Offlineimap does strange things.

Nicolas Sebrecht nicolas.s-dev at laposte.net
Fri Aug 3 23:50:05 BST 2012


On Fri, Aug 03, 2012 at 03:37:31PM +0200, Thomas Kahle wrote:

> On 21:18 Wed 01 Aug 2012, chris coleman wrote:
> > In my experience... the symptoms and errors that Cyril is reporting
> > here in this detailed message.. are a glaring sign of a procedural
> > codebase that has grown too big and complex for one person or group of
> > people - to manage.
> 
> I think this statement is far to general.  Debugging network dependent
> code is tricky

True. And this is not a matter of having a lot of different IMAP
servers at all. It's far more complicated, even if feasible.

Uploading and fetching mails is a very limited scope of tests not
helping to debug that much. We have to deal with threads, remote
connection surprises, IMAP server changing of state because of
concurrent accesses or whatever conditions of weather.

> > One example: each IMAP mail server flavor - dovecot, exchange, cyrus,
> > zarafa, gmail, yahoo, hotmail, etc., should have its own object -
> > derived from a base IMAPServer class- and tailored to each server's
> > peculiarities.
> 
> I'm not sure if this is a good design choice.  It makes offlineimap much
> less portable.  There are RFC's and specifications for mail servers, so
> in principle there should be no speical IMAPServer classes.  Of course,
> reality is different, but that does not mean that we need to go back to
> the dark days of vendor-specific Internet.

Yes. Actually, work was done to try to avoid this Gmail class without
full success... :-/

-- 
Nicolas Sebrecht




More information about the OfflineIMAP-project mailing list