nushell/crates/nu-protocol/src/debugger/debugger_trait.rs
Ian Manske 6fd854ed9f
Replace ExternalStream with new ByteStream type (#12774)
# Description
This PR introduces a `ByteStream` type which is a `Read`-able stream of
bytes. Internally, it has an enum over three different byte stream
sources:
```rust
pub enum ByteStreamSource {
    Read(Box<dyn Read + Send + 'static>),
    File(File),
    Child(ChildProcess),
}
```

This is in comparison to the current `RawStream` type, which is an
`Iterator<Item = Vec<u8>>` and has to allocate for each read chunk.

Currently, `PipelineData::ExternalStream` serves a weird dual role where
it is either external command output or a wrapper around `RawStream`.
`ByteStream` makes this distinction more clear (via `ByteStreamSource`)
and replaces `PipelineData::ExternalStream` in this PR:
```rust
pub enum PipelineData {
    Empty,
    Value(Value, Option<PipelineMetadata>),
    ListStream(ListStream, Option<PipelineMetadata>),
    ByteStream(ByteStream, Option<PipelineMetadata>),
}
```

The PR is relatively large, but a decent amount of it is just repetitive
changes.

This PR fixes #7017, fixes #10763, and fixes #12369.

This PR also improves performance when piping external commands. Nushell
should, in most cases, have competitive pipeline throughput compared to,
e.g., bash.
| Command | Before (MB/s) | After (MB/s) | Bash (MB/s) |
| -------------------------------------------------- | -------------:|
------------:| -----------:|
| `throughput \| rg 'x'` | 3059 | 3744 | 3739 |
| `throughput \| nu --testbin relay o> /dev/null` | 3508 | 8087 | 8136 |

# User-Facing Changes
- This is a breaking change for the plugin communication protocol,
because the `ExternalStreamInfo` was replaced with `ByteStreamInfo`.
Plugins now only have to deal with a single input stream, as opposed to
the previous three streams: stdout, stderr, and exit code.
- The output of `describe` has been changed for external/byte streams.
- Temporary breaking change: `bytes starts-with` no longer works with
byte streams. This is to keep the PR smaller, and `bytes ends-with`
already does not work on byte streams.
- If a process core dumped, then instead of having a `Value::Error` in
the `exit_code` column of the output returned from `complete`, it now is
a `Value::Int` with the negation of the signal number.

# After Submitting
- Update docs and book as necessary
- Release notes (e.g., plugin protocol changes)
- Adapt/convert commands to work with byte streams (high priority is
`str length`, `bytes starts-with`, and maybe `bytes ends-with`).
- Refactor the `tee` code, Devyn has already done some work on this.

