If we are going to expose php_session_active, we might aswell expose the other values. I will cook up a patch and apply it to trunk shortly

The notice is documented and can be disabled using the display error settings. To detect whether or not session_start was successful, you have to test its return value (as of 5.3.0).

It is actually a documentation problem as the examples in the manual does not check the return values (maybe add one on top of them using 5.3+).

pajoye,

I'm not sure if you understand which problem I'm trying to address here; it's 
not whether or not session_start() was successful, it's whether or not there is 
a current active session.

Many times, in frameworks or other large-scale applications, you have no way of 
knowing whether session_start() has already been called. You cannot rely on 
session_id(), SID, session_start(), $_SESSION, or any other session functions or 
session functionality (that I'm aware of) to tell me "yes, you already have a 
session open" or "no, you do not have a session open" _before_ I call 
session_start().

Any update on this? 

Simple request; I'm not exactly sure what all Kalle was speaking to WRT "expose 
the other values", but what I'm requested here is a tiny piece of functionality, 
and that is *all* that is being requested.

If you're trying to establish whether it's safe to start your own session, this patch is deficient because PS(session_status) is not a boolean - it could also be php_session_disabled.

I'd imagine robust code to handle this (rather unusual) case where you require a session but don't know if it's present, would be something like:

if (session_is_available()) {
    if (!session_is_active()) {
        if (headers_sent()) {
            // bail out
        } else {
            session_start();
        }
    }
} else {
    // bail out
}

We could instead provide a session_status() function which corresponds to the internal enum (returning SESSION_DISABLED, SESSION_NONE or SESSION_ACTIVE.)

I assume this is what Kalle was talking about WRT exposing the other values.

Both options are trivial patches. The former seems friendlier but more prone to error.

Thank you very much, Arpad; my rudimentary C skills were barely enough for me to 
craft the patch (which I somehow managed to compile and use), and so I appreciate 
the background info. I would be happy with any solution that helps me to solve the 
problem!

This bug has been fixed in SVN.

Snapshots of the sources are packaged every three hours; this change
will be in the next snapshot. You can grab the snapshot at
http://snaps.php.net/.

 For Windows:

http://windows.php.net/snapshots/
 
Thank you for the report, and for helping us make PHP better.

Thank you Arpad!

ee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3dee3e3deee3e3de3e3dee3e3dee3e3dee3e3dee3e3d