Steven Fackler

Router should support request body types other than hyper::Body

7

The generated server Service implementations already support arbitrary request body types, but Router is currently hardcoded to work with hyper::Body directly. It'd be great if that were relaxed to work...

Support a #[serde(validate = "some_function")] attribute on fields

19

I occasionally run into fields in a serde-deserializable type that require some extra validation after deserialization, for example "a u32 >= 2". The way to do this now is to...

Avoid lossy buffering in #[serde(flatten)]

13

The `Deserialize` implementation of a type with a field annotated `#[serde(flatten)]` currently buffers up all of the unknown map entries in memory, and then deserializes the flattened value from them...

prost-build tries to install libprotobuf-lite.a to /usr/local?

3

On a CI setup that presumably doesn't have protobuf stuff installed, I see this build failure: stdout: ``` cargo:rerun-if-changed=/usr/local/cargo/registry/src/github.com-1ecc6299db9ec823/prost-build-0.10.4/third-party/protobuf/cmake CMAKE_TOOLCHAIN_FILE_x86_64-unknown-linux-gnu = None CMAKE_TOOLCHAIN_FILE_x86_64_unknown_linux_gnu = None HOST_CMAKE_TOOLCHAIN_FILE = None CMAKE_TOOLCHAIN_FILE =...

A more ergonomic enumeration representation

28

To deal with the "openness" of Protobuf enumeration types, Prost currently represents them in the message struct as `i32` and generates a Rust enum that can convert to and from...

Connection::graceful_shutdown always waits for the first request.

3

**Version** hyper 0.14.16 **Platform** Linux DESKTOP-DHO88R7 4.19.104-microsoft-standard #1 SMP Wed Feb 19 06:37:35 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux **Description** If you gracefully shut down a server connection future before...

It is the replacement for the deprecated `SecTrustGetCertificateAtIndex` function.

``` thread 'secure_transport::test::cipher_configuration' panicked at 'assertion failed: `(left == right)` (left: `[TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384, TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256, TLS_ECDHE_ECDSA_WITH_AES_256_CBC_SHA384, TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA256, TLS_ECDHE_ECDSA_WITH_AES_256_CBC_SHA, TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA, TLS_ECDHE_ECDSA_WITH_3DES_EDE_CBC_SHA, TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384, TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256, TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA384, TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256, TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA, TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA, TLS_ECDHE_RSA_WITH_3DES_EDE_CBC_SHA, TLS_DHE_RSA_WITH_AES_256_GCM_SHA384, TLS_DHE_RSA_WITH_AES_128_GCM_SHA256]`, right: `[TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384, TLS_NULL_WITH_NULL_NULL,...

GCC doesn't yet have support for aarch64 Darwin, but binutils appears to compile just fine.

Floats are encodedwith sign extension while doubles are encoded without

4

I'm working on a Smile implementation in Rust, and am trying to exactly match Jackson's behavior to be able to test against it. While testing floats and doubles, I've noticed...