RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-05-31 11:08:11 +00:00
Code Issues Activity Actions

LibJS: Remove the concept of bytecode optimization levels

Browse source 
While this would be useful in the future for implementing a multi-tiered
optimization strategy, currently a binary on/off is enough for us. This
removes the confusingly on-by-default `OptimizationLevel::None` option
which made the optimization pipeline a no-op even if
`Bytecode::Interpreter::set_optimizations_enabled` had been called.

Fixes #15982
This commit is contained in:
Daniel Bertalan 2023-06-26 20:10:11 +02:00 • committed by Ali Mohammad Pur
parent e012565898
commit cc9ec6693b
2 changed files with 16 additions and 36 deletions
Download patch file Download diff file Expand all files Collapse all files

10
Userland/Libraries/LibJS/Bytecode/Interpreter.h
View file

@ -86,13 +86,7 @@ public:
size_t pc() const;
DeprecatedString debug_position() const;
enum class OptimizationLevel {
None,
Optimize,
__Count,
Default = None,
};
static Bytecode::PassManager& optimization_pipeline(OptimizationLevel = OptimizationLevel::Default);
static Bytecode::PassManager& optimization_pipeline();
VM::InterpreterExecutionScope ast_interpreter_scope(Realm&);
@ -109,8 +103,6 @@ private:
MarkedVector<Value>& registers() { return window().registers; }
static AK::Array<OwnPtr<PassManager>, static_cast<UnderlyingType<Interpreter::OptimizationLevel>>(Interpreter::OptimizationLevel::__Count)> s_optimization_pipelines;
VM& m_vm;
Vector<Variant<NonnullOwnPtr<RegisterWindow>, RegisterWindow*>> m_register_windows;
Optional<BasicBlock const*> m_pending_jump;