- From: Stefan Eissing <stefan.eissing@greenbytes.de>
- Date: Fri, 11 Oct 2002 15:15:17 +0200
- To: Martin Duerst <duerst@w3.org>
- Cc: uri@w3.org
Am Freitag, 11.10.02, um 11:16 Uhr (Europe/Berlin) schrieb Martin
Duerst:
> At 10:33 02/10/11 +0200, Stefan Eissing wrote:
>> [...]
>> While looking at it, I see that segment and rel_segment have different
>> sets of allowed characters. Noticeably, ':' is allowed in a (absolute)
>> path segment and forbidden in the rel_segment.
>>
>> Isn't that calling for trouble? I'd imagine that there is plenty of
>> code around
>> which converts absolute uris to uri references without looking if the
>> starting rel_segment will be free of ':' chars.
>
> Well, yes, but if there is a ':', then the part before it is
> interpreted as a scheme, and it's an absolute URI.
Exactly my concern. If ':' is allowed in path segments, a URI like
http://example.com/my:car.html
is a valid http URI. My (imaginary) HTML editor now
creates relative URIs whenever it can and generates:
http://example.com/index.html as
<html>
<body>
<a href="my:car.html">my beautiful black toyota</a>
</body>
</html>
Oops.
//Stefan
Received on Friday, 11 October 2002 09:15:52 UTC