Some of the reports defined in this section do text matches of character strings provided by the client and are compared to stored calendar data. Since iCalendar data is, by default, encoded in the UTF-8 charset and may include characters outside the US-ASCII charset range in some property and parameter values, there is a need to ensure that text matching follows well-defined rules.
To deal with this, this specification makes use of the IANA Collation Registry defined in [RFC4790] to specify collations that may be used to carry out the text comparison operations with a well-defined rule.
The comparisons used in CalDAV are all "substring" matches, as per [RFC4790], Section 4.2. Collations supported by the server MUST support "substring" match operations.
CalDAV servers are REQUIRED to support the "i;ascii-casemap" and "i;octet" collations, as described in [RFC4790], and MAY support other collations.
Servers MUST advertise the set of collations that they support via the CALDAV:supported-collation-set property defined on any resource that supports reports that use collations.
Clients MUST only use collations from the list advertised by the server.
In the absence of a collation explicitly specified by the client, or if the client specifies the "default" collation identifier (as defined in [RFC4790], Section 3.1), the server MUST default to using "i;ascii-casemap" as the collation.
Wildcards (as defined in [RFC4790], Section 3.2) MUST NOT be used in the collation identifier.
If the client chooses a collation not supported by the server, the server MUST respond with a CALDAV:supported-collation precondition error response.