Gaetano, ... No. A good example for it are the test forms, for example: http://xins.sf.net/demo/specdocs/myproject/MyFunction-testform.html Thing is, we kept...
... Id must be null for notification(or optionally not present in 1.1), Zero(0) is a perfectly valid id for a request. Matching null against not null should...
Been thinking a bit about the spec over the weekend. Point 1: support for notifications: the current spec is extremely vague on the topic, and there are some...
... I don't think so, notifications are realy useful for dispatching events and messages to a peer. << I don't mean to question whether notifications are...
In message #223 <http://groups.yahoo.com/group/json-rpc/message/223> , Gaetano made several suggestions for system methods. ... I think these are a good...
Gaetano, your points and concerns on notifications and [id] are right on track with my observations as well. I think they're a bit ambiguous at the moment and...
I think these are a good starting point. I already implement system.listMethods in Jayrock though I have not been able to take advantage of it as yet. But then...
re-reading the current spec, I think there might be a problem in the http binding: 'To give the serverside peer a chance to send messages to the client side...
I propose the content-type to be "application/json-rpc". The alternative would be "application/jsonrpc". Either would be fine, though my apps are using the...
I may be wrong, but I think the intention that Jan had with the empty POST was to get the effect of a hanging GET to tunnel events. If that's so, why not use...
Using GET instead of POST might be fine, but leaves open the identification question: how can the server correctly identify the client that is asking to be...
... I would only think identification should be done by a session id, never by IP. Just seems like a step backwards without much advantage -- Matthew P. C....
MPCM
WickedLogic@...
May 8, 2006 1:52 pm
235
Of course the second option was only mentioned to scare people a bit... :) Thinking even more about it, I see some problems with session ids inside method...
Or that the session id is a required parameter of the rpc methods being called (in addition to anything else outside the json). That way multiplexing can...
MPCM
WickedLogic@...
May 8, 2006 2:56 pm
237
... I agree with that. I just dropped Douglas a mail to see what he did to get application/json registered. Jan...
... To get application/json-rpc into the IETF tree it looks like an RFC needs to be written for it. I guess I could take the 1.0 spec and put it into RFC...
... PS: btw: my implementation distinguishes between POST and GET, servicing html documentation / debugging pages to get requests. This makes it very friendly...
I agree. I've made a note for the moment of using application/json-rpc for the 1.1 draft. While on the subject, though, if the arguments laid in Appendix A [1]...
A nice way to add session ID to requests / responses with complete backward compat could be to take advantage of the fact that the current ID field can be of...
... I think I have the whole thread here. For the Zopes, we are in a situation where positional and named parameters are both allowed. And very useful. For...
To solve the problem of interface versioning, while adhering to the xmlrpc spec, we have been using an extremely simple solution for a very long time: just...
Hello, let me introduce myself, my name is Balazs Ree and I am also a the co-author of the "jsonserver" json-rpc implementation in Zope. I wanted to make a...
Balazs Ree
ree@...
May 10, 2006 3:12 pm
247
I can see the point in keyword/named arguments, especially if you are doing translation before calling the `real` method (doing security checks and basically...
MPCM
WickedLogic@...
May 10, 2006 3:36 pm
248
... There are some languages there, where a function signature does not define any type for the expected parameters, with PHP being a prime example. In such a...
... I apologize in advance if I misunderstand you, I don't intend to do so. Suppose we have no support for named arguments as it is the situation right now....
Balazs Ree
ree@...
May 10, 2006 4:44 pm
250
Let me recap (i hope I get this all right). I see Atif's proposal (allow "params" to be either object or array) as requiring either: "name all params" (if...
Having problems with message search?
Fill out this form to ensure your group is one of the first to be migrated to the new message search system.
