Re: Sessions: strange behaviour in 4.3.3 ?
| From: | Dino Tsoumakis | Date: | Fri, 16 Jan 2004 17:53:03 +0000 |
| Subject: | Re: Sessions: strange behaviour in 4.3.3 ? | ||
| References: | 1 2 | Groups: | php.general |
| Request: | Send a blank email to php-general+get-174854@lists.php.net to get a copy of this message | ||
"Chris Shiflett" <shiflett@php.net> wrote
news:20040116162954.93129.qmail@web14304.mail.yahoo.com...
> --- Dino Tsoumakis <dimaki@gmx.de> wrote:
> > Warning: session_start(): Cannot send session cache limiter - headers
> > already sent (output started at XXXXXX/class.session.php:61) in
> > XXXXXX/class.session.php on line 38
>
> You need to either put session_start() prior to any output (my preference)
> or use output buffering by putting ob_start() prior to any output.
If you hve a close look at the code http://www.serv-int.de/class.session.txt
you will find the session_start() in the constructor of the class prior to
any output.
The thing is, that session_start() calls the user defined session functions
(defined in session_set_save_handler()) and that's what breaks the whole
thing.
This is not a problem of session_start(), I'm pretty sure.
The problem is, the session id is not set in the open function, but it is
set in the read function.