Chunked string parsing

[thiagomacieira]

This is a new v0.5 feature: the ability to parse chunked CBOR strings per chunk. It also adds the ability to do zero-copy parsing of strings. From the docs:

   if (cbor_value_is_text_string(value)) {
       char *ptr;
       size_t len;
       while (1) {
           err = cbor_value_get_text_string(value, &ptr, &len, &value));
           if (err) return err;
           if (ptr == NULL) return CborNoError;
           consume(ptr, len);
       }
   }

There are two commits in this series:

1. Add support for parsing chunked strings
2. rewrite the iterate_string_chunks function with new chunk code

[thiagomacieira]

@lpereira @otaviobp @rzr @erichkeane @alradmsft

[otaviobp]

The overall code looks good to me, but I am not working with any projects using cbor, so I didn't test it. Better asking for someone else that is using cbor to check that too.
Nice feature btw. I would definitely use that if I was working with cbor right now.

[otaviobp]

There is no cbor_value_get_text_string defined here. Am I missing something or this should be cbor_value_get_text_string_chunk?

[thiagomacieira]

cbor_value_get_xxx_string is an inline function in cbor.h which asserts that the type is correct, then calls _cbor_value_get_string. This documentation (and below) apply to the public function.

[otaviobp]

The inline function is called cbor_value_get_xxxx_string_chunk. I couldn't find a _cbor_value_get_string too. Unless there is a macro somewhere defining them, there is a problem with names in the sample code.

[thiagomacieira]

You're absolutely right. Thanks for catching this.

[otaviobp]

same here

[thiagomacieira]

With no reply, assuming everything is ok.