Paul Battley dijo [Tue, Nov 15, 2011 at 10:14:28AM +0000]:
> Hi,
>
> > $ ruby1.9.1 -e 'require "htmlentities";
> > t=HTMLEntities.new.encode("<élan>"); puts t; puts t.encoding'
> > <élan>
> > UTF-8
>
> That's the expected behaviour: only the basic entities are escaped by
> default, i.e. only that which is necessary to produce valid HTML text.
> If you want the encoder to change é, you must tell it to use :decimal,
> :hexadecimal, or :named entities, thus:
Yup - The problem here is that your test suite requires the resulting
string's encoding to be US_ASCII (see
test_should_encode_utf8_to_ascii_if_possible in ruby_1_9_test.rb), and
'é' is not part of US_ASCII.
> >> HTMLEntities.new.encode("<élan>", :named)
> => "<élan>"
>
> This error isn't a serious failure, and may indicate a subtle change
> in how Ruby marks string encoding. Which version and patch level of
> Ruby is it, and is it really 1.9.1, or is it 1.9.2 with a package name
> of 1.9.1 (i.e. what does ruby -v say)?
Right, I am running Ruby 1.9.3p0 - The reason I'm running it as 1.9.1
as it is the version-explicit binary name we use in Debian, as it's
the binary compatibility level.
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
