- From: Roy T. Fielding <fielding@gbiv.com>
- Date: Sat, 14 Feb 2004 01:15:19 -0800
- To: Gisle Aas <gisle@ActiveState.com>
- Cc: uri@w3.org
On Thursday, July 24, 2003, at 12:02 AM, Gisle Aas wrote: > If the buffer starts out as "a/../../c" then this algorithm ends up > with "a/c" (step 5 kills the "/../.."). I don't think that is the > intention. Shouldn't step 5 and 6 be swapped? Ouch! That revealed a fundamental bug in the way that I was trying to describe the algorithm. After much fiddling and quite a bit of testing, I have rewritten the example algorithms for merge and remove_dot_segments: http://gbiv.com/protocols/uri/rev-2002/rfc2396bis.html#absolutize and the example written in perl is now much simpler: http://gbiv.com/protocols/uri/rev-2002/uri_test.pl So much for trying to finish tonight. Cheers, Roy T. Fielding <http://roy.gbiv.com/> Chief Scientist, Day Software <http://www.day.com/>
Received on Saturday, 14 February 2004 04:14:12 UTC