nushell/crates/nu-command/src/strings/split/chars.rs
Stefan Holderbach a52386e837
Box ShellError in Value::Error (#8375)
# Description

Our `ShellError` at the moment has a `std::mem::size_of<ShellError>` of
136 bytes (on AMD64). As a result `Value` directly storing the struct
also required 136 bytes (thanks to alignment requirements).

This change stores the `Value::Error` `ShellError` on the heap.

Pro:
- Value now needs just 80 bytes
- Should be 1 cacheline less (still at least 2 cachelines)

Con:
- More small heap allocations when dealing with `Value::Error`
  - More heap fragmentation
  - Potential for additional required memcopies

# Further code changes

Includes a small refactor of `try` due to a type mismatch in its large
match.

# User-Facing Changes

None for regular users.

Plugin authors may have to update their matches on `Value` if they use
`nu-protocol`

Needs benchmarking to see if there is a benefit in real world workloads.
**Update** small improvements in runtime for workloads with high volume
of values. Significant reduction in maximum resident set size, when many
values are held in memory.

# Tests + Formatting
2023-03-12 09:57:27 +01:00

138 lines
3.9 KiB
Rust

use crate::grapheme_flags;
use nu_protocol::{
ast::Call,
engine::{Command, EngineState, Stack},
Category, Example, PipelineData, ShellError, Signature, Span, Type, Value,
};
use unicode_segmentation::UnicodeSegmentation;
#[derive(Clone)]
pub struct SubCommand;
impl Command for SubCommand {
fn name(&self) -> &str {
"split chars"
}
fn signature(&self) -> Signature {
Signature::build("split chars")
.input_output_types(vec![(Type::String, Type::List(Box::new(Type::String)))])
.switch("grapheme-clusters", "split on grapheme clusters", Some('g'))
.switch(
"code-points",
"split on code points (default; splits combined characters)",
Some('c'),
)
.vectorizes_over_list(true)
.category(Category::Strings)
}
fn usage(&self) -> &str {
"Split a string into a list of characters."
}
fn search_terms(&self) -> Vec<&str> {
vec!["character", "separate", "divide"]
}
fn examples(&self) -> Vec<Example> {
vec![
Example {
description: "Split the string into a list of characters",
example: "'hello' | split chars",
result: Some(Value::List {
vals: vec![
Value::test_string("h"),
Value::test_string("e"),
Value::test_string("l"),
Value::test_string("l"),
Value::test_string("o"),
],
span: Span::test_data(),
}),
},
Example {
description: "Split on grapheme clusters",
example: "'🇯🇵ほげ' | split chars -g",
result: Some(Value::List {
vals: vec![
Value::test_string("🇯🇵"),
Value::test_string(""),
Value::test_string(""),
],
span: Span::test_data(),
}),
},
]
}
fn run(
&self,
engine_state: &EngineState,
_stack: &mut Stack,
call: &Call,
input: PipelineData,
) -> Result<PipelineData, ShellError> {
split_chars(engine_state, call, input)
}
}
fn split_chars(
engine_state: &EngineState,
call: &Call,
input: PipelineData,
) -> Result<PipelineData, ShellError> {
let span = call.head;
let graphemes = grapheme_flags(call)?;
input.flat_map(
move |x| split_chars_helper(&x, span, graphemes),
engine_state.ctrlc.clone(),
)
}
fn split_chars_helper(v: &Value, name: Span, graphemes: bool) -> Vec<Value> {
match v.span() {
Ok(v_span) => {
if let Ok(s) = v.as_string() {
if graphemes {
s.graphemes(true)
.collect::<Vec<_>>()
.into_iter()
.map(move |x| Value::string(x, v_span))
.collect()
} else {
s.chars()
.collect::<Vec<_>>()
.into_iter()
.map(move |x| Value::string(x, v_span))
.collect()
}
} else {
vec![Value::Error {
error: Box::new(ShellError::PipelineMismatch {
exp_input_type: "string".into(),
dst_span: name,
src_span: v_span,
}),
}]
}
}
Err(error) => vec![Value::Error {
error: Box::new(error),
}],
}
}
#[cfg(test)]
mod test {
use super::*;
#[test]
fn test_examples() {
use crate::test_examples;
test_examples(SubCommand {})
}
}