darn close to PCs. The only real
difference is in the BIOS, and to get Windows to install Apple had to
get that emulated. I suspect that Windows is doing exactly the same
thing here, though with very slightly different results.

Regardless of what the exact nature of the issue is, clearly it is
something that slipped through the cracks. As you mention, this is a
closed system, very closed in fact. There's basically only two Apple
x86 systems out there and they use very similar hardware. As such,
I'm sure this has absolutely nothing to do with the hardware side of
things since that would come pretty darn quickly.

What's much more likely is that there is something fairly specific
(though probably not obvious) that these users are doing differently
from what everyone at Apple tested. Maybe this bug should have been
caught before a beta, but you're never going to catch all bugs in the
development stage, that's why we have beta releases. I'm sure this
will be fixed by the next spin of the software.

As for the users who got caught out by this, well you know what the
say about those who live on the bleeding edge...

 

[Tony Hill]

You just jogged my memory about an experience I had with WinNT 3.51(?) on a
Digital Alpha PC years ago. I had two separate disks, one for each OS and
WinNT noticed the 2nd disk on its 1st boot after install. It asked if I
wanted to "initialize" it - obviously I chose no but it still clobbered the
Unix OS on that other disk. Plus ça change..........

Yup. To say that Windows is not very friendly in a multiple-OS
environment is a HUGE understatement! It has taken me more than a few
tries to figure out how to get Windows to respect a Linux install
sufficiently. More recently I've been mostly avoiding this sorts of
headaches by having my Linux box setup on a fully separate computer,
however that is not really ideal in many situations.

 

[chrisv]

To say that Windows is not very friendly in a multiple-OS
environment is a HUGE understatement! It has taken me more than a few
tries to figure out how to get Windows to respect a Linux install
sufficiently. More recently I've been mostly avoiding this sorts of
headaches by having my Linux box setup on a fully separate computer,
however that is not really ideal in many situations.

I dual-boot Windows and Mepis Linux at home. It's a bit of a
minefield, since some issue with the way Linux interprets your
harddrive makes the "easy way" a bit risky. (The "easy way" is to
install Windows first, leaving space on the HD for Linux, then install
Linux with Grub on the MBR, with Grub automatically recognizing
Windows and offering the two boot-options upon start-up.) The last
time I tried this (over a year ago), it left my Windows partition
unbootable.

So, now I do it the "hard but safe way", still installing Linux
second, but not installing Grub on the MBR. I then copy the Linux
boot-sector to C: (using the "dd" trick) and edit boot.ini so that
ntldr offers the OS choice on start-up. A little more work, but it
also prevents any unfortunate consequences should M$ re-write the MBR
for any reason (like a service pack install).

 

[Keith]

The new Apple's are pretty darn close to PCs. The only real
difference is in the BIOS, and to get Windows to install Apple had to
get that emulated. I suspect that Windows is doing exactly the same
thing here, though with very slightly different results.

No, they are *not*, for the reasons you point out below (snipped for
brevity). Who cares what the hardware looks like, the fact is that it's a
*closed* system and is more easily tested. This is a major f'up. ...and
people rag on M$ for far less. Unless these people can be proven to have
done somethign really strange to have shot themselves in the foor, this
sort of error is inexcusable.

<snip>

 

[Robert Redelmeier]

Who cares what the hardware looks like, the fact is that it's a
*closed* system and is more easily tested. This is a major f'up.
...and people rag on M$ for far less. Unless these people can
be proven to have done somethign really strange to have shot
themselves in the foor, this sort of error is inexcusable.

Agreed. The hardware is a single target. The software is "known
troublesome" [rewrites MBR] MS-Windows (specified versions _and_
SP levels, including installs). Testing should be easy for a
professional lab.

Frankly, I expected boot camp to be a ROM monitor/loader.
A custom MBR is not robust enough, and remapping the HD too risky.

Can the x86 Apples boot from removeable media (CD/USB/floppy)?
If they can't, then a software failure requires hardware service.
I hope they licenced Kirk McKusick's SoftUpdates for the OS
X kernel. Otherwise, the inevitable powerfails will make the
machine unbootable, and unfixable short of disk-swap.

-- Robert