Piotr Sikora
Piotr Sikora
@johnlanni thank you for digging into it! I agree that under no circumstances Wasm should be able to crash the host, so this is definitely a serious issues, but I...
@phlax Is this finished? I thought there were a few more things that you wanted to add here, but I don't recall details.
The original ABI had a function exactly like that, and [Proxy-Wasm C++ SDK](https://github.com/proxy-wasm/proxy-wasm-cpp-sdk) still has it (see: [proxy_validate_configuration](https://github.com/proxy-wasm/proxy-wasm-cpp-sdk/blob/95bb82ce45c41d9100fd1ec15d2ffc67f7f3ceee/proxy_wasm_externs.h#L161), [RootContext::validateConfiguration](https://github.com/proxy-wasm/proxy-wasm-cpp-sdk/blob/95bb82ce45c41d9100fd1ec15d2ffc67f7f3ceee/proxy_wasm_api.h#L328)), but I'm not aware of any plugins using it. Note there...
> Another consideration here is that we access to those content types that > are being inspected and only buffer when inspection. Could you elaborate? I don't understand how this...
You can add trailers from the request/response body callbacks (e.g. upon receiving `end_of_stream=True`, which implies that there are no trailers).
It's more likely that you'll get answer to that question in the Envoy issue tracker than here, since it doesn't sound like this is specific to the Rust SDK. cc...
> There is no way for ProxyWasm to 'consume and replace' the header/body data. The whole ABI is about modifying streamed headers, body data, and trailers. > Did I represent...
> is there any issue here ? It doesn't pass tests (see Clippy failures). Also, I believe that @mpwarres and @martijneken are currently working on the documentation for the whole...
The issue is that this SDK originally defined `get_http_call_response_headers()` as returning `Vec`, and `String` is always UTF-8 in Rust, but HTTP header values are not. You should be using `get_http_call_response_headers_bytes()`...
> @PiotrSikora I can also add the examples from the mentioned repo to the `examples/` directory... I took a look at them, and they seem to be mostly duplicates of...