Closed GoogleCodeExporter closed 8 years ago
I'm not sure if i like this patch\idea. I understand the requirement - but in
the case that the number (and type) of extra params are known ahead of time it
would be just as easy to overload the given method with those extra params.
Can you point me in the direction of the other implementations where this is
the default behavior? I'm sure you can understand my reluctance to add this
patch without further investigation. Also, sorry for the major delay in
picking this up :)
Original comment by brian.di...@gmail.com
on 20 Sep 2011 at 4:12
Original comment by brian.di...@gmail.com
on 20 Sep 2011 at 4:12
Hi, sorry for late reply, we've been busy with other projects lately.
The rationale for this switch is that we have a third party system which sends
requests to us which according to the spec can add new parameters at any time
in the future. Our end should keep responding to these requests as it did
before the extra parameters were added.
When I wrote this patch I had a look at how two of the Python implementations
found at one of the spec sites (jsonrpc.org? can't find the page now) handles
this. We could check with the JSON-RPC Google group to get some other opinions
on this.
Original comment by hansj...@gmail.com
on 17 Nov 2011 at 4:13
What I don't like about this is the possibility of calling a method that you
didn't intend on calling. ie: you specifically wanted the v2 version of
doSomethingImportant that has 3 parameters but since the API you're actually
calling only supports v1 with 2 parameters. That said, the ability to turn it
on\off makes it a little more acceptable. What do other people think?
Original comment by brian.di...@gmail.com
on 18 Nov 2011 at 3:46
Ok, i applied the patch and committed it. The fact that it's optional and
defaults to false makes me feel better about it.
Thanks
Original comment by brian.di...@gmail.com
on 18 Nov 2011 at 8:03
Great, thanks. I've verified the patches in version 0.18 (both this
and issue 9).
Original comment by hansj...@gmail.com
on 21 Nov 2011 at 1:20
thanks
Original comment by brian.di...@gmail.com
on 21 Nov 2011 at 11:15
Original issue reported on code.google.com by
hansj...@gmail.com
on 7 Jun 2011 at 12:58Attachments: