|Title|Named links get invalid names when Swedish characters are used
|Date|09-Feb-2005 18:18:37 EET
|[Bug criticality]|[JSPWiki:LightBug]
|Browser version|Firefox 1.0
|[Bug status]|ClosedBug
|[PageProvider] used|-
|Servlet Container|Tomcat 5.5.4
|Operating System|OS/2 Warp 4.5
|Java version|1.4.2_05

When I create a header such as "Göteborg", the "ö" is replaced by the numeric character reference "%F6" in the named link end, which is invalid in any XHTML version. The name should use another naming method instead:

Valid alternatives:



I'm sorry, I don't get it.  What JSPWiki does should be quite according to [RFC 3987|http://www.ietf.org/rfc/rfc3987.txt]...  If you use UTF-8, that is.  What exactly is the problem?  Could you post an example?

-- JanneJalkanen


It is true that RFC 3986 and 3987 allow for pct-encoded.

But in XHTML, fragment identifiers is of type NMTOKEN:
[XML|http://www.w3.org/TR/2000/REC-xml-20001006]. In HTML 4, CDATA was used.

As I understand it, NMTOKEN does not include any pct-encoded (%F6) so numeric character references can't be used. 

See also [XHTML 1.0|http://www.w3.org/TR/xhtml1/#C_8]
"When defining fragment identifiers to be backward-compatible, only strings matching the pattern A-Za-z A-Za-z0-9:_.-* should be used".

Example page [TestingGöteborg]
-- Per

Fixed in 2.2.17.