fix segfault when thread is finalised without golang's runtime gc as …

[niedbalski]

Description

• Enforces a runtime garbage collector when the context is done, to ensure that
  no object remains without a finaliser before the thread gets shutdown by fluent-bit.
• Also, switching Message.Record type to any.

Fix

Without this fix, when cancelling the context and the channel has been written to:

[2022/05/12 09:42:13] [debug] [out flush] cb_destroy coro_id=0
[2022/05/12 09:42:13] [debug] [task] destroy task=0x4033ff63f0 (task_id=0)
^C[2022/05/12 09:42:14] [engine] caught signal (SIGINT)
[2022/05/12 09:42:14] [ warn] [engine] service will shutdown in max 5 seconds
[2022/05/12 09:42:14] [ info] [engine] service has stopped (0 pending tasks)
[2022/05/12 09:42:14] [debug] [GO] running exit callback
2022/05/12 09:42:14 calling FLBPluginExit(): name="go-s3-replay-plugin"
[2022/05/12 09:42:14] [engine] caught signal (SIGSEGV)
qemu: uncaught target signal 6 (Aborted) - core dumped

with this fix, same operation, no segfault.

[2022/05/12 09:43:43] [ warn] [engine] service will shutdown in max 5 seconds
[2022/05/12 09:43:43] [ info] [engine] service has stopped (0 pending tasks)
[2022/05/12 09:43:43] [debug] [GO] running exit callback
2022/05/12 09:43:43 calling FLBPluginExit(): name="go-s3-replay-plugin"
[2022/05/12 09:43:43] [ info] [output:stdout:test-output] thread worker #0 stopping...
[2022/05/12 09:43:43] [ info] [output:stdout:test-output] thread worker #0 stopped

Signed-off-by: Jorge Niedbalski [email protected]

[nicolasparada]

Thanks.