From 06b6b564ae2436c853a8a360243e4044f0b64db9 Mon Sep 17 00:00:00 2001
From: Apertis package maintainers <packagers@lists.apertis.org>
Date: Fri, 11 Jun 2021 14:37:30 +0200
Subject: [PATCH] d/patches: backport improvements to the `sort` command
The post-installation script of at least one packages makes use of the
`-k` option for `sort`, which isn't implemented in release 0.0.6 of
`rust-coreutils`.
As this has since been implemented upstream, this commit picks the
corresponding patches so we can benefit of those improvements.
Signed-off-by: Arnaud Ferraris <arnaud.ferraris@collabora.com>
---
debian/patches/Ignore-a-test.patch | 60 +
...table-sort-ignore-non-printing-month.patch | 2303 +++++++++++++++++
...s-fixes-and-performance-improvements.patch | 502 ++++
debian/patches/series | 6 +
.../sort-implement-k-and-t-support.patch | 1333 ++++++++++
5 files changed, 4204 insertions(+)
create mode 100644 debian/patches/Ignore-a-test.patch
create mode 100644 debian/patches/Sort-Implement-stable-sort-ignore-non-printing-month.patch
create mode 100644 debian/patches/Sort-Various-fixes-and-performance-improvements.patch
create mode 100644 debian/patches/sort-implement-k-and-t-support.patch
diff --git a/debian/patches/Ignore-a-test.patch b/debian/patches/Ignore-a-test.patch
new file mode 100644
index 0000000..1584ee6
--- /dev/null
+++ b/debian/patches/Ignore-a-test.patch
@@ -0,0 +1,60 @@
+From: Sylvestre Ledru <sylvestre@debian.org>
+Date: Fri, 9 Apr 2021 10:14:41 +0200
+Subject: Ignore a test (#2053)
+
+* Disable chksum: test_arg_overrides_stdin
+fails often with:
+
+---- test_cksum::test_arg_overrides_stdin stdout ----
+current_directory_resolved:
+touch: /tmp/.tmpv9hydc/a
+run: /target/x86_64-unknown-linux-gnu/debug/coreutils cksum a
+thread 'test_cksum::test_arg_overrides_stdin' panicked at 'Broken pipe (os error 32)', tests/common/util.rs:742:37
+note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
+
+* rustfmt the recent change
+
+Origin: upstream, https://github.com/uutils/coreutils/commit/9ae4928b7b4f43495921fe8ac6128ea226d20dbd
+---
+ src/uu/sort/src/sort.rs | 6 ++----
+ src/uu/stdbuf/src/stdbuf.rs | 3 +--
+ 2 files changed, 3 insertions(+), 6 deletions(-)
+
+diff --git a/src/uu/sort/src/sort.rs b/src/uu/sort/src/sort.rs
+index 08936ff..cf6c258 100644
+--- a/src/uu/sort/src/sort.rs
++++ b/src/uu/sort/src/sort.rs
+@@ -665,9 +665,7 @@ fn get_leading_gen(a: &str) -> String {
+ for c in p_iter.to_owned() {
+ let next_char_numeric = p_iter.peek().unwrap_or(&'\0').is_numeric();
+ // Only general numeric recognizes e notation and, see block below, the '+' sign
+- if (c.eq(&'e') && !next_char_numeric)
+- || (c.eq(&'E') && !next_char_numeric)
+- {
++ if (c.eq(&'e') && !next_char_numeric) || (c.eq(&'E') && !next_char_numeric) {
+ r = a.split(c).next().unwrap_or("").to_owned();
+ break;
+ // If positive sign and next char is not numeric, split at postive sign at keep trailing numbers
+@@ -813,7 +811,7 @@ fn human_numeric_convert(a: &str) -> f64 {
+ 'E' => 1E18,
+ 'Z' => 1E21,
+ 'Y' => 1E24,
+- _ => 1f64,
++ _ => 1f64,
+ };
+ num_part * suffix
+ }
+diff --git a/src/uu/stdbuf/src/stdbuf.rs b/src/uu/stdbuf/src/stdbuf.rs
+index a61ba96..67ed9a8 100644
+--- a/src/uu/stdbuf/src/stdbuf.rs
++++ b/src/uu/stdbuf/src/stdbuf.rs
+@@ -80,8 +80,7 @@ fn print_version() {
+ fn print_usage(opts: &Options) {
+ let brief = "Run COMMAND, with modified buffering operations for its standard streams\n \
+ Mandatory arguments to long options are mandatory for short options too.";
+- let explanation =
+- "If MODE is 'L' the corresponding stream will be line buffered.\n \
++ let explanation = "If MODE is 'L' the corresponding stream will be line buffered.\n \
+ This option is invalid with standard input.\n\n \
+ If MODE is '0' the corresponding stream will be unbuffered.\n\n \
+ Otherwise MODE is a number which may be followed by one of the following:\n\n \
diff --git a/debian/patches/Sort-Implement-stable-sort-ignore-non-printing-month.patch b/debian/patches/Sort-Implement-stable-sort-ignore-non-printing-month.patch
new file mode 100644
index 0000000..f023d25
--- /dev/null
+++ b/debian/patches/Sort-Implement-stable-sort-ignore-non-printing-month.patch
@@ -0,0 +1,2303 @@
+From: electricboogie <32370782+electricboogie@users.noreply.github.com>
+Date: Thu, 8 Apr 2021 15:07:09 -0500
+Subject: Sort: Implement stable sort, ignore non-printing, month sort dedup,
+ auto parallel sort through rayon, zero terminated sort, check silent (#2008)
+
+Origin: upstream, https://github.com/uutils/coreutils/commit/8474249e5f301068565c2d62f04b04d40b0b5817
+---
+ Cargo.lock | 20 +-
+ src/uu/sort/Cargo.toml | 3 +-
+ src/uu/sort/src/sort.rs | 558 ++++++++++++++++-----
+ tests/by-util/test_sort.rs | 252 ++++++++--
+ .../sort/exponents-positive-general.expected | 12 +
+ tests/fixtures/sort/exponents-positive-general.txt | 12 +
+ .../sort/exponents-positive-numeric.expected | 12 +
+ tests/fixtures/sort/exponents-positive-numeric.txt | 12 +
+ .../sort/human-mixed-inputs-reverse.expected | 37 ++
+ tests/fixtures/sort/human-mixed-inputs-reverse.txt | 37 ++
+ .../sort/human-mixed-inputs-stable.expected | 37 ++
+ tests/fixtures/sort/human-mixed-inputs-stable.txt | 37 ++
+ .../sort/human-mixed-inputs-unique.expected | 13 +
+ tests/fixtures/sort/human-mixed-inputs-unique.txt | 37 ++
+ tests/fixtures/sort/human-mixed-inputs.expected | 37 ++
+ tests/fixtures/sort/human-mixed-inputs.txt | 46 ++
+ .../sort/mixed_floats_ints_chars_numeric.expected | 30 ++
+ .../sort/mixed_floats_ints_chars_numeric.txt | 30 ++
+ ...ixed_floats_ints_chars_numeric_reverse.expected | 30 ++
+ ...oats_ints_chars_numeric_reverse_stable.expected | 30 ++
+ ...ed_floats_ints_chars_numeric_reverse_stable.txt | 30 ++
+ ...mixed_floats_ints_chars_numeric_stable.expected | 30 ++
+ .../mixed_floats_ints_chars_numeric_stable.txt | 30 ++
+ ...mixed_floats_ints_chars_numeric_unique.expected | 20 +
+ .../mixed_floats_ints_chars_numeric_unique.txt | 30 ++
+ ...oats_ints_chars_numeric_unique_reverse.expected | 20 +
+ ...ed_floats_ints_chars_numeric_unique_reverse.txt | 30 ++
+ ...loats_ints_chars_numeric_unique_stable.expected | 20 +
+ ...xed_floats_ints_chars_numeric_unique_stable.txt | 30 ++
+ tests/fixtures/sort/months-dedup.expected | 6 +
+ tests/fixtures/sort/months-dedup.txt | 37 ++
+ .../sort/numeric-floats-with-nan2.expected | 23 +
+ tests/fixtures/sort/numeric-floats-with-nan2.txt | 23 +
+ tests/fixtures/sort/zero-terminated.expected | 1 +
+ tests/fixtures/sort/zero-terminated.txt | 1 +
+ 35 files changed, 1422 insertions(+), 191 deletions(-)
+ create mode 100644 tests/fixtures/sort/exponents-positive-general.expected
+ create mode 100644 tests/fixtures/sort/exponents-positive-general.txt
+ create mode 100644 tests/fixtures/sort/exponents-positive-numeric.expected
+ create mode 100644 tests/fixtures/sort/exponents-positive-numeric.txt
+ create mode 100644 tests/fixtures/sort/human-mixed-inputs-reverse.expected
+ create mode 100644 tests/fixtures/sort/human-mixed-inputs-reverse.txt
+ create mode 100644 tests/fixtures/sort/human-mixed-inputs-stable.expected
+ create mode 100644 tests/fixtures/sort/human-mixed-inputs-stable.txt
+ create mode 100644 tests/fixtures/sort/human-mixed-inputs-unique.expected
+ create mode 100644 tests/fixtures/sort/human-mixed-inputs-unique.txt
+ create mode 100644 tests/fixtures/sort/human-mixed-inputs.expected
+ create mode 100644 tests/fixtures/sort/human-mixed-inputs.txt
+ create mode 100644 tests/fixtures/sort/mixed_floats_ints_chars_numeric.expected
+ create mode 100644 tests/fixtures/sort/mixed_floats_ints_chars_numeric.txt
+ create mode 100644 tests/fixtures/sort/mixed_floats_ints_chars_numeric_reverse.expected
+ create mode 100644 tests/fixtures/sort/mixed_floats_ints_chars_numeric_reverse_stable.expected
+ create mode 100644 tests/fixtures/sort/mixed_floats_ints_chars_numeric_reverse_stable.txt
+ create mode 100644 tests/fixtures/sort/mixed_floats_ints_chars_numeric_stable.expected
+ create mode 100644 tests/fixtures/sort/mixed_floats_ints_chars_numeric_stable.txt
+ create mode 100644 tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique.expected
+ create mode 100644 tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique.txt
+ create mode 100644 tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique_reverse.expected
+ create mode 100644 tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique_reverse.txt
+ create mode 100644 tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique_stable.expected
+ create mode 100644 tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique_stable.txt
+ create mode 100644 tests/fixtures/sort/months-dedup.expected
+ create mode 100644 tests/fixtures/sort/months-dedup.txt
+ create mode 100644 tests/fixtures/sort/numeric-floats-with-nan2.expected
+ create mode 100644 tests/fixtures/sort/numeric-floats-with-nan2.txt
+ create mode 100644 tests/fixtures/sort/zero-terminated.expected
+ create mode 100644 tests/fixtures/sort/zero-terminated.txt
+
+diff --git a/Cargo.lock b/Cargo.lock
+index 88ac18c..53eb8b4 100644
+--- a/Cargo.lock
++++ b/Cargo.lock
+@@ -994,12 +994,6 @@ dependencies = [
+ "maybe-uninit",
+ ]
+
+-[[package]]
+-name = "static_assertions"
+-version = "1.1.0"
+-source = "registry+https://github.com/rust-lang/crates.io-index"
+-checksum = "a2eb9349b6444b326872e140eb1cf5e7c522154d69e7a0ffb0fb81c06b37543f"
+-
+ [[package]]
+ name = "strsim"
+ version = "0.8.0"
+@@ -1126,17 +1120,6 @@ dependencies = [
+ "winapi 0.3.9",
+ ]
+
+-[[package]]
+-name = "twox-hash"
+-version = "1.6.0"
+-source = "registry+https://github.com/rust-lang/crates.io-index"
+-checksum = "04f8ab788026715fa63b31960869617cba39117e520eb415b0139543e325ab59"
+-dependencies = [
+- "cfg-if 0.1.10",
+- "rand 0.7.3",
+- "static_assertions",
+-]
+-
+ [[package]]
+ name = "typenum"
+ version = "1.13.0"
+@@ -2028,10 +2011,11 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
+ checksum = "81feaf23fcf9fd3df43afd736c6785292d6f8ce9d202ac0dc7585c2a49879762"
+ dependencies = [
+ "clap",
++ "fnv",
+ "itertools",
+ "rand 0.7.3",
++ "rayon",
+ "semver",
+- "twox-hash",
+ "uucore",
+ "uucore_procs",
+ ]
+diff --git a/src/uu/sort/Cargo.toml b/src/uu/sort/Cargo.toml
+index 16ae0cb..cfcd74d 100644
+--- a/src/uu/sort/Cargo.toml
++++ b/src/uu/sort/Cargo.toml
+@@ -15,9 +15,10 @@ edition = "2018"
+ path = "src/sort.rs"
+
+ [dependencies]
++rayon = "1.5"
+ rand = "0.7"
+ clap = "2.33"
+-twox-hash = "1.6.0"
++fnv = "1.0.7"
+ itertools = "0.9"
+ semver = "0.9.0"
+ uucore = { version=">=0.0.8", package="uucore", path="../../uucore", features=["fs"] }
+diff --git a/src/uu/sort/src/sort.rs b/src/uu/sort/src/sort.rs
+index 6c29ad9..08936ff 100644
+--- a/src/uu/sort/src/sort.rs
++++ b/src/uu/sort/src/sort.rs
+@@ -7,23 +7,29 @@
+ // * file that was distributed with this source code.
+ #![allow(dead_code)]
+
++// Although these links don't always seem to describe reality, check out the POSIX and GNU specs:
++// https://pubs.opengroup.org/onlinepubs/9699919799/utilities/sort.html
++// https://www.gnu.org/software/coreutils/manual/html_node/sort-invocation.html
++
+ // spell-checker:ignore (ToDO) outfile nondictionary
+ #[macro_use]
+ extern crate uucore;
+
+ use clap::{App, Arg};
++use fnv::FnvHasher;
+ use itertools::Itertools;
+ use rand::distributions::Alphanumeric;
+ use rand::{thread_rng, Rng};
++use rayon::prelude::*;
+ use semver::Version;
+ use std::cmp::Ordering;
+ use std::collections::BinaryHeap;
++use std::env;
+ use std::fs::File;
+ use std::hash::{Hash, Hasher};
+ use std::io::{stdin, stdout, BufRead, BufReader, BufWriter, Lines, Read, Write};
+ use std::mem::replace;
+ use std::path::Path;
+-use twox_hash::XxHash64;
+ use uucore::fs::is_stdin_interactive; // for Iterator::dedup()
+
+ static NAME: &str = "sort";
+@@ -33,27 +39,37 @@ static VERSION: &str = env!("CARGO_PKG_VERSION");
+ static OPT_HUMAN_NUMERIC_SORT: &str = "human-numeric-sort";
+ static OPT_MONTH_SORT: &str = "month-sort";
+ static OPT_NUMERIC_SORT: &str = "numeric-sort";
++static OPT_GENERAL_NUMERIC_SORT: &str = "general-numeric-sort";
+ static OPT_VERSION_SORT: &str = "version-sort";
+
+ static OPT_DICTIONARY_ORDER: &str = "dictionary-order";
+ static OPT_MERGE: &str = "merge";
+ static OPT_CHECK: &str = "check";
++static OPT_CHECK_SILENT: &str = "check-silent";
+ static OPT_IGNORE_CASE: &str = "ignore-case";
+ static OPT_IGNORE_BLANKS: &str = "ignore-blanks";
++static OPT_IGNORE_NONPRINTING: &str = "ignore-nonprinting";
+ static OPT_OUTPUT: &str = "output";
+ static OPT_REVERSE: &str = "reverse";
+ static OPT_STABLE: &str = "stable";
+ static OPT_UNIQUE: &str = "unique";
+ static OPT_RANDOM: &str = "random-sort";
++static OPT_ZERO_TERMINATED: &str = "zero-terminated";
++static OPT_PARALLEL: &str = "parallel";
++static OPT_FILES0_FROM: &str = "files0-from";
+
+ static ARG_FILES: &str = "files";
+
+ static DECIMAL_PT: char = '.';
+ static THOUSANDS_SEP: char = ',';
++static NEGATIVE: char = '-';
++static POSITIVE: char = '+';
++
+ #[derive(Eq, Ord, PartialEq, PartialOrd)]
+ enum SortMode {
+ Numeric,
+ HumanNumeric,
++ GeneralNumeric,
+ Month,
+ Version,
+ Default,
+@@ -67,10 +83,13 @@ struct Settings {
+ stable: bool,
+ unique: bool,
+ check: bool,
++ check_silent: bool,
+ random: bool,
+- compare_fns: Vec<fn(&str, &str) -> Ordering>,
++ compare_fn: fn(&str, &str) -> Ordering,
+ transform_fns: Vec<fn(&str) -> String>,
++ threads: String,
+ salt: String,
++ zero_terminated: bool,
+ }
+
+ impl Default for Settings {
+@@ -83,10 +102,13 @@ impl Default for Settings {
+ stable: false,
+ unique: false,
+ check: false,
++ check_silent: false,
+ random: false,
+- compare_fns: Vec::new(),
++ compare_fn: default_compare,
+ transform_fns: Vec::new(),
++ threads: String::new(),
+ salt: String::new(),
++ zero_terminated: false,
+ }
+ }
+ }
+@@ -206,6 +228,12 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
+ .long(OPT_NUMERIC_SORT)
+ .help("compare according to string numerical value"),
+ )
++ .arg(
++ Arg::with_name(OPT_GENERAL_NUMERIC_SORT)
++ .short("g")
++ .long(OPT_GENERAL_NUMERIC_SORT)
++ .help("compare according to string general numerical value"),
++ )
+ .arg(
+ Arg::with_name(OPT_VERSION_SORT)
+ .short("V")
+@@ -230,12 +258,24 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
+ .long(OPT_CHECK)
+ .help("check for sorted input; do not sort"),
+ )
++ .arg(
++ Arg::with_name(OPT_CHECK_SILENT)
++ .short("C")
++ .long(OPT_CHECK_SILENT)
++ .help("exit successfully if the given file is already sorted, and exit with status 1 otherwise. "),
++ )
+ .arg(
+ Arg::with_name(OPT_IGNORE_CASE)
+ .short("f")
+ .long(OPT_IGNORE_CASE)
+ .help("fold lower case to upper case characters"),
+ )
++ .arg(
++ Arg::with_name(OPT_IGNORE_NONPRINTING)
++ .short("-i")
++ .long(OPT_IGNORE_NONPRINTING)
++ .help("ignore nonprinting characters"),
++ )
+ .arg(
+ Arg::with_name(OPT_IGNORE_BLANKS)
+ .short("b")
+@@ -274,18 +314,65 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
+ .long(OPT_UNIQUE)
+ .help("output only the first of an equal run"),
+ )
++ .arg(
++ Arg::with_name(OPT_ZERO_TERMINATED)
++ .short("z")
++ .long(OPT_ZERO_TERMINATED)
++ .help("line delimiter is NUL, not newline"),
++ )
++ .arg(
++ Arg::with_name(OPT_PARALLEL)
++ .long(OPT_PARALLEL)
++ .help("change the number of threads running concurrently to N")
++ .takes_value(true)
++ .value_name("NUM_THREADS"),
++ )
++ .arg(
++ Arg::with_name(OPT_FILES0_FROM)
++ .long(OPT_FILES0_FROM)
++ .help("read input from the files specified by NUL-terminated NUL_FILES")
++ .takes_value(true)
++ .value_name("NUL_FILES")
++ .multiple(true),
++ )
+ .arg(Arg::with_name(ARG_FILES).multiple(true).takes_value(true))
+ .get_matches_from(args);
+
+- let mut files: Vec<String> = matches
+- .values_of(ARG_FILES)
+- .map(|v| v.map(ToString::to_string).collect())
+- .unwrap_or_default();
++ // check whether user specified a zero terminated list of files for input, otherwise read files from args
++ let mut files: Vec<String> = if matches.is_present(OPT_FILES0_FROM) {
++ let files0_from: Vec<String> = matches
++ .values_of(OPT_FILES0_FROM)
++ .map(|v| v.map(ToString::to_string).collect())
++ .unwrap_or_default();
++
++ let mut files = Vec::new();
++ for path in &files0_from {
++ let (reader, _) = open(path.as_str()).expect("Could not read from file specified.");
++ let buf_reader = BufReader::new(reader);
++ for line in buf_reader.split(b'\0') {
++ if let Ok(n) = line {
++ files.push(
++ std::str::from_utf8(&n)
++ .expect("Could not parse zero terminated string from input.")
