Marcin Soltysiak wrote:
Ok, Rambo - I got something for you :-)

in /NA/Functions there is function auth_hack_session with hardcoded replace
of

auth_sessiondata|O:12:"auth_session":10:

with

a:10

Mean time I got such session file

auth_sessiondata|O:12:"auth_session":11:{s:11:"start_utime";s:21:"0.57296900
1044723043";s:4:"seed";s:15:"144930435945608";s:4:"pad1";s:32:"d7e3fc9961957
323fb289f65180a75f7";s:4:"pad2";s:32:"6e2d7d6d6c651417ea7d7a8491ed5e7f";s:11
:"last_rerand";s:21:"0.57809600
1044723043";s:27:"__PHP_Incomplete_Class_Name";s:12:"auth_session";s:9:"logi
ntype";s:0:"";s:8:"username";s:5:"D?�AT";s:7:"expires";i:1044732614;s:6:"pas
swd";s:8:"�s�����Z";s:9:"sitegroup";s:0:"";}

What causes inconsistency in auth procedure that led to "cookie problem"
because auth returned empty array.

Somehow I feel it has to do with

s:27:"__PHP_Incomplete_Class_Name"

that seems to be an Alien - 11th of Nostromo here

What'd you say?

The __PHP_yaydya is pretty alien here, of course we could check if unserialize works on objects nowadays too (and then decide to skip the mangling). Or we could change it to use RegExp matching/mangling...

The RegExp math/mangle is propably best quick solution, I'll make the change and acknowledge when it's committed to CVS.

What is his PHP version ??

/Rambo


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to