From: David Vallner <david@...>
Date: 2006-10-20T09:51:47+09:00
Subject: Re: challenge: correctly handling Accept-Language HTTP header

--------------enig13035FB3AA5E2804F508CD71
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

khaines@enigo.com wrote:
> If one has a list of en, en-gb and es, and one gets an Accept-Language
> header like this:
>=20
> es-mx,es;q=3D0.8,en-us;q=3D0.5,en;q=3D0.3
>=20
> the code would return the 'es' as the best matching language option.
>=20

I have some doubts you'll ever see that sort of header in the wild.

Besides, parsing that one is hardly something that needs a C library.

Here's an almost-one-liner, if you're willing to forgive the each block
being three statements and remove all whitespace (it actually parses...)

"es-mx,es;q=3D0.8,en-us;q=3D0.5,en;q=3D0.3".split(",").map { | i |
	i.split(";")
}.each { | i |
	i[1] =3D~ /q=3D(\d+\.\d+)/;
	i[1] =3D $1 && $1.to_f || 1.0
}.sort_by{ | i |
	i[1]
}.reverse[0][0]

=3D> "es-mx"

(The quality value for es-mx defaults to 1, es is only 0.8)

David Vallner
(That's by far the ugliest code I've written in a year.)


--------------enig13035FB3AA5E2804F508CD71
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: OpenPGP digital signature
Content-Disposition: attachment; filename="signature.asc"

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (MingW32)

iD8DBQFFOB2jy6MhrS8astoRAtLvAJ9Y2622jFm67RtzAwAxNZNhjsc1WwCcC9vx
g7u0k6RYdoZ8zeIktQEBBKo=
=DJeg
-----END PGP SIGNATURE-----

--------------enig13035FB3AA5E2804F508CD71--