--===============0019163958213925802==
Content-Type: multipart/signed; micalg=pgp-sha1;
	protocol="application/pgp-signature"; boundary="=-3PY/nspsjPh2wgIAESpw"


--=-3PY/nspsjPh2wgIAESpw
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

On Sat, 2003-08-02 at 18:41, Waldo Bastian wrote:
> On Saturday 02 August 2003 17:47, Petter E. Stokke wrote:
> > Is that a good idea? As I mentioned earlier, there are many non-standar=
d
> > URL constructs out there which are mangled by this behaviour in
> > Konqueror.=20
>=20
> You actually only mentioned "foo"

Well, one that's been giving me grief is ed2k, which is a pretty nastily
constructed, but KURL parseable, URL format for the eDonkey P2P network.
With the new KURL behaviour, filenames and hashes embedded in ed2k links
come out of Konqueror in all lower case.

The reason I brought it up, though, is that I'm writing a kioslave for
accessing file systems defined by symbolic names in a database, where
the host portion of the URL refers to this name, which is case
sensitive. This works just fine in KDE 3.1, but obviously breaks with
the new KURL behaviour in CVS.

> > At the very least, it should be limited to protocols which=20
> > are known to refer to DNS names in the host part, like http.
>=20
> Possibly, unfortunately that requires knowledge on a per-protocol basis=20
> whether it refers to DNS names or not.

I'll admit that's a pretty bad hack, but it's at least a compromise. But
why does KURL now convert host names to lower case in the first place?
Why is this an improvement over the previous lossless behaviour? I'll
admit it's potentially a useful thing to do to a DNS name, but isn't it
preferable for KURL to be more flexible about it, especially given that
it's used all over the kio subsystem?

The only argument I can come up with in favour of the new behaviour is
its usefulness in comparing KURLs to each other, where eg. the host part
"www.foobar.com" should match "www.FOOBAR.com". But isn't it preferable
to do a case insensitive match of the host part over actually enforcing
lower case, given the breakage that could result from such mangling?

--=20
Petter E. Stokke <gibreel@project23.no> http://www.gibreel.net/
PGP key: http://www.gibreel.net/key.asc
Fingerprint: 4FF3 12BD 692A 0FFF 984F  78DA 4776 81FB 1906 3A9F

--=-3PY/nspsjPh2wgIAESpw
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQA/K/btR3aB+xkGOp8RAvEnAJ0eBmP144uqnd7NzAdPKSFlVDogawCePsdo
D/lqj5cYvR0BkeDpoKTgj6g=
=8MFQ
-----END PGP SIGNATURE-----

--=-3PY/nspsjPh2wgIAESpw--


--===============0019163958213925802==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

 
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<

--===============0019163958213925802==--