DECISION: what reconstruct -f should do

[Bron Gondwana]

On Wed, 27 Apr 2011 08:19:10 +0200
Joost Roeleveld <joost at antarean.org> wrote:
> In my experience, most IMAP-mail clients have problems inserting the 
> "INBOX.sub" folder. In these cases we'd probably need to do this manually 
> using cyradm.

I think at this point it comes down to "what is most likely" - remember this
is a pretty weird edge case anyway.  The folder exists - has sub folders,
is missing ALL its cyrus.* files, and doesn't have any messages in it.

The flip side, intermediate directory that legitimately isn't a folder is
much more likely.

> Would a second flag, for instance "--full-dir-tree" (or similar), to change the 
> behaviour to create the directory regardless work?

It would, but we don't do long flag names, and besides I hate flag proliferation.
It's hard enough remembering what all the flags to reconstruct do as it is!  This
is a super rare edge case as it is...

Bron.