1 Name usage types: as parameters, as base classes, as callables. This potentially restricts
2 attribute usage effects because names mentioned as base classes are not propagated and
3 made freely available for use in attribute accesses.
4
5 Low-Level Instructions and Macro Instructions
6 =============================================
7
8 Have contexts and values stored separately in memory. This involves eliminating DataValue
9 and storing attributes using two words.
10
11 Migrate macro instructions such as the *Index instructions to library code implemented
12 using low-level instructions.
13
14 Consider introducing classic machine level instructions (word addition, subtraction, and
15 so on) in order to implement all current RSVP instructions.
16
17 Move common code sequences to a library routine, such as the context checking that occurs
18 in functions and methods.
19
20 Dataflow Optimisations
21 ======================
22
23 Assignments, particularly now that no result register exists, may cause StoreTemp/LoadTemp
24 instruction pairs to be produced and these could be eliminated.
25
26 Class and Module Attribute Assignment
27 =====================================
28
29 Allow unrestricted class and module assignment (but not new external binding of
30 attributes), eliminating run-time checks on object types in instructions like
31 StoreAttrIndex. This may involve less specific objects being identified during inspection.
32
33 Potentially re-evaluate class bases in order to see if they are non-constant.
34
35 Verify that the context information is correctly set, particularly for the unoptimised
36 cases.
37
38 Update docs/assignment.txt.
39
40 Prevent assignments within classes, such as method aliasing, from causing the source of an
41 assignment from being automatically generated. Instead, only external references should be
42 registered.
43
44 Prevent "from <module> import ..." statements from registering references to such local
45 aliases such that they cause the source of each alias to be automatically generated.
46
47 Consider attribute assignment observations, along with the possibility of class and module
48 attribute assignment.
49
50 (Note direct assignments as usual, indirect assignments via the attribute usage
51 mechanism. During attribute collection and inference, add assigned values to all
52 inferred targets.)
53
54 (Since class attributes can be assigned, StoreAttrIndex would no longer need to reject
55 static attributes, although this might still be necessary where attribute usage analysis
56 has not been performed.)
57
58 Potentially consider changing static attribute details to use object-relative offsets in
59 order to simplify the instruction implementations. This might allow us to eliminate the
60 static attribute flag for attributes in the object table, at least at run-time.
61
62 Dynamic Attribute Access
63 ========================
64
65 Consider explicit accessor initialisation:
66
67 attr = accessor("attr")
68 getattr(C, attr)
69
70 Attribute Usage
71 ===============
72
73 To consider: is it useful to distinguish between attribute name sets when the same names
74 are mentioned, but where one path through the code sets different values on attributes
75 than another? The _attrtypes collapses observations in order to make a list of object
76 types for a name, and the final set of names leading to such type deductions might be a
77 useful annotation to be added alongside _attrcombined.
78
79 Update the reports to group identical sets of attribute names.
80
81 Interface/Type Generalisation
82 -----------------------------
83
84 Consolidate interface observations by taking all cached table accesses and determining
85 which usage patterns lead to the same types. For example, if full usage of {a, b} and
86 {a, b, c} leads to A and B in both cases, either {a, b} can be considered as partial usage
87 of the complete interface {a, b, c}, or the latter can be considered as an
88 overspecification of the former.
89
90 Consider type deduction and its consequences where types belong to the same hierarchy
91 and where a guard could be generated for the most general type.
92
93 Consider permitting multiple class alternatives where the attributes are all identical.
94
95 Support class attribute positioning similar to instance attribute positioning, potentially
96 (for both) based on usage observations. For example, if __iter__ is used on two classes,
97 the class attribute could be exposed at a similar relative position to the class (and
98 potentially accessible using a LoadAttr-style instruction).
99
100 **** Constant attribute users need not maintain usage since they are already resolved. ****
101
102 Self-related Usage
103 ------------------
104
105 Perform attribute usage on attributes of self as names, potentially combining observations
106 across methods.
107
108 Additional Guards
109 -----------------
110
111 Consider handling branches of values within namespaces in order to support more precise value usage.
112
113 Loop entry points and other places where usage becomes more specific might be used as
114 places to impose guards. See tests/attribute_access_type_restriction_loop_list.py for an
115 example. (Such information is already shown in the reports.)
