Re: [Evolution-hackers] Evolution functionality



On Mon, 2005-03-28 at 10:29 -0800, James Couzens wrote:
> Everyone,
> 
> Perhaps this has been addressed in the past, but during a brief scan I
> was unable to come up with anything so please in advance forgive this
> message if it is a rehash of a dead/solved issue.
> 
> I am curious about two things with respect to functionality.  
> 
> #1) A filter to remove duplicate e-mails.  Boy would this be a treat.
> Does it exist currently?

no

>   If not, would it be a
> hard/difficult/excessively time consuming thing to facilitate?

wouldn't necessarily be difficult to implement, but is it really needed?
I've seen a few people ask for this but generally it's because of 1 of
two reasons:

1. evo had a bug downloading pop mail where it would refetch some
messages (server changed the headers usually, by changing/adding
[X-]Status: headers)

2. people get multiple copies of the message through different paths
(one via a mailing list and one direct).

the first one is easier to match for (assuming we know which headers to
ignore when byte comparing), but I'd rather fix the pop download code
(and I'm pretty sure that code has been fixed now) :)

the second one is harder to match for... in addition to ignore Status:
and the like, we'd have to ignore Received: headers and other non-common
header fields (meaning headers that both messages don't share, since a
mailing list might add some of its own, etc) probably. We'd also have to
do smart subject matching because mailing-lists tend to munge that
header. we probably also could not match against From/Reply-To since
some mailing-list software likes to munge those too. body matching is
also a little more cumbersome because mailing lists like to add their
own footers.

"why can't you just match against the message-id header?"

because it's not guarenteed to be unique (yes, I know what the rfc says
but that isn't real-world)

so while it can be used for a quick "is it worth spending processing
time to see if these 2 messages really match?" compare, it can't be the
only compare done to decide if they match.

>   If not
> I'd be happy to even write it if someone can shed some light on where to
> start.

feel free. I'd guess you'd start by constructing some matching algorithm
and then you'd want to look at Camel and possibly EPlugin to start
implementing it.

> 
> #2) Maildir was removed some time ago as best I can tell.  Anyone happen
> to know the subject of the thread(s) where I can read up on why this was
> the case and or discern the status of this apparently former
> functionality?

it wasn't removed. the change was simply that the "On This COmputer"
folder heirarchy was changed such that it was always mbox - it made the
directory layout much simpler.

could we have used maildir instead? sure, but it was decided to go with
mbox because that's what other unix mailers use. (mbox and maildir both
have their drawbacks)

-- 
Jeffrey Stedfast
Evolution Hacker - Novell, Inc.
fejj ximian com  - www.novell.com

Attachment: smime.p7s
Description: S/MIME cryptographic signature



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]