Merge branch 'master' of https://github.com/uutils/coreutils

.github/workflows/GNU.yml

@@ -80,6 +80,9 @@ jobs:
                -e '/tests\/misc\/help-version-getopt.sh/ D' \
                Makefile
 
+        # printf doesn't limit the values used in its arg, so this produced ~2GB of output
+        sed -i '/INT_OFLOW/ D' tests/misc/printf.sh
+
         # Use the system coreutils where the test fails due to error in a util that is not the one being tested
         sed -i 's|stat|/usr/bin/stat|' tests/chgrp/basic.sh tests/cp/existing-perm-dir.sh tests/touch/60-seconds.sh tests/misc/sort-compress-proc.sh
         sed -i 's|ls -|/usr/bin/ls -|' tests/chgrp/posix-H.sh tests/chown/deref.sh tests/cp/same-file.sh tests/misc/mknod.sh tests/mv/part-symlink.sh tests/du/8gb.sh

Cargo.lock

@@ -28,6 +28,15 @@ dependencies = [
  "winapi 0.3.9",
 ]
 
+[[package]]
+name = "ansi_term"
+version = "0.12.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d52a9bb7ec0cf484c551830a7ce27bd20d67eac647e1befb56b0be4ee39a55d2"
+dependencies = [
+ "winapi 0.3.9",
+]
+
 [[package]]
 name = "arrayvec"
 version = "0.4.12"
@@ -127,9 +136,9 @@ checksum = "14c189c53d098945499cdfa7ecc63567cf3886b3332b312a5b4585d8d3a6a610"
 
 [[package]]
 name = "cast"
-version = "0.2.3"
+version = "0.2.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4b9434b9a5aa1450faa3f9cb14ea0e8c53bb5d2b3c1bfd1ab4fc03e9f33fbfb0"
+checksum = "cc38c385bfd7e444464011bb24820f40dd1c76bcdfa1b78611cb7c2e5cafab75"
 dependencies = [
  "rustc_version",
 ]
@@ -169,7 +178,7 @@ version = "2.33.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "37e58ac78573c40708d45522f0d80fa2f01cc4f9b4e2bf749807255454312002"
 dependencies = [
- "ansi_term",
+ "ansi_term 0.11.0",
  "atty",
  "bitflags",
  "strsim",
@@ -212,6 +221,7 @@ dependencies = [
  "lazy_static",
  "libc",
  "nix 0.20.0",
+ "pretty_assertions",
  "rand 0.7.3",
  "regex",
  "sha1",
@@ -452,9 +462,9 @@ dependencies = [
 
 [[package]]
 name = "crossbeam-channel"
-version = "0.5.0"
+version = "0.5.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "dca26ee1f8d361640700bde38b2c37d8c22b3ce2d360e1fc1c74ea4b0aa7d775"
+checksum = "06ed27e177f16d65f0f0c22a213e17c696ace5dd64b14258b52f9417ccb52db4"
 dependencies = [
  "cfg-if 1.0.0",
  "crossbeam-utils",
@@ -517,6 +527,16 @@ dependencies = [
  "memchr 2.3.4",
 ]
 
+[[package]]
+name = "ctor"
+version = "0.1.20"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5e98e2ad1a782e33928b96fc3948e7c355e5af34ba4de7670fe8bac2a3b2006d"
+dependencies = [
+ "quote 1.0.9",
+ "syn",
+]
+
 [[package]]
 name = "custom_derive"
 version = "0.1.7"
@@ -529,6 +549,12 @@ version = "2.1.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "f4f47ca1860a761136924ddd2422ba77b2ea54fe8cc75b9040804a0d9d32ad97"
 
+[[package]]
+name = "diff"
+version = "0.1.12"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0e25ea47919b1560c4e3b7fe0aaab9becf5b84a10325ddf7db0f0ba5e1026499"
+
 [[package]]
 name = "digest"
 version = "0.6.2"
@@ -580,7 +606,7 @@ checksum = "1d34cfa13a63ae058bfa601fe9e313bbdb3746427c1459185464ce0fcf62e1e8"
 dependencies = [
  "cfg-if 1.0.0",
  "libc",
- "redox_syscall 0.2.5",
+ "redox_syscall 0.2.6",
  "winapi 0.3.9",
 ]
 
@@ -783,6 +809,15 @@ dependencies = [
  "cfg-if 1.0.0",
 ]
 
