OS X: Resource fork, Rsync and Eudora 6.2.4 - Eudora

On Fri, 28 Mar 2008 04:23:21 -0500, Joel West wrote:

> I'm having problems that sound like the "resource fork" problem
> mentioned earlier. I have been using Eudora for about 12 years and
> moved mailboxes 6 or 7 times, but this time it didn't work.

> What I did was:
> 1. Backup all my HDD using a command-line rsync, including 3.7gb of
> Eudora 6.1.1 (paid mode) mailboxes
> 2. Watch my HDD crash
> 3. Move the Eudora files back to my new HDD using rsync
> 4. Notice that Eudora 6.1.1 doesn't launch on my MacBook Air, so
> install 6.2.4
> 5. What I see is:
> a. error message "not a properly formatted mailbox" for In, Out, Junk
> etc.
> b. my custom mailboxes don't show up in the menu, and
> c. the settings (e.g. e-mail accounts are also lost).
>
> So the problem could be the rsync, the 6.1.1 -> 6.2.4, or the PowerPC
> to Intel transition.
>
> I opened up the mailboxes in a text editor, and they look like
> standard *nix text file mailboxes.
>
> Any suggestions on what to try? I don't want to abandon my mailboxes
> and 10+ email accounts, and of course don't want to lose any data

[assume a suitable tune playing :]

"Got no rsync, got no Mac, all I do is to read the flak"

A history of rsync on Mac OS X [December 4, 2007]
http://patternbuffer.wordpress.com/2...c-on-mac-os-x/

"I got the Sun in the Morning and the Moon at Night"
[composer: Irving Berlin, show: Annie Get Your Gun]
http://www.youtube.com/watch?v=E3ZcmIZz1as [Reba McEntire!]
http://www.youtube.com/watch?v=2fbgzk6gvkg [Barbara Cook]
http://www.youtube.com/watch?v=odol-YtN9S0 [new talent]

http://en.wikipedia.org/wiki/I_Got_t..._Moon_at_Night)

--

In article
<7298c66d-2f9d-41c0-af78-51c7588ef306@13g2000hsb.googlegroups.com>,
<openitstrategies@gmail.com> wrote:

> I'm having problems that sound like the "resource fork" problem
> mentioned earlier. I have been using Eudora for about 12 years and
> moved mailboxes 6 or 7 times, but this time it didn't work.
>
> What I did was:
> 1. Backup all my HDD using a command-line rsync, including 3.7gb of
> Eudora 6.1.1 (paid mode) mailboxes
> 2. Watch my HDD crash
> 3. Move the Eudora files back to my new HDD using rsync
> 4. Notice that Eudora 6.1.1 doesn't launch on my MacBook Air, so
> install 6.2.4
> 5. What I see is:
> a. error message "not a properly formatted mailbox" for In, Out, Junk
> etc.
> b. my custom mailboxes don't show up in the menu, and
> c. the settings (e.g. e-mail accounts are also lost).
>
> So the problem could be the rsync, the 6.1.1 -> 6.2.4, or the PowerPC
> to Intel transition.
>
> I opened up the mailboxes in a text editor, and they look like
> standard *nix text file mailboxes.
>
> Any suggestions on what to try? I don't want to abandon my mailboxes
> and 10+ email accounts, and of course don't want to lose any data
>
> Joel West



Sounds like you are moving from one Mac to the other?

Why not try this:

..zip your whole "Eudora Folder" on the PowerPC mac.

Transfer the .zip file (maybe using rsync?)

Unzip it on the Air.

See if it works that way?

I've not used rsync to move Eudora mailboxes (and usually just copy the
folder directly to an external hard disk when moving them back/forth
between machines), so I can't say if something about rsync isn't
respecting something.

- Steve

In article
<7298c66d-2f9d-41c0-af78-51c7588ef306@13g2000hsb.googlegroups.com>,
openitstrategies@gmail.com wrote:

> I'm having problems that sound like the "resource fork" problem
> mentioned earlier. I have been using Eudora for about 12 years and
> moved mailboxes 6 or 7 times, but this time it didn't work.
>
> What I did was:
> 1. Backup all my HDD using a command-line rsync, including 3.7gb of
> Eudora 6.1.1 (paid mode) mailboxes

What OS version? Which version/build of rsync? What command line flags?
It has mostly been unsafe to use rsync as a backup tool on MacOS X
because the version shipped by Apple prior to Tiger completely ignored
resource forks (as did the open source base) and Apple's variant had
some more subtle problems until about a year ago.

*In general* rsync is a risky choice as a backup tool. It is great for
what it is designed for (synchronizing big files with small changes
across slow links) but it has a known edge case where the target can end
up mangled if the source changes mid-sync. If you used rsync to back up
those mailboxes while Eudora was running, there's a significant chance
that the data in them got mangled, instead of just having the resource
fork dropped.

> 2. Watch my HDD crash

That's always bad...

If you still have the old disk accessible, it might be possible to
salvage some data even if the disk will not mount. If you can get 'dd'
to read any data from the raw device node, you might be able to use it
to dump what can be read from the disk to someplace where you can treat
it like a logically mangled disk image that needs repair. I've done that
successfully, and while it is a major chore requiring patience and lot
of free disk space, it can be worthwhile if you have no good backup.

