Hi Max, it appears to me that mod_ldap is trying to manage a global (cross process + cross thread) lock outside of rmm, which is why it tells RMM to not use a lock.
Windows is a special case where there is only 1 process and N threads. But passing in a thread mutex only is probably wrong outside of windows.
Not having looked too closely, it seems likely that an LDAP_CACHE_LOCK() is missing somewhere where we call into rmm.
Hi Max, it appears to me that mod_ldap is trying to manage a global (cross process + cross thread) lock outside of rmm, which is why it tells RMM to not use a lock.
Windows is a special case where there is only 1 process and N threads. But passing in a thread mutex only is probably wrong outside of windows.
Not having looked too closely, it seems likely that an LDAP_CACHE_LOCK() is missing somewhere where we call into rmm.
Have you been able to verify on a later release?