alamb commented on code in PR #2587:
URL: https://github.com/apache/arrow-datafusion/pull/2587#discussion_r879772780


##########
datafusion/jit/src/api.rs:
##########
@@ -604,6 +606,15 @@ impl<'a> CodeBlock<'a> {
             internal_err!("No func with the name {} exist", fn_name)
         }
     }
+
+    pub fn deref(&self, ptr: Expr, ty: JITType) -> Result<Expr> {
+        Ok(Expr::Deref(Box::new(ptr), ty))
+    }
+
+    pub fn store(&mut self, value: Expr, ptr: Expr) -> Result<()> {

Review Comment:
   ```suggestion
       /// Store the value in `value` to the address in `ptr`
       pub fn store(&mut self, value: Expr, ptr: Expr) -> Result<()> {
   ```



##########
datafusion/jit/src/compile.rs:
##########
@@ -0,0 +1,184 @@
+// Licensed to the Apache Software Foundation (ASF) under one
+// or more contributor license agreements.  See the NOTICE file
+// distributed with this work for additional information
+// regarding copyright ownership.  The ASF licenses this file
+// to you under the Apache License, Version 2.0 (the
+// "License"); you may not use this file except in compliance
+// with the License.  You may obtain a copy of the License at
+//
+//   http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing,
+// software distributed under the License is distributed on an
+// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, either express or implied.  See the License for the
+// specific language governing permissions and limitations
+// under the License.
+
+//! Compile DataFusion Expr to JIT'd function.
+
+use datafusion_common::Result;
+
+use crate::api::Assembler;
+use crate::{
+    api::GeneratedFunction,
+    ast::{Expr as JITExpr, I64, PTR_SIZE},
+};
+
+/// Wrap JIT Expr to array compute function.
+pub fn build_calc_fn(
+    assembler: &Assembler,
+    jit_expr: JITExpr,
+    input_names: Vec<String>,
+) -> Result<GeneratedFunction> {
+    let mut builder = assembler.new_func_builder("calc_fn");
+    for input in &input_names {
+        builder = builder.param(format!("{}_array", input), I64);
+    }
+    let mut builder = builder.param("result", I64).param("len", I64);
+
+    let mut fn_body = builder.enter_block();
+    fn_body.declare_as("index", fn_body.lit_i(0))?;
+    fn_body.while_block(
+        |cond| cond.lt(cond.id("index")?, cond.id("len")?),
+        |w| {
+            w.declare_as("offset", w.mul(w.id("index")?, w.lit_i(PTR_SIZE as 
i64))?)?;
+            for input in &input_names {
+                w.declare_as(
+                    format!("{}_ptr", input),
+                    w.add(w.id(format!("{}_array", input))?, w.id("offset")?)?,
+                )?;
+                w.declare_as(input, w.deref(w.id(format!("{}_ptr", input))?, 
I64)?)?;
+            }
+            w.declare_as("res_ptr", w.add(w.id("result")?, w.id("offset")?)?)?;
+            w.declare_as("res", jit_expr.clone())?;
+            w.store(w.id("res")?, w.id("res_ptr")?)?;
+
+            w.assign("index", w.add(w.id("index")?, w.lit_i(1))?)?;
+            Ok(())
+        },
+    )?;
+
+    let gen_func = fn_body.build();
+    Ok(gen_func)
+}
+
+#[cfg(test)]
+mod test {
+    use std::{collections::HashMap, sync::Arc};
+
+    use arrow::{
+        array::{Array, PrimitiveArray},
+        datatypes::{DataType, Int64Type},
+    };
+    use datafusion_common::{DFField, DFSchema, DataFusionError};
+    use datafusion_expr::Expr as DFExpr;
+
+    use crate::ast::BinaryExpr;
+
+    use super::*;
+
+    fn run_df_expr(

Review Comment:
   In the longer term I would like to see this type of logic encapsulated 
somehow
   
   So we would have a function or struct that took an `Expr` and several 
`ArrayRefs` and then called a JIT or non-JIT version of evaluation depending on 
flags or options. 