++ .to_string(),
++ );
++ }
++ }
++ }
++ files
++ } else {
++ matches
++ .values_of(ARG_FILES)
++ .map(|v| v.map(ToString::to_string).collect())
++ .unwrap_or_default()
++ };
+
+ settings.mode = if matches.is_present(OPT_HUMAN_NUMERIC_SORT) {
+ SortMode::HumanNumeric
+ } else if matches.is_present(OPT_MONTH_SORT) {
+ SortMode::Month
++ } else if matches.is_present(OPT_GENERAL_NUMERIC_SORT) {
++ SortMode::GeneralNumeric
+ } else if matches.is_present(OPT_NUMERIC_SORT) {
+ SortMode::Numeric
+ } else if matches.is_present(OPT_VERSION_SORT) {
+@@ -294,12 +381,29 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
+ SortMode::Default
+ };
+
++ if matches.is_present(OPT_PARALLEL) {
++ // "0" is default - threads = num of cores
++ settings.threads = matches
++ .value_of(OPT_PARALLEL)
++ .map(String::from)
++ .unwrap_or("0".to_string());
++ env::set_var("RAYON_NUM_THREADS", &settings.threads);
++ }
++
+ if matches.is_present(OPT_DICTIONARY_ORDER) {
+ settings.transform_fns.push(remove_nondictionary_chars);
++ } else if matches.is_present(OPT_IGNORE_NONPRINTING) {
++ settings.transform_fns.push(remove_nonprinting_chars);
+ }
+
++ settings.zero_terminated = matches.is_present(OPT_ZERO_TERMINATED);
+ settings.merge = matches.is_present(OPT_MERGE);
++
+ settings.check = matches.is_present(OPT_CHECK);
++ if matches.is_present(OPT_CHECK_SILENT) {
++ settings.check_silent = matches.is_present(OPT_CHECK_SILENT);
++ settings.check = true;
++ };
+
+ if matches.is_present(OPT_IGNORE_CASE) {
+ settings.transform_fns.push(|s| s.to_uppercase());
+@@ -327,20 +431,14 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
+ crash!(1, "sort: extra operand `{}' not allowed with -c", files[1])
+ }
+
+- settings.compare_fns.push(match settings.mode {
++ settings.compare_fn = match settings.mode {
+ SortMode::Numeric => numeric_compare,
++ SortMode::GeneralNumeric => general_numeric_compare,
+ SortMode::HumanNumeric => human_numeric_size_compare,
+ SortMode::Month => month_compare,
+ SortMode::Version => version_compare,
+ SortMode::Default => default_compare,
+- });
+-
+- if !settings.stable {
+- match settings.mode {
+- SortMode::Default => {}
+- _ => settings.compare_fns.push(default_compare),
+- }
+- }
++ };
+
+ exec(files, &mut settings)
+ }
+@@ -359,67 +457,79 @@ fn exec(files: Vec<String>, settings: &mut Settings) -> i32 {
+
+ if settings.merge {
+ file_merger.push_file(buf_reader.lines());
+- } else if settings.check {
+- return exec_check_file(buf_reader.lines(), &settings);
++ } else if settings.zero_terminated {
++ for line in buf_reader.split(b'\0') {
++ if let Ok(n) = line {
++ lines.push(
++ std::str::from_utf8(&n)
++ .expect("Could not parse string from zero terminated input.")
++ .to_string(),
++ );
++ }
++ }
+ } else {
+ for line in buf_reader.lines() {
+ if let Ok(n) = line {
+ lines.push(n);
+- } else {
+- break;
+ }
+ }
+ }
+ }
+
+- sort_by(&mut lines, &settings);
++ if settings.check {
++ return exec_check_file(lines, &settings);
++ } else {
++ sort_by(&mut lines, &settings);
++ }
+
+ if settings.merge {
+ if settings.unique {
+- print_sorted(file_merger.dedup(), &settings.outfile)
++ print_sorted(file_merger.dedup(), &settings)
+ } else {
+- print_sorted(file_merger, &settings.outfile)
++ print_sorted(file_merger, &settings)
+ }
+- } else if settings.unique && settings.mode == SortMode::Numeric {
++ } else if settings.mode == SortMode::Month && settings.unique {
+ print_sorted(
+ lines
+ .iter()
+- .dedup_by(|a, b| num_sort_dedup(a) == num_sort_dedup(b)),
+- &settings.outfile,
++ .dedup_by(|a, b| get_months_dedup(a) == get_months_dedup(b)),
++ &settings,
+ )
+ } else if settings.unique {
+- print_sorted(lines.iter().dedup(), &settings.outfile)
++ print_sorted(
++ lines
++ .iter()
++ .dedup_by(|a, b| get_nums_dedup(a) == get_nums_dedup(b)),
++ &settings,
++ )
+ } else {
+- print_sorted(lines.iter(), &settings.outfile)
++ print_sorted(lines.iter(), &settings)
+ }
+
+ 0
+ }
+
+-fn exec_check_file(lines: Lines<BufReader<Box<dyn Read>>>, settings: &Settings) -> i32 {
++fn exec_check_file(unwrapped_lines: Vec<String>, settings: &Settings) -> i32 {
+ // errors yields the line before each disorder,
+ // plus the last line (quirk of .coalesce())
+- let unwrapped_lines = lines.filter_map(|maybe_line| {
+- if let Ok(line) = maybe_line {
+- Some(line)
+- } else {
+- None
+- }
+- });
+- let mut errors = unwrapped_lines
+- .enumerate()
+- .coalesce(|(last_i, last_line), (i, line)| {
+- if compare_by(&last_line, &line, &settings) == Ordering::Greater {
+- Err(((last_i, last_line), (i, line)))
+- } else {
+- Ok((i, line))
+- }
+- });
++ let mut errors =
++ unwrapped_lines
++ .iter()
++ .enumerate()
++ .coalesce(|(last_i, last_line), (i, line)| {
++ if compare_by(&last_line, &line, &settings) == Ordering::Greater {
++ Err(((last_i, last_line), (i, line)))
++ } else {
++ Ok((i, line))
++ }
++ });
+ if let Some((first_error_index, _line)) = errors.next() {
+ // Check for a second "error", as .coalesce() always returns the last
+ // line, no matter what our merging function does.
+ if let Some(_last_line_or_next_error) = errors.next() {
+- println!("sort: disorder in line {}", first_error_index);
++ if !settings.check_silent {
++ println!("sort: disorder in line {}", first_error_index);
++ };
+ 1
+ } else {
+ // first "error" was actually the last line.
+@@ -431,8 +541,9 @@ fn exec_check_file(lines: Lines<BufReader<Box<dyn Read>>>, settings: &Settings)
+ }
+ }
+
++#[inline(always)]
+ fn transform(line: &str, settings: &Settings) -> String {
+- let mut transformed = line.to_string();
++ let mut transformed = line.to_owned();
+ for transform_fn in &settings.transform_fns {
+ transformed = transform_fn(&transformed);
+ }
+@@ -440,8 +551,9 @@ fn transform(line: &str, settings: &Settings) -> String {
+ transformed
+ }
+
++#[inline(always)]
+ fn sort_by(lines: &mut Vec<String>, settings: &Settings) {
+- lines.sort_by(|a, b| compare_by(a, b, &settings))
++ lines.par_sort_by(|a, b| compare_by(a, b, &settings))
+ }
+
+ fn compare_by(a: &str, b: &str, settings: &Settings) -> Ordering {
+@@ -454,72 +566,198 @@ fn compare_by(a: &str, b: &str, settings: &Settings) -> Ordering {
+ (a, b)
+ };
+
+- for compare_fn in &settings.compare_fns {
+- let cmp: Ordering = if settings.random {
+- random_shuffle(a, b, settings.salt.clone())
++ // 1st Compare
++ let mut cmp: Ordering = if settings.random {
++ random_shuffle(a, b, settings.salt.clone())
++ } else {
++ (settings.compare_fn)(a, b)
++ };
++
++ // Call "last resort compare" on any equal
++ if cmp == Ordering::Equal {
++ if settings.random || settings.stable || settings.unique {
++ cmp = Ordering::Equal
+ } else {
+- compare_fn(a, b)
++ cmp = default_compare(a, b)
+ };
+- if cmp != Ordering::Equal {
+- if settings.reverse {
+- return cmp.reverse();
+- } else {
+- return cmp;
+- }
+- }
++ };
++
++ if settings.reverse {
++ return cmp.reverse();
++ } else {
++ return cmp;
+ }
+- Ordering::Equal
+ }
+
++// Test output against BSDs and GNU with their locale
++// env var set to lc_ctype=utf-8 to enjoy the exact same output.
++#[inline(always)]
+ fn default_compare(a: &str, b: &str) -> Ordering {
+ a.cmp(b)
+ }
+
+-fn get_leading_number(a: &str) -> &str {
++// This function does the initial detection of numeric lines.
++// Lines starting with a number or positive or negative sign.
++// It also strips the string of any thing that could never
++// be a number for the purposes of any type of numeric comparison.
++#[inline(always)]
++fn leading_num_common(a: &str) -> &str {
+ let mut s = "";
+- for c in a.chars() {
+- if !c.is_numeric() && !c.eq(&'-') && !c.eq(&' ') && !c.eq(&'.') && !c.eq(&',') {
+- s = a.trim().split(c).next().unwrap();
++ for (idx, c) in a.char_indices() {
++ // check whether char is numeric, whitespace or decimal point or thousand seperator
++ if !c.is_numeric()
++ && !c.is_whitespace()
++ && !c.eq(&DECIMAL_PT)
++ && !c.eq(&THOUSANDS_SEP)
++ // check for e notation
++ && !c.eq(&'e')
++ && !c.eq(&'E')
++ // check whether first char is + or -
++ && !a.chars().nth(0).unwrap_or('\0').eq(&POSITIVE)
++ && !a.chars().nth(0).unwrap_or('\0').eq(&NEGATIVE)
++ {
++ // Strip string of non-numeric trailing chars
++ s = &a[..idx];
+ break;
+ }
+- s = a.trim();
++ // If line is not a number line, return the line as is
++ s = a;
+ }
+- return s;
++ s
+ }
+
+-// Matches GNU behavior, see:
+-// https://www.gnu.org/software/coreutils/manual/html_node/sort-invocation.html
+-// Specifically *not* the same as sort -n | uniq
+-fn num_sort_dedup(a: &str) -> &str {
+- // Empty lines are dumped
+- if a.is_empty() {
+- return "0";
+- // And lines that don't begin numerically are dumped
+- } else if !a.trim().chars().nth(0).unwrap_or('\0').is_numeric() {
+- return "0";
++// This function cleans up the initial comparison done by leading_num_common for a numeric compare.
++// GNU sort does its numeric comparison through strnumcmp. However, we don't have or
++// may not want to use libc. Instead we emulate the GNU sort numeric compare by ignoring
++// those leading number lines GNU sort would not recognize. GNU numeric compare would
++// not recognize a positive sign or scientific/E notation so we strip those elements here.
++fn get_leading_num(a: &str) -> &str {
++ let mut s = "";
++ let b = leading_num_common(a);
++
++ // GNU numeric sort doesn't recognize '+' or 'e' notation so we strip
++ for (idx, c) in b.char_indices() {
++ if c.eq(&'e') || c.eq(&'E') || b.chars().nth(0).unwrap_or('\0').eq(&POSITIVE) {
++ s = &b[..idx];
++ break;
++ }
++ // If no further processing needed to be done, return the line as-is to be sorted
++ s = b;
++ }
++
++ // And empty number or non-number lines are to be treated as ‘0’ but only for numeric sort
++ // All '0'-ed lines will be sorted later, but only amongst themselves, during the so-called 'last resort comparison.'
++ if s.is_empty() {
++ s = "0";
++ };
++ s
++}
++
++// This function cleans up the initial comparison done by leading_num_common for a general numeric compare.
++// In contrast to numeric compare, GNU general numeric/FP sort *should* recognize positive signs and
++// scientific notation, so we strip those lines only after the end of the following numeric string.
++// For example, 5e10KFD would be 5e10 or 5x10^10 and +10000HFKJFK would become 10000.
++fn get_leading_gen(a: &str) -> String {
++ // Make this iter peekable to see if next char is numeric
++ let mut p_iter = leading_num_common(a).chars().peekable();
++ let mut r = String::new();
++ // Cleanup raw stripped strings
++ for c in p_iter.to_owned() {
++ let next_char_numeric = p_iter.peek().unwrap_or(&'\0').is_numeric();
++ // Only general numeric recognizes e notation and, see block below, the '+' sign
++ if (c.eq(&'e') && !next_char_numeric)
++ || (c.eq(&'E') && !next_char_numeric)
++ {
++ r = a.split(c).next().unwrap_or("").to_owned();
++ break;
++ // If positive sign and next char is not numeric, split at postive sign at keep trailing numbers
++ // There is a more elegant way to do this in Rust 1.45, std::str::strip_prefix
++ } else if c.eq(&POSITIVE) && !next_char_numeric {
++ let mut v: Vec<&str> = a.split(c).collect();
++ let x = v.split_off(1);
++ r = x.join("");
++ break;
++ // If no further processing needed to be done, return the line as-is to be sorted
++ } else {
++ r = a.to_owned();
++ }
++ }
++ r
++}
++
++fn get_months_dedup(a: &str) -> String {
++ let pattern = if a.trim().len().ge(&3) {
++ // Split at 3rd char and get first element of tuple ".0"
++ a.split_at(3).0
+ } else {
+- // Prepare lines for comparison of only the numerical leading numbers
+- return get_leading_number(a);
++ ""
+ };
++
++ let month = match pattern.to_uppercase().as_ref() {
++ "JAN" => Month::January,
++ "FEB" => Month::February,
++ "MAR" => Month::March,
++ "APR" => Month::April,
++ "MAY" => Month::May,
++ "JUN" => Month::June,
++ "JUL" => Month::July,
++ "AUG" => Month::August,
++ "SEP" => Month::September,
++ "OCT" => Month::October,
++ "NOV" => Month::November,
++ "DEC" => Month::December,
++ _ => Month::Unknown,
++ };
++
++ if month == Month::Unknown {
++ "".to_owned()
++ } else {
++ pattern.to_uppercase()
++ }
++}
++
++// *For all dedups/uniques we must compare leading numbers*
++// Also note numeric compare and unique output is specifically *not* the same as a "sort | uniq"
++// See: https://www.gnu.org/software/coreutils/manual/html_node/sort-invocation.html
++fn get_nums_dedup(a: &str) -> &str {
++ // Trim and remove any leading zeros
++ let s = a.trim().trim_start_matches('0');
++
++ // Get first char
++ let c = s.chars().nth(0).unwrap_or('\0');
++
++ // Empty lines and non-number lines are treated as the same for dedup
++ if s.is_empty() {
++ ""
++ } else if !c.eq(&NEGATIVE) && !c.is_numeric() {
++ ""
++ // Prepare lines for comparison of only the numerical leading numbers
++ } else {
++ get_leading_num(s)
++ }
+ }
+
+ /// Parse the beginning string into an f64, returning -inf instead of NaN on errors.
++#[inline(always)]
+ fn permissive_f64_parse(a: &str) -> f64 {
++ // Remove thousands seperators
++ let a = a.replace(THOUSANDS_SEP, "");
++
+ // GNU sort treats "NaN" as non-number in numeric, so it needs special care.
+- match a.parse::<f64>() {
++ // *Keep this trim before parse* despite what POSIX may say about -b and -n
++ // because GNU and BSD both seem to require it to match their behavior
++ match a.trim().parse::<f64>() {
+ Ok(a) if a.is_nan() => std::f64::NEG_INFINITY,
+ Ok(a) => a,
+ Err(_) => std::f64::NEG_INFINITY,
+ }
+ }
+
+-/// Compares two floats, with errors and non-numerics assumed to be -inf.
