Re: JSON and unicode surrogate pairs

On 06/11/2013 12:07 AM, Robert Haas wrote:
> On Mon, Jun 10, 2013 at 10:18 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> Well, if we have to break backwards compatibility when we try to do
>> binary storage, we're not going to be happy either.  So I think we'd
>> better have a plan in mind for what will happen then.
> Who says we're ever going to do any such thing?  This was extensively
> debated when we added the original type, and I thought that it was
> agreed that we might ultimately need both a type that stored JSON as
> text and another that stored it as binary.  
This is where the compatibility comes in - we do want both to
accept the same textual format.
> And we might need an
> XML-binary type as well.  But there are also cases where storing the
> data as text is *better*, 
Then use text :)
> and I don't see us ever getting rid of that.
While JSON is a "serialisation format" most things people want
to used it for are actually structured types, not their serialisation
to text. The serialisation should happen automatically.



-- 
Hannu Krosing
PostgreSQL Consultant
Performance, Scalability and High Availability
2ndQuadrant Nordic OÜ