16 Aug
2012
16 Aug
'12
7:02 a.m.
On 08/16/2012 08:00 AM, Alan DeKok wrote:
Phil Mayers wrote:
Bear in mind that rlm_ldap has some quite complex password processing rules hard-coded into the source. In particular, I think that code still, incorrectly, uses PW_USER_PASSWORD as the attribute, then overrides it based on the presence and value of a {type} header:
That's what has been working for ~10 years. I'm wary of breaking peoples systems in a minor release.
Oh totally - I'm not suggesting fiddling with the existing code, just noting that the reason the OP was getting the warning about User-Password versus Cleartext-Password was (I guess) rlm_ldap doing it, rather than anything he did.