On Wed, Dec 30, 2009 at 12:38 PM, David E. Wheeler <da...@kineticode.com> wrote: > On Dec 29, 2009, at 6:14 PM, Robert Haas wrote: > >> I've been mulling this over and I think this is a pretty good idea. >> If we could get it done in time for 8.5, we could actually change the >> output type of EXPLAIN (FORMAT JSON) to the new type. If not, I'm >> inclined to say that we should postpone adding any more functions that >> generate json output until such time as we have a real type for it. I >> wouldn't feel too bad about changing the output type of EXPLAIN >> (FORMAT JSON) from text to json in 8.6, because it's relatively >> difficult to be depending on that for anything very important. It's >> much easier to be depending on something like this, and changing it >> later could easily break working applications. > > +1 > >> Anyone have an interest in taking a crack at this? > > There are a bunch of C libraries listed on http://www.json.org/. Perhaps one > has a suitable license and clean enough implementation to be used?
It looks like they are all very permissive, though I wonder what the legal effect of a license clause that the software be used for Good and not Evil might be. I guess the question is whether we would slurp one of these into our code base, or whether we would add an analog of --with-libxml and provide only a stub implementation when the library is not present. Any opinions? Does anyone know whether any of these implementations are commonly packaged already? ...Robert -- Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-hackers