Dragos Carp

Hi, What is the status of your work? Is it ready to be merged in CMake? Please send me a pointer to the branch you are working on. My problem...

As far as I know he wanted to wait for dmd .dll support on windows and just afterwards do the PR/merge. But this was 3 years ago :(.

Thank you for the contribution! Could you add some unit-tests to document and verify the use-case?

At the moment `protoc_gen_d` generates just classes. For small messages, as an optimization, you can replace `class` with `struct` and use value semantics. [Options](https://developers.google.com/protocol-buffers/docs/proto3#options) probably could be used to control...

I think that plugin parameters are passed in the [`parameter` member of `CodeGeneratorRequest`](https://github.com/dcarp/protobuf-d/blob/master/protoc_gen_d/proto/google/protobuf/compiler/plugin.proto#L74).

From what I saw, I don't think that the `options` work with `proto3`. I manually ported `descriptor.proto` to `proto3`. Protobuf documentation describes the option in context of `proto3` using `extend`...

I think that as long as libasync is used, the fix in libasync in necessary (the two functions are called internally by libasync). Actually, it is not necessary to add...

Extending `sockAddrLen()` and `toAddressString()` in AsyncTCPConnection to support UNIX sockeis in libasync is pretty harmless. The API will still not support them. Anyhow, it is great that you implemented the...

Hi Eugene, this was on my plan also, but if you have the resources to add it, please do so. Actually this will be completely awesome! It is much easier...

2) I think that Tuples are good enough: as good as a struct. Writing `ei1!("b", "c")(8, 9);` instead of `ei1.b = 8; ei1.c = 9;` is more ore less a...