##########
datafusion/jit/src/compile.rs:
##########
@@ -0,0 +1,184 @@
+// Licensed to the Apache Software Foundation (ASF) under one
+// or more contributor license agreements.  See the NOTICE file
+// distributed with this work for additional information
+// regarding copyright ownership.  The ASF licenses this file
+// to you under the Apache License, Version 2.0 (the
+// "License"); you may not use this file except in compliance
+// with the License.  You may obtain a copy of the License at
+//
+//   http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing,
+// software distributed under the License is distributed on an
+// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, either express or implied.  See the License for the
+// specific language governing permissions and limitations
+// under the License.
+
+//! Compile DataFusion Expr to JIT'd function.
+
+use datafusion_common::Result;
+
+use crate::api::Assembler;
+use crate::{
+    api::GeneratedFunction,
+    ast::{Expr as JITExpr, I64, PTR_SIZE},
+};
+
+/// Wrap JIT Expr to array compute function.
+pub fn build_calc_fn(
+    assembler: &Assembler,
+    jit_expr: JITExpr,
+    input_names: Vec<String>,
+) -> Result<GeneratedFunction> {
+    let mut builder = assembler.new_func_builder("calc_fn");
+    for input in &input_names {
+        builder = builder.param(format!("{}_array", input), I64);
+    }
+    let mut builder = builder.param("result", I64).param("len", I64);
+
+    let mut fn_body = builder.enter_block();
+    fn_body.declare_as("index", fn_body.lit_i(0))?;
+    fn_body.while_block(
+        |cond| cond.lt(cond.id("index")?, cond.id("len")?),
+        |w| {
+            w.declare_as("offset", w.mul(w.id("index")?, w.lit_i(PTR_SIZE as 
i64))?)?;
+            for input in &input_names {
+                w.declare_as(
+                    format!("{}_ptr", input),
+                    w.add(w.id(format!("{}_array", input))?, w.id("offset")?)?,
+                )?;
+                w.declare_as(input, w.deref(w.id(format!("{}_ptr", input))?, 
I64)?)?;
+            }
+            w.declare_as("res_ptr", w.add(w.id("result")?, w.id("offset")?)?)?;
+            w.declare_as("res", jit_expr.clone())?;
+            w.store(w.id("res")?, w.id("res_ptr")?)?;
+
+            w.assign("index", w.add(w.id("index")?, w.lit_i(1))?)?;
+            Ok(())
+        },
+    )?;
+
+    let gen_func = fn_body.build();
+    Ok(gen_func)
+}
+
+#[cfg(test)]
+mod test {
+    use std::{collections::HashMap, sync::Arc};
+
+    use arrow::{
+        array::{Array, PrimitiveArray},
+        datatypes::{DataType, Int64Type},
+    };
+    use datafusion_common::{DFField, DFSchema, DataFusionError};
+    use datafusion_expr::Expr as DFExpr;
+
+    use crate::ast::BinaryExpr;
+
+    use super::*;
+
+    fn run_df_expr(
+        df_expr: DFExpr,
+        schema: Arc<DFSchema>,
+        lhs: PrimitiveArray<Int64Type>,
+        rhs: PrimitiveArray<Int64Type>,
+    ) -> Result<PrimitiveArray<Int64Type>> {
+        if lhs.null_count() != 0 || rhs.null_count() != 0 {
+            return Err(DataFusionError::NotImplemented(
+                "Computing on nullable array not yet supported".to_string(),
+            ));
+        }
+        if lhs.len() != rhs.len() {
+            return Err(DataFusionError::NotImplemented(
+                "Computing on different length arrays not yet 
supported".to_string(),
+            ));
+        }
+
+        // translate DF Expr to JIT Expr
+        let input_fields = schema.field_names();
+        let jit_expr: JITExpr = (df_expr, schema).try_into()?;
+
+        // allocate memory for calc result
+        let len = lhs.len();
+        let result = vec![0i64; len];
+
+        // compile and run JIT code
+        let assembler = Assembler::default();
+        let gen_func = build_calc_fn(&assembler, jit_expr, input_fields)?;
+        let mut jit = assembler.create_jit();
+        let code_ptr = jit.compile(gen_func)?;
+        let code_fn =
+            unsafe { core::mem::transmute::<_, fn(i64, i64, i64, i64) -> 
()>(code_ptr) };

