Immutable artifact markers

[xFrednet]

This PR adds markers for the artifact, the output roughly looks like this:

@mjp41 Are there any interesting implementation details I forgot?

Tests

• The collection of tests for immutability:
  • ./snapshots/cpython-patched/Lib/test/test_freeze/README.md Line 4

Implementation

• The definition of the Py_CHECKWRITE macro:
  • ./snapshots/cpython-patched/Include/refcount.h Line 145
• The atomic RC branch for immutable objects in Py_INCREF:
  • ./snapshots/cpython-patched/Include/refcount.h Line 392
• The atomic RC branch for immutable objects in Py_DECREF:
  • ./snapshots/cpython-patched/Include/refcount.h Line 564
• The implementation of the InterpreterLocal type:
  • ./snapshots/cpython-patched/Modules/_immutablemodule.c Line 143
• The implementation of the SharedField type:
  • ./snapshots/cpython-patched/Modules/_immutablemodule.c Line 318
• The pre-freeze hook of function objects:
  • ./snapshots/cpython-patched/Objects/funcobject.c Line 1208
• The pre-freeze hook of module objects:
  • ./snapshots/cpython-patched/Objects/moduleobject.c Line 1559
• Explanation how weak references work for immutable objects:
  • ./snapshots/cpython-patched/Objects/weakrefobject.c Line 36
• The branch that allows direct sharing for immutable object across sub-interpreters:
  • ./snapshots/cpython-patched/Python/crossinterp.c Line 491
• Explanation how a stack is used to implement the DFS based SCC algorithm:
  • ./snapshots/cpython-patched/Python/immutability.c Line 248
• The state used to track a single freeze call and construct SCCs:
  • ./snapshots/cpython-patched/Python/immutability.c Line 294
• The function that rolls back immutability on failure:
  • ./snapshots/cpython-patched/Python/immutability.c Line 1979
• The entry point to the freeze() function in C:
  • ./snapshots/cpython-patched/Python/immutability.c Line 2202

[mjp41]

This looks like a great set. I guess we could split up the freeze implementation a bit, but it is a tricky piece of code, and I am not sure these sign posts would be sufficient.

[xFrednet]

I find the code pretty readable, and therefore the markers sufficient, but this is very likely due to too much kool-aid