From: Aaron Patterson <aaron@...>
Date: 2010-05-22T02:08:45+09:00
Subject: [ruby-core:30351] Re: [Backport #3329] Segfault using nokogiri


--45Z9DzgjV8m4Oswq
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Fri, May 21, 2010 at 09:50:54AM +0900, John Rowell wrote:
> Issue #3329 has been updated by John Rowell.
>=20
>=20
> This may be happening due to character encodings. A
>=20
>     res =3D xslt.transform(page.search('//body'))
>=20
> (where 'page' is a Mechanize instance) causes a segfault, while a
>=20
>     res =3D xslt.transform(Nokogiri::HTML(page.content, nil, page.encodin=
g))
>=20
> does not. The original page is encoded with ISO-8859-1, and Mechanize doe=
sn't always convert text to UTF-8 (#text is converted, #content is not). Ma=
ybe libxslt only accepts UTF-8 and Nokogiri is not properly converting the =
encodings before sending the text.

This sounds like it may be a bug in Nokogiri and not Ruby.  Can you
please add a ticket to our tracker here:

  http://github.com/tenderlove/nokogiri/issues

Also, if you provide the output of `nokogiri -v` and a script to
reproduce the problem, that would be extremely helpful.  Thanks!

--=20
Aaron Patterson
http://tenderlovemaking.com/

--45Z9DzgjV8m4Oswq
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (Darwin)

iQEcBAEBAgAGBQJL9r4gAAoJEJUxcLy0/6/GE90H+wTaFDww4E0Tuth2fFPtOq/1
8u3f0v8i7L+tjcLNLKqTqtmLoq4z3htisfz3wWO9jmH56A8ye7hWFEvO+gdprnbS
Wkee8UdgDnAN/bNF61vofr26IEOQrs7KzHKPTFS8RoBhYiwo+Bj4ZWfUH8jThh2T
lc7DKPntSDxWwpz/bGbmIBvl97vwmF6jbF7keJ4E5DuzRNGJa5gcIQIlfahHt+1N
3G4H+5OrvrTDSr5noYVToBbIP/KxycVuHfnA1g2xOVyXVH2twWBCmV8/St/PuLPt
mrVmtC2MEHxmSKzk24P6zoKFi8rgcsxVQrthp2xQJIXo8bY5oC7OIMIOJmBSN04=
=/jhi
-----END PGP SIGNATURE-----

--45Z9DzgjV8m4Oswq--