Pruning Duplicates

[Jorey Bump]

Wesley Craig wrote, at 08/25/2008 10:45 PM:
> I've seen this before with Thunderbird.  As I recall, Thunderbird 
> requests a lengthy operation but times out (or fills a buffer?) before 
> getting a result back.  It then tries the operation again, until the 
> mailbox is woefully full.

Interesting. Aside from webmail, Thunderbird is indeed the only other 
client used to access this account.

> To clean up, we typically calculate checksums on the files and find 
> duplicates that way.

In this case, I can easily find the inodes that reference more than one 
file:

  ls -li /var/spool/imap/user/bob/folder | cut -b 1-8 | sort | uniq -d

I was hoping to somehow leverage this information to delete the 
duplicates only, but haven't quite figured it out myself.