LibJS/Bytecode: Keep instruction source mappings in Executable

Instead of storing source offsets with each instruction, we now keep
them in a side table in Executable.

This shrinks each instruction by 8 bytes, further improving locality.

Userland/Libraries/LibJS/Bytecode/BasicBlock.h

@@ -8,6 +8,7 @@
 
 #include <AK/Badge.h>
 #include <AK/String.h>
+#include <LibJS/Bytecode/Executable.h>
 #include <LibJS/Forward.h>
 #include <LibJS/Heap/Handle.h>
 
@@ -45,6 +46,9 @@ public:
     BasicBlock const* handler() const { return m_handler; }
     BasicBlock const* finalizer() const { return m_finalizer; }
 
+    auto const& source_map() const { return m_source_map; }
+    void add_source_map_entry(size_t bytecode_offset, SourceRecord const& source_record) { m_source_map.set(bytecode_offset, source_record); }
+
 private:
     explicit BasicBlock(String name);
 
@@ -53,6 +57,8 @@ private:
     BasicBlock const* m_finalizer { nullptr };
     String m_name;
     bool m_terminated { false };
+
+    HashMap<size_t, SourceRecord> m_source_map;
 };
 
 }

Userland/Libraries/LibJS/Bytecode/Executable.cpp

@@ -106,11 +106,13 @@ UnrealizedSourceRange Executable::source_range_at(size_t offset) const
         return {};
     auto it = InstructionStreamIterator(bytecode.span().slice(offset), this);
     VERIFY(!it.at_end());
-    auto& instruction = *it;
+    auto mapping = source_map.get(offset);
+    if (!mapping.has_value())
+        return {};
     return UnrealizedSourceRange {
         .source_code = source_code,
-        .start_offset = instruction.source_record().source_start_offset,
+        .start_offset = mapping->source_start_offset,
-        .end_offset = instruction.source_record().source_end_offset,
+        .end_offset = mapping->source_end_offset,
     };
 }
 

Userland/Libraries/LibJS/Bytecode/Executable.h

@@ -84,6 +84,8 @@ public:
     Vector<ExceptionHandlers> exception_handlers;
     Vector<size_t> basic_block_start_offsets;
 
+    HashMap<size_t, SourceRecord> source_map;
+
     ByteString const& get_string(StringTableIndex index) const { return string_table->get(index); }
     DeprecatedFlyString const& get_identifier(IdentifierTableIndex index) const { return identifier_table->get(index); }
 

Userland/Libraries/LibJS/Bytecode/Generator.cpp

@@ -96,6 +96,8 @@ CodeGenerationErrorOr<NonnullGCPtr<Executable>> Generator::generate(VM& vm, ASTN
     };
     Vector<UnlinkedExceptionHandlers> unlinked_exception_handlers;
 
+    HashMap<size_t, SourceRecord> source_map;
+
     for (auto& block : generator.m_root_basic_blocks) {
         basic_block_start_offsets.append(bytecode.size());
         if (block->handler() || block->finalizer()) {
@@ -108,6 +110,11 @@ CodeGenerationErrorOr<NonnullGCPtr<Executable>> Generator::generate(VM& vm, ASTN
         }
 
         block_offsets.set(block.ptr(), bytecode.size());
+
+        for (auto& [offset, source_record] : block->source_map()) {
+            source_map.set(bytecode.size() + offset, source_record);
+        }
+
         Bytecode::InstructionStreamIterator it(block->instruction_stream());
         while (!it.at_end()) {
             auto& instruction = const_cast<Instruction&>(*it);
@@ -161,6 +168,7 @@ CodeGenerationErrorOr<NonnullGCPtr<Executable>> Generator::generate(VM& vm, ASTN
 
     executable->exception_handlers = move(linked_exception_handlers);
     executable->basic_block_start_offsets = move(basic_block_start_offsets);
+    executable->source_map = move(source_map);
 
     return executable;
 }

Userland/Libraries/LibJS/Bytecode/Generator.h

@@ -77,8 +77,7 @@ public:
         new (slot) OpType(forward<Args>(args)...);
         if constexpr (OpType::IsTerminator)
             m_current_basic_block->terminate({});
-        auto* op = static_cast<OpType*>(slot);
+        m_current_basic_block->add_source_map_entry(slot_offset, { m_current_ast_node->start_offset(), m_current_ast_node->end_offset() });
-        op->set_source_record({ m_current_ast_node->start_offset(), m_current_ast_node->end_offset() });
     }
 
     template<typename OpType, typename ExtraSlotType, typename... Args>
@@ -93,8 +92,7 @@ public:
         new (slot) OpType(forward<Args>(args)...);
         if constexpr (OpType::IsTerminator)
             m_current_basic_block->terminate({});
-        auto* op = static_cast<OpType*>(slot);
+        m_current_basic_block->add_source_map_entry(slot_offset, { m_current_ast_node->start_offset(), m_current_ast_node->end_offset() });
-        op->set_source_record({ m_current_ast_node->start_offset(), m_current_ast_node->end_offset() });
     }
 
     template<typename OpType, typename... Args>

Userland/Libraries/LibJS/Bytecode/Instruction.cpp

@@ -52,7 +52,7 @@ void Instruction::visit_labels(Function<void(JS::Bytecode::Label&)> visitor)
 UnrealizedSourceRange InstructionStreamIterator::source_range() const
 {
     VERIFY(m_executable);
-    auto record = dereference().source_record();
+    auto record = m_executable->source_map.get(offset()).value();
     return {
         .source_code = m_executable->source_code,
         .start_offset = record.source_start_offset,

Userland/Libraries/LibJS/Bytecode/Instruction.h

@@ -7,6 +7,7 @@
 #pragma once
 
 #include <AK/Forward.h>
+#include <AK/Function.h>
 #include <AK/Span.h>
 #include <LibJS/Bytecode/Executable.h>
 #include <LibJS/Forward.h>
@@ -140,17 +141,12 @@ public:
     void visit_labels(Function<void(Label&)> visitor);
     static void destroy(Instruction&);
 
-    // FIXME: Find a better way to organize this information
-    void set_source_record(SourceRecord rec) { m_source_record = rec; }
-    SourceRecord source_record() const { return m_source_record; }
-
 protected:
     Instruction(Type, size_t length);
 
     void visit_labels_impl(Function<void(Label&)>) { }
 
 private:
-    SourceRecord m_source_record {};
     Type m_type {};
     u32 m_length {};
 };

Userland/Libraries/LibJS/Bytecode/Label.h

@@ -7,10 +7,11 @@
 #pragma once
 
 #include <AK/Format.h>
-#include <LibJS/Bytecode/BasicBlock.h>
 
 namespace JS::Bytecode {
 
+class BasicBlock;
+
 class Label {
 public:
     explicit Label(BasicBlock const& block)