Re: [apps-discuss] Review of draft-ietf-appsawg-json-pointer

["Martin J. Dürst" <[e-mail]>]

On 2012/11/16 0:24, Paul Hoffman wrote:

> - One of the security considerations is too limited.
>     Note that JSON pointers can contain the NUL (Unicode U+0000)
>     character, which may not be representable in all programming
>     languages.
> The same is true for all control characters, and for some programming languages, all non-ASCII characters. Proposed rewording:
>     Note that JSON pointers can contain the NUL (Unicode U+0000)
>     character, control characters, non-ASCII characters, and so
>     on. These characters may not be representable in all programming
>     languages.

I have to disagree here. There may indeed be programming languages that 
cannot handle non-ASCII characters. But these days, this means that they 
are very much fringe languages. On the other hand, there are also 
programming languages, or implementations of programming languages, that 
cannot handle ASCII (think EBCDIC machines).

The original warning, although it doesn't say so explicitly, is more or 
less specifically about string representation in C. C is an extremely 
widely used programming language, and its usual string representation 
takes the null byte to signal the end of a string.

So the warning essentially just reads: If you language happens to use a 
null byte to signal the end of a string (e.g. like in C) then you may 
have to be careful and invest additional effort (e.g. use a string 
library that handles null bytes, too). Same applies to wide strings and 
null 16-bit or 32-bit units.

So I don't think it's a good idea to widen this warning.

Regards,   Martin.