+[[package]]
+name = "lscolors"
+version = "0.7.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d24b894c45c9da468621cdd615a5a79ee5e5523dd4f75c76ebc03d458940c16e"
+dependencies = [
+ "ansi_term 0.12.1",
+]
+
 [[package]]
 name = "match_cfg"
 version = "0.1.0"
@@ -897,6 +932,12 @@ version = "0.1.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "b8f8bdf33df195859076e54ab11ee78a1b208382d3a26ec40d142ffc1ecc49ef"
 
+[[package]]
+name = "once_cell"
+version = "1.7.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "af8b08b04175473088b46763e51ee54da5f9a164bc162f615b91bc179dbf15a3"
+
 [[package]]
 name = "onig"
 version = "4.3.3"
@@ -925,6 +966,15 @@ version = "11.1.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "0ab1bc2a289d34bd04a330323ac98a1b4bc82c9d9fcb1e66b63caa84da26b575"
 
+[[package]]
+name = "output_vt100"
+version = "0.1.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "53cdc5b785b7a58c5aad8216b3dfa114df64b0b06ae6e1501cef91df2fbdf8f9"
+dependencies = [
+ "winapi 0.3.9",
+]
+
 [[package]]
 name = "paste"
 version = "0.1.18"
@@ -994,6 +1044,18 @@ version = "0.2.10"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "ac74c624d6b2d21f425f752262f42188365d7b8ff1aff74c82e45136510a4857"
 
+[[package]]
+name = "pretty_assertions"
+version = "0.7.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "1cab0e7c02cf376875e9335e0ba1da535775beb5450d21e1dffca068818ed98b"
+dependencies = [
+ "ansi_term 0.12.1",
+ "ctor",
+ "diff",
+ "output_vt100",
+]
+
 [[package]]
 name = "proc-macro-hack"
 version = "0.5.19"
@@ -1176,9 +1238,9 @@ checksum = "41cc0f7e4d5d4544e8861606a285bb08d3e70712ccc7d2b84d7c0ccfaf4b05ce"
 
 [[package]]
 name = "redox_syscall"
-version = "0.2.5"
+version = "0.2.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "94341e4e44e24f6b591b59e47a8a027df12e008d73fd5672dbea9cc22f4507d9"
+checksum = "8270314b5ccceb518e7e578952f0b72b88222d02e8f77f5ecf7abbb673539041"
 dependencies = [
  "bitflags",
 ]
@@ -1189,14 +1251,14 @@ version = "0.1.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "8440d8acb4fd3d277125b4bd01a6f38aee8d814b3b5fc09b3f2b825d37d3fe8f"
 dependencies = [
- "redox_syscall 0.2.5",
+ "redox_syscall 0.2.6",
 ]
 
 [[package]]
 name = "regex"
-version = "1.4.5"
+version = "1.4.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "957056ecddbeba1b26965114e191d2e8589ce74db242b6ea25fc4062427a5c19"
+checksum = "2a26af418b574bd56588335b3a3659a65725d4e636eb1016c2f9e3b38c7cc759"
 dependencies = [
  "aho-corasick",
  "memchr 2.3.4",
@@ -1400,9 +1462,9 @@ dependencies = [
 
 [[package]]
 name = "syn"
-version = "1.0.69"
+version = "1.0.70"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "48fe99c6bd8b1cc636890bcc071842de909d902c81ac7dab53ba33c421ab8ffb"
+checksum = "b9505f307c872bab8eb46f77ae357c8eba1fdacead58ee5a850116b1d7f82883"
 dependencies = [
  "proc-macro2",
  "quote 1.0.9",
@@ -1460,7 +1522,7 @@ checksum = "077185e2eac69c3f8379a4298e1e07cd36beb962290d4a51199acf0fdc10607e"
 dependencies = [
  "libc",
  "numtoa",
- "redox_syscall 0.2.5",
+ "redox_syscall 0.2.6",
  "redox_termios",
 ]
 
@@ -1996,11 +2058,12 @@ dependencies = [
  "clap",
  "globset",
  "lazy_static",
+ "lscolors",
  "number_prefix",
+ "once_cell",
  "term_grid",
  "termsize",
  "time",
- "unicode-width",
  "uucore",
  "uucore_procs",
 ]
@@ -2306,6 +2369,7 @@ dependencies = [
  "serde_json",
  "smallvec 1.6.1",
  "tempdir",
+ "unicode-width",
  "uucore",
  "uucore_procs",
 ]

Cargo.toml

@@ -335,6 +335,7 @@ filetime = "0.2"
 glob = "0.3.0"
 libc = "0.2"
 nix = "0.20.0"
+pretty_assertions = "0.7.2"
 rand = "0.7"
 regex = "1.0"
 sha1 = { version="0.6", features=["std"] }

src/uu/cat/src/cat.rs

@@ -22,6 +22,12 @@ use std::io::{self, Read, Write};
 use thiserror::Error;
 use uucore::fs::is_stdin_interactive;
 
