You don't need to be this complicated. To index only the messages in a
MHonArc archive, simply tell SWISH-E *not* to index the indexes:

filename contains thrd date auth

This is how I do our Hypermail archives.
Roy Tennant

On Tue, 7 Mar 2000, Julian Perry wrote:

> Louis
> 
> > I am using MHonArc to archive email messages. The actual messages are
> > prefixed msg0001.html, etc. while the indexes are prefixed thrd, date,
> > auth. I am only interested in indexing the msg's. In looking through the
> > swish-e archives, I notice that there is a reference to doing this using
> > "regexp", but it is a little confusing. Assuming that my goal is to screen
> > for "msg*.html" only, what should go into the configuration file? Also,
> > what other provisions must I make outside the config file to enable this?
> 
> The regexp for that would be:
>   msg[0-9]*\.html
> or, if you know there will be 4 digits:
>   msg[0-9][0-9][0-9][0-9]\.html
> 
> [0-9]*   means zero or more occurences of the digits 0-9
> \.       means match a full-stop (as a . on it's own is the
>          wildcard to match any single character)
> 
> You could get even more specific and specify anchors for
> the name so that 'msg' must follow a '/' - and that
> the .html must be at the end:
>   /msg[0-9][0-9][0-9][0-9]\.html$
> -- 
> Cheers
> Jules.
>