Review Comment:
   I wonder why not cast to the types you really want, like `fn(*i64, *i64, 
*i64, i64)` and then you can avoid the `as i64` stuff below



##########
datafusion/jit/src/compile.rs:
##########
@@ -0,0 +1,184 @@
+// Licensed to the Apache Software Foundation (ASF) under one
+// or more contributor license agreements.  See the NOTICE file
+// distributed with this work for additional information
+// regarding copyright ownership.  The ASF licenses this file
+// to you under the Apache License, Version 2.0 (the
+// "License"); you may not use this file except in compliance
+// with the License.  You may obtain a copy of the License at
+//
+//   http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing,
+// software distributed under the License is distributed on an
+// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, either express or implied.  See the License for the
+// specific language governing permissions and limitations
+// under the License.
+
+//! Compile DataFusion Expr to JIT'd function.
+
+use datafusion_common::Result;
+
+use crate::api::Assembler;
+use crate::{
+    api::GeneratedFunction,
+    ast::{Expr as JITExpr, I64, PTR_SIZE},
+};
+
+/// Wrap JIT Expr to array compute function.
+pub fn build_calc_fn(
+    assembler: &Assembler,
+    jit_expr: JITExpr,
+    input_names: Vec<String>,
+) -> Result<GeneratedFunction> {
+    let mut builder = assembler.new_func_builder("calc_fn");
+    for input in &input_names {
+        builder = builder.param(format!("{}_array", input), I64);
+    }
+    let mut builder = builder.param("result", I64).param("len", I64);
+
+    let mut fn_body = builder.enter_block();
+    fn_body.declare_as("index", fn_body.lit_i(0))?;
+    fn_body.while_block(
+        |cond| cond.lt(cond.id("index")?, cond.id("len")?),
+        |w| {
+            w.declare_as("offset", w.mul(w.id("index")?, w.lit_i(PTR_SIZE as 
i64))?)?;
+            for input in &input_names {
+                w.declare_as(
+                    format!("{}_ptr", input),
+                    w.add(w.id(format!("{}_array", input))?, w.id("offset")?)?,
+                )?;
+                w.declare_as(input, w.deref(w.id(format!("{}_ptr", input))?, 
I64)?)?;
+            }
+            w.declare_as("res_ptr", w.add(w.id("result")?, w.id("offset")?)?)?;
+            w.declare_as("res", jit_expr.clone())?;
+            w.store(w.id("res")?, w.id("res_ptr")?)?;
+
+            w.assign("index", w.add(w.id("index")?, w.lit_i(1))?)?;
+            Ok(())
+        },
+    )?;
+
+    let gen_func = fn_body.build();
+    Ok(gen_func)
+}
+
+#[cfg(test)]
+mod test {
+    use std::{collections::HashMap, sync::Arc};
+
+    use arrow::{
+        array::{Array, PrimitiveArray},
+        datatypes::{DataType, Int64Type},
+    };
+    use datafusion_common::{DFField, DFSchema, DataFusionError};
+    use datafusion_expr::Expr as DFExpr;
+
+    use crate::ast::BinaryExpr;
+
+    use super::*;
+
+    fn run_df_expr(
+        df_expr: DFExpr,
+        schema: Arc<DFSchema>,
+        lhs: PrimitiveArray<Int64Type>,
+        rhs: PrimitiveArray<Int64Type>,
+    ) -> Result<PrimitiveArray<Int64Type>> {
+        if lhs.null_count() != 0 || rhs.null_count() != 0 {
+            return Err(DataFusionError::NotImplemented(
+                "Computing on nullable array not yet supported".to_string(),
+            ));
+        }
+        if lhs.len() != rhs.len() {
+            return Err(DataFusionError::NotImplemented(
+                "Computing on different length arrays not yet 
supported".to_string(),
+            ));
+        }
+
+        // translate DF Expr to JIT Expr
+        let input_fields = schema.field_names();
+        let jit_expr: JITExpr = (df_expr, schema).try_into()?;
+
+        // allocate memory for calc result
+        let len = lhs.len();
+        let result = vec![0i64; len];
+
+        // compile and run JIT code
+        let assembler = Assembler::default();
+        let gen_func = build_calc_fn(&assembler, jit_expr, input_fields)?;
+        let mut jit = assembler.create_jit();
+        let code_ptr = jit.compile(gen_func)?;
+        let code_fn =
+            unsafe { core::mem::transmute::<_, fn(i64, i64, i64, i64) -> 
()>(code_ptr) };
+        code_fn(
+            lhs.values().as_ptr() as i64,
+            rhs.values().as_ptr() as i64,
+            result.as_ptr() as i64,
+            len as i64,
+        );
+
+        let result_array = PrimitiveArray::<Int64Type>::from_iter(result);
+        Ok(result_array)
+    }
+
+    #[test]
+    fn array_add() {
+        let array_a: PrimitiveArray<Int64Type> =

