1 Fix the attribute names in the report.
2
3 Consider type deduction and its consequences where types belong to the same hierarchy
4 and where a guard could be generated for the most general type.
5
6 Consider attribute merging where many attributes are just aliases for the same underlying
7 definition.
8
9 Consider merging the InspectedModule.store tests with the scope conflict handling.
10
11 Consider permitting multiple class alternatives where the attributes are all identical.
12
13 Support class attribute positioning similar to instance attribute positioning, potentially
14 (for both) based on usage observations. For example, if __iter__ is used on two classes,
15 the class attribute could be exposed at a similar relative position to the class (and
16 potentially accessible using a LoadAttr-style instruction).
17
18 Consider references to defaults as occurring only within the context of a particular
19 function, thus eliminating default value classes if such functions are not themselves
20 invoked.
21
22 Consider labelling _scope on assignments and dealing with the assignment of removed
23 attributes, possibly removing the entire assignment, and distinguishing between such cases
24 and unknown names.
25
26 Check name origin where multiple branches could yield multiple scope interpretations:
27
28 ----
29 try:
30 set # built-in name
31 except NameError:
32 from sets import Set as set # local definition of name
33
34 set # could be confused by the local definition at run-time
35 ----
36
37 Support __init__ traversal (and other implicit names) more effectively.
38
39 Check context_value initialisation (avoiding or handling None effectively).
40
41 __getitem__ could be written in Python, using a native method only to access fragments.
42
43 Support slicing.
44
45 Consider better "macro" support where new expressions need to be generated and processed.
46
47 **** Constant attribute users need not maintain usage since they are already resolved. ****
48
49 Loop entry points should capture usage to update later assignments in the loop.
50 The continue and break statements should affect usage propagation.
51
52 Consider handling CallFunc in micropython.inspect in order to produce instances of specific classes.
53 Then, consider adding support for guard removal/verification where known instances are involved.
54 Consider handling branches of values within namespaces in order to support more precise value usage.