- From: Sebastian Pipping <webmaster@hartwork.org>
- Date: Tue, 31 Jul 2007 21:34:29 +0200
- To: Mike Brown <mike@skew.org>
- CC: uri@w3.org
Mike Brown wrote: >> 'example://A/b/c/%7bfoo%7d' --> 'example://A/b/c/%7bfoo%7d' > > Per RFC 3986 sec. 6.2.2.2, normalization of percent-encoding is achieved by > decoding those percent-encoded octets that correspond to unreserved > characters. %7b and %7d correspond to "{" and "}" which are reserved > characters. --------------------------------------------------------------- Yes, but shouldn't the host "A" become lowercase and "%7b" and "%7d" become uppercase? --------------------------------------------------------------- >> 'a/b/../../c' --> 'a/b/../../c' >> 'a/b/././c' --> 'a/b/././c' >> 'a/b/../c/././d' --> 'a/b/../c/././d' > > Per RFC 3986 sec. 6.2.2.3, normalization of dot segments in relative paths > (such as these) occurs during the URI reference resolution process. Only dot > segments in absolute paths would need to be normalized outside of that > process. --------------------------------------------------------------- I see, will have to fix that for uriparser. --------------------------------------------------------------- >>> We also do some special-casing to make sure the [Relativize] algorithm >>> isn't tripped up by empty path segments. >> Doesn't that change the "content" of URI? > > I'm not sure what you mean. --------------------------------------------------------------- I meant that maybe "a/b///c" and "a/b/c" are different for some applications and therefore fixing it would not work for them. Sebastian
Received on Tuesday, 31 July 2007 19:40:41 UTC