Michael Ilyin

Am I correct that this feature is required to write: ```Python with zenoh.open(...) as session: session.declare_subscriber(...) ``` and be sure that subscriber works inside the whole `with` block, not dropped...

> This feature is not "required", as I can find a way to make it work in my implementation. However, it is required to have a semantic consistency between the...

> I think we can go ahead with this feature under an `unstable` feature gate. Similar approach should be applied to `declare_queryable`. Agree. @wyfo please implement it on Rust for...

Initial problem fixed in draft pr above. But there are multiple problems: - test doesn't fail if no connection to zenoh router, so it's just skipped. That's why this problem...

replier_id is declared unstable now, so this issue is not a priority. Source info is unstable too https://github.com/eclipse-zenoh/zenoh/issues/709

Let me summarize the problem with Influx_db: The concept of "Timestamp" is that it's some unqiue and ordered identifier of sample. It's uniqueness is guaranteed by the fact, that at...

Currenly we could safely make the standalone `new_timestamp()` function internal


I think we have two separate problems here - how to improve our logging to make it convenient and uniform for all bindings - how log initialization API should look...

The approach looks good for me. The naming `init_logging` I think is definitely better than `try_init_log_from_env`. Usage of `test-log` crate seems to be right thing too