On 06/08/2012 02:35, Peter Snow wrote:
Well you can continue to kid yourselves that the documentation is good if you like. The facts say differently. For example, I visit http://wiki.dovecot.org/MainConfig for help with the main config and at the top of the page it tells me that this page is for version 1.x, so I click the link to view the page for 2.x, which takes me to a page saying that the page I want has yet to be created. I therefore have no option but to refer to the version 1.x documentation. I copy "mechanisms = plain" from it
Google is *such* a useful tool.. http://lmgtfy.com/?q=dovecot+%22mechanisms+plain%22
but when I restart dovecot, it fails, telling me that it is not recognized!
No, probably it says something different. Please quote the error message, not your interpretation of the error message?
I noticed that also and did indeed follow many of them. Many of them though are for version 1.x but don't say so.
There are many useful differences between 1.x and 2.x, but its a gradual evolution, not a big change. As such the configuration changed a "large amount" between 1.x and 2.x, in that it's now stored in multiple files rather than a few big files, but for your concern such a change is relatively minor and the configuration options are largely the same.
By the way, I've now got it running. It wasn't failing due to the user being used to run the processes. It was due to misconfiguration of the way that the virtual users were setup, which in the end I managed to fix by interrogating a server with a working implementation (albeit ver 1.x) which was similar to what I needed and copying parts of it's config.
Please always post details of your problem and solution - us technical folks learn from people's mistakes, but it's not possible to learn and make things better without knowing what your problem and eventual solution were?
Additionally note that this is an opensource project and the documentation is written by people like yourself. Please consider clarifying whatever original document put you on the wrong track?
Although mutt now connects to it fine, roundcube doesn't, but don't worry. I'm not planning to bother you further.
Well, IMAP is just IMAP no matter which server you are using, so don't
treat this as some big black box that you can't open up and inspect.
IMAP is a plain text protocol and it shouldn't scare a technical person
to debug things. roundcube is also an extremely flexible beast and you
will need to get certain key settings correct before it connects
correctly, it likely also feels very "brittle" in that there aren't that
many settings to get right, but if any are wrong you will get major breakage
Good luck
Ed W
P.S. You came here with all guns blazing and seems like you are going to leave the same way? Why not try a more softly softly approach?