UTF-8 impact on authentication

The use of UTF-8 for internal data handling has impacts on the processing of authentication requests by WebSEAL.

The following list describes the impact of the usage of UTF-8 for internal data handling.

• UTF-8 logins over basic authentication are not supported.

  Use of UTF-8 with basic authentication login is not supported. UTF-8 logins with basic authentication cannot be supported because browsers transmit data in inconsistent ways. WebSEAL does not support multi-byte basic authentication logins because of browser inconsistency.

  WebSEAL uses basic authentication login strings with the expectation that they are in the local code page. WebSEAL supports 7-bit ASCII and single-byte Latin code pages. For example, a server that wants to allow French users to use basic authentication logins must run in a Latin locale. WebSEAL uses the basic authentication login string and converts it to UTF-8 internally. However, if the French user has a UTF-8 code page, basic authentication login is not available because the login string is multi-byte.

• Forms login.

  In previous versions of WebSEAL, forms login data was always used by WebSEAL with the auto function. WebSEAL examined the login data to see whether it was in UTF-8 format. If the data was not in UTF-8 format, the data was processed as local code page.

  For WebSEAL version 5.1 and greater, this setting is configurable as described in UTF-8 support in POST body information (forms).

• Cross-domain single signon, e-community single signon, and failover authentication

  Each of these authentication methods employs encoded tokens. The encoding of these tokens must be configured to use either UTF-8 encoding or non-UTF-8 encoding.