1
Fork 0
mirror of https://github.com/RGBCube/uutils-coreutils synced 2025-07-29 12:07:46 +00:00
This commit is contained in:
electricboogie 2021-04-18 13:43:41 -05:00
parent da94e35044
commit dad7761be9
3 changed files with 83 additions and 20 deletions

View file

@ -74,6 +74,18 @@ impl ExternalSorter {
self
}
/// Sorts a given iterator, returning a new iterator with items
pub fn sort<T, I>(
&self,
iterator: I,
) -> Result<SortedIterator<T, impl Fn(&T, &T) -> Ordering + Send + Sync>, Error>
where
T: Sortable + Ord,
I: Iterator<Item = T>,
{
self.sort_by(iterator, |a, b| a.cmp(b))
}
/// Sorts a given iterator with a comparator function, returning a new iterator with items
pub fn sort_by<T, I, F>(&self, iterator: I, cmp: F) -> Result<SortedIterator<T, F>, Error>
where
@ -255,3 +267,63 @@ impl<T: Sortable, F: Fn(&T, &T) -> Ordering> Iterator for SortedIterator<T, F> {
})
}
}
#[cfg(test)]
pub mod test {
use super::*;
use byteorder::{ReadBytesExt, WriteBytesExt};
#[test]
fn test_smaller_than_segment() {
let sorter = ExternalSorter::new();
let data: Vec<u32> = (0..100u32).collect();
let data_rev: Vec<u32> = data.iter().rev().cloned().collect();
let sorted_iter = sorter.sort(data_rev.into_iter()).unwrap();
// should not have used any segments (all in memory)
assert_eq!(sorted_iter.segments_file.len(), 0);
let sorted_data: Vec<u32> = sorted_iter.collect();
assert_eq!(data, sorted_data);
}
#[test]
fn test_multiple_segments() {
let sorter = ExternalSorter::new().with_segment_size(100);
let data: Vec<u32> = (0..1000u32).collect();
let data_rev: Vec<u32> = data.iter().rev().cloned().collect();
let sorted_iter = sorter.sort(data_rev.into_iter()).unwrap();
assert_eq!(sorted_iter.segments_file.len(), 10);
let sorted_data: Vec<u32> = sorted_iter.collect();
assert_eq!(data, sorted_data);
}
#[test]
fn test_parallel() {
let sorter = ExternalSorter::new()
.with_segment_size(100)
.with_parallel_sort();
let data: Vec<u32> = (0..1000u32).collect();
let data_rev: Vec<u32> = data.iter().rev().cloned().collect();
let sorted_iter = sorter.sort(data_rev.into_iter()).unwrap();
assert_eq!(sorted_iter.segments_file.len(), 10);
let sorted_data: Vec<u32> = sorted_iter.collect();
assert_eq!(data, sorted_data);
}
impl Sortable for u32 {
fn encode<W: Write>(&self, writer: &mut W) {
writer.write_u32::<byteorder::LittleEndian>(*self).unwrap();
}
fn decode<R: Read>(reader: &mut R) -> Option<u32> {
reader.read_u32::<byteorder::LittleEndian>().ok()
}
}
}

View file

@ -90,7 +90,7 @@ static NEGATIVE: char = '-';
static POSITIVE: char = '+';
static DEFAULT_TMPDIR: &str = r"/tmp";
// 16GB buffer for Vec<Lines> before we dump to disk
// 16GB buffer for Vec<Line> before we dump to disk
static DEFAULT_BUF_SIZE: usize = 16000000000;
#[derive(Eq, Ord, PartialEq, PartialOrd, Clone)]
@ -890,11 +890,7 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
.map(String::from)
.unwrap_or(format!("{}", DEFAULT_BUF_SIZE));
if human_numeric_convert(&input) < 128000 {
panic!("sort will not operate with less than 128K of memory.");
} else {
human_numeric_convert(&input)
}
}
}

View file

@ -8,6 +8,16 @@ fn test_helper(file_name: &str, args: &str) {
.stdout_is_fixture(format!("{}.expected", file_name));
}
#[test]
fn test_larger_than_specified_segment() {
new_ucmd!()
.arg("-n")
.arg("-S 100")
.arg("numeric_unsorted_ints.txt")
.succeeds()
.stdout_is_fixture(format!("{}", "numeric_unsorted_ints.expected"));
}
#[test]
fn test_months_whitespace() {
test_helper("months-whitespace", "-M");
@ -100,21 +110,6 @@ fn test_random_shuffle_two_runs_not_the_same() {
assert_ne!(result, unexpected);
}
#[test]
fn test_random_shuffle_contains_two_runs_not_the_same() {
// check to verify that two random shuffles are not equal; this has the
// potential to fail in the unlikely event that random order is the same
// as the starting order, or if both random sorts end up having the same order.
const FILE: &'static str = "default_unsorted_ints.expected";
let (at, _ucmd) = at_and_ucmd!();
let result = new_ucmd!().arg("-R").arg(FILE).run().stdout_move_str();
let expected = at.read(FILE);
let unexpected = new_ucmd!().arg("-R").arg(FILE).run().stdout_move_str();
assert_ne!(result, expected);
assert_ne!(result, unexpected);
}
#[test]
fn test_numeric_floats_and_ints() {
test_helper("numeric_floats_and_ints", "-n");