--WYTEVAkct0FjGQmd Content-Type: text/plain; charset=utf-8 Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Jos=C3=A9 Valim: > I also disagree that ruby hashes should be with indifferent access. +1, they shouldn=E2=80=99t (IMHO). Hash keys can comprise any objects that = have sane #hash and #eql? methods (stdlib=E2=80=99s Set is fundamentally based on this premise) and I don=E2=80=99t think any special casing should be made f= or Strings and Symbols (except for String freezing when they become keys). > The only places where [Rails] accept[s] both types > are in params and cookies and we do that as a security > measure. If we automatically converted such keys to > symbols, someone could use it to cause a DoS attack. Also, the above will at some point in time stop being a security concern (IIRC the DoS is MRI-only, right?). > Overall, I agree with the proposed syntax and with the behavior > that it should return a symbol. For consistency, I also think > interpolation should be allowed when double quotes are used. +1 =E2=80=93 and thanks a lot for your great EuRuKo talk! :) =E2=80=94 Piotr Szotkowski --=20 My McDonald=E2=80=99s order was 28 and the next one was 29 and I thought, =E2=80=98Aha! Insecure predictable sequence numbers.=E2=80=99 [Mark Pilgrim] --WYTEVAkct0FjGQmd Content-Type: application/pgp-signature; name="signature.asc" Content-Description: Digital signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (GNU/Linux) iEYEARECAAYFAk3jnFkACgkQi/mCfdEo8UpH9QCfUo08IEn/hneUrifkZBEFdMoD 1goAn11j6klNuXVrwHSVG0bJSoSgD0IX =CXwF -----END PGP SIGNATURE----- --WYTEVAkct0FjGQmd-- Jos Valim: > I also disagree that ruby hashes should be with indifferent access. +1, they shouldn¡Çt (IMHO). Hash keys can comprise any objects that have sane #hash and #eql? methods (stdlib¡Çs Set is fundamentally based on this premise) and I don¡Çt think any special casing should be made for Strings and Symbols (except for String freezing when they become keys). > The only places where [Rails] accept[s] both types > are in params and cookies and we do that as a security > measure. If we automatically converted such keys to > symbols, someone could use it to cause a DoS attack. Also, the above will at some point in time stop being a security concern (IIRC the DoS is MRI-only, right?). > Overall, I agree with the proposed syntax and with the behavior > that it should return a symbol. For consistency, I also think > interpolation should be allowed when double quotes are used. +1 and thanks a lot for your great EuRuKo talk! :) ¡½ Piotr Szotkowski -- My McDonald¡Çs order was 28 and the next one was 29 and I thought, ¡ÆAha! Insecure predictable sequence numbers.¡Ç [Mark Pilgrim] -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (GNU/Linux) iEYEARECAAYFAk3jnFkACgkQi/mCfdEo8UpH9QCfUo08IEn/hneUrifkZBEFdMoD 1goAn11j6klNuXVrwHSVG0bJSoSgD0IX =CXwF -----END PGP SIGNATURE-----