Fwd: "International Domain Names" support in KDE
bastian at kde.org
Tue Jan 28 11:56:28 GMT 2003
On Tuesday 28 January 2003 00:28, Thiago Macieira wrote:
> Waldo Bastian wrote:
> >On Monday 27 January 2003 23:08, Thiago Macieira wrote:
> >> The resolver part of the code is more or less settled. I wonder how
> >> support goes in the URL part now...
> >Trying to figure that out... I think KURL should convert alls hosts to
> > their ascii representation, but with prettyURL and prettyHost returning
> > the unicode representation. I just need to figure out if there are
> > urls/ioslaves around that use the host-name part of an URL for something
> > that isn't a DNS-style host.
> >And then all use of KURL::host() must be reviewed to see if it should be
> >changed to KURL::prettyHost()
> From what I gather, the ACE form of the host should never be used, except
> when passing down to an ASCII-only service, such as DNS. Other 8-bit only
> services might want to use UTF-8 in the future.
> So, IMO, KURL should normalise the hostnames (encode then decode -- see my
> normalizeDomain function in the Resolver class), but other than that do
> nothing else with it. All ioslaves and apps should accept Unicode hostnames
> and pass them down to the resolution API as such. Only the resolution
> process itself should transform into ACE.
The problem is that here is no decent standard for including non-ascii in
URLs. Basically the encoding of an URL is very poorly defined, unlike the
encoding used for e.g. subject lines in mail you more or less have to guess
the encoding based on the context in which they are used. Not surprisingly
thinks break horribly when they are transfered to another context.
By including the ACE form in the URL we at least don't hit that problem with
the hostname part and programs that aren't IDN aware will be able to proces
bastian at kde.org -=|[ SuSE, The Linux Desktop Experts ]|=- bastian at suse.com
More information about the kde-core-devel