Describe the bug
In RDKit 2024.09.1 I am seeing different HetAtomTautomerv2 hashes being generated depending on the order of that atoms and/or bonds in the input. It appears to be related to the recent fix #7502 that shrinks tautomeric zones in the v2 tautomer/protomer hash - I don't see the issue if I revert that change.

To Reproduce

>>> from rdkit import Chem
>>> from rdkit.Chem import rdMolHash

>>> # Same molecule with atoms in different order in input SMILES
>>> mol1 = Chem.MolFromSmiles("CNC(=O)N[C@@H](C)c1ccccc1")
>>> mol2 = Chem.MolFromSmiles("C[C@H](NC(=O)NC)c1ccccc1")
>>> Chem.MolToSmiles(mol1) == Chem.MolToSmiles(mol2)
True
>>> hash1 = rdMolHash.MolHash(mol1, rdMolHash.HashFunction.HetAtomTautomerv2)
>>> hash2 = rdMolHash.MolHash(mol2, rdMolHash.HashFunction.HetAtomTautomerv2)
>>> hash1 == hash2
False
>>> hash1
'[CH3]-[N]:[C](:[O]):[N]:[C](-[CH3])-[c]1:[cH]:[cH]:[cH]:[cH]:[cH]:1_3_0'
>>> hash2
'[C]:[N]:[C](:[O]):[N]-[C@@H](-[CH3])-[c]1:[cH]:[cH]:[cH]:[cH]:[cH]:1_5_0'

Interestingly, using Chem.RenumberAtoms with _smilesAtomOutputOrder on both molecules in this example to get a consistent atom order does not seem to fix this - they still produce different hashes. The bonds are still in a different order so I presume that is the issue. In general I don't think it is straightforward to consistently renumber atoms/bonds for all different tautomers/protomers anyway, so I don't think it is possible to fix this just by renumbering atoms/bonds ahead of hash generation.

Expected behavior
Expect atom order not to affect hashes, i.e. hash1 == hash2 in the above example.

Configuration (please complete the following information):

• RDKit version: 2024.09.1
• Python version (if relevant): 3.10
• Are you using conda? No
• If you are not using conda: how did you install the RDKit? Source compile