Re: test xmlbase/test012

I probably have my facts wrong - I haven't checked the documents:

There are two possible behaviours:

a) leave extra ".."'s in the path
b) delete extra ".."'s.

a is legal in 2396 and not in 2396bis
b is legal in 2396 (but not recommended) and in 2396bis

You are recommending we fix on a?

Brian




Jeremy Carroll wrote:
> See
> http://lists.w3.org/Archives/Public/www-rdf-comments/2003OctDec/0047.html
> 
> I suggest we modify the comment in the test case as follows:
> 
> See
> 
> http://www.w3.org/2000/10/rdf-tests/rdfcore/xmlbase/test012
> 
> OLD
> [[
>   Description: Example of relative uri with too many ..'s in path.
>   Note: RFC 2396 appears to permit implementations that discard
>         excess .. components, "compensating for obvious author
>         errors".
>         Such behaviour is not correct for RDF/XML.
>   Author: Jeremy Carroll (jjc@hpl.hp.com)
> ]]
> 
> NEW
> [[
>   Description: Example of relative uri with too many ..'s in path.
>   Note: The behavior recommended in the forthcoming revision 
>     of RFC 2396 may be different from in RFC 2396, hence  
>     document authors are advised to avoid such relative URIs.
>   Note: RFC 2396 appears to permit implementations that discard
>         excess .. components, "compensating for obvious author
>         errors".
>         To avoid ambiguity RDF/XML requires that no such compensation
>         is performed.
>   Author: Jeremy Carroll (jjc@hpl.hp.com)
> ]]
> 
> And then the following response ...
> 
> <<<
> Thank you for your comment.
> 
> We were unaware of this planned changes to RFC 2396.
> We have modified the comment in the test case to read:
> [[
>   ..
> ]]
> 
> In answer to your question: RDF/XML and RDF Concepts depend normatively on RFC 
> 2396. RFC 2396bis has not yet completed its standardization.
> Thus the test case is correct.
> 
> 
> Jeremy

Received on Thursday, 30 October 2003 05:02:01 UTC