+-/// Stops coercing at the first non-numeric char.
+ fn numeric_compare(a: &str, b: &str) -> Ordering {
+ #![allow(clippy::comparison_chain)]
+
+- let sa = get_leading_number(a);
+- let sb = get_leading_number(b);
++ let sa = get_leading_num(a);
++ let sb = get_leading_num(b);
+
+ let fa = permissive_f64_parse(sa);
+ let fb = permissive_f64_parse(sb);
+@@ -534,19 +772,50 @@ fn numeric_compare(a: &str, b: &str) -> Ordering {
+ }
+ }
+
++/// Compares two floats, with errors and non-numerics assumed to be -inf.
++/// Stops coercing at the first non-numeric char.
++fn general_numeric_compare(a: &str, b: &str) -> Ordering {
++ #![allow(clippy::comparison_chain)]
++
++ let sa = get_leading_gen(a);
++ let sb = get_leading_gen(b);
++
++ let fa = permissive_f64_parse(&sa);
++ let fb = permissive_f64_parse(&sb);
++
++ // f64::cmp isn't implemented (due to NaN issues); implement directly instead
++ if fa > fb {
++ Ordering::Greater
++ } else if fa < fb {
++ Ordering::Less
++ } else {
++ Ordering::Equal
++ }
++}
++
++// GNU/BSD does not handle converting numbers to an equal scale
++// properly. GNU/BSD simply recognize that there is a human scale and sorts
++// those numbers ahead of other number inputs. There are perhaps limits
++// to the type of behavior we should emulate, and this might be such a limit.
++// Properly handling these units seems like a value add to me. And when sorting
++// these types of numbers, we rarely care about pure performance.
+ fn human_numeric_convert(a: &str) -> f64 {
+- let int_str = get_leading_number(a);
+- let (_, s) = a.split_at(int_str.len());
+- let int_part = permissive_f64_parse(int_str);
+- let suffix: f64 = match s.parse().unwrap_or('\0') {
+- 'K' => 1000f64,
++ let num_str = get_leading_num(a);
++ let suffix = a.trim_start_matches(num_str);
++ let num_part = permissive_f64_parse(num_str);
++ let suffix: f64 = match suffix.parse().unwrap_or('\0') {
++ // SI Units
++ 'K' => 1E3,
+ 'M' => 1E6,
+ 'G' => 1E9,
+ 'T' => 1E12,
+ 'P' => 1E15,
+- _ => 1f64,
++ 'E' => 1E18,
++ 'Z' => 1E21,
++ 'Y' => 1E24,
++ _ => 1f64,
+ };
+- int_part * suffix
++ num_part * suffix
+ }
+
+ /// Compare two strings as if they are human readable sizes.
+@@ -555,6 +824,7 @@ fn human_numeric_size_compare(a: &str, b: &str) -> Ordering {
+ #![allow(clippy::comparison_chain)]
+ let fa = human_numeric_convert(a);
+ let fb = human_numeric_convert(b);
++
+ // f64::cmp isn't implemented (due to NaN issues); implement directly instead
+ if fa > fb {
+ Ordering::Greater
+@@ -565,16 +835,6 @@ fn human_numeric_size_compare(a: &str, b: &str) -> Ordering {
+ }
+ }
+
+-fn random_shuffle(a: &str, b: &str, salt: String) -> Ordering {
+- #![allow(clippy::comparison_chain)]
+- let salt_slice = salt.as_str();
+-
+- let da = hash(&[a, salt_slice].concat());
+- let db = hash(&[b, salt_slice].concat());
+-
+- da.cmp(&db)
+-}
+-
+ fn get_rand_string() -> String {
+ thread_rng()
+ .sample_iter(&Alphanumeric)
+@@ -583,12 +843,22 @@ fn get_rand_string() -> String {
+ .collect::<String>()
+ }
+
+-fn hash<T: Hash>(t: &T) -> u64 {
+- let mut s: XxHash64 = Default::default();
++fn get_hash<T: Hash>(t: &T) -> u64 {
++ let mut s: FnvHasher = Default::default();
+ t.hash(&mut s);
+ s.finish()
+ }
+
++fn random_shuffle(a: &str, b: &str, x: String) -> Ordering {
++ #![allow(clippy::comparison_chain)]
++ let salt_slice = x.as_str();
++
++ let da = get_hash(&[a, salt_slice].concat());
++ let db = get_hash(&[b, salt_slice].concat());
++
++ da.cmp(&db)
++}
++
+ #[derive(Eq, Ord, PartialEq, PartialOrd)]
+ enum Month {
+ Unknown,
+@@ -608,13 +878,15 @@ enum Month {
+
+ /// Parse the beginning string into a Month, returning Month::Unknown on errors.
+ fn month_parse(line: &str) -> Month {
+- match line
+- .split_whitespace()
+- .next()
+- .unwrap()
+- .to_uppercase()
+- .as_ref()
+- {
++ // GNU splits at any 3 letter match "JUNNNN" is JUN
++ let pattern = if line.trim().len().ge(&3) {
++ // Split a 3 and get first element of tuple ".0"
++ line.split_at(3).0
++ } else {
++ ""
++ };
++
++ match pattern.to_uppercase().as_ref() {
+ "JAN" => Month::January,
+ "FEB" => Month::February,
+ "MAR" => Month::March,
+@@ -632,7 +904,16 @@ fn month_parse(line: &str) -> Month {
+ }
+
+ fn month_compare(a: &str, b: &str) -> Ordering {
+- month_parse(a).cmp(&month_parse(b))
++ let ma = month_parse(a);
++ let mb = month_parse(b);
++
++ if ma > mb {
++ Ordering::Greater
++ } else if ma < mb {
++ Ordering::Less
++ } else {
++ Ordering::Equal
++ }
+ }
+
+ fn version_compare(a: &str, b: &str) -> Ordering {
+@@ -650,19 +931,26 @@ fn version_compare(a: &str, b: &str) -> Ordering {
+ }
+
+ fn remove_nondictionary_chars(s: &str) -> String {
+- // Using 'is_ascii_whitespace()' instead of 'is_whitespace()', because it
+- // uses only symbols compatible with UNIX sort (space, tab, newline).
+- // 'is_whitespace()' uses more symbols as whitespace (e.g. vertical tab).
++ // According to GNU, dictionary chars are those of ASCII
++ // and a blank is a space or a tab
++ s.chars()
++ .filter(|c| c.is_ascii_alphanumeric() || c.is_ascii_whitespace())
++ .collect::<String>()
++}
++
++fn remove_nonprinting_chars(s: &str) -> String {
++ // However, GNU says nonprinting chars are more permissive.
++ // All of ASCII except control chars ie, escape, newline
+ s.chars()
+- .filter(|c| c.is_alphanumeric() || c.is_ascii_whitespace())
++ .filter(|c| c.is_ascii() && !c.is_ascii_control())
+ .collect::<String>()
+ }
+
+-fn print_sorted<S, T: Iterator<Item = S>>(iter: T, outfile: &Option<String>)
++fn print_sorted<S, T: Iterator<Item = S>>(iter: T, settings: &Settings)
+ where
+ S: std::fmt::Display,
+ {
+- let mut file: Box<dyn Write> = match *outfile {
++ let mut file: Box<dyn Write> = match settings.outfile {
+ Some(ref filename) => match File::create(Path::new(&filename)) {
+ Ok(f) => Box::new(BufWriter::new(f)) as Box<dyn Write>,
+ Err(e) => {
+@@ -673,9 +961,16 @@ where
+ None => Box::new(stdout()) as Box<dyn Write>,
+ };
+
+- for line in iter {
+- let str = format!("{}\n", line);
+- crash_if_err!(1, file.write_all(str.as_bytes()))
++ if settings.zero_terminated {
++ for line in iter {
++ let str = format!("{}\0", line);
++ crash_if_err!(1, file.write_all(str.as_bytes()));
++ }
++ } else {
++ for line in iter {
++ let str = format!("{}\n", line);
++ crash_if_err!(1, file.write_all(str.as_bytes()));
++ }
+ }
+ }
+
+@@ -700,6 +995,22 @@ mod tests {
+
+ use super::*;
+
++ #[test]
++ fn test_get_hash() {
++ let a = "Ted".to_string();
++
++ assert_eq!(2646829031758483623, get_hash(&a));
++ }
++
++ #[test]
++ fn test_random_shuffle() {
++ let a = "Ted";
++ let b = "Ted";
++ let c = get_rand_string();
++
++ assert_eq!(Ordering::Equal, random_shuffle(a, b, c));
++ }
++
+ #[test]
+ fn test_default_compare() {
+ let a = "your own";
+@@ -746,13 +1057,4 @@ mod tests {
+
+ assert_eq!(Ordering::Less, version_compare(a, b));
+ }
+-
+- #[test]
+- fn test_random_compare() {
+- let a = "9";
+- let b = "9";
+- let c = get_rand_string();
+-
+- assert_eq!(Ordering::Equal, random_shuffle(a, b, c));
+- }
+ }
+diff --git a/tests/by-util/test_sort.rs b/tests/by-util/test_sort.rs
+index 2bac71d..43aaf1d 100644
+--- a/tests/by-util/test_sort.rs
++++ b/tests/by-util/test_sort.rs
+@@ -1,58 +1,228 @@
+ use crate::common::util::*;
+
++#[test]
++fn test_check_zero_terminated_failure() {
++ new_ucmd!()
++ .arg("-z")
++ .arg("-c")
++ .arg("zero-terminated.txt")
++ .fails()
++ .stdout_is("sort: disorder in line 0\n");
++}
++
++#[test]
++fn test_check_zero_terminated_success() {
++ new_ucmd!()
++ .arg("-z")
++ .arg("-c")
++ .arg("zero-terminated.expected")
++ .succeeds();
++}
++
++#[test]
++fn test_random_shuffle_len() {
++ // check whether output is the same length as the input
++ const FILE: &'static str = "default_unsorted_ints.expected";
++ let (at, _ucmd) = at_and_ucmd!();
++ let result = new_ucmd!().arg("-R").arg(FILE).run().stdout;
++ let expected = at.read(FILE);
++
++ assert_ne!(result, expected);
++ assert_eq!(result.len(), expected.len());
++}
++
++#[test]
++fn test_random_shuffle_contains_all_lines() {
++ // check whether lines of input are all in output
++ const FILE: &'static str = "default_unsorted_ints.expected";
++ let (at, _ucmd) = at_and_ucmd!();
++ let result = new_ucmd!().arg("-R").arg(FILE).run().stdout;
++ let expected = at.read(FILE);
++ let result_sorted = new_ucmd!().pipe_in(result.clone()).run().stdout;
++
++ assert_ne!(result, expected);
++ assert_eq!(result_sorted, expected);
++}
++
++#[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;
++ let expected = at.read(FILE);
++ let unexpected = new_ucmd!().arg("-R").arg(FILE).run().stdout;
++
++ assert_ne!(result, expected);
++ assert_ne!(result, unexpected);
++}
++
+ #[test]
+ fn test_numeric_floats_and_ints() {
+- for numeric_sort_param in vec!["-n", "--numeric-sort"] {
+- let input = "1.444\n8.013\n1\n-8\n1.04\n-1";
++ test_helper("numeric_floats_and_ints", "-n");
++}
++
++#[test]
++fn test_numeric_floats() {
++ test_helper("numeric_floats", "-n");
++}
++
++#[test]
++fn test_numeric_floats_with_nan() {
++ test_helper("numeric_floats_with_nan", "-n");
++}
++
++#[test]
++fn test_numeric_unfixed_floats() {
++ test_helper("numeric_unfixed_floats", "-n");
++}
++
++#[test]
++fn test_numeric_fixed_floats() {
++ test_helper("numeric_fixed_floats", "-n");
++}
++
++#[test]
++fn test_numeric_unsorted_ints() {
++ test_helper("numeric_unsorted_ints", "-n");
++}
++
++#[test]
++fn test_human_block_sizes() {
++ test_helper("human_block_sizes", "-h");
++}
++
++#[test]
++fn test_month_default() {
++ test_helper("month_default", "-M");
++}
++
++#[test]
++fn test_month_stable() {
++ test_helper("month_stable", "-Ms");
++}
++
++#[test]
++fn test_default_unsorted_ints() {
++ test_helper("default_unsorted_ints", "");
++}
++
++#[test]
++fn test_numeric_unique_ints() {
++ test_helper("numeric_unsorted_ints_unique", "-nu");
++}
++
++#[test]
++fn test_version() {
++ test_helper("version", "-V");
++}
++
++#[test]
++fn test_ignore_case() {
++ test_helper("ignore_case", "-f");
++}
++
++#[test]
++fn test_dictionary_order() {
++ test_helper("dictionary_order", "-d");
++}
++
++#[test]
++fn test_dictionary_order2() {
++ for non_dictionary_order2_param in vec!["-d"] {
+ new_ucmd!()
+- .arg(numeric_sort_param)
+- .pipe_in(input)
++ .pipe_in("a👦ðŸ»aa b\naaaa b")
++ .arg(non_dictionary_order2_param)
+ .succeeds()
+- .stdout_only("-8\n-1\n1\n1.04\n1.444\n8.013\n");
++ .stdout_only("a👦ðŸ»aa b\naaaa b\n");
+ }
+ }
+
+ #[test]
+-fn test_numeric_floats() {
+- for numeric_sort_param in vec!["-n", "--numeric-sort"] {
+- let input = "1.444\n8.013\n1.58590\n-8.90880\n1.040000000\n-.05";
++fn test_non_printing_chars() {
++ for non_printing_chars_param in vec!["-i"] {
+ new_ucmd!()
+- .arg(numeric_sort_param)
+- .pipe_in(input)
++ .pipe_in("a👦ðŸ»aa b\naaaa b")
++ .arg(non_printing_chars_param)
+ .succeeds()
+- .stdout_only("-8.90880\n-.05\n1.040000000\n1.444\n1.58590\n8.013\n");
++ .stdout_only("aaaa b\na👦ðŸ»aa b\n");
+ }
+ }
+
+ #[test]
+-fn test_numeric_floats_with_nan() {
+- for numeric_sort_param in vec!["-n", "--numeric-sort"] {
+- let input = "1.444\n1.0/0.0\n1.58590\n-8.90880\n1.040000000\n-.05";
++fn test_exponents_positive_general_fixed() {
++ for exponents_positive_general_param in vec!["-g"] {
+ new_ucmd!()
+- .arg(numeric_sort_param)
+- .pipe_in(input)
++ .pipe_in("100E6\n\n50e10\n+100000\n\n10000K78\n10E\n\n\n1000EDKLD\n\n\n100E6\n\n50e10\n+100000\n\n")
++ .arg(exponents_positive_general_param)
+ .succeeds()
+- .stdout_only("-8.90880\n-.05\n1.0/0.0\n1.040000000\n1.444\n1.58590\n");
++ .stdout_only("\n\n\n\n\n\n\n\n10000K78\n1000EDKLD\n10E\n+100000\n+100000\n100E6\n100E6\n50e10\n50e10\n");
+ }
+ }
+
+ #[test]
+-fn test_numeric_unfixed_floats() {
+- test_helper("numeric_fixed_floats", "-n");
++fn test_exponents_positive_numeric() {
++ test_helper("exponents-positive-numeric", "-n");
+ }
+
+ #[test]
+-fn test_numeric_fixed_floats() {
+- test_helper("numeric_fixed_floats", "-n");
++fn test_months_dedup() {
++ test_helper("months-dedup", "-Mu");
+ }
+
+ #[test]
+-fn test_numeric_unsorted_ints() {
+- test_helper("numeric_unsorted_ints", "-n");
++fn test_mixed_floats_ints_chars_numeric() {
++ test_helper("mixed_floats_ints_chars_numeric", "-n");
+ }
+
+ #[test]
+-fn test_human_block_sizes() {
++fn test_mixed_floats_ints_chars_numeric_unique() {
++ test_helper("mixed_floats_ints_chars_numeric_unique", "-nu");
++}
++
++#[test]
++fn test_mixed_floats_ints_chars_numeric_reverse() {
++ test_helper("mixed_floats_ints_chars_numeric_unique_reverse", "-nur");
++}
++
++#[test]
++fn test_mixed_floats_ints_chars_numeric_stable() {
++ test_helper("mixed_floats_ints_chars_numeric_stable", "-ns");
++}
++
++#[test]
++fn test_numeric_floats_and_ints2() {
++ for numeric_sort_param in vec!["-n", "--numeric-sort"] {
++ let input = "1.444\n8.013\n1\n-8\n1.04\n-1";
++ new_ucmd!()
++ .arg(numeric_sort_param)
++ .pipe_in(input)
++ .succeeds()
++ .stdout_only("-8\n-1\n1\n1.04\n1.444\n8.013\n");
++ }
++}
++
++#[test]
++fn test_numeric_floats2() {
++ for numeric_sort_param in vec!["-n", "--numeric-sort"] {
++ let input = "1.444\n8.013\n1.58590\n-8.90880\n1.040000000\n-.05";
++ new_ucmd!()
++ .arg(numeric_sort_param)
++ .pipe_in(input)
++ .succeeds()
++ .stdout_only("-8.90880\n-.05\n1.040000000\n1.444\n1.58590\n8.013\n");
++ }
++}
++
++#[test]
++fn test_numeric_floats_with_nan2() {
++ test_helper("numeric-floats-with-nan2", "-n");
++}
++
++#[test]
++fn test_human_block_sizes2() {
+ for human_numeric_sort_param in vec!["-h", "--human-numeric-sort"] {
+ let input = "8981K\n909991M\n-8T\n21G\n0.8M";
+ new_ucmd!()
+@@ -64,7 +234,7 @@ fn test_human_block_sizes() {
+ }
+
