I promised a few quibbles. GET seems troublesome. There seems no way in WD1.1 to distinguish a GET that wants a JSON response from a GET that wants a JSON-RPC...
I wonder if in that case the right solution wouldn't be to simply send back the data from the "result" member of the Procedure Return Object (and the "error"...
I just realized a problem with what I proposed earlier. FWIW, JSON text cannot be composed of a single String, Number, Boolean or Null value. It must be either...
... I would like to back up Jim's problem statement. When a request arrives to the server, it must be possible to decide if we have a JSON-RPC request or not...
Balazs Ree
ree@...
Sep 7, 2006 5:51 am
319
Sections 6.6 and 6.6.1 are probably more relevant for this discussion because that's what allows your case of optional versus required parameters to pass. I...
Atif, Thank you for the response. I was just thinking in terms of how the data is likely to be used. "system.describe" is a MUST for a conforming JSON-RPC...
Point taken, Jim. I agree that there may be room for misunderstanding. I'll see what I can do to improve the wording or discussion around that topic. Perhaps a...
... 1.1 WD only specifies "application/json" as an Accept parameter in the header to distinguish this from any other GET. Would "application/json-rpc" be a...
Hi Atif, ... It *is* also about content negotiation. If a client requires a json-rpc response it won't accept a non-json-rpc response and vice versa. But I can...
Balazs Ree
ree@...
Sep 11, 2006 12:05 pm
324
I found the following comment in the new draft: Atif Aziz: Strictly disallow the use of cookies. Disallowing cookies will make it impossible to use JSON-RPC...
... Why not just leave this outside of the spec, instead of `must`? I definitely see lots of places where sessions would to be started outside of the rpc call...
Matt
WickedLogic@...
Sep 14, 2006 2:39 am
326
Matt + Wilfred, I wouldn't that note to heart. It was put in there to provoke some reaction and I have that reaction now. :) I'll be removing it moving forward...
... Hi, Atif Nice to hear! I was wondering about that comment as well. Do you see an important distinction between a JSON-RPC "Service" and JSON-RPC as a...
Jim, Perhaps I'm missing a finer point here. If a json-rpc service is listening, it either gathers what it needs from the environmental variables just like the...
Matt
WickedLogic@...
Sep 14, 2006 2:18 pm
329
I agree with what you are saying, because ideally the service should function with all of it's requirements being fulfilled in the request channel and not rely...
Matt
WickedLogic@...
Sep 14, 2006 2:33 pm
330
... That would be nice, except, From 6.3.1 in the WD, "The client MUST NOT use the JSON syntax for the parameter values." and "It is specifically not possible...
On 9/14/06, Jim Washington <jwashin@...> wrote: That would be nice, except, ... Would someone be willing to go over the reasons behind not being able to...
Matt
WickedLogic@...
Sep 14, 2006 3:51 pm
332
Hi, it seems the current WD does not allow the specification of the structure of complex data types as parameter or result, basically because there is no way...
Andreas Hofmeister
andi@...
Sep 15, 2006 3:44 pm
333
Hi Balazs, It's really a hard problem to solve entirely through Content-Type anyhow. The spec suggests "application/json" because that is the best-fit...
Hi Jim, Isn't this the same problem as trying to disambiguate a JSON-RPC request from other types so that it could be handled off the same end-point? If so,...
... Hi, Atif Yes. Thanks for the response. ... Yes. That is what we are aiming for as well. As I read the WD a bit closer, I am now understanding it to say...
... Am I to understand from that, that a server implementation would not necessarily be nonconforming if it did not go to great lengths to support JSON-RPC GET...
... Would someone be willing to go over the reasons behind not being able to pass a json-rpc request in json syntax as a `request object`? It seems somewhat...
Andread, I did consider this, but unfortunately, it seemed that baking anything short of an entire schema language for JSON would be a kludge and a problem...
Atif, Thank you for the reply. I completely understand your points, it just caught me off guard that the option to pass a raw json-rpc request object was...
Matt
WickedLogic@...
Sep 19, 2006 2:23 pm
340
Hi Matt, ... Even if it seems simple from client/server implementation perspective, the problem is that it's not very GET-proper. Moreover, if the spec allows...
Atif, Thank you for the reply. Seemingly as I am the only one concerned I'll drop the topic. ; ) (Seems like the sort of thing best discussed over a cup of...
Matt
WickedLogic@...
Sep 22, 2006 11:41 am
343
I am currently implementing JSON-RPC server/client and I have decided not to follow some of the requirements from the specification. I am one of the developers...
... Why exactly would this be difficult? ... Would you mind pointing out the reference in the working draft? (On a separate note, in October I'll be at...