We need a per-task copy of decimal.Context() so that mutations are isolated between asyncio tasks and threads using
sys.flags.thread_inherit_context. This is done by adding a "depth" counter to PyContext and copying the decimal.Context() instance whenever it doesn't match the depth of the current PyContext.

This is an alternative to GH-146482. It has the following differences:

• it avoids the potential problem of keeping context vars alive longer, via the ctx_vars_origin reference. Instead, we are just using a counter here to track depth.
• it avoids copying decimal.Context in the case that the context is reused many times, e.g. for asyncio tasks
• it creates a bit more work for the _pydecimal and _decimal.c code since it has to track the depth on it's Context instance

• Issue: data races in decimal module with global context #141148