Re: [GeneX-dev] Embperl vs. Mason: Conclusions

[<ja...@op...>]

"Michael Pear" <mic...@ho...> writes:

> In Embperl, there is a way to set the expiration time for the 
> session cookie. I used 1 day in the Embperl session handling.
> I presume there is something in Apache::Session
> as well. I would say that it is a time period longer than the
> session expiration.

There is also a nice way to expire cookies with CGI.pm, too, but I'm
not sure how to do it with straight mod_perl. Mostly I wanted to know
what time frames sessions should last. Have you gotten any feedback
yet about that? 

One day seems a bit short to me, but I suppose that in the future we
could enable users to configure the timeout and store those values in
the DB, along with other user-specific preferences. 

One thought I just had about cleaning up sessions is what we want is
to only store on session per user. So if we have a us_fk column in the
session table, the INSERT trigger can automatically delete any
sessions with the same user ID. We can still use timestamps, but they
aren't as necessary.

> I think it best to have sessions that persist the browser session
> itself.  

Sorry, I didn't parse this very well. Try again?

> It is nice to come back to working on something without having to
> supply a login.

Agreed. Especially if we start storing more things in the session than
just login info.

jas.