---------

Co-authored-by: Devyn Cairns <devyn.cairns@gmail.com>
2024-05-16 07:11:18 -07:00

151 lines
5.5 KiB
Rust

//! Traits related to debugging
//!
//! The purpose of DebugContext is achieving static dispatch on `eval_xxx()` calls.
//! The main Debugger trait is intended to be used as a trait object.
//!
//! The debugging information is stored in `EngineState` as the `debugger` field storing a `Debugger`
//! trait object behind `Arc` and `Mutex`. To evaluate something (e.g., a block), first create a
//! `Debugger` trait object (such as the `Profiler`). Then, add it to engine state via
//! `engine_state.activate_debugger()`. This sets the internal state of EngineState to the debugging
//! mode and calls `Debugger::activate()`. Now, you can call `eval_xxx::<WithDebug>()`. When you're
//! done, call `engine_state.deactivate_debugger()` which calls `Debugger::deactivate()`, sets the
//! EngineState into non-debugging mode, and returns the original mutated `Debugger` trait object.
//! (`NoopDebugger` is placed in its place inside `EngineState`.) After deactivating, you can call
//! `Debugger::report()` to get some output from the debugger, if necessary.
use crate::{
ast::{Block, PipelineElement},
engine::EngineState,
PipelineData, ShellError, Span, Value,
};
use std::{fmt::Debug, ops::DerefMut};
/// Trait used for static dispatch of `eval_xxx()` evaluator calls
///
/// DebugContext implements the same interface as Debugger (except activate() and deactivate(). It
/// is intended to be implemented only by two structs
/// * WithDebug which calls down to the Debugger methods
/// * WithoutDebug with default implementation, i.e., empty calls to be optimized away
pub trait DebugContext: Clone + Copy + Debug {
/// Called when the evaluator enters a block
#[allow(unused_variables)]
fn enter_block(engine_state: &EngineState, block: &Block) {}
/// Called when the evaluator leaves a block
#[allow(unused_variables)]
fn leave_block(engine_state: &EngineState, block: &Block) {}
/// Called when the evaluator enters a pipeline element
#[allow(unused_variables)]
fn enter_element(engine_state: &EngineState, element: &PipelineElement) {}
/// Called when the evaluator leaves a pipeline element
#[allow(unused_variables)]
fn leave_element(
engine_state: &EngineState,
element: &PipelineElement,
result: &Result<PipelineData, ShellError>,
) {
}
}
/// Marker struct signalizing that evaluation should use a Debugger
///
/// Trait methods call to Debugger trait object inside the supplied EngineState.
#[derive(Clone, Copy, Debug)]
pub struct WithDebug;
impl DebugContext for WithDebug {
fn enter_block(engine_state: &EngineState, block: &Block) {
if let Ok(mut debugger) = engine_state.debugger.lock() {
debugger.deref_mut().enter_block(engine_state, block);
}
}
fn leave_block(engine_state: &EngineState, block: &Block) {
if let Ok(mut debugger) = engine_state.debugger.lock() {
debugger.deref_mut().leave_block(engine_state, block);
}
}
fn enter_element(engine_state: &EngineState, element: &PipelineElement) {
if let Ok(mut debugger) = engine_state.debugger.lock() {
debugger.deref_mut().enter_element(engine_state, element);
}
}
fn leave_element(
engine_state: &EngineState,
element: &PipelineElement,
result: &Result<PipelineData, ShellError>,
) {
if let Ok(mut debugger) = engine_state.debugger.lock() {
debugger
.deref_mut()
.leave_element(engine_state, element, result);
}
}
}
/// Marker struct signalizing that evaluation should NOT use a Debugger
///
/// Trait methods are empty calls to be optimized away.
#[derive(Clone, Copy, Debug)]
pub struct WithoutDebug;
impl DebugContext for WithoutDebug {}
/// Debugger trait that every debugger needs to implement.
///
/// By default, its methods are empty. Not every Debugger needs to implement all of them.
pub trait Debugger: Send + Debug {
/// Called by EngineState::activate_debugger().
///
/// Intended for initializing the debugger.
fn activate(&mut self) {}
/// Called by EngineState::deactivate_debugger().
///
/// Intended for wrapping up the debugger after a debugging session before returning back to
/// normal evaluation without debugging.
fn deactivate(&mut self) {}
/// Called when the evaluator enters a block
#[allow(unused_variables)]
fn enter_block(&mut self, engine_state: &EngineState, block: &Block) {}
/// Called when the evaluator leaves a block
#[allow(unused_variables)]
fn leave_block(&mut self, engine_state: &EngineState, block: &Block) {}
/// Called when the evaluator enters a pipeline element
#[allow(unused_variables)]
fn enter_element(&mut self, engine_state: &EngineState, pipeline_element: &PipelineElement) {}
/// Called when the evaluator leaves a pipeline element
#[allow(unused_variables)]
fn leave_element(
&mut self,
engine_state: &EngineState,
element: &PipelineElement,
result: &Result<PipelineData, ShellError>,
) {
}
/// Create a final report as a Value
///
/// Intended to be called after deactivate()
#[allow(unused_variables)]
fn report(&self, engine_state: &EngineState, debugger_span: Span) -> Result<Value, ShellError> {
Ok(Value::nothing(debugger_span))
}
}
/// A debugger that does nothing
///
/// Used as a placeholder debugger when not debugging.
#[derive(Debug)]
pub struct NoopDebugger;
impl Debugger for NoopDebugger {}