failed bind vs. authorization identity
Hallvard B Furuseth <[email protected]> Tue, 27 Sep 2005 07:43:05 +0200
| Newsgroups | gmane.ietf.ldapbis |
|---|---|
| Message-ID | <[email protected]> |
It's dangerous to look to closely at an internet-draft... authmeth 4.2 (Anonymous Authorization After Failed Bind) says: > Upon receipt of a Bind request, the LDAP session is moved to an > anonymous state and only upon completion of the authentication > exchange (and the Bind operation) with a resultCode of success is > the LDAP session moved to an authenticated state. Thus, a failed > Bind operation produces an anonymous authorization state. If an already authenticated user does a Bind and receives non-success, does he then know that the session has reverted to anonymous? Even if he got unavailableCriticalExtension, which means the server "MUST NOT perform the operation" ([Prococol] 4.1.11. Controls)? Or invalidDNSyntax? How about protocolError - that can be either a valid Bind request with an unsupported version number, or genuine protocol error. What is a Bind request anyway? Any LDAPMessage with the [APPLICATION 0] tag? Any message which parses as a valid BindRequest from the ASN.1 in [protocol] 4.2(Bind Operation)? Is it still a Bind request if the LDAPMessage is itself is valid, but it contains a control which is invalid according to its control spec? -- Hallvard