From kde-core-devel  Tue Sep 27 17:18:00 2005
From: Thiago Macieira <thiago () kde ! org>
Date: Tue, 27 Sep 2005 17:18:00 +0000
To: kde-core-devel
Subject: Re: KURL problem
Message-Id: <200509271418.10872.thiago () kde ! org>
X-MARC-Message: https://marc.info/?l=kde-core-devel&m=112784151703454
MIME-Version: 1
Content-Type: multipart/mixed; boundary="--nextPart2661668.8vgEPhm3zy"

--nextPart2661668.8vgEPhm3zy
Content-Type: text/plain;
  charset="utf-8"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

David Faure wrote:
>However the real question is: do you remember why you asked for whether
> "%E1.foo" was a valid URL? It is parsed by the current KURL, but Thiago
> tells me that this url doesn't have a defined interpretation.

I guess we were trying to test =C3=A1.foo, but you changed it to %E1.foo to=
=20
avoid encoding issues.

Let me explain why %E1 is not valid in the hostname part: URLs are=20
supposed to be UTF-8 binary byte sequences. That means non-ASCII bytes=20
are supposed to be converted into characters when they form a valid UTF-8=20
sequence (e.g., %C3%A1), but invalid sequences are not supposed to be=20
discarded (e.g. %E1).

However, hostnames are Unicode strings, so you can't have invalid high-bit=
=20
sequences in UTF-8 representation (and they are impossible in UTF-16=20
representation). /usr/bin/idn behaves the same way:

(run on UTF-8 environment)
$ echo =C3=A1.foo | iconv -t latin1 | idn -a --quiet
idn: idna_to_ascii_4z: String preparation failed

BTW, Qt has for some time allowed those non-UTF8 sequences to be converted=
=20
into QStrings and back, by using some reserved characters in UTF-16.

=2D-=20
  Thiago Macieira  -  thiago (AT) macieira.info - thiago (AT) kde.org
    PGP/GPG: 0x6EF45358; fingerprint:
    E067 918B B660 DBD1 105C  966C 33F5 F005 6EF4 5358

2. T=C3=B3 cennan his weorc gearu, ymbe se circolwyrde, wear=C3=B0 se c=C3=
=A6gbord and se=20
leohtspeccabord, and =C3=BEa m=C3=BDs c=C3=B3mon lator. On =C3=BEone d=C3=
=A6g, he hine reste.

--nextPart2661668.8vgEPhm3zy
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)

iD8DBQBDOX7SM/XwBW70U1gRAiy/AJ4qsJz2xX/gI8Og7jtA5h00Rh6VRwCdHk5F
Q/y99fPwwf1mSgMU35wF1Xg=
=prpo
-----END PGP SIGNATURE-----

--nextPart2661668.8vgEPhm3zy--