Re: [Json] The text in draft-ietf-json-text-sequence

[Nico Williams <[e-mail]>]

On Tue, Jul 08, 2014 at 11:50:54AM +0000, Joe Hildebrand (jhildebr) wrote:
> On 7/7/14, 10:57 PM, "Paul Hoffman" <[e-mail]> wrote:
> >- Do you feel that the protocol will be significantly broken if it uses
> >LF as the text separator?
> 
> Somewhat.

Considering that...

 - ...LFs cannot appear unescaped in strings,
 - therefore LFs in JSON texts are just whitespace,
 - therefore LFs in JSON texts may safely be removed or replaced with
   other JSON whitespace,

and considering that if we used RS a logger would still have to check
that a text to be logged does not include RS,...

...what might be broken if we use LF instead of RS?

Loggers would have to s/LF/SP/g or s/LF//g or equivalent.  But in the
common case this is equivalent to checking that a text to be logged has
no RS.

For loggers that produce their own texts, all they have to do is select
a "compact" encoding, as provided by most libraries, or s/LF/SP/g if no
such option is available.

(My apologies, I'm now repeating myself somewhat.)

Nico
--