David Ackroyd

The stats handler approach also allows cases to be captured that the interceptors do not, like a failure to decode the message. We ran into that case with OpenTracing using...

Managing to _mostly_ work around this by applying a similar configuration to @arunk-s, but disabling the relevant linting rules: ``` protoc: version: 3.6.1 # could do with an update here......

Fair point, I can't say I'd really considered it prior to now, as we've been using these middlewares and our monitoring is built on OpenTracing, Prometheus and Zap. With OpenTracing...

Got further with this by running a custom build of the `oci-build-task`, where its `--output` is set to `"type=oci,dest="+imagePath`. The image builds successfully with `buildctl`, however the final steps of...

I believe this is resolved by https://github.com/concourse/oci-build-task/pull/93 and https://github.com/concourse/registry-image-resource/pull/321. The later appears to have made it into `registry-image-resource` release v1.6.0: https://github.com/concourse/registry-image-resource/compare/v1.5.1...v1.6.0, while `oci-build-task` has not yet cut a release since...

This is tripping up our code formatting, as we've moved that to GolangCI Lint (running `gofmt` and `goimports`) rather than running those tools using locally installed versions. This is to...

In the setup I'm working with, we're executing GolangCI Lint in a docker image, where the code is mounted into the container. We're finding that linting has become cripplingly slow:...

Nearly a year on, and life has been busy. Perhaps we can pick this back up soon @pavelnikolov ?

@jmoiron any reason not to move ahead with this one? Similar PRs have been raised before without any feedback. Not having this in places means that the transaction type can't...

@slimsag would like to see this one go in too. I can see this failed a DCO check, but I’m not sure the current PR builds are checking for this?