On Fri, Nov 25, 2005 at 09:08:52AM +0100, jeremiah_at_easywebsite.se wrote:
> The URL is a valid in the way that the resource physically exists and is
> reachable as an URL through other means.

The URL may be syntactically valid, but it does not specify a reachable
resource. Strictly speaking, wget and other clients that do give you
a resource for that URL are giving you invalid data: they're "broken".

> Yes, the current way is an attempt to adhere to RFC-1738. Also, there have
> been servers (in the past) for which the current method works while the full
>
> path method don't.
>
> Since there are also servers where the full path works while the separate
> path
> parts don't work, I certainly wouldn't mind seeing a change to libcurl that
> would make it attempt the other method if the first method fails.
>
> This sounds like it is worth doing, though perhaps of limited appeal.

There are definitely performance gains to be had implementing full-path ftp
CWDs, especially on ftp servers behind a high-latency link, even if it
breaks the specs. Falling back to the standard method on failure is
probably a reasonable compromise in today's Internet.

>>> Dan

-- 
http://www.MoveAnnouncer.com              The web change of address service
          Let webmasters know that your web site has moved