Bug 49199 (suse34199)

<!-- SBZ_reproduce  -->
use whole disk on a mac partition label.

Please provide the output of parted /dev/hda print of such a disk.
Additionally I need tha YaST2 log files.

dd if=/dev/zero of=/dev/hda count=4k bs=4k
parted /dev/hda
mklabel mac
quit


real world example:

mac:~ # pdisk -l /dev/sda

Partition map (with 512 byte blocks) on '/dev/sda'
 #:                type name         length   base      ( size )
 1: Apple_partition_map 'Apple    '        63 @ 1        
 2:           Apple_HFS 'boot     '    204800 @ 64        (100.0M)
 3:     Apple_UNIX_SVR2 'swap     '   2048000 @ 204864    (1000.0M)
 4:     Apple_UNIX_SVR2 'root     '  20480000 @ 2252864   (  9.8G)
 5:     Apple_UNIX_SVR2 'abuild   '   4096000 @ 22732864  (  2.0G)
 6:     Apple_UNIX_SVR2 'blah     '  20480000 @ 26828864  (  9.8G)
 7:          Apple_Free 'Extra    ' 108992624 @ 47308864  ( 52.0G)

Device block size=512, Number of Blocks=156301488 (74.5G)
DeviceType=0x0, DeviceId=0x0

mac:~ # parted /dev/sda print
Disk geometry for /dev/sda: 0.000-76319.085 megabytes
Disk label type: mac
Minor    Start       End     Filesystem  Name                  Flags
1          0.000      0.031              Apple                
type=Apple_partition_map
2          0.031    100.031  hfs         Apple_HFS             type=Apple_HFS
3        100.031   1100.031  linux-swap  swap                  swap,
type=Apple_UNIX_SVR2
4       1100.031  11100.031  reiserfs    root                  root,
type=Apple_UNIX_SVR2
5      11100.031  13100.031  reiserfs    abuild                type=Apple_UNIX_SVR2
6      13100.031  23100.031  ext2        blah                  type=Apple_UNIX_SVR2
7      23100.031  76319.085              Extra                 type=Apple_Free
Information: Don't forget to update /etc/fstab, if necessary.             

Is there really code to start at number 2? 

Ok, I see.
Mac partition labels are currently not fully supported.
If you want to have them fully supported, channel this as a feature
request thruogh the proper channels (Klaus Kaempf and/or Ralf Flaxa)

Looks like a past decision "No, we won't support Macs any more, concentrate on 
IBM PPC servers instead" has been reverted once more. 
Thomas, there should be old code supporting MAC partitions, please have a look 
if it can be resurrected. Please send me an effort estimation. 

I this really coordinated with Ralf, as far as I can remember mac partition
table support was one of Olafs toy projects. No problem to implement such
things if I have the time, currently this is definitely not the case.
If this is really demanded by the decision makers, I hope they did not forget 
to move the release date for SLES9 accordingly into the future. There are 
already features for SLES9 that I am  pretty sure cannot be implemented in the
timeframe left for SLES9 anyway.
Of course I have no problems to add another such feature as long as nobody 
expect it to be really available on SLES9. If someone really wants to have 
it he has to tell me which feature to drop for it.

Well, maybe I am completly wrong, but we had Mac support in the past? Who said
that we will drop this support? Could you please give me a pointer?

There are several pointers (mails) inside and outside of SuSE about not 
shipping a ppc/mac linux distribution any more. It's the same with DEC Alpha 
support, it's simply ignored currently and completely untested. 

Klaus, the question was: Did we have this support? Who said that we drop it? I
know that we had no commitment to do another "Mac-Distribution" - but does this
really mean that we have to drop allready given features?

This is no "given feature". To my knowledge using "whole disk" for disk
with mac labels never worked. Seems I really _should_ actively have dropped
all support for mac disklabel since you deduce from the rudimentary support
that still works by chance, that you may demand any additional functionality.

Calm down Thomas, I just want to understand! I know that we once were able to
install on a Mac. As long as we did not lost any functionality here, it is fine
for the moment. I was just under the impression, that we lost some things ... 

Next question:

To improve the behavior, would it help if we would assign this task to one of
the PPC team members? Is this a job which can be done without deeper Yast2
knowledge?

Installation should still work, partitioning of a mac disklabel was
IMHO never supported very well (if at all, I am not sure about this).
The changes would have to be done inside the ycp code of yast2-storage.
Of course this can be done by anyone but IMHO there is a pretty thorough
understanding of the existing ycp code necessary to prevent breakage of 
other partition labels and the partitions suggestion.

To comment #9: Of course, we did not _actively_ drop features, but we surely 
didn't care for unused/unneeded code pathes. 
 
We will add this feature to the planning, although with a low priority (e.g. 
EVMS support is more important). Maybe it'll be implemented in time, maybe with 
a service pack in the future. Thats up to Ralf. 

Too late for SLES9 now, dealied until next release.

reopen to mark as dup

this is still (or again) the case in sles10.

If you want support for mac disklabel you should make this a feature for SLES10
in ccordination with SLES project management. And pe prepared it to have 
implemented in parted by a person that has knowledge about mac partition table.

sure we will do this by ourselves (as discussed with kkaempf), keep cool :-)

works better in 10.1