+ #[test]
+-fn test_month_default() {
++fn test_month_default2() {
+ for month_sort_param in vec!["-M", "--month-sort"] {
+ let input = "JAn\nMAY\n000may\nJun\nFeb";
+ new_ucmd!()
+@@ -76,12 +246,7 @@ fn test_month_default() {
+ }
+
+ #[test]
+-fn test_month_stable() {
+- test_helper("month_stable", "-Ms");
+-}
+-
+-#[test]
+-fn test_default_unsorted_ints() {
++fn test_default_unsorted_ints2() {
+ let input = "9\n1909888\n000\n1\n2";
+ new_ucmd!()
+ .pipe_in(input)
+@@ -90,7 +255,7 @@ fn test_default_unsorted_ints() {
+ }
+
+ #[test]
+-fn test_numeric_unique_ints() {
++fn test_numeric_unique_ints2() {
+ for numeric_unique_sort_param in vec!["-nu"] {
+ let input = "9\n9\n8\n1\n";
+ new_ucmd!()
+@@ -102,18 +267,8 @@ fn test_numeric_unique_ints() {
+ }
+
+ #[test]
+-fn test_version() {
+- test_helper("version", "-V");
+-}
+-
+-#[test]
+-fn test_ignore_case() {
+- test_helper("ignore_case", "-f");
+-}
+-
+-#[test]
+-fn test_dictionary_order() {
+- test_helper("dictionary_order", "-d");
++fn test_zero_terminated() {
++ test_helper("zero-terminated", "-z");
+ }
+
+ #[test]
+@@ -192,6 +347,15 @@ fn test_check() {
+ .stdout_is("");
+ }
+
++#[test]
++fn test_check_silent() {
++ new_ucmd!()
++ .arg("-C")
++ .arg("check_fail.txt")
++ .fails()
++ .stdout_is("");
++}
++
+ fn test_helper(file_name: &str, args: &str) {
+ new_ucmd!()
+ .arg(args)
+diff --git a/tests/fixtures/sort/exponents-positive-general.expected b/tests/fixtures/sort/exponents-positive-general.expected
+new file mode 100644
+index 0000000..3dbc92f
+--- /dev/null
++++ b/tests/fixtures/sort/exponents-positive-general.expected
+@@ -0,0 +1,12 @@
++
++
++
++
++
++
++10E
++1000EDKLD
++10000K78
+++100000
++100E6
++50e10
+diff --git a/tests/fixtures/sort/exponents-positive-general.txt b/tests/fixtures/sort/exponents-positive-general.txt
+new file mode 100644
+index 0000000..23ea527
+--- /dev/null
++++ b/tests/fixtures/sort/exponents-positive-general.txt
+@@ -0,0 +1,12 @@
++10000K78
++10E
++
++
++1000EDKLD
++
++
++100E6
++
++50e10
+++100000
++
+diff --git a/tests/fixtures/sort/exponents-positive-numeric.expected b/tests/fixtures/sort/exponents-positive-numeric.expected
+new file mode 100644
+index 0000000..174088f
+--- /dev/null
++++ b/tests/fixtures/sort/exponents-positive-numeric.expected
+@@ -0,0 +1,12 @@
++
++
++
++
++
++
+++100000
++10E
++50e10
++100E6
++1000EDKLD
++10000K78
+diff --git a/tests/fixtures/sort/exponents-positive-numeric.txt b/tests/fixtures/sort/exponents-positive-numeric.txt
+new file mode 100644
+index 0000000..23ea527
+--- /dev/null
++++ b/tests/fixtures/sort/exponents-positive-numeric.txt
+@@ -0,0 +1,12 @@
++10000K78
++10E
++
++
++1000EDKLD
++
++
++100E6
++
++50e10
+++100000
++
+diff --git a/tests/fixtures/sort/human-mixed-inputs-reverse.expected b/tests/fixtures/sort/human-mixed-inputs-reverse.expected
+new file mode 100644
+index 0000000..37e1762
+--- /dev/null
++++ b/tests/fixtures/sort/human-mixed-inputs-reverse.expected
+@@ -0,0 +1,37 @@
++.2T
++2G
++100M
++7800900K
++51887300-
++1890777
++56908-90078
++6780.0009866
++6780.000986
++789----009999 90-0 90-0
++1
++0001
++apr
++MAY
++JUNNNN
++JAN
++AUG
++APR
++0000000
++00
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++-1.4
+diff --git a/tests/fixtures/sort/human-mixed-inputs-reverse.txt b/tests/fixtures/sort/human-mixed-inputs-reverse.txt
+new file mode 100644
+index 0000000..d8b5124
+--- /dev/null
++++ b/tests/fixtures/sort/human-mixed-inputs-reverse.txt
+@@ -0,0 +1,37 @@
++JAN
++
++0000000
++
++00
++
++0001
++
++1
++
++-1.4
++
++JUNNNN
++AUG
++
++apr
++
++APR
++
++
++MAY
++1890777
++
++56908-90078
++
++51887300-
++
++6780.0009866
++
++789----009999 90-0 90-0
++
++6780.000986
++
++100M
++7800900K
++2G
++.2T
+diff --git a/tests/fixtures/sort/human-mixed-inputs-stable.expected b/tests/fixtures/sort/human-mixed-inputs-stable.expected
+new file mode 100644
+index 0000000..0fcdcdb
+--- /dev/null
++++ b/tests/fixtures/sort/human-mixed-inputs-stable.expected
+@@ -0,0 +1,37 @@
++-1.4
++JAN
++
++0000000
++
++00
++
++
++
++
++JUNNNN
++AUG
++
++apr
++
++APR
++
++
++MAY
++
++
++
++
++
++
++0001
++1
++789----009999 90-0 90-0
++6780.000986
++6780.0009866
++56908-90078
++1890777
++51887300-
++7800900K
++100M
++2G
++.2T
+diff --git a/tests/fixtures/sort/human-mixed-inputs-stable.txt b/tests/fixtures/sort/human-mixed-inputs-stable.txt
+new file mode 100644
+index 0000000..d8b5124
+--- /dev/null
++++ b/tests/fixtures/sort/human-mixed-inputs-stable.txt
+@@ -0,0 +1,37 @@
++JAN
++
++0000000
++
++00
++
++0001
++
++1
++
++-1.4
++
++JUNNNN
++AUG
++
++apr
++
++APR
++
++
++MAY
++1890777
++
++56908-90078
++
++51887300-
++
++6780.0009866
++
++789----009999 90-0 90-0
++
++6780.000986
++
++100M
++7800900K
++2G
++.2T
+diff --git a/tests/fixtures/sort/human-mixed-inputs-unique.expected b/tests/fixtures/sort/human-mixed-inputs-unique.expected
+new file mode 100644
+index 0000000..50f53b6
+--- /dev/null
++++ b/tests/fixtures/sort/human-mixed-inputs-unique.expected
+@@ -0,0 +1,13 @@
++-1.4
++JAN
++0001
++789----009999 90-0 90-0
++6780.000986
++6780.0009866
++56908-90078
++1890777
++51887300-
++7800900K
++100M
++2G
++.2T
+diff --git a/tests/fixtures/sort/human-mixed-inputs-unique.txt b/tests/fixtures/sort/human-mixed-inputs-unique.txt
+new file mode 100644
+index 0000000..d8b5124
+--- /dev/null
++++ b/tests/fixtures/sort/human-mixed-inputs-unique.txt
+@@ -0,0 +1,37 @@
++JAN
++
++0000000
++
++00
++
++0001
++
++1
++
++-1.4
++
++JUNNNN
++AUG
++
++apr
++
++APR
++
++
++MAY
++1890777
++
++56908-90078
++
++51887300-
++
++6780.0009866
++
++789----009999 90-0 90-0
++
++6780.000986
++
++100M
++7800900K
++2G
++.2T
+diff --git a/tests/fixtures/sort/human-mixed-inputs.expected b/tests/fixtures/sort/human-mixed-inputs.expected
+new file mode 100644
+index 0000000..1f900a8
+--- /dev/null
++++ b/tests/fixtures/sort/human-mixed-inputs.expected
+@@ -0,0 +1,37 @@
++-1.4
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++00
++0000000
++APR
++AUG
++JAN
++JUNNNN
++MAY
++apr
++0001
++1
++789----009999 90-0 90-0
++6780.000986
++6780.0009866
++56908-90078
++1890777
++51887300-
++7800900K
++100M
++2G
++.2T
+diff --git a/tests/fixtures/sort/human-mixed-inputs.txt b/tests/fixtures/sort/human-mixed-inputs.txt
+new file mode 100644
+index 0000000..cfd8a9d
+--- /dev/null
++++ b/tests/fixtures/sort/human-mixed-inputs.txt
+@@ -0,0 +1,46 @@
++JAN
++
++0000000
++
++00
++
++0001
++
++1
++
++-1.4
++
++JUNNNN
++AUG
++
++apr
++
++APR
++
++
++MAY
++1890777
++
++56908-90078
++
++51887300-
++
++6780.0009866
++
++789----009999 90-0 90-0
++
++6780.000986
++
++1M
++10M
++100M
++1000M
++10000M
++
++7800900K
++780090K
++78009K
++7800K
++780K
++2G
++.2T
+diff --git a/tests/fixtures/sort/mixed_floats_ints_chars_numeric.expected b/tests/fixtures/sort/mixed_floats_ints_chars_numeric.expected
+new file mode 100644
+index 0000000..a781a36
+--- /dev/null
++++ b/tests/fixtures/sort/mixed_floats_ints_chars_numeric.expected
+@@ -0,0 +1,30 @@
++-2028789030
++-896689
++-8.90880
++-1
++-.05
++
++
++
++
++
++
++
++
++000
++CARAvan
++00000001
++1
++1.040000000
++1.444
++1.58590
++8.013
++45
++46.89
++ 4567.
++ 37800
++576,446.88800000
++576,446.890
++4798908.340000000000
++4798908.45
++4798908.8909800
+diff --git a/tests/fixtures/sort/mixed_floats_ints_chars_numeric.txt b/tests/fixtures/sort/mixed_floats_ints_chars_numeric.txt
+new file mode 100644
+index 0000000..a5813ea
+--- /dev/null
++++ b/tests/fixtures/sort/mixed_floats_ints_chars_numeric.txt
+@@ -0,0 +1,30 @@
++576,446.890
++576,446.88800000
++
++
++ 4567.
++45
++46.89
++-1
++1
++00000001
++4798908.340000000000
++4798908.45
++4798908.8909800
++
++
++ 37800
++
++-2028789030
++-896689
++CARAvan
++
++-8.90880
++-.05
++1.444
++1.58590
++1.040000000
++
++8.013
++
++000
+diff --git a/tests/fixtures/sort/mixed_floats_ints_chars_numeric_reverse.expected b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_reverse.expected
+new file mode 100644
+index 0000000..6b02421
+--- /dev/null
++++ b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_reverse.expected
+@@ -0,0 +1,30 @@
++4798908.8909800
++4798908.45
++4798908.340000000000
++576,446.890
++576,446.88800000
++ 37800
++ 4567.
++46.89
++45
++8.013
++1.58590
++1.444
++1.040000000
++1
++00000001
++CARAvan
++000
++
++
++
++
++
++
++
++
++-.05
++-1
++-8.90880
++-896689
++-2028789030
+diff --git a/tests/fixtures/sort/mixed_floats_ints_chars_numeric_reverse_stable.expected b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_reverse_stable.expected
+new file mode 100644
+index 0000000..cb1028f
+--- /dev/null
++++ b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_reverse_stable.expected
+@@ -0,0 +1,30 @@
++4798908.8909800
++4798908.45
++4798908.340000000000
++576,446.890
++576,446.88800000
++ 37800
++ 4567.
++46.89
++45
++8.013
++1.58590
++1.444
++1.040000000
++1
++00000001
++
++
++
++
++
++CARAvan
++
++
++
++000
++-.05
++-1
++-8.90880
++-896689
++-2028789030
+diff --git a/tests/fixtures/sort/mixed_floats_ints_chars_numeric_reverse_stable.txt b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_reverse_stable.txt
+new file mode 100644
+index 0000000..a5813ea
+--- /dev/null
++++ b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_reverse_stable.txt
+@@ -0,0 +1,30 @@
++576,446.890
++576,446.88800000
++
++
++ 4567.
++45
++46.89
++-1
++1
++00000001
++4798908.340000000000
++4798908.45
++4798908.8909800
++
++
++ 37800
++
++-2028789030
++-896689
++CARAvan
++
++-8.90880
++-.05
++1.444
++1.58590
++1.040000000
++
++8.013
++
++000
+diff --git a/tests/fixtures/sort/mixed_floats_ints_chars_numeric_stable.expected b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_stable.expected
+new file mode 100644
+index 0000000..63a3e64
+--- /dev/null
++++ b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_stable.expected
+@@ -0,0 +1,30 @@
++-2028789030
++-896689
++-8.90880
++-1
++-.05
++
++
++
++
++
++CARAvan
++
++
++
++000
++1
++00000001
++1.040000000
++1.444
++1.58590
++8.013
++45
++46.89
++ 4567.
++ 37800
++576,446.88800000
++576,446.890
++4798908.340000000000
++4798908.45
++4798908.8909800
+diff --git a/tests/fixtures/sort/mixed_floats_ints_chars_numeric_stable.txt b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_stable.txt
+new file mode 100644
+index 0000000..a5813ea
+--- /dev/null
++++ b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_stable.txt
+@@ -0,0 +1,30 @@
++576,446.890
++576,446.88800000
++
++
++ 4567.
++45
++46.89
++-1
++1
++00000001
++4798908.340000000000
++4798908.45
++4798908.8909800
++
++
++ 37800
++
++-2028789030
++-896689
++CARAvan
++
++-8.90880
++-.05
++1.444
++1.58590
++1.040000000
++
++8.013
++
++000
+diff --git a/tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique.expected b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique.expected
+new file mode 100644
+index 0000000..cb27c66
+--- /dev/null
++++ b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique.expected
+@@ -0,0 +1,20 @@
++-2028789030
++-896689
++-8.90880
++-1
++-.05
++
++1
++1.040000000
++1.444
++1.58590
++8.013
++45
++46.89
++ 4567.
++ 37800
++576,446.88800000
++576,446.890
++4798908.340000000000
++4798908.45
++4798908.8909800
+diff --git a/tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique.txt b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique.txt
+new file mode 100644
+index 0000000..a5813ea
+--- /dev/null
++++ b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique.txt
+@@ -0,0 +1,30 @@
++576,446.890
++576,446.88800000
++
++
++ 4567.
++45
++46.89
++-1
++1
++00000001
++4798908.340000000000
++4798908.45
++4798908.8909800
++
++
++ 37800
++
++-2028789030
++-896689
++CARAvan
++
++-8.90880
++-.05
++1.444
++1.58590
++1.040000000
++
++8.013
++
++000
+diff --git a/tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique_reverse.expected b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique_reverse.expected
+new file mode 100644
+index 0000000..bbce169
+--- /dev/null
++++ b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique_reverse.expected
+@@ -0,0 +1,20 @@
++4798908.8909800
++4798908.45
++4798908.340000000000
++576,446.890
++576,446.88800000
++ 37800
++ 4567.
++46.89
++45
++8.013
++1.58590
++1.444
++1.040000000
++1
++
++-.05
++-1
++-8.90880
++-896689
++-2028789030
+diff --git a/tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique_reverse.txt b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique_reverse.txt
+new file mode 100644
+index 0000000..a5813ea
+--- /dev/null
++++ b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique_reverse.txt
+@@ -0,0 +1,30 @@
++576,446.890
++576,446.88800000
++
++
++ 4567.
++45
++46.89
++-1
++1
++00000001
++4798908.340000000000
++4798908.45
++4798908.8909800
++
++
++ 37800
++
++-2028789030
++-896689
++CARAvan
++
++-8.90880
++-.05
++1.444
++1.58590
++1.040000000
++
++8.013
++
++000
+diff --git a/tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique_stable.expected b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique_stable.expected
+new file mode 100644
+index 0000000..bbce169
+--- /dev/null
++++ b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique_stable.expected
+@@ -0,0 +1,20 @@
++4798908.8909800
++4798908.45
++4798908.340000000000
++576,446.890
++576,446.88800000
++ 37800
++ 4567.
++46.89
++45
++8.013
++1.58590
++1.444
++1.040000000
++1
++
++-.05
++-1
++-8.90880
++-896689
++-2028789030
+diff --git a/tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique_stable.txt b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique_stable.txt
+new file mode 100644
+index 0000000..a5813ea
+--- /dev/null
++++ b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique_stable.txt
+@@ -0,0 +1,30 @@
++576,446.890
++576,446.88800000
++
++
++ 4567.