Review Comment:
   I recommend using different values for `array_a` and `array_b` so issues in 
argument handling would be evident
   
   Like maybe
   
   ```rust
           let array_b: PrimitiveArray<Int64Type> =
               PrimitiveArray::from_iter_values((10..20).map(|x| x + 1));
   ```



##########
datafusion/jit/src/api.rs:
##########
@@ -604,6 +606,15 @@ impl<'a> CodeBlock<'a> {
             internal_err!("No func with the name {} exist", fn_name)
         }
     }
+
+    pub fn deref(&self, ptr: Expr, ty: JITType) -> Result<Expr> {

Review Comment:
   What do you think about adding docstrings? 
   
   ```suggestion
       /// Return the value pointed to by the ptr stored in `ptr`
       pub fn deref(&self, ptr: Expr, ty: JITType) -> Result<Expr> {
   ```
   



##########
datafusion/jit/src/compile.rs:
##########
@@ -0,0 +1,184 @@
+// Licensed to the Apache Software Foundation (ASF) under one
+// or more contributor license agreements.  See the NOTICE file
+// distributed with this work for additional information
+// regarding copyright ownership.  The ASF licenses this file
+// to you under the Apache License, Version 2.0 (the
+// "License"); you may not use this file except in compliance
+// with the License.  You may obtain a copy of the License at
+//
+//   http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing,
+// software distributed under the License is distributed on an
+// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, either express or implied.  See the License for the
+// specific language governing permissions and limitations
+// under the License.
+
+//! Compile DataFusion Expr to JIT'd function.
+
+use datafusion_common::Result;
+
+use crate::api::Assembler;
+use crate::{
+    api::GeneratedFunction,
+    ast::{Expr as JITExpr, I64, PTR_SIZE},
+};
+
+/// Wrap JIT Expr to array compute function.
+pub fn build_calc_fn(
+    assembler: &Assembler,
+    jit_expr: JITExpr,
+    input_names: Vec<String>,
+) -> Result<GeneratedFunction> {
+    let mut builder = assembler.new_func_builder("calc_fn");
+    for input in &input_names {
+        builder = builder.param(format!("{}_array", input), I64);
+    }
+    let mut builder = builder.param("result", I64).param("len", I64);
+
+    let mut fn_body = builder.enter_block();
+    fn_body.declare_as("index", fn_body.lit_i(0))?;
+    fn_body.while_block(
+        |cond| cond.lt(cond.id("index")?, cond.id("len")?),
+        |w| {
+            w.declare_as("offset", w.mul(w.id("index")?, w.lit_i(PTR_SIZE as 
i64))?)?;
+            for input in &input_names {
+                w.declare_as(
+                    format!("{}_ptr", input),
+                    w.add(w.id(format!("{}_array", input))?, w.id("offset")?)?,
+                )?;
+                w.declare_as(input, w.deref(w.id(format!("{}_ptr", input))?, 
I64)?)?;
+            }
+            w.declare_as("res_ptr", w.add(w.id("result")?, w.id("offset")?)?)?;
+            w.declare_as("res", jit_expr.clone())?;
+            w.store(w.id("res")?, w.id("res_ptr")?)?;
+
+            w.assign("index", w.add(w.id("index")?, w.lit_i(1))?)?;
+            Ok(())
+        },
+    )?;
+
+    let gen_func = fn_body.build();
+    Ok(gen_func)
+}
+
+#[cfg(test)]
+mod test {
+    use std::{collections::HashMap, sync::Arc};
+
