Approach suggested by Irit

python · JelleZijlstra · Apr 28, 2025 · Mar 21, 2025 · Mar 21, 2025 · Mar 21, 2025
commit ee8e8570e756d719ed7fa3ec577fb2919eccd7bc
diff --git a/Include/internal/pycore_instruction_sequence.h b/Include/internal/pycore_instruction_sequence.h
@@ -45,6 +45,9 @@ typedef struct instruction_sequence {
 
     /* PyList of instruction sequences of nested functions */
     PyObject *s_nested;
+
+    /* Code for creating annotations, spliced into the main sequence later */
+    struct instruction_sequence *s_annotations_code;
 } _PyInstructionSequence;
 
 typedef struct {
@@ -66,8 +69,8 @@ _PyJumpTargetLabel _PyInstructionSequence_NewLabel(_PyInstructionSequence *seq);
 int _PyInstructionSequence_ApplyLabelMap(_PyInstructionSequence *seq);
 int _PyInstructionSequence_InsertInstruction(_PyInstructionSequence *seq, int pos,
                                              int opcode, int oparg, _Py_SourceLocation loc);
-int _PyInstructionSequence_InjectSequence(_PyInstructionSequence *seq, int pos,
-                                          _PyInstructionSequence *injected);
+int _PyInstructionSequence_SetAnnotationsCode(_PyInstructionSequence *seq,
+                                              _PyInstructionSequence *annotations);
 int _PyInstructionSequence_AddNested(_PyInstructionSequence *seq, _PyInstructionSequence *nested);
 void PyInstructionSequence_Fini(_PyInstructionSequence *seq);
 

diff --git a/Include/internal/pycore_opcode_metadata.h b/Include/internal/pycore_opcode_metadata.h
diff --git a/Include/opcode_ids.h b/Include/opcode_ids.h
diff --git a/Lib/_opcode_metadata.py b/Lib/_opcode_metadata.py
diff --git a/Lib/test/test_dis.py b/Lib/test/test_dis.py
@@ -1007,7 +1007,8 @@ def test_boundaries(self):
     def test_widths(self):
         long_opcodes = set(['JUMP_BACKWARD_NO_INTERRUPT',
                             'LOAD_FAST_BORROW_LOAD_FAST_BORROW',
-                            'INSTRUMENTED_CALL_FUNCTION_EX'])
+                            'INSTRUMENTED_CALL_FUNCTION_EX',
+                            'ANNOTATIONS_PLACEHOLDER'])
         for op, opname in enumerate(dis.opname):
             if opname in long_opcodes or opname.startswith("INSTRUMENTED"):
                 continue

diff --git a/Misc/NEWS.d/next/Core_and_Builtins/2025-03-21-08-47-36.gh-issue-130907.rGg-ge.rst b/Misc/NEWS.d/next/Core_and_Builtins/2025-03-21-08-47-36.gh-issue-130907.rGg-ge.rst
@@ -1,2 +1,3 @@
-Raise an error if the ``__annotations__`` attribute is accessed on a module
-that has not finished executing.
+If the ``__annotations__`` of a module object are accessed while the
+module is executing, return the annotations that have been defined so far,
+without caching them.
@@ -2026,6 +2026,10 @@ dummy_func(
             }
         }
 
+        pseudo(ANNOTATIONS_PLACEHOLDER, (--)) = {
+            NOP,
+        };
+
         inst(DICT_UPDATE, (dict, unused[oparg - 1], update -- dict, unused[oparg - 1])) {
             PyObject *dict_o = PyStackRef_AsPyObjectBorrow(dict);
             PyObject *update_o = PyStackRef_AsPyObjectBorrow(update);

@@ -654,6 +654,9 @@ codegen_enter_scope(compiler *c, identifier name, int scope_type,
         loc.lineno = 0;
     }
     ADDOP_I(c, loc, RESUME, RESUME_AT_FUNC_START);
+    if (scope_type == COMPILE_SCOPE_MODULE) {
+        ADDOP(c, loc, ANNOTATIONS_PLACEHOLDER);
+    }
     return SUCCESS;
 }
 
@@ -832,9 +835,8 @@ codegen_process_deferred_annotations(compiler *c, location loc)
 
     if (nested_instr_seq != NULL) {
         RETURN_IF_ERROR(
-            _PyInstructionSequence_InjectSequence(old_instr_seq, 1, nested_instr_seq));
+            _PyInstructionSequence_SetAnnotationsCode(old_instr_seq, nested_instr_seq));
         _PyCompile_SetInstrSequence(c, old_instr_seq);
-        PyInstructionSequence_Fini(nested_instr_seq);
     }
 
     return SUCCESS;

