added JSON serializers for JSONiJ and Argo; added Jackson Databind with S

[ProgrammerBruce]

added JSON serializers for JSONiJ and Argo; added Jackson Databind with Strings serializer; changed serializer names to lower case in reports and per preferences discussed in mailing list

[ProgrammerBruce]

This brings the JSON implementations to 17.

Here are the latest numbers.

                             create     ser   +same   deser   +shal   +deep   total   size  +dfl

json/jackson-manual 179 3049 2977 4798 4821 5140 8188 468 253 json/jackson-databind 177 4165 3879 6559 6759 6860 11025 503 271 json/jackson-databind-strings 177 5962 5644 7783 7547 7868 13830 503 271 json/protostuff-manual 177 3303 3081 4875 5064 5257 8560 449 233 json/protostuff-runtime 174 4060 3872 5843 5934 6084 10144 469 243 json/protobuf 317 22795 22259 140944 142096 148504 171299 488 253 json/google-gson/manual 176 45162 45102 14870 15125 15138 60300 468 253 json/google-gson/databind 179 63433 62824 67441 68222 68162 131595 486 259 json/svenson-databind 201 12692 12097 32897 32809 33357 46049 495 268 json/json-lib-databind 189 72804 72101 171915 170857 170664 243468 485 263 json/fastjson-databind 194 5536 5182 8737 8779 8797 14333 478 254 json/json.simple/manual/tree 191 16009 15581 21521 21846 22293 38302 495 269 json/json-smart/manual/tree 192 13299 13155 9267 9793 9996 23295 495 269 json/org.json/manual/tree 192 16388 16212 19840 19934 19978 36366 485 259 json/jsonij-jpath 190 76539 76373 31816 31885 31876 108415 478 254 json/jsonij-manual/tree 188 77141 76481 17222 17383 17672 94813 478 254 json/argo-manual/tree 192 179156 178227 51734 51625 52176 231332 485 263

[cowtowncoder]

One idea for going forward is that maybe String-as-input taking tests should be a separate test type; because:

• Binary codecs can not use String as input, and
• When allowing/expecting String input, textual parsers have different handling; and trying to shoe horn input via byte[] adds overhead unevenly (depending on what kind of input library supports).

But for now I'll just merge additional jackson codec.

Actually I think that separate test would work even without having to add new sets of codecs; can just extend base interface, have bogus impl (throw not-supported?) for existing codecs.