> 3. Move the Eudora files back to my new HDD using rsync
> 4. Notice that Eudora 6.1.1 doesn't launch on my MacBook Air, so
> install 6.2.4
> 5. What I see is:
> a. error message "not a properly formatted mailbox" for In, Out, Junk
> etc.
> b. my custom mailboxes don't show up in the menu, and
> c. the settings (e.g. e-mail accounts are also lost).
>
> So the problem could be the rsync, the 6.1.1 -> 6.2.4, or the PowerPC
> to Intel transition.

Indeed. If there was a OS version change involved as well, it could also
be related.

> I opened up the mailboxes in a text editor, and they look like
> standard *nix text file mailboxes.
>
> Any suggestions on what to try? I don't want to abandon my mailboxes
> and 10+ email accounts, and of course don't want to lose any data

It sounds like you have already lost your settings. The "Eudora
Settings" file has *only* a resource fork, so if you backed it up using
a version of rsync (or command line flags) that ignored resource forks,
you never backed it up. You will probably need to re-enter the account
settings and other options by hand unless you have some better sort of
backup.

The mailboxes are almost certainly recoverable, but the lack of resource
forks is probably confusing Eudora and will mean that any meta-data that
isn't discoverable from the messages (i.e. read/unread/replied status)
is lost. The Eudora mailboxes are not actually identical to any other
variant of the traditional mbox format, since they have these oddities:

1. All linebreaks are bare carriage returns
2. The message-delimiter "From " lines all use '???@???' in the sender
address part.

Assuming the mailboxes you have are not internally damaged, it should be
possible to cause Eudora to recognize a file as a mailbox and
reconstruct a sane resource fork TOC from the contents by setting the
type and creator for the file before you try to open it with Eudora. The
easiest command-line way to do this is with the SetFile tool that is
part of the Apple Developer Tools package:

/Developer/Tools/SetFile -c CSOm -t TEXT filename

--
Now where did I hide that website...

> > 2. Watch my HDD crash
>
> That's always bad...
>
> If you still have the old disk accessible, it might be possible to
> salvage some data even if the disk will not mount. If you can get 'dd'
> to read any data from the raw device node, you might be able to use it
> to dump what can be read from the disk to someplace where you can treat
> it like a logically mangled disk image that needs repair. I've done that
> successfully, and while it is a major chore requiring patience and lot
> of free disk space, it can be worthwhile if you have no good backup.

dd is good, but a bit manual. ddrescue automates some of the repetitive
parts of copying a damaged disk and works well. (Yes, I used it over a
period of about 3 months to recover more than 90% of my files from a
crashed disk. It can take a long time, but it may be worth it.)

http://www.gnu.org/software/ddrescue/ddrescue.html

Note that dd_rescue was an early attempt to do the same thing, but is
nowhere near as developed as ddrescue.

Cheers

DAvid

In article <davidmor-E789D5.18171431032008@news.internode.on.net>,
David Morrison <davidmor@excite.com> wrote:

> > > 2. Watch my HDD crash
> >
> > That's always bad...
> >
> > If you still have the old disk accessible, it might be possible to
> > salvage some data even if the disk will not mount. If you can get 'dd'
> > to read any data from the raw device node, you might be able to use it
> > to dump what can be read from the disk to someplace where you can treat
> > it like a logically mangled disk image that needs repair. I've done that
> > successfully, and while it is a major chore requiring patience and lot
> > of free disk space, it can be worthwhile if you have no good backup.
>
> dd is good, but a bit manual. ddrescue automates some of the repetitive
> parts of copying a damaged disk and works well. (Yes, I used it over a
> period of about 3 months to recover more than 90% of my files from a
> crashed disk. It can take a long time, but it may be worth it.)
>
> http://www.gnu.org/software/ddrescue/ddrescue.html
>
> Note that dd_rescue was an early attempt to do the same thing, but is
> nowhere near as developed as ddrescue.

Just using dd with the right args (conv=noerror,sync if I recall
correctly) did the trick for me without much hassle, and in combination
with fsck_hfs got me back over 95% of the files on a nearly-full 30GB
disk back in 2003. It took about 2 days start to finish and was mostly a
matter of waiting for dd to give up on the really bad blocks, not
actually doing anything.

--
Now where did I hide that website...

> Just using dd with the right args (conv=noerror,sync if I recall
> correctly) did the trick for me without much hassle, and in combination
> with fsck_hfs got me back over 95% of the files on a nearly-full 30GB
> disk back in 2003. It took about 2 days start to finish and was mostly a
> matter of waiting for dd to give up on the really bad blocks, not
> actually doing anything.

Where ddrescue helps is that it starts off copying large chunks, so as
to copy the disk as quickly as possible before it fails totally. After
it has got through the entire disk, skipping the bits with errors, it
goes back and starts again on the skipped bits reading a block at a time
until it hits the bad block(s). Then it goes to the end of the skipped
section and starts reading backwards until it hits the bad blocks. It
then goes on to the next skipped section and repeats the process.

The idea is to maximise the number of blocks retrieved.

Cheers

David