Re: Problem due to lack of HTML date-stamp

On 15 Mar 1999 HyperMail_www_at_associate.com wrote:

> Has anyone else noticed that HTML files generated by HyperMail are not
> date-stamped in the name? That's obvious. But on a site where the HTML
> archives are rotated daily, the lack of date-stamping can make new HTML
> files appear to be read, even though they are new (if they've been read
> within the last few days, for example).

I don't full get what you mean here, but I think I get your main point.

> I've never done wildcards in C before (is it possible?)

Yes it is. Although hypermail doesn't "look" for which files to load really, it just start with 0000 and counts upwards until there's no more such files to read. Adding "wildcards" would mean it would have to use a different method. I'm not saying that is bad or anything, it is just some facts.

> But my thought would be to add a "seconds" stamp, like that in the lock
> file, to the HTML file names. Something like 0000-12657890.html

In what way would that help? Wouldn't it be better if you could tell hypermail to only keep the last XX hours/days/monts mails instead?

It doesn't really do that yet, but perhaps we could consider adding that instead of adding a kludgy work-around?  

> Just an idea.

We need ideas! Keep 'em coming!

--
             Daniel Stenberg - http://www.fts.frontec.se/~dast
   ech`echo xiun|tr nu oc|sed 'sx\([sx]\)\([xoi]\)xo un\2\1 is xg'`ol