[ Brian L. Heess writes: ]
>
> I am currently only zubzcribed to the workers list, so if you could reply
> or cc me directly, I'd appreciate it!
Me too, so I won't see any replies on the Mj-users list unless I'm also
Cc'ed.
> I am configuring Wilma 1.4 (seems even nicer than 1.3). I archive all of
> my lists daily in majordomo (I switched to daily years ago since monthly
> was WAY too big for most users), so have file names such as:
>
> mantis.19980822
> mantis.19980823
> mantis.19980824
> mantis.19980825
> mantis.19980826
>
> I guess I'd like either to get the wilma engine to treat each month as a
> whole based on the available archives or add each DAY from the MJ archives
> into a month as time goes on in the Wilma archives (no, I don't think
> breaking the Wilma archives into daily searchable chunks is that helpful
> as there is not THAT much traffic on the lists, so no need to re-write the
> regexp or hack the code for that).
>
> Has anybody approached this problem or solved it?
Jason and I discussed this long ago in relation to Mj2 and reducing
the intimate ties to monthly archives, but nothing was really settled
and nothing's been done in development of Wilma 2.0. I guess your
choices for the present are hacking the code or using a wrapper to
collect the daily files into a temporary monthly archive. Note that a
wrapper needs to set the modification time of the monthly archive to the
latest modification time of all the dailies that comprise it, otherwise
everything will get rebuilt every run. Or it needs to be smart enough to
not create monthly archives that haven't changed since the last run.
Maybe this is a good opportunity to solicit input for what Wilma should
do in the context of nebulous archiving periods. I wouldn't be adverse
to some variability in the quanta of the searchable "chunks", but I
think it probably should have some limits, say quarterly, monthly, and
weekly, although the problem is how does one concisely express them in a
meaningful way?
--
Dave Wolfe
Follow-Ups:
References:
|
|
