[Dovecot] Auto create folder(s)

Bill Cole dovecot-20061108 at billmail.scconsult.com
Thu Sep 27 15:52:14 EEST 2007


At 4:49 PM -0400 9/26/07, Charles Marcus  imposed 
structure on a stream of electrons, yielding:
>On 9/26/2007, Fábio M. Catunda (catunda at contactnet.com.br) wrote:
>>  >
>>>  Could it could be written to simply create 
>>>the folder if the target folder doesn't exist 
>>>(ie, if the initial save fails due to 
>>>non-existent folder)? Then there would be no 
>>>performance hit...
>>>
>>There will be a performance hit couse you have 
>>to check if the folder exists or not.
>
>This is not good... but not what I wrote either.
>
>I'm not a programmer, so this may not be possible, but...
>
>What I wrote was, do NOT test BEFORE delivery is 
>attempted - only create the folder IF delivery 
>attempt FAILS *DUE* to non-existent folder.
>
>Ie:
>
>Deliver to user+foo
>    Error: Folder 'foo' doesn't exist for user
>   Ok, Do: Create folder 'foo' for user
>     ReDo: Deliver to user+foo
>
>Maybe this is not reliable - and if not, 
>definitely shouldn't be done. But if there is a 
>way to catch the error 'destination folder 
>doesn't exist', then it would NOT be a 
>performance hit, because no test is performed, 
>only an error handled differently.


I don't think there would be much real benefit on 
most OS's and filesystems given what is going to 
actually need to happen underneath the surface 
logic. It is simpler in English, and maybe in C, 
but the difference in what actually gets done is 
trivial, because whether the code says 'check for 
existence' or 'use this' the OS is still going to 
have to do roughly the same work.

More significantly, I think such a feature as you 
describe would have to be used with great 
caution. I certainly don't want to find that some 
spammer has created a bazillion new IMAP folders 
by sending to random tags.


-- 
Bill Cole                                  
bill at scconsult.com



More information about the dovecot mailing list