RE: [PEAR] Inconsisteny problem with HTTP_Session 0.4

Hi Stefan, thanks for your quick answer and for taking care of my problem. > On 8 Feb 2004 at 1:43, Torsten Roehr wrote: > > > I'm using HTTP_Session 0.4 with the PEAR::DB container and experience > > a strange inconsistency problem. Sometimes the session data just gets > > lost although the session is not expired. I've tested my expiry check > > successfully with a session limit of a few minutes and than set the > > value to 4 hours for the production environment (it's not a > > high-security website). I've set idle time to 0 to avoid it being > > checked. > > Hi Torsten, > > I haven't yet heard of problems like this. But I'll promise to > investigate further into this subject when I'll work on HTTP_Session > again (hopefully beginning of March). Currently time's a bit short. > > If you manage to get additional information on the problem or can > maybe even narrow it down more specific please leave me an email. I just searched the HTTP_Session source code for session.gc_maxlifetime and only found it in /Session/Container/DB.php. But there it is only read with ini_get(). So it seems that it is nowhere set and therefore always the standard time of 1440 seconds. Isn't this a crucial value for making sure how long the session data is supposed to be available or am I wrong here? If so I could set it manually to 4 hours before setting up my HTTP_Session stuff. And maybe someone could include it into the package with the next release. What do you think, Stefan? Big thanks in advance. > > > Regards, > Stefan Neufeind