++45
++46.89
++-1
++1
++00000001
++4798908.340000000000
++4798908.45
++4798908.8909800
++
++
++ 37800
++
++-2028789030
++-896689
++CARAvan
++
++-8.90880
++-.05
++1.444
++1.58590
++1.040000000
++
++8.013
++
++000
+diff --git a/tests/fixtures/sort/months-dedup.expected b/tests/fixtures/sort/months-dedup.expected
+new file mode 100644
+index 0000000..a2821f2
+--- /dev/null
++++ b/tests/fixtures/sort/months-dedup.expected
+@@ -0,0 +1,6 @@
++
++JAN
++apr
++MAY
++JUNNNN
++AUG
+diff --git a/tests/fixtures/sort/months-dedup.txt b/tests/fixtures/sort/months-dedup.txt
+new file mode 100644
+index 0000000..d8b5124
+--- /dev/null
++++ b/tests/fixtures/sort/months-dedup.txt
+@@ -0,0 +1,37 @@
++JAN
++
++0000000
++
++00
++
++0001
++
++1
++
++-1.4
++
++JUNNNN
++AUG
++
++apr
++
++APR
++
++
++MAY
++1890777
++
++56908-90078
++
++51887300-
++
++6780.0009866
++
++789----009999 90-0 90-0
++
++6780.000986
++
++100M
++7800900K
++2G
++.2T
+diff --git a/tests/fixtures/sort/numeric-floats-with-nan2.expected b/tests/fixtures/sort/numeric-floats-with-nan2.expected
+new file mode 100644
+index 0000000..51c9985
+--- /dev/null
++++ b/tests/fixtures/sort/numeric-floats-with-nan2.expected
+@@ -0,0 +1,23 @@
++-8.90880
++-.05
++
++
++
++
++
++
++
++
++
++
++
++
++
++
++Karma
++1
++1.0/0.0
++1.040000000
++1.2
++1.444
++1.58590
+diff --git a/tests/fixtures/sort/numeric-floats-with-nan2.txt b/tests/fixtures/sort/numeric-floats-with-nan2.txt
+new file mode 100644
+index 0000000..9b78741
+--- /dev/null
++++ b/tests/fixtures/sort/numeric-floats-with-nan2.txt
+@@ -0,0 +1,23 @@
++Karma
++
++1.0/0.0
++
++
++-8.90880
++
++
++-.05
++
++
++1.040000000
++
++1.444
++
++
++1.58590
++
++
++1
++
++1.2
++
+diff --git a/tests/fixtures/sort/zero-terminated.expected b/tests/fixtures/sort/zero-terminated.expected
+new file mode 100644
+index 0000000..4e53b30
+--- /dev/null
++++ b/tests/fixtures/sort/zero-terminated.expected
+@@ -0,0 +1 @@
++../..�../../by-util�../../common�../../fixtures�../../fixtures/cat�../../fixtures/cksum�../../fixtures/comm�../../fixtures/cp�../../fixtures/cp/dir_with_mount�../../fixtures/cp/dir_with_mount/copy_me�../../fixtures/cp/hello_dir�../../fixtures/cp/hello_dir_with_file�../../fixtures/csplit�../../fixtures/cut�../../fixtures/cut/sequences�../../fixtures/dircolors�../../fixtures/du�../../fixtures/du/subdir�../../fixtures/du/subdir/deeper�../../fixtures/du/subdir/links�../../fixtures/env�../../fixtures/expand�../../fixtures/fmt�../../fixtures/fold�../../fixtures/hashsum�../../fixtures/head�../../fixtures/join�../../fixtures/mv�../../fixtures/nl�../../fixtures/numfmt�../../fixtures/od�../../fixtures/paste�../../fixtures/ptx�../../fixtures/shuf�../../fixtures/sort�../../fixtures/sum�../../fixtures/tac�../../fixtures/tail�../../fixtures/tsort�../../fixtures/unexpand�../../fixtures/uniq�../../fixtures/wc�
+\ No newline at end of file
+diff --git a/tests/fixtures/sort/zero-terminated.txt b/tests/fixtures/sort/zero-terminated.txt
+new file mode 100644
+index 0000000..5c547c8
+--- /dev/null
++++ b/tests/fixtures/sort/zero-terminated.txt
+@@ -0,0 +1 @@
++../../fixtures/paste�../../fixtures/du�../../fixtures/fold�../../fixtures�../../fixtures/cp/dir_with_mount/copy_me�../../fixtures/sum�../../fixtures/expand�../../fixtures/mv�../../fixtures/shuf�../../fixtures/od�../../fixtures/env�../../fixtures/cut�../../fixtures/cp/hello_dir�../../fixtures/hashsum�../../common�../../fixtures/du/subdir/links�../../fixtures/dircolors�../../fixtures/nl�../../fixtures/wc�../../fixtures/cut/sequences�../../fixtures/numfmt�../../fixtures/comm�../../fixtures/du/subdir�../../fixtures/cp/hello_dir_with_file�../../fixtures/ptx�../../fixtures/cp/dir_with_mount�../../fixtures/cat�../../fixtures/cp�../..�../../fixtures/tail�../../fixtures/du/subdir/deeper�../../fixtures/head�../../fixtures/join�../../by-util�../../fixtures/csplit�../../fixtures/cksum�../../fixtures/fmt�../../fixtures/tsort�../../fixtures/tac�../../fixtures/unexpand�../../fixtures/uniq�../../fixtures/sort�
+\ No newline at end of file
diff --git a/debian/patches/Sort-Various-fixes-and-performance-improvements.patch b/debian/patches/Sort-Various-fixes-and-performance-improvements.patch
new file mode 100644
index 0000000..6023dbe
--- /dev/null
+++ b/debian/patches/Sort-Various-fixes-and-performance-improvements.patch
@@ -0,0 +1,502 @@
+From: electricboogie <32370782+electricboogie@users.noreply.github.com>
+Date: Sat, 10 Apr 2021 04:56:20 -0500
+Subject: Sort: Various fixes and performance improvements (#2057)
+
+* Various fixes and performance improvements
+
+* fix a typo
+
+Co-authored-by: Michael Debertol <michael.debertol@gmail.com>
+
+Co-authored-by: Sylvestre Ledru <sledru@mozilla.com>
+Co-authored-by: Michael Debertol <michael.debertol@gmail.com>
+
+Origin: upstream, https://github.com/uutils/coreutils/commit/e5113ad00ef76032cc15b0052877c23a48783d4a
+---
+ src/uu/sort/src/sort.rs | 114 ++++++++++++++-------
+ tests/.DS_Store | Bin 0 -> 6160 bytes
+ tests/by-util/test_sort.rs | 49 +++++++--
+ tests/fixtures/.DS_Store | 1 +
+ ...ts_chars_numeric_unique_reverse_stable.expected | 20 ++++
+ tests/fixtures/sort/multiple_decimals.expected | 33 ++++++
+ tests/fixtures/sort/multiple_decimals_general.txt | 35 +++++++
+ tests/fixtures/sort/multiple_decimals_numeric.txt | 35 +++++++
+ 8 files changed, 243 insertions(+), 44 deletions(-)
+ create mode 100644 tests/.DS_Store
+ create mode 100644 tests/fixtures/.DS_Store
+ create mode 100644 tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique_reverse_stable.expected
+ create mode 100644 tests/fixtures/sort/multiple_decimals.expected
+ create mode 100644 tests/fixtures/sort/multiple_decimals_general.txt
+ create mode 100644 tests/fixtures/sort/multiple_decimals_numeric.txt
+
+diff --git a/src/uu/sort/src/sort.rs b/src/uu/sort/src/sort.rs
+index cf6c258..88f7bc7 100644
+--- a/src/uu/sort/src/sort.rs
++++ b/src/uu/sort/src/sort.rs
+@@ -22,6 +22,7 @@ use rand::distributions::Alphanumeric;
+ use rand::{thread_rng, Rng};
+ use rayon::prelude::*;
+ use semver::Version;
++use std::borrow::Cow;
+ use std::cmp::Ordering;
+ use std::collections::BinaryHeap;
+ use std::env;
+@@ -262,7 +263,7 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
+ Arg::with_name(OPT_CHECK_SILENT)
+ .short("C")
+ .long(OPT_CHECK_SILENT)
+- .help("exit successfully if the given file is already sorted, and exit with status 1 otherwise. "),
++ .help("exit successfully if the given file is already sorted, and exit with status 1 otherwise."),
+ )
+ .arg(
+ Arg::with_name(OPT_IGNORE_CASE)
+@@ -353,7 +354,7 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
+ if let Ok(n) = line {
+ files.push(
+ std::str::from_utf8(&n)
+- .expect("Could not parse zero terminated string from input.")
++ .expect("Could not parse string from zero terminated input.")
+ .to_string(),
+ );
+ }
+@@ -488,6 +489,8 @@ fn exec(files: Vec<String>, settings: &mut Settings) -> i32 {
+ } else {
+ print_sorted(file_merger, &settings)
+ }
++ } else if settings.mode == SortMode::Default && settings.unique {
++ print_sorted(lines.iter().dedup(), &settings)
+ } else if settings.mode == SortMode::Month && settings.unique {
+ print_sorted(
+ lines
+@@ -499,7 +502,7 @@ fn exec(files: Vec<String>, settings: &mut Settings) -> i32 {
+ print_sorted(
+ lines
+ .iter()
+- .dedup_by(|a, b| get_nums_dedup(a) == get_nums_dedup(b)),
++ .dedup_by(|a, b| get_num_dedup(a, &settings) == get_num_dedup(b, &settings)),
+ &settings,
+ )
+ } else {
+@@ -603,12 +606,13 @@ fn default_compare(a: &str, b: &str) -> Ordering {
+ #[inline(always)]
+ fn leading_num_common(a: &str) -> &str {
+ let mut s = "";
++
++ // check whether char is numeric, whitespace or decimal point or thousand separator
+ for (idx, c) in a.char_indices() {
+- // check whether char is numeric, whitespace or decimal point or thousand seperator
+ if !c.is_numeric()
+ && !c.is_whitespace()
+- && !c.eq(&DECIMAL_PT)
+ && !c.eq(&THOUSANDS_SEP)
++ && !c.eq(&DECIMAL_PT)
+ // check for e notation
+ && !c.eq(&'e')
+ && !c.eq(&'E')
+@@ -621,7 +625,7 @@ fn leading_num_common(a: &str) -> &str {
+ break;
+ }
+ // If line is not a number line, return the line as is
+- s = a;
++ s = &a;
+ }
+ s
+ }
+@@ -633,16 +637,17 @@ fn leading_num_common(a: &str) -> &str {
+ // not recognize a positive sign or scientific/E notation so we strip those elements here.
+ fn get_leading_num(a: &str) -> &str {
+ let mut s = "";
+- let b = leading_num_common(a);
+
+- // GNU numeric sort doesn't recognize '+' or 'e' notation so we strip
+- for (idx, c) in b.char_indices() {
+- if c.eq(&'e') || c.eq(&'E') || b.chars().nth(0).unwrap_or('\0').eq(&POSITIVE) {
+- s = &b[..idx];
++ let a = leading_num_common(a);
++
++ // GNU numeric sort doesn't recognize '+' or 'e' notation so we strip trailing chars
++ for (idx, c) in a.char_indices() {
++ if c.eq(&'e') || c.eq(&'E') || a.chars().nth(0).unwrap_or('\0').eq(&POSITIVE) {
++ s = &a[..idx];
+ break;
+ }
+ // If no further processing needed to be done, return the line as-is to be sorted
+- s = b;
++ s = &a;
+ }
+
+ // And empty number or non-number lines are to be treated as ‘0’ but only for numeric sort
+@@ -657,30 +662,32 @@ fn get_leading_num(a: &str) -> &str {
+ // In contrast to numeric compare, GNU general numeric/FP sort *should* recognize positive signs and
+ // scientific notation, so we strip those lines only after the end of the following numeric string.
+ // For example, 5e10KFD would be 5e10 or 5x10^10 and +10000HFKJFK would become 10000.
+-fn get_leading_gen(a: &str) -> String {
++fn get_leading_gen(a: &str) -> &str {
+ // Make this iter peekable to see if next char is numeric
+- let mut p_iter = leading_num_common(a).chars().peekable();
+- let mut r = String::new();
++ let raw_leading_num = leading_num_common(a);
++ let mut p_iter = raw_leading_num.chars().peekable();
++ let mut result = "";
+ // Cleanup raw stripped strings
+ for c in p_iter.to_owned() {
+ let next_char_numeric = p_iter.peek().unwrap_or(&'\0').is_numeric();
+- // Only general numeric recognizes e notation and, see block below, the '+' sign
+- if (c.eq(&'e') && !next_char_numeric) || (c.eq(&'E') && !next_char_numeric) {
+- r = a.split(c).next().unwrap_or("").to_owned();
++ // Only general numeric recognizes e notation and the '+' sign
++ if (c.eq(&'e') && !next_char_numeric)
++ || (c.eq(&'E') && !next_char_numeric)
++ // Only GNU (non-general) numeric recognize thousands seperators, takes only leading #
++ || c.eq(&THOUSANDS_SEP)
++ {
++ result = a.split(c).next().unwrap_or("");
+ break;
+ // If positive sign and next char is not numeric, split at postive sign at keep trailing numbers
+ // There is a more elegant way to do this in Rust 1.45, std::str::strip_prefix
+ } else if c.eq(&POSITIVE) && !next_char_numeric {
+- let mut v: Vec<&str> = a.split(c).collect();
+- let x = v.split_off(1);
+- r = x.join("");
++ result = a.trim().trim_start_matches('+');
+ break;
+- // If no further processing needed to be done, return the line as-is to be sorted
+- } else {
+- r = a.to_owned();
+ }
++ // If no further processing needed to be done, return the line as-is to be sorted
++ result = a;
+ }
+- r
++ result
+ }
+
+ fn get_months_dedup(a: &str) -> String {
+@@ -714,10 +721,10 @@ fn get_months_dedup(a: &str) -> String {
+ }
+ }
+
+-// *For all dedups/uniques we must compare leading numbers*
++// *For all dedups/uniques expect default we must compare leading numbers*
+ // Also note numeric compare and unique output is specifically *not* the same as a "sort | uniq"
+ // See: https://www.gnu.org/software/coreutils/manual/html_node/sort-invocation.html
+-fn get_nums_dedup(a: &str) -> &str {
++fn get_num_dedup<'a>(a: &'a str, settings: &&mut Settings) -> &'a str {
+ // Trim and remove any leading zeros
+ let s = a.trim().trim_start_matches('0');
+
+@@ -731,20 +738,50 @@ fn get_nums_dedup(a: &str) -> &str {
+ ""
+ // Prepare lines for comparison of only the numerical leading numbers
+ } else {
+- get_leading_num(s)
++ let result = match settings.mode {
++ SortMode::Numeric => get_leading_num(s),
++ SortMode::GeneralNumeric => get_leading_gen(s),
++ SortMode::HumanNumeric => get_leading_num(s),
++ SortMode::Version => get_leading_num(s),
++ _ => s,
++ };
++ result
++ }
++}
++
++#[inline(always)]
++fn remove_thousands_sep<'a, S: Into<Cow<'a, str>>>(input: S) -> Cow<'a, str> {
++ let input = input.into();
++ if input.contains(THOUSANDS_SEP) {
++ let output = input.replace(THOUSANDS_SEP, "");
++ Cow::Owned(output)
++ } else {
++ input
++ }
++}
++
++#[inline(always)]
++fn remove_trailing_dec<'a, S: Into<Cow<'a, str>>>(input: S) -> Cow<'a, str> {
++ let input = input.into();
++ if let Some(s) = input.find(DECIMAL_PT) {
++ let (leading, trailing) = input.split_at(s);
++ let output = [leading, ".", trailing.replace(DECIMAL_PT, "").as_str()].concat();
++ Cow::Owned(output)
++ } else {
++ input
+ }
+ }
+
+ /// Parse the beginning string into an f64, returning -inf instead of NaN on errors.
+ #[inline(always)]
+ fn permissive_f64_parse(a: &str) -> f64 {
+- // Remove thousands seperators
+- let a = a.replace(THOUSANDS_SEP, "");
+-
+ // GNU sort treats "NaN" as non-number in numeric, so it needs special care.
+ // *Keep this trim before parse* despite what POSIX may say about -b and -n
+ // because GNU and BSD both seem to require it to match their behavior
+- match a.trim().parse::<f64>() {
++ //
++ // Remove any trailing decimals, ie 4568..890... becomes 4568.890
++ // Then, we trim whitespace and parse
++ match remove_trailing_dec(a).trim().parse::<f64>() {
+ Ok(a) if a.is_nan() => std::f64::NEG_INFINITY,
+ Ok(a) => a,
+ Err(_) => std::f64::NEG_INFINITY,
+@@ -757,8 +794,13 @@ fn numeric_compare(a: &str, b: &str) -> Ordering {
+ let sa = get_leading_num(a);
+ let sb = get_leading_num(b);
+
+- let fa = permissive_f64_parse(sa);
+- let fb = permissive_f64_parse(sb);
++ // Avoids a string alloc for every line to remove thousands seperators here
++ // instead of inside the get_leading_num function, which is a HUGE performance benefit
++ let ta = remove_thousands_sep(sa);
++ let tb = remove_thousands_sep(sb);
++
++ let fa = permissive_f64_parse(&ta);
++ let fb = permissive_f64_parse(&tb);
+
+ // f64::cmp isn't implemented (due to NaN issues); implement directly instead
+ if fa > fb {
+@@ -799,8 +841,8 @@ fn general_numeric_compare(a: &str, b: &str) -> Ordering {
+ // these types of numbers, we rarely care about pure performance.
