sort: distinguish blanks from control whitespace

src/uu/sort/src/custom_str_cmp.rs

@@ -9,8 +9,13 @@
 
 use std::cmp::Ordering;
 
+// Newlines also act as blanks inside NUL-terminated records.
+pub fn is_blank(c: u8) -> bool {
+    matches!(c, b' ' | b'\t' | b'\n')
+}
+
 fn filter_char(c: u8, ignore_non_printing: bool, ignore_non_dictionary: bool) -> bool {
-    if ignore_non_dictionary && !(c.is_ascii_alphanumeric() || c.is_ascii_whitespace()) {
+    if ignore_non_dictionary && !(c.is_ascii_alphanumeric() || is_blank(c)) {
         return false;
     }
     !(ignore_non_printing && (c.is_ascii_control() || !c.is_ascii()))

src/uu/sort/src/numeric_str_cmp.rs

@@ -14,6 +14,7 @@
 //! More specifically, exponent can be understood so that the original number is in `(1..10)*10^exponent`.
 //! From that follows the constraints of this algorithm: It is able to compare numbers in ±(1*10^[`i64::MIN`]..10*10^[`i64::MAX`]).
 
+use crate::custom_str_cmp::is_blank;
 use std::{cmp::Ordering, ops::Range};
 
 #[derive(Debug, PartialEq, Eq, PartialOrd, Ord, Clone)]
@@ -63,7 +64,7 @@ impl NumInfo {
         let mut first_char = true;
 
         for (idx, &char) in num.iter().enumerate() {
-            if first_char && char.is_ascii_whitespace() {
+            if first_char && is_blank(char) {
                 continue;
             }
 

src/uu/sort/src/sort.rs

@@ -24,7 +24,7 @@ use bigdecimal::BigDecimal;
 use chunks::LineData;
 use clap::builder::ValueParser;
 use clap::{Arg, ArgAction, ArgMatches, Command};
-use custom_str_cmp::custom_str_cmp;
+use custom_str_cmp::{custom_str_cmp, is_blank};
 use ext_sort::ext_sort;
 use foldhash::fast::FoldHasher;
 use foldhash::{HashMap, SharedSeed};
@@ -934,7 +934,7 @@ fn tokenize_default(
     // pretend that there was whitespace in front of the line
     let mut previous_was_whitespace = true;
     for (idx, char) in line.iter().enumerate() {
-        let is_whitespace = char.is_ascii_whitespace();
+        let is_whitespace = is_blank(*char);
         let treat_as_separator = if is_whitespace {
             if blank_thousands_sep && *char == b' ' {
                 !is_blank_thousands_sep(line, idx, allow_unit_after_blank)

tests/by-util/test_sort.rs

@@ -3,7 +3,7 @@
 // For the full copyright and license information, please view the LICENSE
 // file that was distributed with this source code.
 
-// spell-checker:ignore (words) ints (linux) Nofile dfgi abmon avril
+// spell-checker:ignore (words) ints (linux) Nofile dfgi abmon avril naab
 
 #![allow(clippy::cast_possible_wrap)]
 
@@ -4123,3 +4123,22 @@ sort: invalid suffix in --buffer-size argument '8zz'
 }
 
 // spell-checker:enable
+
+#[rstest::rstest]
+#[case(&["-k2,3"], "aa\n0aa\r\n", "0aa\r\naa\n")]
+#[case(&["-k2,3"], "aa\n0aa\x0c\n", "0aa\x0c\naa\n")]
+#[case(&["-d"], "a\rb\naab\n", "aab\na\rb\n")]
+#[case(&["-d"], "a\x0cb\naab\n", "aab\na\x0cb\n")]
+#[case(&["-n"], "\r4\na\n", "\r4\na\n")]
+#[case(&["-n"], "\x0c4\na\n", "\x0c4\na\n")]
+#[case(&["-z", "-k2"], "b\n1\0a\n2\0", "b\n1\0a\n2\0")]
+#[case(&["-z", "-d"], "a\nb\0aab\0", "a\nb\0aab\0")]
+#[case(&["-z", "-n"], "\n4\0a\0", "a\0\n4\0")]
+fn test_control_whitespace(#[case] args: &[&str], #[case] input: &str, #[case] expected: &str) {
+    new_ucmd!()
+        .env("LC_ALL", "C")
+        .args(args)
+        .pipe_in(input)
+        .succeeds()
+        .stdout_only(expected);
+}