Re: Spam robots and browser history prob

On Wed, 12 May 1999, Daniel Stenberg wrote:

> Glen Stewart wrote:
>
> > 2. I purge mailing list messages to my HyperMail listings DAILY. When
> > a reader comes back on day 2 to read new messages, their browser says
> > and thinks that 0001.html (etc etc) have already been read because the
> > file names are not date-specific. Can we PLEASE switch HyperMail to use
> > of a time stamp in the file name? Something a wildcard delete can
> > handle, like 0001-05111999224500.html ???
>
> The very moment we introduce a system where hypermail stores a special
> "state-file" (as John Finlay has previously described) in which it can
> store its currently used file names I'd have no problem what so ever to
> allow any kind of file names (like some kind of %-configuration in the
> config file). Then hypermail would find the current file names from the
> "state-file" anyway.

You could (almost) do this now using the htmlsuffix config setting - instead of the default setting of "html" use "DDMMYYYYY.html" (generating the DDMMYYYY bit in a the script that calls hypermail). Problem is this also affects the index files (index.html, subject.html etc) but you could get round that with symbolic links if you needed to.

At the moment hypermail needs all generated message files to have the same form - if you're regenerating the archive each day that shouldn't be a problem. As Daniel says, a state file for each archive would free us from that restriction.

Paul Received on Wed 12 May 1999 06:35:24 PM GMT