chore: improve new parser arch

Signed-off-by: Henry Gressmann <mail@henrygressmann.de>

5 files changed, 37 insertions, 24 deletions

diff --git a/Cargo.lock b/Cargo.lock
index 4fddf4d..a6fdbbd 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -293,9 +293,9 @@ checksum = "37b2a672a2cb129a2e41c10b1224bb368f9f37a2b16b612598138befd7b37eb5"
 
 [[package]]
 name = "cc"
-version = "1.0.87"
+version = "1.0.88"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3286b845d0fccbdd15af433f61c5970e711987036cb468f437ff6badd70f4e24"
+checksum = "02f341c093d19155a6e41631ce5971aac4e9a868262212153124c15fa22d1cdc"
 
 [[package]]
 name = "cfg-if"
diff --git a/crates/parser/src/conversion.rs b/crates/parser/src/conversion.rs
index 03b5f82..53cceb6 100644
--- a/crates/parser/src/conversion.rs
+++ b/crates/parser/src/conversion.rs
@@ -1,5 +1,5 @@
-use crate::visit::process_operators;
 use crate::Result;
+use crate::{module::Code, visit::process_operators};
 use alloc::{boxed::Box, format, string::ToString, vec::Vec};
 use tinywasm_types::*;
 use wasmparser::{FuncValidator, OperatorsReader, ValidatorResources};
@@ -159,7 +159,7 @@ pub(crate) fn convert_module_export(export: wasmparser::Export<'_>) -> Result<Ex
 pub(crate) fn convert_module_code(
     func: wasmparser::FunctionBody<'_>,
     mut validator: FuncValidator<ValidatorResources>,
-) -> Result<(Box<[Instruction]>, Box<[ValType]>)> {
+) -> Result<Code> {
     let locals_reader = func.get_locals_reader()?;
     let count = locals_reader.get_count();
     let pos = locals_reader.original_position();
@@ -173,7 +173,7 @@ pub(crate) fn convert_module_code(
         }
     }
 
-    let body = process_operators(&mut validator, &func)?;
+    let body = process_operators(Some(&mut validator), &func)?;
     let locals = locals.into_boxed_slice();
     Ok((body, locals))
 }
diff --git a/crates/parser/src/lib.rs b/crates/parser/src/lib.rs
index dd4b931..5de4b03 100644
--- a/crates/parser/src/lib.rs
+++ b/crates/parser/src/lib.rs
@@ -52,13 +52,13 @@ impl Parser {
         let features = WasmFeatures {
             bulk_memory: true,
             floats: true,
-            function_references: true,
             multi_value: true,
             mutable_global: true,
             reference_types: true,
             sign_extension: true,
             saturating_float_to_int: true,
 
+            function_references: false,
             component_model: false,
             component_model_nested_names: false,
             component_model_values: false,
diff --git a/crates/parser/src/module.rs b/crates/parser/src/module.rs
index 813a65d..8414c17 100644
--- a/crates/parser/src/module.rs
+++ b/crates/parser/src/module.rs
@@ -4,6 +4,8 @@ use alloc::{boxed::Box, format, vec::Vec};
 use tinywasm_types::{Data, Element, Export, FuncType, Global, Import, Instruction, MemoryType, TableType, ValType};
 use wasmparser::{Payload, Validator};
 
+pub(crate) type Code = (Box<[Instruction]>, Box<[ValType]>);
+
 #[derive(Default)]
 pub(crate) struct ModuleReader {
     pub(crate) version: Option<u16>,
@@ -11,7 +13,7 @@ pub(crate) struct ModuleReader {
     pub(crate) func_types: Vec<FuncType>,
     pub(crate) code_type_addrs: Vec<u32>,
     pub(crate) exports: Vec<Export>,
-    pub(crate) code: Vec<(Box<[Instruction]>, Box<[ValType]>)>,
+    pub(crate) code: Vec<Code>,
     pub(crate) globals: Vec<Global>,
     pub(crate) table_types: Vec<TableType>,
     pub(crate) memory_types: Vec<MemoryType>,
diff --git a/crates/parser/src/visit.rs b/crates/parser/src/visit.rs
index b42e462..15024f1 100644
--- a/crates/parser/src/visit.rs
+++ b/crates/parser/src/visit.rs
@@ -10,6 +10,7 @@ struct ValidateThenVisit<'a, T, U>(T, &'a mut U);
 macro_rules! validate_then_visit {
     ($( @$proposal:ident $op:ident $({ $($arg:ident: $argty:ty),* })? => $visit:ident)*) => {
         $(
+            #[inline]
             fn $visit(&mut self $($(,$arg: $argty)*)?) -> Self::Output {
                 self.0.$visit($($($arg.clone()),*)?)?;
                 Ok(self.1.$visit($($($arg),*)?))
@@ -24,23 +25,29 @@ where
     U: VisitOperator<'a>,
 {
     type Output = Result<U::Output>;
-
     wasmparser::for_each_operator!(validate_then_visit);
 }
 
