Percent Encoding the URI Template

[David Fuelling]

In section 5.1 of the spec, I think we should revert back to using Hex
representations of the opening and closing brace. I say this because
we instruct people to percent-encode these values using their hex
equivalent (per section 2.1 of the URI spec - http://tools.ietf.org/html/rfc3986#section-2.1).

Having just had to do this, I don't see the value in providing ASCII
values here, since the spec user is only going to care about the hex
equivalent.

Thoughts?

David

[David Fuelling]

Ok, just reading over draft 5 a bit more, and it does provide the hex
values in section 5.3. I can see the benefits in using the ASCII
equivalent in section 5.1, but I wonder if we should (in section 5.1)
use both ASCII and Hex? Throwing ASCII in there might confuse the
casual reader who doesn't keep reading on to section 5.3.

Thoughts?

david

On Jul 21, 4:24 pm, David Fuelling <sappe...@gmail.com> wrote:
> In section 5.1 of the spec, I think we should revert back to using Hex
> representations of the opening and closing brace.  I say this because
> we instruct people to percent-encode these values using their hex

> equivalent (per section 2.1 of the URI spec -http://tools.ietf.org/html/rfc3986#section-2.1).

[Will Norris]

I switched it to decimal because both the OpenID and OAuth specs use
decimal ASCII representations exclusively. We could add hex as well
if we want, but it needs to be clearly shown as hexidecimal to ensure
unambiguity -- "0x40" instead of just "40" as it was before for the
"@" sign.

-will

[David Fuelling]

I agree -- whatever we put in there should be really clear, especially if we mix hex and decimal together.