> I will run my simple tests in due course, but I do not have a > sophisticated LDAP test. I rely on you real LDAP users...
The quality management part in me would be very glad if your LDAP tests
tried three kinds of bind: anonymous, correct and wrong credentials.
That way it would ensure that whatever happens, authentication is not
broken by accident.
Btw, nobody has to feel bad about the original code, because it resembled
what is used in many free projects. So far I saw lots of code that
sets the network timeout and then uses *_s calls, thus timing out a hung
connect(), but blocking forever on a hung daemon, probably because the
authors thought that the timeout had an effect on *_s calls.
