-
Notifications
You must be signed in to change notification settings - Fork 31
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Support for calls with no parameters #170
Conversation
This actually looks fine. This "empty params" feels like a corner case of the spec to be fair, but so it is Can you declare an "empty/void" call in |
I agree. I had a very quick look last night to try to see what the spec said, but I don't think it considers this case. I guess the Python implementations used on servers define the spec 🤷. |
I've added those tests. They uncovered a bug in |
There are a few more fixes needed. If you run
You should be able to reproduce the issue. |
Done! I've also added a makefile target to build/test the examples. |
Extended example: https://github.com/tbrk/neos-server-ocaml-client |
One last thing, sorry. Can you update the changelog? |
Thanks |
…c, rpc and ppx_deriving_rpc (8.1.2) CHANGES: * Add the `noargs` constructor for declaring interfaces that do not take any parameters. (tbrk mirage/ocaml-rpc#170) * Allow Xmlrpc callers to override the base64 decoding function. (tbrk mirage/ocaml-rpc#171)
Somehow this breaks the tests interfaces for xapi-idl. Was this meant to be a non-breaking change? example:
|
Yes, this was meant as non-breaking (I would have waited and opted for a future 9 release othwerise). I hoped the tests would pick up any breakage. |
Adding methods to signature is going to be a breaking change for code that reimplements interfaces for testing purposes, I think a similar case happened with the calls not needing a response. In any case I'll change the code when I upgrade, I just wanted to make you aware of this ;) |
It is a good point. I did not think carefully enough about that. Sorry again for the disruption and thanks for the notification |
Sorry for the hassle @psafont. I think @mseri was just trying to help me out with a problem I had using this library to call methods on a Python server. I wonder why the opam-repository regression testing didn't detect a problem with xapi-idl as a reverse dependency. (I looked back to the pull request but CI log is no longer available.) |
xapi-idl is not on opam-repository and instead it's in a dedicated opam repository, so it's normal it got missed. Now that most of the xapi code is in the same repository it should be possible to easily upload it to the main opam-repository. I don't think it would be useful for any other reason but to check for breaking dependencies,and it would add a significant load to the CI, and add burden to the repository maintainers. I prefer the occasional hassle, it part of my job anyway :) |
OK. Thanks for the explanation! |
Proposition for solving #169.