[Cp-tools-discuss] Re: Gjdoc URI Dilemma

[Chris Burdess]

Julian Scheid wrote:
> Sascha Brawer wrote:
> >Url-escaping wouldn't work?
> >
> >http://developer.classpath.org/doc/java/io/InputStream.html#read%28byte%5b%5d,int,int%29
> > 
> >
> 
> No, unfortunately that's not enough.  Here are the relevant parts of the 
> spec again:
> 
> http://www.w3.org/TR/xhtml1/#C_8
> http://www.w3.org/TR/REC-xml/#id
> http://www.w3.org/TR/html4/types.html#h-6.2
> 
> To quote from the XML 1.0 spec:
> 
> [4] NameChar ::= Letter | Digit | '.' | '-' | '_' | ':' | CombiningChar 
> | Extender
> [5] Name ::= (Letter | '_' | ':') (NameChar)*
> 
> And from the HTML 4 spec:
> 
> ID and NAME tokens must begin with a letter ([A-Za-z]) and may be 
> followed by any number of letters, digits ([0-9]), hyphens ("-"), 
> underscores ("_"), colons (":"), and periods (".").

In that case I think the anchor name should (1) be URL-encoded as Sascha
suggests above, and then (2) instances of '%' in the resulting name
should be replaced by '-'. Since '-' is not a valid Java name character,
there can be no ambiguity (which would be the case with '_', for
instance).

I recommend that ':' not be used as a replacement character; although quite
valid in XML 1.0 Names, the colon is used extensively to denote a namespace
prefix.
-- 
Chris Burdess