Jason Lunn
Jason Lunn
@DamyanBG - can you provide a test or a series of steps we can reproduce that works on Python 3.12 but fails on Python 3.13?
@lekoala - the CLA must be completed before this can be reviewed.
@dorner - I remembered that there was a [discussion](https://github.com/protocolbuffers/protobuf/pull/15817#pullrequestreview-1879136663) about the shadowing problem in #15817 that introduced `ServiceDescriptor` support in Ruby in the first place. The result was that the...
> lol... I had been going back to the DescriptorPool to get the ServiceDescriptor, but I changed it and forgot I had. 🤦 I got it working now. > >...
Looking into this now, ETA for a fix unknown. A workaround should be to explicitly `require 'google/protobuf/well_known_types'`
@jez - could you rebase this? I doubt your changes are the reason C++ tests failed
The Ruby test failures look real - `ruby/ext/google/protobuf_c/message.c` only impacts the CRuby implementation when used without FFI. FFI on CRuby and JRuby both get their `Message[]` implementation from [`message.rb`](https://github.com/protocolbuffers/protobuf/blob/c49d529607650fdc97a59e48f36ba0d1bbf292af/ruby/lib/google/protobuf/ffi/message.rb#L162), and...
> @JasonLunn I've added alternative implementations in the two files you mentioned. > > I tested that the FFI version in `message.rb` fixed the test failure. > > I was...
I took a look at this yesterday, and I find very few test expectations for `resolve`, which leads me to believe that the current behavior of raising an exception for...