http://bugzilla.spamassassin.org/show_bug.cgi?id=3936





------- Additional Comments From [EMAIL PROTECTED]  2004-10-31 14:55 -------
Subject: Re:  ArchiveIterator uses opt_n for two different things

On Sun, Oct 31, 2004 at 02:25:04PM -0800, [EMAIL PROTECTED] wrote:
> > I've enabled -n in my nightly mass-check runs for non-Bayes, because
> > the sorting is irrelevent without Bayes.
> 
> Not really true.  Sorting is also used for --tail and --head.  I use -n
> and --tail every night.

Except that with -n *there is no sorting*.  It's whatever order the
message path(s) are passed to AI.  There's no data to sort on since the
receive date isn't determined before the run with -n.

> Please post your proposal in the bug before committing it, I'd like to
> hear what's planned.

There's two operations: "sort run" (order the message array in date order)
and "check receive date" (figure out the receive date and determine if
the message is useful -- gives sort run data to use.)

There's really no need to change the API here after thinking about it some
more.  What I plan to do is:

Internally, use "opt_n" to specify whether or not the message array should be
sorted by date -- same as current.  However, add a "use_receive_date" option
(internal only) which is set whenever any of "after", "before", or "!opt_n" is
set.

This lets those options work as expected, and disables the receive date
determination if its not actually needed.





------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

Reply via email to