Bug 2074412 - Update to ringbuf 0.5.2

[badboy]

Mechanical translation to the newer trait methods.

---

Similar to mozilla/cubeb-pulse-rs#115
I don't know if this is all actually correct.
Tests look unchanged

I have two tests not working on the default branch:

backend::tests::api::test_get_channel_count_of_input_for_a_output_only_deivce
backend::tests::api::test_set_buffer_size_sync

they still don't work with this patch.

[badboy]

thread '<unnamed>' (34480) panicked at src/backend/mod.rs:3264:13:
can't push to ringbuf
stack backtrace:
   0: std::panicking::begin_panic::<&str>
             at /rustc/48a229ceaefd4985c50990b14116b6d856af0985/library/std/src/panicking.rs:751:5
   1: <cubeb_coreaudio::backend::InputCallbackLogger>::push
             at ./src/backend/mod.rs:3264:13
   2: cubeb_coreaudio::backend::audiounit_input_callback::{closure#0}
             at ./src/backend/mod.rs:656:18
   3: cubeb_coreaudio::backend::audiounit_input_callback
             at ./src/backend/mod.rs:706:9
   4: _AUGenericOutputFactory
   5: <unknown>
   6: <unknown>
   7: <unknown>
   8: __pthread_joiner_wake
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.

well, I guess I was wrong.

[kinetiknz]

Thanks!

[kinetiknz]

push was fallible in ringbuf 0.2, so switching to try_push here without panicking is sufficient.

[larseggert]

0.5.2 had some soundness issues that were fixed in 0.5.3. I don't think this crate or Gecko were affected, but better to update. (We can then also drop the cargo audit suppression for Gecko.)