That's very interesting. The expires header will still be
computationally expensive to process however you perform the
operation.
Keeping an index makes a lot of sense once you have seen a message
though. You could mark the timestamp of the last message that has been
indexed and then only reprocess the new ones but I don't think there's
any getting around the fact you have to perform this calculation. At
least the headers are the easiest thing to parse.
Steve
On Feb 15, 11:48 am, Joshua Baer <
joshuab...@otherinbox.com> wrote:
> Oh so basically just filter them out at display time. That makes sense.
>
> ~Josh
>
> On Feb 14, 2011, at 9:15 PM, Ian Ragsdale <
ianragsd...@otherinbox.com> wrote:
>
>
>
>
>
>
>
> > If disk space isn't really a concern, another option is to not to actively delete them at all, but to filter them out when iterating the mailbox. That means mailbox counts could be an issue, but is almost cost free and easy to implement.
>