Create a LocalScope for CollectionComprehensions and generate a Declaration

[KuechA]

Fixes #2005 by

• Entering a LocalScope for list, set, dict comprehensions and generators. This somewhat simulates PEP 709 where the loop variable is available only inside the loop.
• Generating a VariableDeclaration inside this scope in the PythonAddDeclarationsPass for the variables.
• Adds test that the local variables are not accessible outside the comprehension and that the refersTo edges are set
• For AssignExpressions with the := operator inside a list/set/dict comprehension, the variable is looked up or created outside of the comprehension's LocalScope to address PEP 572

[codecov]

Codecov Report

Attention: Patch coverage is 90.00000% with 4 lines in your changes missing coverage. Please review.

Project coverage is 78.29%. Comparing base (541dea1) to head (1ebbc6f).

Files with missing lines	Patch %	Lines
.../main/kotlin/de/fraunhofer/aisec/cpg/graph/Node.kt	60.00%	0 Missing and 2 partials ⚠️
...s/kotlin/de/fraunhofer/aisec/cpg/test/TestUtils.kt	50.00%	2 Missing ⚠️

Additional details and impacted files

Files with missing lines	Coverage Δ
...ain/kotlin/de/fraunhofer/aisec/cpg/ScopeManager.kt	82.70% <100.00%> (+0.05%)	⬆️
...otlin/de/fraunhofer/aisec/cpg/graph/TypeBuilder.kt	56.41% <100.00%> (+2.56%)	⬆️
...er/aisec/cpg/frontends/python/ExpressionHandler.kt	83.04% <100.00%> (ø)
...ofer/aisec/cpg/passes/PythonAddDeclarationsPass.kt	92.78% <100.00%> (+2.11%)	⬆️
.../main/kotlin/de/fraunhofer/aisec/cpg/graph/Node.kt	86.17% <60.00%> (-1.48%)	⬇️
...s/kotlin/de/fraunhofer/aisec/cpg/test/TestUtils.kt	71.27% <50.00%> (-1.55%)	⬇️

... and 5 files with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[maximiliankaul]

Looks good in general. However, there is still an issue with AssignExpressions leaking to the containing scope. See the added tests. I'm in favor of this PR once these issues are resolved.

[KuechA]

We should discuss what we expect in this case. My version of Python3.13 crashes with the error "UnboundLocalError: cannot access local variable 'a' where it is not associated with a value".

We may interpret this by setting the refersTo to same as the reference a in the second item and later, an analysis must figure out that this variable is not initialized yet (we should check the EOG for this too).

We may also interpret this that refersTo should be null

[oxisto]

I would say that refersTo should be null. I think we did that in other cases where we have an unbound error as well

[KuechA]

Ok, since the SymbolResolver is in charge of this, I added issue #2035 which blocks this PR. I also added the asserts again.