 pub(crate) fn process_operators<R: WasmModuleResources>(
-    validator: &mut FuncValidator<R>,
+    validator: Option<&mut FuncValidator<R>>,
     body: &FunctionBody<'_>,
 ) -> Result<Box<[Instruction]>> {
     let mut reader = body.get_operators_reader()?;
-    let mut builder = FunctionBuilder::new(1024);
+    let remaining = reader.get_binary_reader().bytes_remaining();
+    let mut builder = FunctionBuilder::new(remaining);
 
-    while !reader.eof() {
-        let validate = validator.visitor(reader.original_position());
-        reader.visit_operator(&mut ValidateThenVisit(validate, &mut builder))???;
+    if let Some(validator) = validator {
+        while !reader.eof() {
+            let validate = validator.visitor(reader.original_position());
+            reader.visit_operator(&mut ValidateThenVisit(validate, &mut builder))???;
+        }
+        validator.finish(reader.original_position())?;
+    } else {
+        while !reader.eof() {
+            reader.visit_operator(&mut builder)??;
+        }
     }
 
-    validator.finish(reader.original_position())?;
     Ok(builder.instructions.into_boxed_slice())
 }
 
@@ -114,7 +121,7 @@ pub(crate) struct FunctionBuilder {
 
 impl FunctionBuilder {
     pub(crate) fn new(instr_capacity: usize) -> Self {
-        Self { instructions: Vec::with_capacity(instr_capacity), label_ptrs: Vec::with_capacity(64) }
+        Self { instructions: Vec::with_capacity(instr_capacity), label_ptrs: Vec::with_capacity(128) }
     }
 
     #[cold]
@@ -124,7 +131,8 @@ impl FunctionBuilder {
 
     #[inline]
     fn visit(&mut self, op: Instruction) -> Result<()> {
-        Ok(self.instructions.push(op))
+        self.instructions.push(op);
+        Ok(())
     }
 }
 
@@ -337,6 +345,7 @@ impl<'a> wasmparser::VisitOperator<'a> for FunctionBuilder {
         self.visit(Instruction::Else(0))
     }
 
+    #[inline]
     fn visit_end(&mut self) -> Self::Output {
         let Some(label_pointer) = self.label_ptrs.pop() else {
             return self.visit(Instruction::EndFunc);
@@ -348,16 +357,19 @@ impl<'a> wasmparser::VisitOperator<'a> for FunctionBuilder {
             Instruction::Else(ref mut else_instr_end_offset) => {
                 *else_instr_end_offset = current_instr_ptr - label_pointer;
 
+                #[cold]
+                fn error() -> crate::ParseError {
+                    crate::ParseError::UnsupportedOperator(
+                        "Expected to end an if block, but the last label was not an if".to_string(),
+                    )
+                }
+
                 // since we're ending an else block, we need to end the if block as well
-                let if_label_pointer = self.label_ptrs.pop().ok_or(crate::ParseError::UnsupportedOperator(
-                    "Expected to end an if block, but the last label was not an if".to_string(),
-                ))?;
+                let if_label_pointer = self.label_ptrs.pop().ok_or_else(error)?;
 
                 let if_instruction = &mut self.instructions[if_label_pointer];
                 let Instruction::If(_, ref mut else_offset, ref mut end_offset) = if_instruction else {
-                    return Err(crate::ParseError::UnsupportedOperator(
-                        "Expected to end an if block, but the last label was not an if".to_string(),
-                    ));
+                    return Err(error());
                 };
 
                 *else_offset = Some(label_pointer - if_label_pointer);
@@ -386,8 +398,7 @@ impl<'a> wasmparser::VisitOperator<'a> for FunctionBuilder {
             .collect::<Result<Vec<Instruction>, wasmparser::BinaryReaderError>>()
             .expect("BrTable targets are invalid, this should have been caught by the validator");
 
-        self.instructions
-            .extend(IntoIterator::into_iter([Instruction::BrTable(def, instrs.len())]).chain(instrs.into_iter()));
+        self.instructions.extend(IntoIterator::into_iter([Instruction::BrTable(def, instrs.len())]).chain(instrs));
         Ok(())
     }