+ fn human_numeric_convert(a: &str) -> f64 {
+ let num_str = get_leading_num(a);
+- let suffix = a.trim_start_matches(num_str);
+- let num_part = permissive_f64_parse(num_str);
++ let suffix = a.trim_start_matches(&num_str);
++ let num_part = permissive_f64_parse(&num_str);
+ let suffix: f64 = match suffix.parse().unwrap_or('\0') {
+ // SI Units
+ 'K' => 1E3,
+diff --git a/tests/.DS_Store b/tests/.DS_Store
+new file mode 100644
+index 0000000..3cbbb78
+--- /dev/null
++++ b/tests/.DS_Store
+@@ -0,0 +1 @@
++���Bud1������������%���������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������� ������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������ ������@���������������������������������������� ������@������������������������������������������ ������@������������������������������������������ ������@�������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������� ���E���%�������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������DSDB�����������������������������`����������������������������������������������� ������@������������������������������������������ ������@������������������������������������������ ������@��������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������
+\ No newline at end of file
+diff --git a/tests/by-util/test_sort.rs b/tests/by-util/test_sort.rs
+index 43aaf1d..6455d83 100644
+--- a/tests/by-util/test_sort.rs
++++ b/tests/by-util/test_sort.rs
+@@ -1,5 +1,31 @@
+ use crate::common::util::*;
+
++fn test_helper(file_name: &str, args: &str) {
++ new_ucmd!()
++ .arg(args)
++ .arg(format!("{}.txt", file_name))
++ .succeeds()
++ .stdout_is_fixture(format!("{}.expected", file_name));
++}
++
++#[test]
++fn test_multiple_decimals_general() {
++ new_ucmd!()
++ .arg("-g")
++ .arg("multiple_decimals_general.txt")
++ .succeeds()
++ .stdout_is("\n\n\n\n\n\n\n\nCARAvan\n-2028789030\n-896689\n-8.90880\n-1\n-.05\n000\n00000001\n1\n1.040000000\n1.444\n1.58590\n8.013\n45\n46.89\n576,446.88800000\n576,446.890\n 4567.\n4567.1\n4567.34\n\t\t\t\t\t\t\t\t\t\t4567..457\n\t\t\t\t37800\n\t\t\t\t\t\t45670.89079.098\n\t\t\t\t\t\t45670.89079.1\n4798908.340000000000\n4798908.45\n4798908.8909800\n");
++}
++
++#[test]
++fn test_multiple_decimals_numeric() {
++ new_ucmd!()
++ .arg("-n")
++ .arg("multiple_decimals_numeric.txt")
++ .succeeds()
++ .stdout_is("-2028789030\n-896689\n-8.90880\n-1\n-.05\n\n\n\n\n\n\n\n\n000\nCARAvan\n00000001\n1\n1.040000000\n1.444\n1.58590\n8.013\n45\n46.89\n 4567.\n4567.1\n4567.34\n\t\t\t\t\t\t\t\t\t\t4567..457\n\t\t\t\t37800\n\t\t\t\t\t\t45670.89079.098\n\t\t\t\t\t\t45670.89079.1\n576,446.88800000\n576,446.890\n4798908.340000000000\n4798908.45\n4798908.8909800\n");
++}
++
+ #[test]
+ fn test_check_zero_terminated_failure() {
+ new_ucmd!()
+@@ -44,6 +70,21 @@ fn test_random_shuffle_contains_all_lines() {
+ assert_eq!(result_sorted, expected);
+ }
+
++#[test]
++fn test_random_shuffle_two_runs_not_the_same() {
++ // check to verify that two random shuffles are not equal; this has the
++ // potential to fail in the very unlikely event that the 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;
++ let expected = at.read(FILE);
++ let unexpected = new_ucmd!().arg("-R").arg(FILE).run().stdout;
++
++ assert_ne!(result, expected);
++ 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
+@@ -355,11 +396,3 @@ fn test_check_silent() {
+ .fails()
+ .stdout_is("");
+ }
+-
+-fn test_helper(file_name: &str, args: &str) {
+- new_ucmd!()
+- .arg(args)
+- .arg(format!("{}{}", file_name, ".txt"))
+- .succeeds()
+- .stdout_is_fixture(format!("{}{}", file_name, ".expected"));
+-}
+diff --git a/tests/fixtures/.DS_Store b/tests/fixtures/.DS_Store
+new file mode 100644
+index 0000000..05adc2c
+--- /dev/null
++++ b/tests/fixtures/.DS_Store
+@@ -0,0 +1 @@
++���Bud1����������� ����������������������������������������������������������bwspblob�����������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������s�o�r�tbwspblob����bplist00�
+diff --git a/tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique_reverse_stable.expected b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique_reverse_stable.expected
+new file mode 100644
+index 0000000..bbce169
+--- /dev/null
++++ b/tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique_reverse_stable.expected
+@@ -0,0 +1,20 @@
++4798908.8909800
++4798908.45
++4798908.340000000000
++576,446.890
++576,446.88800000
++ 37800
++ 4567.
++46.89
++45
++8.013
++1.58590
++1.444
++1.040000000
++1
++
++-.05
++-1
++-8.90880
++-896689
++-2028789030
+diff --git a/tests/fixtures/sort/multiple_decimals.expected b/tests/fixtures/sort/multiple_decimals.expected
+new file mode 100644
+index 0000000..6afbdca
+--- /dev/null
++++ b/tests/fixtures/sort/multiple_decimals.expected
+@@ -0,0 +1,33 @@
++-2028789030
++-896689
++-8.90880
++-1
++-.05
++
++
++
++
++
++
++
++
++000
++CARAvan
++00000001
++1
++1.040000000
++1.444
++1.58590
++8.013
++45
++46.89
++ 4567..457
++ 4567.
++4567.1
++4567.34
++ 37800
++576,446.88800000
++576,446.890
++4798908.340000000000
++4798908.45
++4798908.8909800
+diff --git a/tests/fixtures/sort/multiple_decimals_general.txt b/tests/fixtures/sort/multiple_decimals_general.txt
+new file mode 100644
+index 0000000..4e65ecf
+--- /dev/null
++++ b/tests/fixtures/sort/multiple_decimals_general.txt
+@@ -0,0 +1,35 @@
++576,446.890
++576,446.88800000
++
++4567.1
++ 4567..457
++ 45670.89079.1
++ 45670.89079.098
++4567.34
++ 4567.
++45
++46.89
++-1
++1
++00000001
++4798908.340000000000
++4798908.45
++4798908.8909800
++
++
++ 37800
++
++-2028789030
++-896689
++CARAvan
++
++-8.90880
++-.05
++1.444
++1.58590
++1.040000000
++
++8.013
++
++000
++
+diff --git a/tests/fixtures/sort/multiple_decimals_numeric.txt b/tests/fixtures/sort/multiple_decimals_numeric.txt
+new file mode 100644
+index 0000000..4e65ecf
+--- /dev/null
++++ b/tests/fixtures/sort/multiple_decimals_numeric.txt
+@@ -0,0 +1,35 @@
++576,446.890
++576,446.88800000
++
++4567.1
++ 4567..457
++ 45670.89079.1
++ 45670.89079.098
++4567.34
++ 4567.
++45
++46.89
++-1
++1
++00000001
++4798908.340000000000
++4798908.45
++4798908.8909800
++
++
++ 37800
++
++-2028789030
++-896689
++CARAvan
++
++-8.90880
++-.05
++1.444
++1.58590
++1.040000000
++
++8.013
++
++000
++
diff --git a/debian/patches/series b/debian/patches/series
index 79e2c98..d0b7b0f 100644
--- a/debian/patches/series
+++ b/debian/patches/series
@@ -10,3 +10,9 @@ revert-file-diff-0.diff
revert-file-diff.diff
revert-twox-hash.dif
revert-retain-mut.diff
+
+# Backported from upstream
+Sort-Implement-stable-sort-ignore-non-printing-month.patch
+Ignore-a-test.patch
+Sort-Various-fixes-and-performance-improvements.patch
+sort-implement-k-and-t-support.patch
diff --git a/debian/patches/sort-implement-k-and-t-support.patch b/debian/patches/sort-implement-k-and-t-support.patch
new file mode 100644
index 0000000..85f0d6d
--- /dev/null
+++ b/debian/patches/sort-implement-k-and-t-support.patch
@@ -0,0 +1,1333 @@
+From: Michael Debertol <michael.debertol@gmail.com>
+Date: Sat, 10 Apr 2021 14:54:58 +0200
+Subject: sort: implement -k and -t support (#1996)
+
+* sort: implement basic -k and -t support
+
+This allows to specify keys after the -k flag and a custom field
+separator using -t.
+
+Support for options for specific keys is still missing, and the -b flag
+is not passed down correctly.
+
+* sort: implement support for key options
+
+* remove unstable feature use
+
+* don't pipe in input when we expect a failure
+
+* only tokenize when needed, remove a clone()
+
+* improve comments
+
+* fix clippy lints
+
+* re-add test
+
+* buffer writes to stdout
+
+* fix ignore_non_printing
+
+and make the test fail in case it is broken :)
+
+* move attribute to the right position
+
+* add more tests
+
+* add my name to the copyright section
+
+* disallow dead code
+
+* move a comment
+
+* re-add a loc
+
+* use smallvec for a perf improvement in the common case
+
+* add BENCHMARKING.md
+
+* add ignore_case to benchmarks
+
+Origin: upstream, https://github.com/uutils/coreutils/commit/49c9d8c9018eeeb2ae68a43f3763e087c44b1653
+---
+ Cargo.lock | 10 +
+ src/uu/sort/BENCHMARKING.md | 33 +++
+ src/uu/sort/Cargo.toml | 1 +
+ src/uu/sort/src/sort.rs | 672 +++++++++++++++++++++++++++++++++++---------
+ tests/by-util/test_sort.rs | 164 ++++++++++-
+ 5 files changed, 741 insertions(+), 139 deletions(-)
+ create mode 100644 src/uu/sort/BENCHMARKING.md
+
+diff --git a/Cargo.lock b/Cargo.lock
+index 53eb8b4..6c311d9 100644
+--- a/Cargo.lock
++++ b/Cargo.lock
+@@ -1,5 +1,7 @@
+ # This file is automatically @generated by Cargo.
+ # It is not intended for manual editing.
++version = 3
++
+ [[package]]
+ name = "advapi32-sys"
+ version = "0.2.0"
+@@ -994,6 +996,12 @@ dependencies = [
+ "maybe-uninit",
+ ]
+
++[[package]]
++name = "smallvec"
++version = "1.6.1"
++source = "registry+https://github.com/rust-lang/crates.io-index"
++checksum = "fe0f37c9e8f3c5a4a66ad655a93c74daac4ad00c441533bf5c6e7990bb42604e"
++
+ [[package]]
+ name = "strsim"
+ version = "0.8.0"
+@@ -1456,6 +1464,7 @@ checksum = "f5220db7016a5929da5cbb7eaca38a282ec1c98681c0a08cba3dfd8a83cd4053"
+ dependencies = [
+ "num-traits",
+ "rand 0.7.3",
++ "rand_chacha",
+ "smallvec",
+ "uucore",
+ "uucore_procs",
+@@ -2016,6 +2025,7 @@ dependencies = [
+ "rand 0.7.3",
+ "rayon",
+ "semver",
++ "smallvec 1.6.1",
+ "uucore",
+ "uucore_procs",
+ ]
+diff --git a/src/uu/sort/BENCHMARKING.md b/src/uu/sort/BENCHMARKING.md
+new file mode 100644
+index 0000000..b20db01
+--- /dev/null
++++ b/src/uu/sort/BENCHMARKING.md
+@@ -0,0 +1,33 @@
++# Benchmarking sort
++
++Most of the time when sorting is spent comparing lines. The comparison functions however differ based
++on which arguments are passed to `sort`, therefore it is important to always benchmark multiple scenarios.
++This is an overwiew over what was benchmarked, and if you make changes to `sort`, you are encouraged to check
++how performance was affected for the workloads listed below. Feel free to add other workloads to the
++list that we should improve / make sure not to regress.
++
++Run `cargo build --release` before benchmarking after you make a change!
++
++## Sorting a wordlist
++- Get a wordlist, for example with [words](https://en.wikipedia.org/wiki/Words_(Unix)) on Linux. The exact wordlist
++ doesn't matter for performance comparisons. In this example I'm using `/usr/share/dict/american-english` as the wordlist.
++- Shuffle the wordlist by running `sort -R /usr/share/dict/american-english > shuffled_wordlist.txt`.
++- Benchmark sorting the wordlist with hyperfine: `hyperfine "target/release/coreutils sort shuffled_wordlist.txt -o output.txt"`.
++
++## Sorting a wordlist with ignore_case
++- Same wordlist as above
++- Benchmark sorting the wordlist ignoring the case with hyperfine: `hyperfine "target/release/coreutils sort shuffled_wordlist.txt -f -o output.txt"`.
++
++## Sorting numbers
++- Generate a list of numbers: `seq 0 100000 | sort -R > shuffled_numbers.txt`.
++- Benchmark numeric sorting with hyperfine: `hyperfine "target/release/coreutils sort shuffled_numbers.txt -n -o output.txt"`.
++
++## Stdout and stdin performance
++Try to run the above benchmarks by piping the input through stdin (standard input) and redirect the
++output through stdout (standard output):
++- Remove the input file from the arguments and add `cat [inputfile] | ` at the beginning.
++- Remove `-o output.txt` and add `> output.txt` at the end.
++
++Example: `hyperfine "target/release/coreutils sort shuffled_numbers.txt -n -o output.txt"` becomes
++`hyperfine "cat shuffled_numbers.txt | target/release/coreutils sort -n > output.txt`
++- Check that performance is similar to the original benchmark.
+\ No newline at end of file
+diff --git a/src/uu/sort/Cargo.toml b/src/uu/sort/Cargo.toml
+index cfcd74d..ddea900 100644
+--- a/src/uu/sort/Cargo.toml
++++ b/src/uu/sort/Cargo.toml
+@@ -21,6 +21,7 @@ clap = "2.33"
+ fnv = "1.0.7"
+ itertools = "0.9"
+ semver = "0.9.0"
++smallvec = "1.6.1"
+ uucore = { version=">=0.0.8", package="uucore", path="../../uucore", features=["fs"] }
+ uucore_procs = { version=">=0.0.5", package="uucore_procs", path="../../uucore_procs" }
+
+diff --git a/src/uu/sort/src/sort.rs b/src/uu/sort/src/sort.rs
+index 88f7bc7..7b1fe07 100644
+--- a/src/uu/sort/src/sort.rs
++++ b/src/uu/sort/src/sort.rs
+@@ -2,10 +2,10 @@
+ // *
+ // * (c) Michael Yin <mikeyin@mikeyin.org>
+ // * (c) Robert Swinford <robert.swinford..AT..gmail.com>
++// * (c) Michael Debertol <michael.debertol..AT..gmail.com>
+ // *
+ // * For the full copyright and license information, please view the LICENSE
+ // * file that was distributed with this source code.
+-#![allow(dead_code)]
+
+ // Although these links don't always seem to describe reality, check out the POSIX and GNU specs:
+ // https://pubs.opengroup.org/onlinepubs/9699919799/utilities/sort.html
+@@ -22,6 +22,7 @@ use rand::distributions::Alphanumeric;
+ use rand::{thread_rng, Rng};
+ use rayon::prelude::*;
+ use semver::Version;
++use smallvec::SmallVec;
+ use std::borrow::Cow;
+ use std::cmp::Ordering;
+ use std::collections::BinaryHeap;
+@@ -30,6 +31,7 @@ use std::fs::File;
+ use std::hash::{Hash, Hasher};
+ use std::io::{stdin, stdout, BufRead, BufReader, BufWriter, Lines, Read, Write};
+ use std::mem::replace;
++use std::ops::{Range, RangeInclusive};
+ use std::path::Path;
+ use uucore::fs::is_stdin_interactive; // for Iterator::dedup()
+
+@@ -37,6 +39,16 @@ static NAME: &str = "sort";
+ static ABOUT: &str = "Display sorted concatenation of all FILE(s).";
+ static VERSION: &str = env!("CARGO_PKG_VERSION");
+
++const LONG_HELP_KEYS: &str = "The key format is FIELD[.CHAR][OPTIONS][,FIELD[.CHAR]][OPTIONS].
++
++Fields by default are separated by the first whitespace after a non-whitespace character. Use -t to specify a custom separator.
++In the default case, whitespace is appended at the beginning of each field. Custom separators however are not included in fields.
++
++FIELD and CHAR both start at 1 (i.e. they are 1-indexed). If there is no end specified after a comma, the end will be the end of the line.
++If CHAR is set 0, it means the end of the field. CHAR defaults to 1 for the start position and to 0 for the end position.
++
++Valid options are: MbdfhnRrV. They override the global options for this key.";
++
+ static OPT_HUMAN_NUMERIC_SORT: &str = "human-numeric-sort";
+ static OPT_MONTH_SORT: &str = "month-sort";
+ static OPT_NUMERIC_SORT: &str = "numeric-sort";
+@@ -54,6 +66,8 @@ static OPT_OUTPUT: &str = "output";
+ static OPT_REVERSE: &str = "reverse";
+ static OPT_STABLE: &str = "stable";
+ static OPT_UNIQUE: &str = "unique";
++static OPT_KEY: &str = "key";
++static OPT_SEPARATOR: &str = "field-separator";
+ static OPT_RANDOM: &str = "random-sort";
+ static OPT_ZERO_TERMINATED: &str = "zero-terminated";
+ static OPT_PARALLEL: &str = "parallel";
+@@ -63,10 +77,11 @@ static ARG_FILES: &str = "files";
+
+ static DECIMAL_PT: char = '.';
+ static THOUSANDS_SEP: char = ',';
++
+ static NEGATIVE: char = '-';
+ static POSITIVE: char = '+';
+
+-#[derive(Eq, Ord, PartialEq, PartialOrd)]
++#[derive(Eq, Ord, PartialEq, PartialOrd, Clone)]
+ enum SortMode {
+ Numeric,
+ HumanNumeric,
+@@ -76,8 +91,12 @@ enum SortMode {
+ Default,
+ }
+
+-struct Settings {
++struct GlobalSettings {
+ mode: SortMode,
++ ignore_blanks: bool,
++ ignore_case: bool,
++ dictionary_order: bool,
++ ignore_non_printing: bool,
+ merge: bool,
+ reverse: bool,
+ outfile: Option<String>,
+@@ -86,17 +105,21 @@ struct Settings {
+ check: bool,
+ check_silent: bool,
+ random: bool,
+- compare_fn: fn(&str, &str) -> Ordering,
+- transform_fns: Vec<fn(&str) -> String>,
+- threads: String,
+ salt: String,
++ selectors: Vec<FieldSelector>,
++ separator: Option<char>,
++ threads: String,
+ zero_terminated: bool,
+ }
+
+-impl Default for Settings {
+- fn default() -> Settings {
+- Settings {
++impl Default for GlobalSettings {
++ fn default() -> GlobalSettings {
++ GlobalSettings {
+ mode: SortMode::Default,
++ ignore_blanks: false,
++ ignore_case: false,
++ dictionary_order: false,
++ ignore_non_printing: false,
+ merge: false,
+ reverse: false,
+ outfile: None,
+@@ -105,19 +128,330 @@ impl Default for Settings {
+ check: false,
+ check_silent: false,
+ random: false,
+- compare_fn: default_compare,
+- transform_fns: Vec::new(),
+- threads: String::new(),
+ salt: String::new(),
++ selectors: vec![],
++ separator: None,
++ threads: String::new(),
+ zero_terminated: false,
+ }
+ }
+ }
+
++struct KeySettings {
++ mode: SortMode,
++ ignore_blanks: bool,
++ ignore_case: bool,
++ dictionary_order: bool,
++ ignore_non_printing: bool,
++ random: bool,
++ reverse: bool,
++}
++
++impl From<&GlobalSettings> for KeySettings {
++ fn from(settings: &GlobalSettings) -> Self {
++ Self {
++ mode: settings.mode.clone(),
++ ignore_blanks: settings.ignore_blanks,
++ ignore_case: settings.ignore_case,
++ ignore_non_printing: settings.ignore_non_printing,
++ random: settings.random,
++ reverse: settings.reverse,
++ dictionary_order: settings.dictionary_order,
++ }
++ }
++}
++
++/// Represents the string selected by a FieldSelector.
