Panos advises me that indeed arrayToJSON does not work out of the box in lcserver, although it does work in the IDE and standalones. And that this is not a bug.
Let’s call it a lacuna, I suggest in code, Panos concedes in documentation. The story is that the mergJSON library, which is accessed from LC on all platforms using the mergJSONEncode function, only handles 1-dimensional arrays. The arrayToJSON encoder, and the corresponding decoder, use recursive loops to extend to multidimensional arrays. These convenience handlers are in a stack mergJSONLibrary.livescodecript, which is included in the IDE and standalones, but is not (*even* - my emphasis) distributed with LCserver. So to use arrayToJSON from LCserver one can upload the library stack - available at https://github.com/montegoulding/mergJSON/blob/master/mergJSON/mergJSONLibrary.livecodescript - and then call “start using…”. Or, more simply, copy the handlers from that stack into your own code. An addition to the dictionary entries for arrayToJSON and JSONToArray is definitely in order. Actually, the two handlers in question can be found in the mergJSONEncode and mergJSONDecode dictionary entries, if you first find those entries. I remain rather puzzled that the multi-dimensional array handlers were not embedded in the mergJSON C-code library in the first place by Monte or LC when they took it over. Blazing speed when communicating with javascript, which is where these calls would most likely be used particularly with LCServer, would seem to be imperative. Maybe the internal implementation of arrays was yet not settled? Whatever, a fix for this code lacuna would be nice, if unlikely to happen. [Actually in my case I wasn't talking to javascript. All I wanted was a way to transport an array from the server back to a client. The serialisation arrayEncode worked for me once I realised the keys and elements of the array all had to be in utf-16 encoding, anything in utf-8 could cause the decoding to fail. And for safety in sending over the net I also base64encoded - was this last step necessary or just paranoia?] Neville Smythe _______________________________________________ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
