1 Dynamic Attribute Access
2 ========================
3
4 Ensure that getattr (or other dynamic attribute usage) causes propagation of coverage to
5 the potential attributes.
6
7 Consider explicit accessor initialisation.
8
9 Attribute Usage
10 ===============
11
12 Consider attribute assignment observations, along with the possibility of class attribute
13 assignment.
14
15 Note direct assignments as usual, indirect assignments via the attribute usage
16 mechanism. During attribute collection and inference, add assigned values to all
17 inferred targets.
18
19 Since class attributes can be assigned, StoreAttrIndex would no longer need to reject
20 static attributes, although this might still be necessary where attribute usage analysis
21 has not been performed.
22
23 Potentially consider changing static attribute details to use object-relative offsets in
24 order to simplify the instruction implementations. This might allow us to eliminate the
25 static attribute flag for attributes in the object table, at least at run-time.
26
27 Consider attribute usage observations being suspended inside blocks where AttributeError
28 may be caught (although this doesn't anticipate such exceptions being caught outside a
29 function altogether).
30
31 Consider type deduction and its consequences where types belong to the same hierarchy
32 and where a guard could be generated for the most general type.
33
34 Consider permitting multiple class alternatives where the attributes are all identical.
35
36 Support class attribute positioning similar to instance attribute positioning, potentially
37 (for both) based on usage observations. For example, if __iter__ is used on two classes,
38 the class attribute could be exposed at a similar relative position to the class (and
39 potentially accessible using a LoadAttr-style instruction).
40
41 **** Constant attribute users need not maintain usage since they are already resolved. ****
42
43 Loop entry points should capture usage to update later assignments in the loop.
44 The continue and break statements should affect usage propagation.
45
46 Consider handling CallFunc in micropython.inspect in order to produce instances of specific classes.
47 Then, consider adding support for guard removal/verification where known instances are involved.
48 Consider handling branches of values within namespaces in order to support more precise value usage.
49
50 Frame Optimisations
51 ===================
52
53 Stack frame replacement where a local frame is unused after a call, such as in a tail call
54 situation.
55
56 Local assignment detection plus frame re-use. Example: slice.__init__ calls
57 xrange.__init__ with the same arguments which are unchanged in xrange.__init__. There is
58 therefore no need to build a new frame for this call.
59
60 Function Specialisation
61 =======================
62
63 Specialisation of certain functions, such as isinstance(x, cls) where cls is a known
64 constant.
65
66 Structure and Object Table Optimisations
67 ========================================
68
69 Fix object table entries for attributes not provided by any known object, or provide an
70 error, potentially overridden by options. For example, the augmented assignment methods
71 are not supported by the built-in objects and thus the operator module functions cause
72 the compilation to fail. Alternatively, just supply the methods since something has to do
73 so in the builtins.
74
75 Consider attribute merging where many attributes are just aliases for the same underlying
76 definition.
77
78 Consider references to defaults as occurring only within the context of a particular
79 function, thus eliminating default value classes if such functions are not themselves
80 invoked.
81
82 Scope Handling
83 ==============
84
85 Consider merging the InspectedModule.store tests with the scope conflict handling.
86
87 Consider labelling _scope on assignments and dealing with the assignment of removed
88 attributes, possibly removing the entire assignment, and distinguishing between such cases
89 and unknown names.
90
91 Check name origin where multiple branches could yield multiple scope interpretations:
92
93 ----
94 try:
95 set # built-in name
96 except NameError:
97 from sets import Set as set # local definition of name
98
99 set # could be confused by the local definition at run-time
100 ----
101
102 Object Coverage
103 ===============
104
105 Support __init__ traversal (and other implicit names) more effectively.
106
107 Other
108 =====
109
110 Support tuple as a function returning any input tuple uncopied.
111
112 Check context_value initialisation (avoiding or handling None effectively).
113
114 __getitem__ could be written in Python, using a native method only to access fragments.
115
116 Consider better "macro" support where new expressions need to be generated and processed.
117
118 Detect TestIdentity results involving constants, potentially optimising status-affected
119 instructions:
120
121 TestIdentity(x, y) # where x is always y
122 JumpIfFalse(...) # would be removed (never false)
123 JumpIfTrue(...) # changed to Jump(...)
124
125 Status-affected blocks could be optimised away for such constant-related results.