++enum Selection {
++ /// If we had to transform this selection, we have to store a new string.
++ String(String),
++ /// If there was no transformation, we can store an index into the line.
++ ByIndex(Range<usize>),
++}
++
++impl Selection {
++ /// Gets the actual string slice represented by this Selection.
++ fn get_str<'a>(&'a self, line: &'a Line) -> &'a str {
++ match self {
++ Selection::String(string) => string.as_str(),
++ Selection::ByIndex(range) => &line.line[range.to_owned()],
++ }
++ }
++}
++
++type Field = Range<usize>;
++
++struct Line {
++ line: String,
++ // The common case is not to specify fields. Let's make this fast.
++ selections: SmallVec<[Selection; 1]>,
++}
++
++impl Line {
++ fn new(line: String, settings: &GlobalSettings) -> Self {
++ let fields = if settings
++ .selectors
++ .iter()
++ .any(|selector| selector.needs_tokens())
++ {
++ // Only tokenize if we will need tokens.
++ Some(tokenize(&line, settings.separator))
++ } else {
++ None
++ };
++
++ let selections = settings
++ .selectors
++ .iter()
++ .map(|selector| {
++ if let Some(range) = selector.get_selection(&line, fields.as_deref()) {
++ if let Some(transformed) =
++ transform(&line[range.to_owned()], &selector.settings)
++ {
++ Selection::String(transformed)
++ } else {
++ Selection::ByIndex(range.start().to_owned()..range.end() + 1)
++ }
++ } else {
++ // If there is no match, match the empty string.
++ Selection::ByIndex(0..0)
++ }
++ })
++ .collect();
++ Self { line, selections }
++ }
++}
++
++/// Transform this line. Returns None if there's no need to transform.
++fn transform(line: &str, settings: &KeySettings) -> Option<String> {
++ let mut transformed = None;
++ if settings.ignore_case {
++ transformed = Some(line.to_uppercase());
++ }
++ if settings.ignore_blanks {
++ transformed = Some(
++ transformed
++ .as_deref()
++ .unwrap_or(line)
++ .trim_start()
++ .to_string(),
++ );
++ }
++ if settings.dictionary_order {
++ transformed = Some(remove_nondictionary_chars(
++ transformed.as_deref().unwrap_or(line),
++ ));
++ }
++ if settings.ignore_non_printing {
++ transformed = Some(remove_nonprinting_chars(
++ transformed.as_deref().unwrap_or(line),
++ ));
++ }
++ transformed
++}
++
++/// Tokenize a line into fields.
++fn tokenize(line: &str, separator: Option<char>) -> Vec<Field> {
++ if let Some(separator) = separator {
++ tokenize_with_separator(line, separator)
++ } else {
++ tokenize_default(line)
++ }
++}
++
++/// By default fields are separated by the first whitespace after non-whitespace.
++/// Whitespace is included in fields at the start.
++fn tokenize_default(line: &str) -> Vec<Field> {
++ let mut tokens = vec![0..0];
++ // pretend that there was whitespace in front of the line
++ let mut previous_was_whitespace = true;
++ for (idx, char) in line.char_indices() {
++ if char.is_whitespace() {
++ if !previous_was_whitespace {
++ tokens.last_mut().unwrap().end = idx;
++ tokens.push(idx..0);
++ }
++ previous_was_whitespace = true;
++ } else {
++ previous_was_whitespace = false;
++ }
++ }
++ tokens.last_mut().unwrap().end = line.len();
++ tokens
++}
++
++/// Split between separators. These separators are not included in fields.
++fn tokenize_with_separator(line: &str, separator: char) -> Vec<Field> {
++ let mut tokens = vec![0..0];
++ let mut previous_was_separator = false;
++ for (idx, char) in line.char_indices() {
++ if previous_was_separator {
++ tokens.push(idx..0);
++ }
++ if char == separator {
++ tokens.last_mut().unwrap().end = idx;
++ previous_was_separator = true;
++ } else {
++ previous_was_separator = false;
++ }
++ }
++ tokens.last_mut().unwrap().end = line.len();
++ tokens
++}
++
++struct KeyPosition {
++ /// 1-indexed, 0 is invalid.
++ field: usize,
++ /// 1-indexed, 0 is end of field.
++ char: usize,
++ ignore_blanks: bool,
++}
++
++impl KeyPosition {
++ fn parse(key: &str, default_char_index: usize, settings: &mut KeySettings) -> Self {
++ let mut field_and_char = key.split('.');
++ let mut field = field_and_char
++ .next()
++ .unwrap_or_else(|| crash!(1, "invalid key `{}`", key));
++ let mut char = field_and_char.next();
++
++ // If there is a char index, we expect options to appear after it. Otherwise we expect them after the field index.
++ let value_with_options = char.as_mut().unwrap_or(&mut field);
++
++ let mut ignore_blanks = settings.ignore_blanks;
++ if let Some(options_start) = value_with_options.chars().position(char::is_alphabetic) {
++ for option in value_with_options[options_start..].chars() {
++ // valid options: MbdfghinRrV
++ match option {
++ 'M' => settings.mode = SortMode::Month,
++ 'b' => ignore_blanks = true,
++ 'd' => settings.dictionary_order = true,
++ 'f' => settings.ignore_case = true,
++ 'g' => settings.mode = SortMode::GeneralNumeric,
++ 'h' => settings.mode = SortMode::HumanNumeric,
++ 'i' => settings.ignore_non_printing = true,
++ 'n' => settings.mode = SortMode::Numeric,
++ 'R' => settings.random = true,
++ 'r' => settings.reverse = true,
++ 'V' => settings.mode = SortMode::Version,
++ c => {
++ crash!(1, "invalid option for key: `{}`", c)
++ }
++ }
++ }
++ // Strip away option characters from the original value so we can parse it later
++ *value_with_options = &value_with_options[..options_start];
++ }
++
++ let field = field
++ .parse()
++ .unwrap_or_else(|e| crash!(1, "failed to parse field index for key `{}`: {}", key, e));
++ if field == 0 {
++ crash!(1, "field index was 0");
++ }
++ let char = char.map_or(default_char_index, |char| {
++ char.parse().unwrap_or_else(|e| {
++ crash!(
++ 1,
++ "failed to parse character index for key `{}`: {}",
++ key,
++ e
++ )
++ })
++ });
++ Self {
++ field,
++ char,
++ ignore_blanks,
++ }
++ }
++}
++
++struct FieldSelector {
++ from: KeyPosition,
++ to: Option<KeyPosition>,
++ settings: KeySettings,
++}
++
++impl FieldSelector {
++ fn needs_tokens(&self) -> bool {
++ self.from.field != 1 || self.from.char == 0 || self.to.is_some()
++ }
++
++ /// Look up the slice that corresponds to this selector for the given line.
++ /// If needs_fields returned false, fields may be None.
++ fn get_selection<'a>(
++ &self,
++ line: &'a str,
++ tokens: Option<&[Field]>,
++ ) -> Option<RangeInclusive<usize>> {
++ enum ResolutionErr {
++ TooLow,
++ TooHigh,
++ }
++
++ // Get the index for this line given the KeyPosition
++ fn resolve_index(
++ line: &str,
++ tokens: Option<&[Field]>,
++ position: &KeyPosition,
++ ) -> Result<usize, ResolutionErr> {
++ if tokens.map_or(false, |fields| fields.len() < position.field) {
++ Err(ResolutionErr::TooHigh)
++ } else if position.char == 0 {
++ let end = tokens.unwrap()[position.field - 1].end;
++ if end == 0 {
++ Err(ResolutionErr::TooLow)
++ } else {
++ Ok(end - 1)
++ }
++ } else {
++ let mut idx = if position.field == 1 {
++ // The first field always starts at 0.
++ // We don't need tokens for this case.
++ 0
++ } else {
++ tokens.unwrap()[position.field - 1].start
++ } + position.char
++ - 1;
++ if idx >= line.len() {
++ Err(ResolutionErr::TooHigh)
++ } else {
++ if position.ignore_blanks {
++ if let Some(not_whitespace) =
++ line[idx..].chars().position(|c| !c.is_whitespace())
++ {
++ idx += not_whitespace;
++ } else {
++ return Err(ResolutionErr::TooHigh);
++ }
++ }
++ Ok(idx)
++ }
++ }
++ }
++
++ if let Ok(from) = resolve_index(line, tokens, &self.from) {
++ let to = self.to.as_ref().map(|to| resolve_index(line, tokens, &to));
++ match to {
++ Some(Ok(to)) => Some(from..=to),
++ // If `to` was not given or the match would be after the end of the line,
++ // match everything until the end of the line.
++ None | Some(Err(ResolutionErr::TooHigh)) => Some(from..=line.len() - 1),
++ // If `to` is before the start of the line, report no match.
++ // This can happen if the line starts with a separator.
++ Some(Err(ResolutionErr::TooLow)) => None,
++ }
++ } else {
++ None
++ }
++ }
++}
++
+ struct MergeableFile<'a> {
+ lines: Lines<BufReader<Box<dyn Read>>>,
+- current_line: String,
+- settings: &'a Settings,
++ current_line: Line,
++ settings: &'a GlobalSettings,
+ }
+
+ // BinaryHeap depends on `Ord`. Note that we want to pop smallest items
+@@ -125,7 +459,7 @@ struct MergeableFile<'a> {
+ // trick it into the right order by calling reverse() here.
+ impl<'a> Ord for MergeableFile<'a> {
+ fn cmp(&self, other: &MergeableFile) -> Ordering {
+- compare_by(&self.current_line, &other.current_line, &self.settings).reverse()
++ compare_by(&self.current_line, &other.current_line, self.settings).reverse()
+ }
+ }
+
+@@ -137,7 +471,7 @@ impl<'a> PartialOrd for MergeableFile<'a> {
+
+ impl<'a> PartialEq for MergeableFile<'a> {
+ fn eq(&self, other: &MergeableFile) -> bool {
+- Ordering::Equal == compare_by(&self.current_line, &other.current_line, &self.settings)
++ Ordering::Equal == compare_by(&self.current_line, &other.current_line, self.settings)
+ }
+ }
+
+@@ -145,11 +479,11 @@ impl<'a> Eq for MergeableFile<'a> {}
+
+ struct FileMerger<'a> {
+ heap: BinaryHeap<MergeableFile<'a>>,
+- settings: &'a Settings,
++ settings: &'a GlobalSettings,
+ }
+
+ impl<'a> FileMerger<'a> {
+- fn new(settings: &'a Settings) -> FileMerger<'a> {
++ fn new(settings: &'a GlobalSettings) -> FileMerger<'a> {
+ FileMerger {
+ heap: BinaryHeap::new(),
+ settings,
+@@ -159,7 +493,7 @@ impl<'a> FileMerger<'a> {
+ if let Some(Ok(next_line)) = lines.next() {
+ let mergeable_file = MergeableFile {
+ lines,
+- current_line: next_line,
++ current_line: Line::new(next_line, &self.settings),
+ settings: &self.settings,
+ };
+ self.heap.push(mergeable_file);
+@@ -174,14 +508,17 @@ impl<'a> Iterator for FileMerger<'a> {
+ Some(mut current) => {
+ match current.lines.next() {
+ Some(Ok(next_line)) => {
+- let ret = replace(&mut current.current_line, next_line);
++ let ret = replace(
++ &mut current.current_line,
++ Line::new(next_line, &self.settings),
++ );
+ self.heap.push(current);
+- Some(ret)
++ Some(ret.line)
+ }
+ _ => {
+ // Don't put it back in the heap (it's empty/erroring)
+ // but its first line is still valid.
+- Some(current.current_line)
++ Some(current.current_line.line)
+ }
+ }
+ }
+@@ -205,7 +542,7 @@ With no FILE, or when FILE is -, read standard input.",
+ pub fn uumain(args: impl uucore::Args) -> i32 {
+ let args = args.collect_str();
+ let usage = get_usage();
+- let mut settings: Settings = Default::default();
++ let mut settings: GlobalSettings = Default::default();
+
+ let matches = App::new(executable!())
+ .version(VERSION)
+@@ -316,7 +653,21 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
+ .help("output only the first of an equal run"),
+ )
+ .arg(
+- Arg::with_name(OPT_ZERO_TERMINATED)
++ Arg::with_name(OPT_KEY)
++ .short("k")
++ .long(OPT_KEY)
++ .help("sort by a key")
++ .long_help(LONG_HELP_KEYS)
++ .multiple(true)
++ .takes_value(true),
++ )
++ .arg(
++ Arg::with_name(OPT_SEPARATOR)
++ .short("t")
++ .long(OPT_SEPARATOR)
++ .help("custom separator for -k")
++ .takes_value(true))
++ .arg(Arg::with_name(OPT_ZERO_TERMINATED)
+ .short("z")
+ .long(OPT_ZERO_TERMINATED)
+ .help("line delimiter is NUL, not newline"),
+@@ -350,14 +701,12 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
+ for path in &files0_from {
+ let (reader, _) = open(path.as_str()).expect("Could not read from file specified.");
+ let buf_reader = BufReader::new(reader);
+- for line in buf_reader.split(b'\0') {
+- if let Ok(n) = line {
+- files.push(
+- std::str::from_utf8(&n)
+- .expect("Could not parse string from zero terminated input.")
+- .to_string(),
+- );
+- }
++ for line in buf_reader.split(b'\0').flatten() {
++ files.push(
++ std::str::from_utf8(&line)
++ .expect("Could not parse string from zero terminated input.")
++ .to_string(),
++ );
+ }
+ }
+ files
+@@ -382,21 +731,17 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
+ SortMode::Default
+ };
+
++ settings.dictionary_order = matches.is_present(OPT_DICTIONARY_ORDER);
++ settings.ignore_non_printing = matches.is_present(OPT_IGNORE_NONPRINTING);
+ if matches.is_present(OPT_PARALLEL) {
+ // "0" is default - threads = num of cores
+ settings.threads = matches
+ .value_of(OPT_PARALLEL)
+ .map(String::from)
+- .unwrap_or("0".to_string());
++ .unwrap_or_else(|| "0".to_string());
+ env::set_var("RAYON_NUM_THREADS", &settings.threads);
+ }
+
+- if matches.is_present(OPT_DICTIONARY_ORDER) {
+- settings.transform_fns.push(remove_nondictionary_chars);
+- } else if matches.is_present(OPT_IGNORE_NONPRINTING) {
+- settings.transform_fns.push(remove_nonprinting_chars);
+- }
+-
+ settings.zero_terminated = matches.is_present(OPT_ZERO_TERMINATED);
+ settings.merge = matches.is_present(OPT_MERGE);
+
+@@ -406,13 +751,9 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
+ settings.check = true;
+ };
+
+- if matches.is_present(OPT_IGNORE_CASE) {
+- settings.transform_fns.push(|s| s.to_uppercase());
+- }
++ settings.ignore_case = matches.is_present(OPT_IGNORE_CASE);
+
+- if matches.is_present(OPT_IGNORE_BLANKS) {
+- settings.transform_fns.push(|s| s.trim_start().to_string());
+- }
++ settings.ignore_blanks = matches.is_present(OPT_IGNORE_BLANKS);
+
+ settings.outfile = matches.value_of(OPT_OUTPUT).map(String::from);
+ settings.reverse = matches.is_present(OPT_REVERSE);
+@@ -424,27 +765,64 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
+ settings.salt = get_rand_string();
+ }
+
+- //let mut files = matches.free;
+ if files.is_empty() {
+ /* if no file, default to stdin */
+ files.push("-".to_owned());
+ } else if settings.check && files.len() != 1 {
+- crash!(1, "sort: extra operand `{}' not allowed with -c", files[1])
++ crash!(1, "extra operand `{}' not allowed with -c", files[1])
+ }
+
+- settings.compare_fn = match settings.mode {
+- SortMode::Numeric => numeric_compare,
+- SortMode::GeneralNumeric => general_numeric_compare,
+- SortMode::HumanNumeric => human_numeric_size_compare,
+- SortMode::Month => month_compare,
+- SortMode::Version => version_compare,
+- SortMode::Default => default_compare,
+- };
++ if let Some(arg) = matches.args.get(OPT_SEPARATOR) {
++ let separator = arg.vals[0].to_string_lossy();
++ let separator = separator;
++ if separator.len() != 1 {
++ crash!(1, "separator must be exactly one character long");
++ }
++ settings.separator = Some(separator.chars().next().unwrap())
++ }
+
+- exec(files, &mut settings)
++ if matches.is_present(OPT_KEY) {
++ for key in &matches.args[OPT_KEY].vals {
++ let key = key.to_string_lossy();
++ let mut from_to = key.split(',');
++ let mut key_settings = KeySettings::from(&settings);
++ let from = KeyPosition::parse(
++ from_to
++ .next()
++ .unwrap_or_else(|| crash!(1, "invalid key `{}`", key)),
++ 1,
++ &mut key_settings,
++ );
++ let to = from_to
++ .next()
++ .map(|to| KeyPosition::parse(to, 0, &mut key_settings));
++ let field_selector = FieldSelector {
++ from,
++ to,
++ settings: key_settings,
++ };
++ settings.selectors.push(field_selector);
++ }
++ }
++
++ if !settings.stable || !matches.is_present(OPT_KEY) {
++ // add a default selector matching the whole line
++ let key_settings = KeySettings::from(&settings);
++ settings.selectors.push(FieldSelector {
++ from: KeyPosition {
++ field: 1,
++ char: 1,
++ ignore_blanks: key_settings.ignore_blanks,
++ },
++ to: None,
++ settings: key_settings,
++ });
++ }
++
++ exec(files, &settings)
+ }
+
+-fn exec(files: Vec<String>, settings: &mut Settings) -> i32 {
++fn exec(files: Vec<String>, settings: &GlobalSettings) -> i32 {
+ let mut lines = Vec::new();
+ let mut file_merger = FileMerger::new(&settings);
+
+@@ -459,26 +837,27 @@ fn exec(files: Vec<String>, settings: &mut Settings) -> i32 {
+ if settings.merge {
+ file_merger.push_file(buf_reader.lines());
+ } else if settings.zero_terminated {
+- for line in buf_reader.split(b'\0') {
+- if let Ok(n) = line {
+- lines.push(
+- std::str::from_utf8(&n)
+- .expect("Could not parse string from zero terminated input.")
