Re: Imap splitting success

[Eric Abrahamsen <eric@ericabrahamsen.net>]

Eric Abrahamsen <eric@ericabrahamsen.net> writes:

> Okay, I've finally got everything working right, and wanted to lay out
> my imap splitting setup here for posterity. It was a pain in the butt to
> wrap my head around, and I assume other people may have similar
> difficulties. It ends up looking pretty simple, though, so maybe I
> really am the only one.
>
> The main headache was BBDB integration. In the normal nnmail setup, BBDB
> intercepts all splitting, and calls nnmail-split-fancy if it can't find
> its own match. That works because nnmail-split-fancy is both a variable
> and a function.
>
> For nnimap, nnimap-split-fancy is only a variable. Thus the BBDB
> integration process is reversed: splitting goes straight to
> nnimap-split-fancy, and bbdb/gnus-nnimap-folder-list-from-bbdb dumps
> splits directly into it. That took a while to figure out.
>
> So the basic arrangement: "imap" fields are added to BBDB records you
> want to create splits from. to-list/to-address/split-regexp parameters
> are added to groups you want to automatically split to. Then:
>
> (setq gnus-secondary-select-methods
>       '((nnimap "acc" 
> 		(nnimap-inbox "INBOX")
> 		(nnimap-split-methods nnimap-split-fancy))
>
> (setq nnimap-split-fancy `(|
> 			   ,@(bbdb/gnus-nnimap-folder-list-from-bbdb)
>                            ;; other splits here
> 			   (: gnus-group-split-fancy nil t "INBOX")))
>

I know everyone already knows this, but another key piece of the puzzle
is:

(setq nnmail-mail-splitting-decodes t)

This ensures that your splits are matched against decoded header values,
as you see them in the *Article* buffer, not as they actually exist
encoded on file. This is particularly important in the case of gmail,
which munges any header with non-ASCII characters. With the above
variable set to `t', your splits match against:

Subject: Try our new Ultra-腥™ Lutefisk [lʉːtfesk] substitute!

Instead of:

Subject: Try our new =?utf-8?B?VWx0cmEt6IWl4oSi?= Lutefisk =?utf-8?B?W2w=?=
 =?utf-8?B?yonLkHRmZXNrXQ==?= substitute!

Onward,
E