Moved to libadalang 23 + followed changes

[pjljvandelaar]

Since libadalang 23, the Ada_Type_Decl is no longer an enumeration value of Ada_Node_Kind_Type, but an enumeration range thereof:

  subtype Ada_Type_Decl is Ada_Node_Kind_Type
        range Ada_Anonymous_Type_Decl .. Ada_Formal_Type_Decl;

This enumeration range includes the following 4 values:

Ada_Anonymous_Type_Decl, Ada_Synth_Anonymous_Type_Decl, Ada_Concrete_Type_Decl, and Ada_Formal_Type_Decl

This change of libadalang results in errors like

extraction-primitive_subps.adb:14:35: error: invalid use of subtype mark in expression or call

This pull request fixes these issues.

[pjljvandelaar]

Solves #5

Still to do - fix "example" - needs rejuvenation-ada 2023 for that!

[jeroenk]

This looks sensible.

After removing the rejuvenation dependency from the tests, and removing all the code from main.adb in the test. I do see errors like the following when extracting a database:

-- ...Dependency_Graph_Extractor-Ada/tests/syntax_examples/src/operator_subprogram_calls.adb --
Encountered Libadalang problem: Property Name.relative_name not implemented on type SyntheticIdentifier
Str[21:71-21:74]: "<"
Encountered Libadalang problem: Property Name.relative_name not implemented on type SyntheticIdentifier
Str[39:71-39:75]: "<="

Is that a regression in libadalang 23, or is that a problem with extractor?

[pjljvandelaar]

Is that a regression in libadalang 23, or is that a problem with extractor?

Good question: to be investigated (when time permits ;-( )

[jeroenk]

I forgot the details, but you might remember: can we assume that if a class (say the Name class) has a certain property (P_...) that the property is also implemented for all its subclasses (SyntheticIdentifier in this case)?