+    use arrow::{
+        array::{Array, PrimitiveArray},
+        datatypes::{DataType, Int64Type},
+    };
+    use datafusion_common::{DFField, DFSchema, DataFusionError};
+    use datafusion_expr::Expr as DFExpr;
+
+    use crate::ast::BinaryExpr;
+
+    use super::*;
+
+    fn run_df_expr(
+        df_expr: DFExpr,
+        schema: Arc<DFSchema>,
+        lhs: PrimitiveArray<Int64Type>,
+        rhs: PrimitiveArray<Int64Type>,
+    ) -> Result<PrimitiveArray<Int64Type>> {
+        if lhs.null_count() != 0 || rhs.null_count() != 0 {
+            return Err(DataFusionError::NotImplemented(
+                "Computing on nullable array not yet supported".to_string(),
+            ));
+        }
+        if lhs.len() != rhs.len() {
+            return Err(DataFusionError::NotImplemented(
+                "Computing on different length arrays not yet 
supported".to_string(),
+            ));
+        }
+
+        // translate DF Expr to JIT Expr
+        let input_fields = schema.field_names();
+        let jit_expr: JITExpr = (df_expr, schema).try_into()?;
+
+        // allocate memory for calc result
+        let len = lhs.len();
+        let result = vec![0i64; len];
+
+        // compile and run JIT code
+        let assembler = Assembler::default();
+        let gen_func = build_calc_fn(&assembler, jit_expr, input_fields)?;
+        let mut jit = assembler.create_jit();
+        let code_ptr = jit.compile(gen_func)?;
+        let code_fn =
+            unsafe { core::mem::transmute::<_, fn(i64, i64, i64, i64) -> 
()>(code_ptr) };
+        code_fn(
+            lhs.values().as_ptr() as i64,
+            rhs.values().as_ptr() as i64,
+            result.as_ptr() as i64,
+            len as i64,
+        );
+
+        let result_array = PrimitiveArray::<Int64Type>::from_iter(result);
+        Ok(result_array)
+    }
+
+    #[test]
+    fn array_add() {
+        let array_a: PrimitiveArray<Int64Type> =
+            PrimitiveArray::from_iter_values((0..10).map(|x| x + 1));
+        let array_b: PrimitiveArray<Int64Type> =
+            PrimitiveArray::from_iter_values((0..10).map(|x| x + 1));
+        let expected =
+            arrow::compute::kernels::arithmetic::add(&array_a, 
&array_b).unwrap();
+
+        let df_expr = datafusion_expr::col("a") + datafusion_expr::col("b");
+        let schema = Arc::new(
+            DFSchema::new_with_metadata(
+                vec![
+                    DFField::new(Some("table1"), "a", DataType::Int64, false),
+                    DFField::new(Some("table1"), "b", DataType::Int64, false),
+                ],
+                HashMap::new(),
+            )
+            .unwrap(),
+        );
+
+        let result = run_df_expr(df_expr, schema, array_a, array_b).unwrap();
+        assert_eq!(result, expected);
+    }
+
+    #[test]
+    fn calc_fn_builder() {
+        let expr = JITExpr::Binary(BinaryExpr::Add(
+            Box::new(JITExpr::Identifier("table1.a".to_string(), I64)),
+            Box::new(JITExpr::Identifier("table1.b".to_string(), I64)),
+        ));
+        let fields = vec!["table1.a".to_string(), "table1.b".to_string()];
+
+        let expected = r#"fn calc_fn_0(table1.a_array: i64, table1.b_array: 
i64, result: i64, len: i64) -> () {
+    let index: i64;
+    index = 0;
+    while index < len {

