Re: URIs vs. half-baked URIs [glib PATCH]

[Daniel Veillard <veillard redhat com>]

On Fri, Aug 24, 2001 at 07:10:25PM -0400, Owen Taylor wrote:
> >    2) Normalize the character sequence according to Normalization Form
> >       C, as defined in [IETFNorm].  (See further discussion in Section
> >       3.1.)
> 
> This step is problematical ... it's easy enough to do in GLib-2.0, but
> since the NFC transformation is not reversible, if a filename is not
> in NFC already, this transformation will loose information that we
> need to uniquely, or at least, efficiently, identify the file.

  Right I should not have quoted that part.

> I don't think we should do this in GLib. You couldn't call open() on
> the result of the filename => URI => filename transformation and
> expect it to work.

  Agreed, too dangerous. There is actually 2 school of thought about
normalization and usually what's prevails is normalize on input so that
you don't get risks of inconsistencies if done internally. In the case
of a filename it's too late the file has been done. Maybe in 10 years when
we switch to the next generation of OSes :-)

Daniel

-- 
Daniel Veillard      | Red Hat Network http://redhat.com/products/network/
veillard redhat com  | libxml Gnome XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/