tvl-depot/tvix/eval/src/vm.rs
Vincent Ambo 293fb0ef53 refactor(tvix/value): encapsulate attrset logic within value::attrs
The internal optimisations of the set representation were previously
leaking into the VM, which is highly undesirable.

Keeping it encapsulated allows us to do additional optimisations
within value::attrs without being concerned about its use in the VM.

Change-Id: I7e7020bb0983b9d355d3db747b049b2faa60131f
Reviewed-on: https://cl.tvl.fyi/c/depot/+/6108
Reviewed-by: eta <tvl@eta.st>
Tested-by: BuildkiteCI
2022-08-24 18:19:52 +00:00

199 lines
6.3 KiB
Rust

//! This module implements the virtual (or abstract) machine that runs
//! Tvix bytecode.
use std::rc::Rc;
use crate::{
chunk::Chunk,
errors::{Error, EvalResult},
opcode::OpCode,
value::{NixAttrs, NixList, NixString, Value},
};
pub struct VM {
ip: usize,
chunk: Chunk,
stack: Vec<Value>,
}
impl VM {
fn inc_ip(&mut self) -> OpCode {
let op = self.chunk.code[self.ip];
self.ip += 1;
op
}
fn pop(&mut self) -> Value {
self.stack.pop().expect("TODO")
}
fn pop_number_pair(&mut self) -> EvalResult<NumberPair> {
let v2 = self.pop();
let v1 = self.pop();
match (v1, v2) {
(Value::Integer(i1), Value::Integer(i2)) => Ok(NumberPair::Integer(i1, i2)),
(Value::Float(f1), Value::Float(f2)) => Ok(NumberPair::Floats(f1, f2)),
(Value::Integer(i1), Value::Float(f2)) => Ok(NumberPair::Floats(i1 as f64, f2)),
(Value::Float(f1), Value::Integer(i2)) => Ok(NumberPair::Floats(f1, i2 as f64)),
(v1, v2) => Err(Error::TypeError {
expected: "number (either int or float)",
actual: if v1.is_number() {
v2.type_of()
} else {
v1.type_of()
},
}),
}
}
fn push(&mut self, value: Value) {
self.stack.push(value)
}
fn run(&mut self) -> EvalResult<Value> {
loop {
match self.inc_ip() {
OpCode::OpConstant(idx) => {
let c = self.chunk.constant(idx).clone();
self.push(c);
}
OpCode::OpAdd => match self.pop_number_pair()? {
NumberPair::Floats(f1, f2) => self.push(Value::Float(f1 + f2)),
NumberPair::Integer(i1, i2) => self.push(Value::Integer(i1 + i2)),
},
OpCode::OpSub => match self.pop_number_pair()? {
NumberPair::Floats(f1, f2) => self.push(Value::Float(f1 - f2)),
NumberPair::Integer(i1, i2) => self.push(Value::Integer(i1 - i2)),
},
OpCode::OpMul => match self.pop_number_pair()? {
NumberPair::Floats(f1, f2) => self.push(Value::Float(f1 * f2)),
NumberPair::Integer(i1, i2) => self.push(Value::Integer(i1 * i2)),
},
OpCode::OpDiv => match self.pop_number_pair()? {
NumberPair::Floats(f1, f2) => self.push(Value::Float(f1 / f2)),
NumberPair::Integer(i1, i2) => self.push(Value::Integer(i1 / i2)),
},
OpCode::OpInvert => {
let v = self.pop().as_bool()?;
self.push(Value::Bool(!v));
}
OpCode::OpNegate => match self.pop() {
Value::Integer(i) => self.push(Value::Integer(-i)),
Value::Float(f) => self.push(Value::Float(-f)),
v => {
return Err(Error::TypeError {
expected: "number (either int or float)",
actual: v.type_of(),
})
}
},
OpCode::OpEqual => {
let v2 = self.pop();
let v1 = self.pop();
let eq = match (v1, v2) {
(Value::Float(f), Value::Integer(i))
| (Value::Integer(i), Value::Float(f)) => f == (i as f64),
(v1, v2) => v1 == v2,
};
self.push(Value::Bool(eq))
}
OpCode::OpNull => self.push(Value::Null),
OpCode::OpTrue => self.push(Value::Bool(true)),
OpCode::OpFalse => self.push(Value::Bool(false)),
OpCode::OpAttrs(count) => self.run_attrset(count)?,
OpCode::OpAttrPath(count) => self.run_attr_path(count)?,
OpCode::OpList(count) => self.run_list(count)?,
OpCode::OpInterpolate(count) => self.run_interpolate(count)?,
}
if self.ip == self.chunk.code.len() {
return Ok(self.pop());
}
}
}
// Construct runtime representation of an attr path (essentially
// just a list of strings).
//
// The difference to the list construction operation is that this
// forces all elements into strings, as attribute set keys are
// required to be strict in Nix.
fn run_attr_path(&mut self, count: usize) -> EvalResult<()> {
debug_assert!(count > 1, "AttrPath needs at least two fragments");
let mut path = Vec::with_capacity(count);
for _ in 0..count {
path.push(self.pop().as_string()?);
}
self.push(Value::AttrPath(path));
Ok(())
}
fn run_attrset(&mut self, count: usize) -> EvalResult<()> {
let attrs = NixAttrs::construct(count, self.stack.split_off(self.stack.len() - count * 2))?;
self.push(Value::Attrs(Rc::new(attrs)));
Ok(())
}
// Interpolate string fragments by popping the specified number of
// fragments of the stack, evaluating them to strings, and pushing
// the concatenated result string back on the stack.
fn run_interpolate(&mut self, count: usize) -> EvalResult<()> {
let mut out = String::new();
for _ in 0..count {
out.push_str(&self.pop().as_string()?.0);
}
self.push(Value::String(NixString(out)));
Ok(())
}
// Construct runtime representation of a list. Because the list
// items are on the stack in reverse order, the vector is created
// initialised and elements are directly assigned to their
// respective indices.
fn run_list(&mut self, count: usize) -> EvalResult<()> {
let mut list = vec![Value::Null; count];
for idx in 0..count {
list[count - idx - 1] = self.pop();
}
self.push(Value::List(NixList(list)));
Ok(())
}
}
#[derive(Clone, Copy, Debug, PartialEq)]
pub enum NumberPair {
Floats(f64, f64),
Integer(i64, i64),
}
pub fn run_chunk(chunk: Chunk) -> EvalResult<Value> {
let mut vm = VM {
chunk,
ip: 0,
stack: vec![],
};
vm.run()
}