Re: Archiving mbox messages - any good utilities?
On Mon, Sep 08, 2003 at 06:49:50AM -0400, David T-G wrote:
> %
> % I read my mail on a user account on an ISP's linux box using mutt, I
> % have a reasonable but not unlimited quota there. I would like to
> % backup my mail on my home machine where I have lots of 'free' disk
> % space.
>
> Makes sense (but do you have a tape backup at home? :-)
>
Yes, I do actually, two even (one on the Linux box and one on the
Win2k box).
>
> %
> % My ideal backup would be a directory structure the same as I maintain
> % on the ISP's Linux machine but with all mail that I have ever saved in
> % mailbxes. If this (sort of) mirrors my 'live' mail then I can delete
> % old mail from the live mail mailboxes secure in the knowledge that the
> % originals are on my home mcahine backup.
>
> There's the trick. I think that that requirement is going to make things
> very difficult. How will your copy method know whether you meant to
> delete that message or not?
>
I was aimimg to delete old messages from the live mail manually. The
copy/archive system *never* deletes a mail message, it just copies new
ones from the live mail to the archive.
> What I do is to use unison, a slightly smarter rsync, to mirror my remote
> tree on my home server. It is clever about only copying the changed bits
> and so data transfer is small, and I thus have an identical copy of "there"
> over "here".
>
Yes, as you say rsync plus a bit.
> Then, when I want to remove an archive from "there", I move it out of the
> tree "here" (into a different tree, of course). Unison notices and
> removes the remote copy for me. After I've built up the archive again
> (just a compressed mbox file; I don't bother with splitting apart by year
> or month) I do the same thing, this time appending the new archive to the
> old one.
>
Hmm, but I don't think this does what I want. All I want to to
able to do is remove old messages (when the fancy takes me, when I'm
nearing quota, whatever) on the live system and know that these old
messages will be archived. I never want to delete chunks or archive
chunks really. All my saved mail is saved by subject/classification
not by date at all. Saving the whole of my live mail at intervals
would guarantee I don't lose messages (as long as I don't delete
anything newer than the last archive save) but it might be quite
difficult finding something if I did ever want to delve back into the
archives. If all the archives were merged with duplicate messages
removed I'd have what I want.
Are there any mbox merging utilities which will merge two mbox files
and remove duplicate messages? That would (rather inefficiently)
allow me to do what I want. Copy my live mail (weekly, say) to hy
home system and merge into the archive there, removing duplicates.
--
Chris Green (chris@xxxxxxxxxxx)