fix: Avoid assertion when hitting shutdown_timeout

[Swatinem]

This reworks the task locking and background worker shutdown logic.
Tasks are now refcounted so they can be executed concurrently while another thread removes them from the queue. In case the background worker can’t be cleanly shut down (due to slow network), the task queue is being dumped, but the background worker itself is not freed from the main thread, to avoid potential use-after-free bugs.
Also, the background worker itself is refcounted and owns its state.

The transport state is now owned by the background worker, and passed explicitly to the task send function. The task state itself was removed in favor of providing the envelope directly.

[codecov-commenter]

Codecov Report

Merging #323 into master will increase coverage by 0.37%.
The diff coverage is 93.10%.

@@            Coverage Diff             @@
##           master     #323      +/-   ##
==========================================
+ Coverage   87.18%   87.55%   +0.37%     
==========================================
  Files          49       49              
  Lines        3971     3993      +22     
==========================================
+ Hits         3462     3496      +34     
+ Misses        509      497      -12     

Impacted Files	Coverage Δ
src/sentry_sync.h	97.36% <ø> (+0.07%)	⬆️
src/sentry_sync.c	93.33% <91.01%> (+2.89%)	⬆️
src/transports/sentry_transport_curl.c	91.59% <95.74%> (+2.25%)	⬆️
src/sentry_core.c	95.33% <100.00%> (+0.93%)	⬆️
src/sentry_logger.h	100.00% <0.00%> (ø)
src/sentry_utils.c	90.11% <0.00%> (+0.03%)	⬆️
src/sentry_value.c	89.81% <0.00%> (+0.18%)	⬆️
src/sentry_options.c	77.77% <0.00%> (+0.22%)	⬆️
... and 2 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 577c9f4...87ee001. Read the comment docs.

[jan-auer]

Notes from an offline discussion: When the crash happens close to startup, it is more likely that it interrupts an ongoing upload. In this case, it would be better to avoid dropping it.

If we refcount the envelopes in the transport queue, we could have multiple threads read and clean the queue concurrently.

[jan-auer]

This looks good now. I'll to a final run through this and the tests tomorrow, to check if we're lacking coverage for anything vital.

[jan-auer]

This is good to go now. Thank you for taking a deep dive into this.