Review Comment:
   I looked at this code and it looks 👍  to me



##########
datafusion/jit/src/compile.rs:
##########
@@ -0,0 +1,184 @@
+// Licensed to the Apache Software Foundation (ASF) under one
+// or more contributor license agreements.  See the NOTICE file
+// distributed with this work for additional information
+// regarding copyright ownership.  The ASF licenses this file
+// to you under the Apache License, Version 2.0 (the
+// "License"); you may not use this file except in compliance
+// with the License.  You may obtain a copy of the License at
+//
+//   http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing,
+// software distributed under the License is distributed on an
+// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, either express or implied.  See the License for the
+// specific language governing permissions and limitations
+// under the License.
+
+//! Compile DataFusion Expr to JIT'd function.
+
+use datafusion_common::Result;
+
+use crate::api::Assembler;
+use crate::{
+    api::GeneratedFunction,
+    ast::{Expr as JITExpr, I64, PTR_SIZE},
+};
+
+/// Wrap JIT Expr to array compute function.
+pub fn build_calc_fn(
+    assembler: &Assembler,
+    jit_expr: JITExpr,
+    input_names: Vec<String>,
+) -> Result<GeneratedFunction> {
+    let mut builder = assembler.new_func_builder("calc_fn");
+    for input in &input_names {
+        builder = builder.param(format!("{}_array", input), I64);
+    }
+    let mut builder = builder.param("result", I64).param("len", I64);
+
+    let mut fn_body = builder.enter_block();
+    fn_body.declare_as("index", fn_body.lit_i(0))?;
+    fn_body.while_block(
+        |cond| cond.lt(cond.id("index")?, cond.id("len")?),
+        |w| {
+            w.declare_as("offset", w.mul(w.id("index")?, w.lit_i(PTR_SIZE as 
i64))?)?;
+            for input in &input_names {
+                w.declare_as(
+                    format!("{}_ptr", input),
+                    w.add(w.id(format!("{}_array", input))?, w.id("offset")?)?,
+                )?;
+                w.declare_as(input, w.deref(w.id(format!("{}_ptr", input))?, 
I64)?)?;
+            }
+            w.declare_as("res_ptr", w.add(w.id("result")?, w.id("offset")?)?)?;
+            w.declare_as("res", jit_expr.clone())?;
+            w.store(w.id("res")?, w.id("res_ptr")?)?;
+
+            w.assign("index", w.add(w.id("index")?, w.lit_i(1))?)?;
+            Ok(())
+        },
+    )?;
+
+    let gen_func = fn_body.build();
+    Ok(gen_func)
+}
+
+#[cfg(test)]
+mod test {
+    use std::{collections::HashMap, sync::Arc};
+
+    use arrow::{
+        array::{Array, PrimitiveArray},
+        datatypes::{DataType, Int64Type},
+    };
+    use datafusion_common::{DFField, DFSchema, DataFusionError};
+    use datafusion_expr::Expr as DFExpr;
+
+    use crate::ast::BinaryExpr;
+
+    use super::*;
+
+    fn run_df_expr(
+        df_expr: DFExpr,
+        schema: Arc<DFSchema>,
+        lhs: PrimitiveArray<Int64Type>,
+        rhs: PrimitiveArray<Int64Type>,
+    ) -> Result<PrimitiveArray<Int64Type>> {
+        if lhs.null_count() != 0 || rhs.null_count() != 0 {
+            return Err(DataFusionError::NotImplemented(
+                "Computing on nullable array not yet supported".to_string(),
+            ));
+        }
+        if lhs.len() != rhs.len() {
+            return Err(DataFusionError::NotImplemented(
+                "Computing on different length arrays not yet 
supported".to_string(),
+            ));
+        }
+
+        // translate DF Expr to JIT Expr
+        let input_fields = schema.field_names();
+        let jit_expr: JITExpr = (df_expr, schema).try_into()?;
+
+        // allocate memory for calc result
+        let len = lhs.len();
+        let result = vec![0i64; len];
+
+        // compile and run JIT code
+        let assembler = Assembler::default();
+        let gen_func = build_calc_fn(&assembler, jit_expr, input_fields)?;
+        let mut jit = assembler.create_jit();
+        let code_ptr = jit.compile(gen_func)?;
+        let code_fn =
+            unsafe { core::mem::transmute::<_, fn(i64, i64, i64, i64) -> 
()>(code_ptr) };

Review Comment:
   code like this is some of my favorite ❤️  -- cast some memory to a function 
pointer and call it ;)
   
   Extremely unsafe but feels very powerful
   
   



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to