protobuf-net / protobuf-net

* Add "--grpc list" and "--grpc get {service}" to protogen
(not yet working in codegen tier)

* wow, it almost works now

* better debugging locally

* experimental support for multiple namespaces; is "opt in" - need to use new SchemaGenerationOptions API with SchemaGenerationFlags,MultipleNamespaceSupport

xref #647

* first stabs at service proto emit

* - make SchemaGenerationOptions simpler / mutable
- make Types an input on SchemaGenerationOptions; remove <T>/Type overloads that take SchemaGenerationOptions

Use correct branch

* Adds IVT attribute to access internal reflection bits in gRPC reflection.

* Let Method and ServiceDescriptor implement IType.

Co-authored-by: Marc Gravell <marc.gravell@gmail.com>

…ample usage (#673)

* Add protobuf-net.AspNetCore with input/output formatters, and show example usage

* nit, unnecessary using declaration

* tweak build script

* add comment on buffering

* tyop

* make ParseIfComplete more obvious

* trying to fix CI

* - remove some tabs and ConfigureAwait (@davidfowl feedback)
- remove all the wwwroot things we aren't using

* RegisterProtobufFormatters() => AddProtoBufNet() - mimic code from MvcNewtonsoftJsonOptionsExtensions etc

* supported chunked usage and file-buffering of over-large inputs

* drop a local since it will rarely be reused

* remove TryRead from the slow path

* - increase in-memory threshold to 256k
- use DI more consistently
- fix namespaces

* remove length from WriteResponseBodyAsync

* respect input/output buffer suppression (note: for input this only applies *after* the memory-buffer threshold)

* release notes

* Update to .NETCOREAPP3.1

* Add .NETCOREAPP3.1 platform target

* Update SDK

* Fix IFDEF

* Remove unused package references

* Downgrade SDK to latest on CI

* Add netcoreapp3.1 target to MessagePipes

- remove some redundant explicit interface implementations

… some members that never should have been public

* investigate protobuf-net/protobuf-net.Grpc#100

- when considering a pre-measured object without sub-type awareness, we need to think in terms of the base type
- add new IObjectSerializer<T> which expresses this; without that: no measurements are trusted
- add test for the specific scenario raised in GRPC 100
- implement the new API is the 4 runtime wrappers
- implement the new API in generated types
- TODO: investigate PEVerify warnings

* fix emit Type (we were returning the token, not the Type)

azure deploy, protoc updates, and due credit to @RemiBou