<DKIM> maxage causes loss of local email

[Janna Martl]

On Wed, Mar 11, 2015 at 12:01:07PM +0100, Nicolas Sebrecht wrote:
>On Wed, Mar 11, 2015 at 03:23:09AM -0400, Janna Martl wrote:
>> On Wed, Mar 11, 2015 at 01:50:13AM -0400, Janna Martl wrote:
>> >
>> >                X  Y      W
>> >|--------------|---------------|
>> >-24             0             24
>> >
>> >                   Y  Z   W
>> >        |--------------|---------------|
>> >       -24             0             24
>> >
>Don't forget the case where there is no Y. This is true at the first run
>and if user didn't have mail for a long time enough (> maxage).

If there's no Y, then messagelist1 = [X,W], messagelist2 = [W], and
nothing changes when you try to exclude things based on the maxage + 1
lists. So X gets deleted, which is correct. If there's no mail in this
range, then the lists are empty, nothing gets excluded, and nothing
happens.


>> The downside, though, is that asking both for messages within maxage
>> and messages within (maxage + 1) means that you're doing two
>> searches instead of (previously) one;
>
>No, you don't have to. Only request for (maxage + 1) ONCE. Everything
>after is "just" local computation.

What I meant is that you have to request a maxage list, and a (maxage
+ 1) list. So that's two IMAP SINCE queries. If you just request
SINCE(maxage + 1), you're getting times > (maxage + 1 in some random
timezone), and can't compute what the result of SINCE(maxage) would
have been, even using internaldates (which I thought you were trying
to avoid using anyway).

-- J.M.