How to use Chars in classes not derived from AR

Rails
1

Hi!

I have a class that derives from nothing and is located in /models. I’m
using " my string ".chars to manipulate strings that potentially contain
special chars and replace them with an US-ASCII equivalent.

It did work until yesterday. Now on a linux system “a”.chars returns
#Enumerable::Enumerator:0xb74e3d6c

While it should return
#<ActiveSupport::Multibyte::Chars:0xb74e1bc0 @string=“a”>

(the class is contained in
active_support/multibyte/chars.rb)

Therefore the whole app crashes because Enumerable does not have methods
like “downcase” or “‘a’.chars[0]” which can handle UTF-8 chars as well.
Using it w/o UTF-8 support ends up in a mess.

Any ideas? How could I require or include
“ActiveSupport::Multibyte::Chars”? Any tests have been negative so far.

regards
David

2

On 18 Jun 2008, at 10:46, David S. wrote:

#Enumerable::Enumerator:0xb74e3d6c

While it should return
#<ActiveSupport::Multibyte::Chars:0xb74e1bc0 @string=“a”>

Have you just upgraded to ruby 1.8.7 by any chance? Are you running
rails 2.1 ? versions < 2.1 aren’t ready for 1.8.7 IIRC.

Fred

3

Yes, that’s it. I’ve just checked on another linux server that is
running with Ruby 1.8.6.144-2 and a vulnerable openssl version. And it
still works there.

It might be connected to the different ruby version. I’m trying to
downgrade now.

Have you just upgraded to ruby 1.8.7 by any chance? Are you running
rails 2.1 ? versions < 2.1 aren’t ready for 1.8.7 IIRC.

Fred

4

This did not solved the problem so far.

Any experiences or ideas regarding this issue?

Thanks
David

5

Finally it did.

I’ve just forgot to undo the changes I did while I was trying to fix it
by myself.

k Thanks!

Any rumors why this happens that often, or why this particular packages
is causing this issue?

I’ve already experienced it several times that our rails application was
broken after upgrading a gem or rails specific deb-package.

Regards
David