Lalit Kumar Bhasin

Option 1 looks good to me, as there could be an inevitable scenario where users would have taken a dependency on nostd::type_traits.

Make sense, but `dynamic_cast` could only be used when RTTI is enabled. Should we keep both the options, something like this (not tested): ```cpp void CleanupTracer() { opentelemetry::nostd::shared_ptr provider =...

> Calling `GetTracerProvider()`, which returns an API level object, and somehow getting at the SDK provider from it, is by definition wrong. > > > > A possible change is...

> Does anyone know how to get the text of the CLA I'm being asked to agree to before I start the process of agreeing to it? https://github.com/cncf/cla

Sorry, ignore the link I shared. But I believe there should be a `pdf` link to download the agreement text before you sign, while selecting the option for individual or...

yeah, the MSVC compiler historically doesn't update the `_cplusplus` macros. Probably existing usage in code fails to catch this error.

> Anyone want to take a guess as to what the correct fix is? Probably the runtime environment is not correct for this [test ](https://github.com/open-telemetry/opentelemetry-cpp/blob/0cd3e88bbb9b927a4e597a824cce61efa1b2e31e/.github/workflows/ci.yml#L310) - the default gcc compiler...

If I remember correctly, Async export was put under feature flag as it is not specs compliant. It calls Export::export simultaneously without waiting for previous call to end.

This was a feature provided for those looking for high throughput at the expense of deviating from the specifications. I don't think the plan was ever to make this compliant.

Thats a good idea. Though won’t be easy as I think this preview flag is both in batch processor and exporter.