[Dovecot] Case-insensitive "detail" mailboxes?
skdovecot at smail.inf.fh-brs.de
Fri Jul 26 08:54:03 EEST 2013
-----BEGIN PGP SIGNED MESSAGE-----
On Thu, 25 Jul 2013, Timo Sirainen wrote:
> On 25.7.2013, at 19.43, Timo Sirainen <tss at iki.fi> wrote:
>> On 25.7.2013, at 19.29, /dev/rob0 <rob0 at gmx.co.uk> wrote:
>>> We're using sieve with LMTP. We want to have lda_mailbox_autocreate
>>> and lmtp_save_to_detail_mailbox. Is there a way to make the "detail"
>>> case-insensitive? If so I have not found it yet.
>>> I suppose we could lowercase the input string for the SQL userdb
>>> query, but that's not what is wanted. The idea being that if a user
>>> makes a mailbox called "Test" is that user+test at example.com and
>>> user+TEST at example.com should both go to that "Test" mailbox. If it
>>> was lowercased, a mailbox called "Test" would be ignored and "test"
>>> With autocreate, this could be a problem if mail is delivered to
>>> autocreated case-sensitive mailboxes that the user won't see.
>> If you used case-insensitive filesystem, it would already work like that. Maybe a new setting to make mailboxes case-insensitive also with case-sensitive filesystems..
> Oh, except then you get into trouble with non-ASCII characters. Which
> unicode chars should be equal? Maybe i;unicode-casemap would work well
> enough, since it's also used for other purposes.
I see two other solvable problems:
"if a user makes a mailbox called "Test" is that user+test at example.com and
user+TEST at example.com should both go to that "Test" mailbox."
Hence, the delivery process would need some sort of mapping (dict):
case-insensitive -> real spelling
i.e. the user creates teSt or teST a.s.o., in order to find the correct
mailbox name in case-sensitive filesystems efficiently.
Secondly, what happens, if the user wants to create two mailboxes with
different case, say Test and tesT and test ? Either this is forbidden or
the delivery process needs to first check, if there is a mailbox with the
very exact name, but which one is to choose, if neither matches?
Personally, I would forbid two mailboxes that differs in case only, if
such mechanism is enabled.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
-----END PGP SIGNATURE-----
More information about the dovecot