Potential Issue with moduleCache Key Handling in _utils.py

[thisisandy]

While reviewing the moduleCache handling logic in _utils.py, I noticed a potential inconsistency in how keys are being checked and used. Specifically, the code appears to check for the presence of "kwargs", "name", and "args" as string keys in the moduleCache dictionary. However, it seems the variables themselves (e.g., kwargs, name, args) are being used elsewhere as keys.

For example, in the code snippet:

if "name" not in moduleCache:
    moduleCache[name] = {}
if "args" not in moduleCache[name]:
    moduleCache[name][args] = {}
if "kwargs" not in moduleCache[name][args]:
    moduleCache[name][args][kwargs_tuple] = {}
moduleCache[name][args][kwargs_tuple] = mod

The checks for "name", "args", and "kwargs" are strings, but the subsequent access and assignment use the variables name, args, and kwargs_tuple as keys.

Potential Issues

1.	Key Mismatch: If the intention was to check the presence of the variables name, args, and kwargs_tuple as keys, using string literals like "name" might lead to incorrect behavior.
2.	Redundant Checks: If these strings ("name", "args", "kwargs") are unrelated to the actual variables name, args, and kwargs_tuple, this could result in redundant checks or unexpected behavior.

Steps to Reproduce

1.	Use the relevant code path that triggers the moduleCache logic.
2.	Populate the moduleCache with keys that demonstrate the mismatch (e.g., strings like "name" vs. variable name).

Expected Behavior

The moduleCache should consistently use either string literals or variables as keys.

Proposed Solution

1.	If the intention is to use the variables name, args, and kwargs_tuple as keys, update the conditional checks to remove string literals:

if name not in moduleCache:
    moduleCache[name] = {}
if args not in moduleCache[name]:
    moduleCache[name][args] = {}
if kwargs_tuple not in moduleCache[name][args]:
    moduleCache[name][args][kwargs_tuple] = {}

2.	If the current logic with string literals is correct, clarify the relationship between the string keys and the variable keys to avoid confusion.

This potential issue might lead to unexpected results, especially in edge cases where the string and variable keys overlap or differ.

[jhgit]

Is this still a problem in https://github.com/RDFLib/html5rdf? This may be a replacement for html5lib (which does not appear to be getting maintenance).