[Savannah-help-public] Re: [sr #103933] Cleaning up mailing list archives?

[Karl Berry]

Hi Ralf,

    Will arch -wipe preserve URLs? 

No, it has no way to do so.  It just regenerates the html from scratch,
using the new mbox files as given.

To preserve url's, the only method I can think of without programming is
to leave the spam msgs in the mboxes, but without any subject or body.
At least then what will be shown on the mailman pages will not be so
distracting.

I guess, in theory, arch could then be taught to not put messages
selected according to some criteria in the html output (e.g., you put
x-spam: yes into the spam msgs), but to increment the counters so the
url's stay the same.  Maybe worth checking the mailman docs/faq/archives
about that?  Perhaps it's already been done.  It doesn't sound too hard,
but I know I'm not going to program it, sorry to say :(.

    gmane has a nice method of both weeding spam, regularly regenerating
    article numbers in order (at long intervals) but still also
    preserving old URLs.  I don't know how appropriate this external
    resource may be for the GNU project, but its features have made me
    use its links.

I think gmane is great, but don't see how it'll help with the mailman
archives?

Best,
Karl