Fix Esp32 issue with Sming tasks not getting handled at startup

[mikee47]

Issue #2971 revealed a problem with the task queue where the onReady callback never gets called.
The callback gets queued but the event loop is created afterwards and lacked a 'kick' to the tcpip thread to start message processing.

This bug has been present since PR #2913.

[what-the-diff]

PR Summary

• Introduced New Function for Message Handling
  A new function named tcpip_kick_thread() has been added. This contains logic that was present in system_os_post() as it is now also called from start_sming_task_loop().

• Enhanced Documentation
  The documentation has been updated with comments for newly introduced functions, these include: tcpip_kick_thread() and tcpip_message_handler(). This addition makes the purpose of these functions more understandable to others.

• Incorporation of tcpip_kick_thread() in Message Queuing
  The start_sming_task_loop() function now uses tcpip_kick_thread(). This means that when the task loop begins, it will also handle any messages that are in the queue.

[mikee47]

@slaff The CI failures in jerryscript really need it updating to a more recent version. I'm prepping v3 for that, will have a PR for you shortly.