+- .to_string(),
+- );
+- }
++ for line in buf_reader.split(b'\0').flatten() {
++ lines.push(Line::new(
++ std::str::from_utf8(&line)
++ .expect("Could not parse string from zero terminated input.")
++ .to_string(),
++ &settings,
++ ));
+ }
+ } else {
+ for line in buf_reader.lines() {
+ if let Ok(n) = line {
+- lines.push(n);
++ lines.push(Line::new(n, &settings));
++ } else {
++ break;
+ }
+ }
+ }
+ }
+
+ if settings.check {
+- return exec_check_file(lines, &settings);
++ return exec_check_file(&lines, &settings);
+ } else {
+ sort_by(&mut lines, &settings);
+ }
+@@ -490,29 +869,31 @@ fn exec(files: Vec<String>, settings: &mut Settings) -> i32 {
+ print_sorted(file_merger, &settings)
+ }
+ } else if settings.mode == SortMode::Default && settings.unique {
+- print_sorted(lines.iter().dedup(), &settings)
++ print_sorted(lines.into_iter().map(|line| line.line).dedup(), &settings)
+ } else if settings.mode == SortMode::Month && settings.unique {
+ print_sorted(
+ lines
+- .iter()
++ .into_iter()
++ .map(|line| line.line)
+ .dedup_by(|a, b| get_months_dedup(a) == get_months_dedup(b)),
+ &settings,
+ )
+ } else if settings.unique {
+ print_sorted(
+ lines
+- .iter()
+- .dedup_by(|a, b| get_num_dedup(a, &settings) == get_num_dedup(b, &settings)),
++ .into_iter()
++ .map(|line| line.line)
++ .dedup_by(|a, b| get_num_dedup(a, settings) == get_num_dedup(b, settings)),
+ &settings,
+ )
+ } else {
+- print_sorted(lines.iter(), &settings)
++ print_sorted(lines.into_iter().map(|line| line.line), &settings)
+ }
+
+ 0
+ }
+
+-fn exec_check_file(unwrapped_lines: Vec<String>, settings: &Settings) -> i32 {
++fn exec_check_file(unwrapped_lines: &[Line], settings: &GlobalSettings) -> i32 {
+ // errors yields the line before each disorder,
+ // plus the last line (quirk of .coalesce())
+ let mut errors =
+@@ -544,51 +925,45 @@ fn exec_check_file(unwrapped_lines: Vec<String>, settings: &Settings) -> i32 {
+ }
+ }
+
+-#[inline(always)]
+-fn transform(line: &str, settings: &Settings) -> String {
+- let mut transformed = line.to_owned();
+- for transform_fn in &settings.transform_fns {
+- transformed = transform_fn(&transformed);
+- }
+-
+- transformed
+-}
+-
+-#[inline(always)]
+-fn sort_by(lines: &mut Vec<String>, settings: &Settings) {
++fn sort_by(lines: &mut Vec<Line>, settings: &GlobalSettings) {
+ lines.par_sort_by(|a, b| compare_by(a, b, &settings))
+ }
+
+-fn compare_by(a: &str, b: &str, settings: &Settings) -> Ordering {
+- let (a_transformed, b_transformed): (String, String);
+- let (a, b) = if !settings.transform_fns.is_empty() {
+- a_transformed = transform(&a, &settings);
+- b_transformed = transform(&b, &settings);
+- (a_transformed.as_str(), b_transformed.as_str())
+- } else {
+- (a, b)
+- };
+-
+- // 1st Compare
+- let mut cmp: Ordering = if settings.random {
+- random_shuffle(a, b, settings.salt.clone())
+- } else {
+- (settings.compare_fn)(a, b)
+- };
++fn compare_by(a: &Line, b: &Line, global_settings: &GlobalSettings) -> Ordering {
++ for (idx, selector) in global_settings.selectors.iter().enumerate() {
++ let a = a.selections[idx].get_str(a);
++ let b = b.selections[idx].get_str(b);
++ let settings = &selector.settings;
+
+- // Call "last resort compare" on any equal
+- if cmp == Ordering::Equal {
+- if settings.random || settings.stable || settings.unique {
+- cmp = Ordering::Equal
++ let cmp: Ordering = if settings.random {
++ random_shuffle(a, b, global_settings.salt.clone())
+ } else {
+- cmp = default_compare(a, b)
++ (match settings.mode {
++ SortMode::Numeric => numeric_compare,
++ SortMode::GeneralNumeric => general_numeric_compare,
++ SortMode::HumanNumeric => human_numeric_size_compare,
++ SortMode::Month => month_compare,
++ SortMode::Version => version_compare,
++ SortMode::Default => default_compare,
++ })(a, b)
+ };
++ if cmp != Ordering::Equal {
++ return if settings.reverse { cmp.reverse() } else { cmp };
++ }
++ }
++
++ // Call "last resort compare" if all selectors returned Equal
++
++ let cmp = if global_settings.random || global_settings.stable || global_settings.unique {
++ Ordering::Equal
++ } else {
++ default_compare(&a.line, &b.line)
+ };
+
+- if settings.reverse {
+- return cmp.reverse();
++ if global_settings.reverse {
++ cmp.reverse()
+ } else {
+- return cmp;
++ cmp
+ }
+ }
+
+@@ -617,8 +992,8 @@ fn leading_num_common(a: &str) -> &str {
+ && !c.eq(&'e')
+ && !c.eq(&'E')
+ // check whether first char is + or -
+- && !a.chars().nth(0).unwrap_or('\0').eq(&POSITIVE)
+- && !a.chars().nth(0).unwrap_or('\0').eq(&NEGATIVE)
++ && !a.chars().next().unwrap_or('\0').eq(&POSITIVE)
++ && !a.chars().next().unwrap_or('\0').eq(&NEGATIVE)
+ {
+ // Strip string of non-numeric trailing chars
+ s = &a[..idx];
+@@ -640,9 +1015,9 @@ fn get_leading_num(a: &str) -> &str {
+
+ let a = leading_num_common(a);
+
+- // GNU numeric sort doesn't recognize '+' or 'e' notation so we strip trailing chars
++ // GNU numeric sort doesn't recognize '+' or 'e' notation so we strip
+ for (idx, c) in a.char_indices() {
+- if c.eq(&'e') || c.eq(&'E') || a.chars().nth(0).unwrap_or('\0').eq(&POSITIVE) {
++ if c.eq(&'e') || c.eq(&'E') || a.chars().next().unwrap_or('\0').eq(&POSITIVE) {
+ s = &a[..idx];
+ break;
+ }
+@@ -670,12 +1045,9 @@ fn get_leading_gen(a: &str) -> &str {
+ // Cleanup raw stripped strings
+ for c in p_iter.to_owned() {
+ let next_char_numeric = p_iter.peek().unwrap_or(&'\0').is_numeric();
+- // Only general numeric recognizes e notation and the '+' sign
+- if (c.eq(&'e') && !next_char_numeric)
+- || (c.eq(&'E') && !next_char_numeric)
+- // Only GNU (non-general) numeric recognize thousands seperators, takes only leading #
+- || c.eq(&THOUSANDS_SEP)
+- {
++ // Only general numeric recognizes e notation and, see block below, the '+' sign
++ // Only GNU (non-general) numeric recognize thousands seperators, takes only leading #
++ if (c.eq(&'e') || c.eq(&'E')) && !next_char_numeric || c.eq(&THOUSANDS_SEP) {
+ result = a.split(c).next().unwrap_or("");
+ break;
+ // If positive sign and next char is not numeric, split at postive sign at keep trailing numbers
+@@ -724,19 +1096,17 @@ fn get_months_dedup(a: &str) -> String {
+ // *For all dedups/uniques expect default we must compare leading numbers*
+ // Also note numeric compare and unique output is specifically *not* the same as a "sort | uniq"
+ // See: https://www.gnu.org/software/coreutils/manual/html_node/sort-invocation.html
+-fn get_num_dedup<'a>(a: &'a str, settings: &&mut Settings) -> &'a str {
++fn get_num_dedup<'a>(a: &'a str, settings: &GlobalSettings) -> &'a str {
+ // Trim and remove any leading zeros
+ let s = a.trim().trim_start_matches('0');
+
+ // Get first char
+- let c = s.chars().nth(0).unwrap_or('\0');
++ let c = s.chars().next().unwrap_or('\0');
+
+ // Empty lines and non-number lines are treated as the same for dedup
+- if s.is_empty() {
+- ""
+- } else if !c.eq(&NEGATIVE) && !c.is_numeric() {
+- ""
+ // Prepare lines for comparison of only the numerical leading numbers
++ if s.is_empty() || (!c.eq(&NEGATIVE) && !c.is_numeric()) {
++ ""
+ } else {
+ let result = match settings.mode {
+ SortMode::Numeric => get_leading_num(s),
+@@ -944,6 +1314,7 @@ fn month_parse(line: &str) -> Month {
+ }
+
+ fn month_compare(a: &str, b: &str) -> Ordering {
++ #![allow(clippy::comparison_chain)]
+ let ma = month_parse(a);
+ let mb = month_parse(b);
+
+@@ -986,32 +1357,29 @@ fn remove_nonprinting_chars(s: &str) -> String {
+ .collect::<String>()
+ }
+
+-fn print_sorted<S, T: Iterator<Item = S>>(iter: T, settings: &Settings)
+-where
+- S: std::fmt::Display,
+-{
++fn print_sorted<T: Iterator<Item = String>>(iter: T, settings: &GlobalSettings) {
+ let mut file: Box<dyn Write> = match settings.outfile {
+ Some(ref filename) => match File::create(Path::new(&filename)) {
+ Ok(f) => Box::new(BufWriter::new(f)) as Box<dyn Write>,
+ Err(e) => {
+- show_error!("sort: {0}: {1}", filename, e.to_string());
++ show_error!("{0}: {1}", filename, e.to_string());
+ panic!("Could not open output file");
+ }
+ },
+- None => Box::new(stdout()) as Box<dyn Write>,
++ None => Box::new(BufWriter::new(stdout())) as Box<dyn Write>,
+ };
+-
+ if settings.zero_terminated {
+ for line in iter {
+- let str = format!("{}\0", line);
+- crash_if_err!(1, file.write_all(str.as_bytes()));
++ crash_if_err!(1, file.write_all(line.as_bytes()));
++ crash_if_err!(1, file.write_all("\0".as_bytes()));
+ }
+ } else {
+ for line in iter {
+- let str = format!("{}\n", line);
+- crash_if_err!(1, file.write_all(str.as_bytes()));
++ crash_if_err!(1, file.write_all(line.as_bytes()));
++ crash_if_err!(1, file.write_all("\n".as_bytes()));
+ }
+ }
++ crash_if_err!(1, file.flush());
+ }
+
+ // from cat.rs
+@@ -1024,7 +1392,7 @@ fn open(path: &str) -> Option<(Box<dyn Read>, bool)> {
+ match File::open(Path::new(path)) {
+ Ok(f) => Some((Box::new(f) as Box<dyn Read>, false)),
+ Err(e) => {
+- show_error!("sort: {0}: {1}", path, e.to_string());
++ show_error!("{0}: {1}", path, e.to_string());
+ None
+ }
+ }
+@@ -1097,4 +1465,34 @@ mod tests {
+
+ assert_eq!(Ordering::Less, version_compare(a, b));
+ }
++
++ #[test]
++ fn test_random_compare() {
++ let a = "9";
++ let b = "9";
++ let c = get_rand_string();
++
++ assert_eq!(Ordering::Equal, random_shuffle(a, b, c));
++ }
++
++ #[test]
++ fn test_tokenize_fields() {
++ let line = "foo bar b x";
++ assert_eq!(tokenize(line, None), vec![0..3, 3..7, 7..9, 9..14,],);
++ }
++
++ #[test]
++ fn test_tokenize_fields_leading_whitespace() {
++ let line = " foo bar b x";
++ assert_eq!(tokenize(line, None), vec![0..7, 7..11, 11..13, 13..18,]);
++ }
++
++ #[test]
++ fn test_tokenize_fields_custom_separator() {
++ let line = "aaa foo bar b x";
++ assert_eq!(
++ tokenize(line, Some('a')),
++ vec![0..0, 1..1, 2..2, 3..9, 10..18,]
++ );
++ }
+ }
+diff --git a/tests/by-util/test_sort.rs b/tests/by-util/test_sort.rs
+index 6455d83..668e783 100644
+--- a/tests/by-util/test_sort.rs
++++ b/tests/by-util/test_sort.rs
+@@ -185,10 +185,10 @@ fn test_dictionary_order2() {
+ fn test_non_printing_chars() {
+ for non_printing_chars_param in vec!["-i"] {
+ new_ucmd!()
+- .pipe_in("a👦ðŸ»aa b\naaaa b")
++ .pipe_in("a👦ðŸ»aa\naaaa")
+ .arg(non_printing_chars_param)
+ .succeeds()
+- .stdout_only("aaaa b\na👦ðŸ»aa b\n");
++ .stdout_only("a👦ðŸ»aa\naaaa\n");
+ }
+ }
+
+@@ -307,6 +307,166 @@ fn test_numeric_unique_ints2() {
+ }
+ }
+
++#[test]
++fn test_keys_open_ended() {
++ let input = "aa bb cc\ndd aa ff\ngg aa cc\n";
++ new_ucmd!()
++ .args(&["-k", "2.2"])
++ .pipe_in(input)
++ .succeeds()
++ .stdout_only("gg aa cc\ndd aa ff\naa bb cc\n");
++}
++
++#[test]
++fn test_keys_closed_range() {
++ let input = "aa bb cc\ndd aa ff\ngg aa cc\n";
++ new_ucmd!()
++ .args(&["-k", "2.2,2.2"])
++ .pipe_in(input)
++ .succeeds()
++ .stdout_only("dd aa ff\ngg aa cc\naa bb cc\n");
++}
++
++#[test]
++fn test_keys_multiple_ranges() {
++ let input = "aa bb cc\ndd aa ff\ngg aa cc\n";
++ new_ucmd!()
++ .args(&["-k", "2,2", "-k", "3,3"])
++ .pipe_in(input)
++ .succeeds()
++ .stdout_only("gg aa cc\ndd aa ff\naa bb cc\n");
++}
++
++#[test]
++fn test_keys_no_field_match() {
++ let input = "aa aa aa aa\naa bb cc\ndd aa ff\n";
++ new_ucmd!()
++ .args(&["-k", "4,4"])
++ .pipe_in(input)
++ .succeeds()
++ .stdout_only("aa bb cc\ndd aa ff\naa aa aa aa\n");
++}
++
++#[test]
++fn test_keys_no_char_match() {
++ let input = "aaa\nba\nc\n";
++ new_ucmd!()
++ .args(&["-k", "1.2"])
++ .pipe_in(input)
++ .succeeds()
++ .stdout_only("c\nba\naaa\n");
++}
++
++#[test]
++fn test_keys_custom_separator() {
++ let input = "aaxbbxcc\nddxaaxff\nggxaaxcc\n";
++ new_ucmd!()
++ .args(&["-k", "2.2,2.2", "-t", "x"])
++ .pipe_in(input)
++ .succeeds()
++ .stdout_only("ddxaaxff\nggxaaxcc\naaxbbxcc\n");
++}
++
++#[test]
++fn test_keys_invalid_field() {
++ new_ucmd!()
++ .args(&["-k", "1."])
++ .fails()
++ .stderr_only("sort: error: failed to parse character index for key `1.`: cannot parse integer from empty string");
++}
++
++#[test]
++fn test_keys_invalid_field_option() {
++ new_ucmd!()
++ .args(&["-k", "1.1x"])
++ .fails()
++ .stderr_only("sort: error: invalid option for key: `x`");
++}
++
++#[test]
++fn test_keys_invalid_field_zero() {
++ new_ucmd!()
++ .args(&["-k", "0.1"])
++ .fails()
++ .stderr_only("sort: error: field index was 0");
++}
++
++#[test]
++fn test_keys_with_options() {
++ let input = "aa 3 cc\ndd 1 ff\ngg 2 cc\n";
++ for param in &[
++ &["-k", "2,2n"][..],
++ &["-k", "2n,2"][..],
++ &["-k", "2,2", "-n"][..],
++ ] {
++ new_ucmd!()
++ .args(param)
++ .pipe_in(input)
++ .succeeds()
++ .stdout_only("dd 1 ff\ngg 2 cc\naa 3 cc\n");
++ }
++}
++
++#[test]
++fn test_keys_with_options_blanks_start() {
++ let input = "aa 3 cc\ndd 1 ff\ngg 2 cc\n";
++ for param in &[&["-k", "2b,2"][..], &["-k", "2,2", "-b"][..]] {
++ new_ucmd!()
++ .args(param)
++ .pipe_in(input)
++ .succeeds()
++ .stdout_only("dd 1 ff\ngg 2 cc\naa 3 cc\n");
++ }
++}
++
++#[test]
++fn test_keys_with_options_blanks_end() {
++ let input = "a b
++a b
++a b
++";
++ new_ucmd!()
++ .args(&["-k", "1,2.1b", "-s"])
++ .pipe_in(input)
++ .succeeds()
++ .stdout_only(
++ "a b
++a b
++a b
++",
++ );
++}
++
++#[test]
++fn test_keys_stable() {
++ let input = "a b
++a b
++a b
++";
++ new_ucmd!()
++ .args(&["-k", "1,2.1", "-s"])
++ .pipe_in(input)
++ .succeeds()
++ .stdout_only(
++ "a b
++a b
++a b
++",
++ );
++}
++
++#[test]
++fn test_keys_empty_match() {
++ let input = "a a a a
++aaaa
++";
++ new_ucmd!()
++ .args(&["-k", "1,1", "-t", "a"])
++ .pipe_in(input)
++ .succeeds()
++ .stdout_only(input);
++}
++
+ #[test]
+ fn test_zero_terminated() {
+ test_helper("zero-terminated", "-z");
--
GitLab