116
117 Strict Interfaces/Types
118 -----------------------
119
120 Make the gathering of usage parameterisable according to the optimisation level so that a
121 choice can be made between control-flow-dependent observations and the simple collection
122 of all attributes used with a name (producing a more static interface observation).
123
124 AttributeError
125 --------------
126
127 Consider attribute usage observations being suspended or optional inside blocks where
128 AttributeError may be caught (although this doesn't anticipate such exceptions being
129 caught outside a function altogether). For example:
130
131 y = a.y
132 try:
133 z = a.z # z is an optional attribute
134 except AttributeError:
135 z = None
136
137 Instantiation Deduction
138 -----------------------
139
140 Consider handling CallFunc in micropython.inspect in order to produce instances of specific classes.
141 Then, consider adding support for guard removal/verification where known instances are involved. For
142 example:
143
144 l = []
145 l.append(123) # type deductions are filtered using instantiation knowledge
146
147 Frame Optimisations
148 ===================
149
150 Stack frame replacement where a local frame is unused after a call, such as in a tail call
151 situation.
152
153 Local assignment detection plus frame re-use. Example: slice.__init__ calls
154 xrange.__init__ with the same arguments which are unchanged in xrange.__init__. There is
155 therefore no need to build a new frame for this call, although in some cases the locals
156 frame might need expanding.
157
158 Reference tracking where objects associated with names are assigned to attributes of other
159 objects may assist in allocation optimisations. Recording whether an object referenced by
160 a name is assigned to an attribute, propagated to another name and assigned to an
161 attribute, or passed to another function or method might, if such observations were
162 combined, allow frame-based or temporary allocation to occur.
163
164 Instantiation
165 =============
166
167 Specific instances could be produced, providing type information and acting somewhat like
168 classes during inspection.
169
170 Inlining
171 ========
172
173 Where a function or method call can always be determined, the body of the target could be
174 inlined - copied into place - within the caller. If the target is only ever called by a
175 single caller it could be moved into place. This could enhance deductions based on
176 attribute usage since observations from the inlined function would be merged into the
177 caller.
178
179 Function Specialisation
180 =======================
181
182 Specialisation of certain functions, such as isinstance(x, cls) where cls is a known
183 constant.
184
185 Structure and Object Table Optimisations
186 ========================================
187
188 Fix object table entries for attributes not provided by any known object, or provide an
189 error, potentially overridden by options. For example, the augmented assignment methods
190 are not supported by the built-in objects and thus the operator module functions cause
191 the compilation to fail. Alternatively, just supply the methods since something has to do
192 so in the builtins.
193
194 Consider attribute merging where many attributes are just aliases for the same underlying
195 definition.
196
197 Consider references to defaults as occurring only within the context of a particular
198 function, thus eliminating default value classes if such functions are not themselves
199 invoked.
200
201 Scope Handling
202 ==============
203
204 Consider merging the InspectedModule.store tests with the scope conflict handling.
205
206 Consider labelling _scope on assignments and dealing with the assignment of removed
207 attributes, possibly removing the entire assignment, and distinguishing between such cases
208 and unknown names.
209
210 Check name origin where multiple branches could yield multiple scope interpretations:
211
212 try:
213 set # built-in name
214 except NameError:
215 from sets import Set as set # local definition of name
216
217 set # could be confused by the local definition at run-time
218
219 Object Coverage
220 ===============
221
222 Support __init__ traversal (and other implicit names) more effectively.
223
224 Importing Modules
225 =================
226
227 Consider supporting relative imports, even though this is arguably a misfeature.
228
229 Other
230 =====
231
232 Consider a separate annotation phase where deductions are added to the AST for the
233 benefit of both the reporting and code generation phases.
234
235 Check context_value initialisation (avoiding or handling None effectively).
236
237 Consider better "macro" support where new expressions need to be generated and processed.
238
239 Detect TestIdentity results involving constants, potentially optimising status-affected
240 instructions:
241
242 TestIdentity(x, y) # where x is always y
243 JumpIfFalse(...) # would be removed (never false)
244 JumpIfTrue(...) # changed to Jump(...)
245
246 Status-affected blocks could be optimised away for such constant-related results.
