Mount HFS+ (was: Re: Converting OS X AddressBook to KAddress?)

[Stefan Bruda]

At 13:11 -1000 on 2002-8-19 Angela Kahealani wrote:
 >
 > The problem occurred when I tried to backtrack from 10 to 9,
 > rebuilding mailboxes in 9 after having imported them into X's Mail
 > program. Therefore, there was no "cat" available.

You can use BBedit's "Concat files" or something like this (don't have
Mac OS to check, but the rightmost menu of BBedit contains such a
command in the default installation).  BBedit also works with Unix
line terminators, yay.

 > And, it ain't just plain ASCII
 > format, as all body lines beginning with "From" have to be
 > preceeded with ">" to not break a message into a truncated
 > message and a headerless piece of garbage. 

Still looks like a plain ASCII file to me, i.e., a file that can be
processed by ASCII utilities such as text editors. ;-)

Sure, here you are:

[bruda(local) ~]% file Mail/in
Mail/in: ASCII text

where Mail/in is my primary mailbox (mbox format).

 > while YDL will mount HFS partitions, and has some hacks to allow
 > messing with an HFS+ partition, I haven't succeeded getting it
 > mounted WRITABLY,

As far as HFS partitions go, make sure that the corresponding line in
/etc/fstab looks like this:

/dev/hda12              /mnt/macos              hfs       defaults     1  2

I have personally changed `defaults' to `noauto,user' of course it
makes no difference--just make sure you have no `ro' in there.  Also
make sure that you have HFS filesystem support compiled in the kernel
(some older benh binaries did not have such).  Usually, HFS is
compiled as a module, so make sure you have module support (modprobe
in the right location between others) up and running.

As for HFS+ filesystems,

 > automounts partitions at boot sees that partition as
 > the HFS wrapper around the HFS+ structure, and mounts it as it
 > presents itself: readonly.

Somebody correct me if I am wrong but I don't believe the kernel has
read-write support for these (yet?).  Instead, I seem to recall that
such a functionality is provided by the hfsutils package (install
hfsutils if needed and see man hfsutils).

When I had to deal with transferring data between HFS+ and ext2, I did
the following though:

o  Fire up MOL (or boot Mac OS X), make a TAR archive (or a tarball if
   you feel like it) containing the data you need to transfer.  There
   are TAR and gzip applications for Mac OS 9 (and below) too. Then either

   - If you have a large enough HFS volume, put the tar[ball] on it,
     boot Linux, mount the HFS volume and simply copy the thing. Or,

   - Boot Linux, fire up MOL (don't forget to install it if not
     already in place), set up Mac OS as directed in the MOL
     documentation for masqueraded networking (it's just a matter of
     running a DHCP server on the Linux part or manually configure the
     TCP/IP pannel as instructed in the documentation), and ftp the
     tarball to the Linux box (and thus filesystem).

o  Finally, get the tarball and expand it.  That's it, it probably
   took longer to write this than doing the actual transfer.

 > Appearantly only by NOT automounting
 > at boot and then manually mounting with the HFS+tools will I be
 > able to access it writably,

Of course, to my knowledge no volume can be mounted with write
privileges twice, no matter its type.

Cheers,
Stefan

-- 
If it was so, it might be; and if it were so, it would be; but as
it isn't, it ain't. That's logic.
    --Lewis Carroll, Through the Looking-Glass