@@ -3847,16 +3847,38 @@ _PyCfg_FromInstructionSequence(_PyInstructionSequence *seq)
             seq->s_instrs[instr->i_oparg].i_target = 1;
         }
     }
+    int offset = 0;
     for (int i = 0; i < seq->s_used; i++) {
         _PyInstruction *instr = &seq->s_instrs[i];
+        if (instr->i_opcode == ANNOTATIONS_PLACEHOLDER) {
+            if (seq->s_annotations_code != NULL) {
+                assert(seq->s_annotations_code->s_labelmap_size == 0
+                    && seq->s_annotations_code->s_nested == NULL);
+                for (int j = 0; j < seq->s_annotations_code->s_used; j++) {
+                    _PyInstruction *ann_instr = &seq->s_annotations_code->s_instrs[j];
+                    assert(!HAS_TARGET(ann_instr->i_opcode));
+                    if (_PyCfgBuilder_Addop(g, ann_instr->i_opcode, ann_instr->i_oparg, ann_instr->i_loc) < 0) {
+                        goto error;
+                    }
+                }
+                offset += seq->s_annotations_code->s_used - 1;
+            }
+            else {
+                offset -= 1;
+            }
+            continue;
+        }
         if (instr->i_target) {
-            jump_target_label lbl_ = {i};
+            jump_target_label lbl_ = {i + offset};
             if (_PyCfgBuilder_UseLabel(g, lbl_) < 0) {
                 goto error;
             }
         }
         int opcode = instr->i_opcode;
         int oparg = instr->i_oparg;
+        if (HAS_TARGET(opcode)) {
+            oparg += offset;
+        }
         if (_PyCfgBuilder_Addop(g, opcode, oparg, instr->i_loc) < 0) {
             goto error;
         }

@@ -34,29 +34,24 @@ typedef _Py_SourceLocation location;
     }
 
 static int
-instr_sequence_grow(instr_sequence *seq, int size) {
+instr_sequence_next_inst(instr_sequence *seq) {
     assert(seq->s_instrs != NULL || seq->s_used == 0);
 
+
     _Py_c_array_t array = {
         .array = (void*)seq->s_instrs,
         .allocated_entries = seq->s_allocated,
         .item_size = sizeof(instruction),
         .initial_num_entries = INITIAL_INSTR_SEQUENCE_SIZE,
     };
 
-    RETURN_IF_ERROR(_Py_CArray_EnsureCapacity(&array, seq->s_used + size));
+    RETURN_IF_ERROR(_Py_CArray_EnsureCapacity(&array, seq->s_used + 1));
     seq->s_instrs = array.array;
     seq->s_allocated = array.allocated_entries;
 
     assert(seq->s_allocated >= 0);
     assert(seq->s_used < seq->s_allocated);
-    seq->s_used += size;
-    return seq->s_used - 1;
-}
-
-static int
-instr_sequence_next_inst(instr_sequence *seq) {
-    return instr_sequence_grow(seq, 1);
+    return seq->s_used++;
 }
 
 _PyJumpTargetLabel
@@ -160,28 +155,11 @@ _PyInstructionSequence_InsertInstruction(instr_sequence *seq, int pos,
 }
 
 int
-_PyInstructionSequence_InjectSequence(instr_sequence *seq, int pos,
-                                      instr_sequence *injected)
+_PyInstructionSequence_SetAnnotationsCode(instr_sequence *seq,
+                                          instr_sequence *annotations)
 {
-    assert(pos >= 0 && pos <= seq->s_used);
-    // Merging labelmaps is not supported
-    assert(injected->s_labelmap_size == 0 && injected->s_nested == NULL);
-    if (injected->s_used == 0) {
-        return SUCCESS;
-    }
-
-    int last_idx = instr_sequence_grow(seq, injected->s_used);
-
-    RETURN_IF_ERROR(last_idx);
-    for (int i = last_idx - injected->s_used; i >= pos; i--) {
-        seq->s_instrs[i + injected->s_used] = seq->s_instrs[i];
-    }
-    for (int i=0; i < injected->s_used; i++) {
-        seq->s_instrs[i + pos] = injected->s_instrs[i];
-    }
-    for(int lbl=0; lbl < seq->s_labelmap_size; lbl++) {
-        seq->s_labelmap[lbl] += injected->s_used;
-    }
+    assert(seq->s_annotations_code == NULL);
+    seq->s_annotations_code = annotations;
     return SUCCESS;
 }
 
@@ -209,6 +187,12 @@ PyInstructionSequence_Fini(instr_sequence *seq) {
 
     PyMem_Free(seq->s_instrs);
     seq->s_instrs = NULL;
+
+    if (seq->s_annotations_code != NULL) {
+        PyInstructionSequence_Fini(seq->s_annotations_code);
+        Py_CLEAR(seq->s_annotations_code);
+    }
+
 }
 
 /*[clinic input]
@@ -231,6 +215,7 @@ inst_seq_create(void)
     seq->s_labelmap = NULL;
     seq->s_labelmap_size = 0;
     seq->s_nested = NULL;
+    seq->s_annotations_code = NULL;
 
     PyObject_GC_Track(seq);
     return seq;
@@ -445,6 +430,7 @@ inst_seq_traverse(PyObject *op, visitproc visit, void *arg)
 {
     _PyInstructionSequence *seq = (_PyInstructionSequence *)op;
     Py_VISIT(seq->s_nested);
+    Py_VISIT((PyObject *)seq->s_annotations_code);
     return 0;
 }
 
@@ -453,6 +439,7 @@ inst_seq_clear(PyObject *op)
 {
     _PyInstructionSequence *seq = (_PyInstructionSequence *)op;
     Py_CLEAR(seq->s_nested);
+    Py_CLEAR(seq->s_annotations_code);
     return 0;
 }