+/// Linux splice support
+#[cfg(any(target_os = "linux", target_os = "android"))]
+mod splice;
+#[cfg(any(target_os = "linux", target_os = "android"))]
+use std::os::unix::io::{AsRawFd, RawFd};
+
 /// Unix domain socket support
 #[cfg(unix)]
 use std::net::Shutdown;
@@ -30,14 +36,6 @@ use std::os::unix::fs::FileTypeExt;
 #[cfg(unix)]
 use unix_socket::UnixStream;
 
-/// Linux splice support
-#[cfg(any(target_os = "linux", target_os = "android"))]
-use nix::fcntl::{splice, SpliceFFlags};
-#[cfg(any(target_os = "linux", target_os = "android"))]
-use nix::unistd::pipe;
-#[cfg(any(target_os = "linux", target_os = "android"))]
-use std::os::unix::io::{AsRawFd, RawFd};
-
 static NAME: &str = "cat";
 static VERSION: &str = env!("CARGO_PKG_VERSION");
 static SYNTAX: &str = "[OPTION]... [FILE]...";
@@ -395,7 +393,7 @@ fn write_fast<R: Read>(handle: &mut InputHandle<R>) -> CatResult<()> {
     {
         // If we're on Linux or Android, try to use the splice() system call
         // for faster writing. If it works, we're done.
-        if !write_fast_using_splice(handle, stdout_lock.as_raw_fd())? {
+        if !splice::write_fast_using_splice(handle, stdout_lock.as_raw_fd())? {
             return Ok(());
         }
     }
@@ -411,75 +409,6 @@ fn write_fast<R: Read>(handle: &mut InputHandle<R>) -> CatResult<()> {
     Ok(())
 }
 
-/// This function is called from `write_fast()` on Linux and Android. The
-/// function `splice()` is used to move data between two file descriptors
-/// without copying between kernel- and userspace. This results in a large
-/// speedup.
-///
-/// The `bool` in the result value indicates if we need to fall back to normal
-/// copying or not. False means we don't have to.
-#[cfg(any(target_os = "linux", target_os = "android"))]
-#[inline]
-fn write_fast_using_splice<R: Read>(handle: &mut InputHandle<R>, writer: RawFd) -> CatResult<bool> {
-    const BUF_SIZE: usize = 1024 * 16;
-
-    let (pipe_rd, pipe_wr) = pipe()?;
-
-    // We only fall back if splice fails on the first call.
-    match splice(
-        handle.file_descriptor,
-        None,
-        pipe_wr,
-        None,
-        BUF_SIZE,
-        SpliceFFlags::empty(),
-    ) {
-        Ok(n) => {
-            if n == 0 {
-                return Ok(false);
-            }
-            splice_exact(pipe_rd, writer, n)?;
-        }
-        Err(_) => {
-            return Ok(true);
-        }
-    }
-
-    loop {
-        let n = splice(
-            handle.file_descriptor,
-            None,
-            pipe_wr,
-            None,
-            BUF_SIZE,
-            SpliceFFlags::empty(),
-        )?;
-        if n == 0 {
-            // We read 0 bytes from the input,
-            // which means we're done copying.
-            break;
-        }
-        splice_exact(pipe_rd, writer, n)?;
-    }
-
-    Ok(false)
-}
-
-/// Splice wrapper which handles short writes
-#[cfg(any(target_os = "linux", target_os = "android"))]
-#[inline]
-fn splice_exact(read_fd: RawFd, write_fd: RawFd, num_bytes: usize) -> nix::Result<()> {
-    let mut left = num_bytes;
-    loop {
-        let written = splice(read_fd, None, write_fd, None, left, SpliceFFlags::empty())?;
-        left -= written;
-        if left == 0 {
-            break;
-        }
-    }
-    Ok(())
-}
-
 /// Outputs file contents to stdout in a line-by-line fashion,
 /// propagating any errors that might occur.
 fn write_lines<R: Read>(

src/uu/cat/src/splice.rs

@@ -0,0 +1,91 @@
+use super::{CatResult, InputHandle};
+
+use nix::fcntl::{splice, SpliceFFlags};
+use nix::unistd::{self, pipe};
+use std::io::Read;
+use std::os::unix::io::RawFd;
+
+const BUF_SIZE: usize = 1024 * 16;
+
+/// This function is called from `write_fast()` on Linux and Android. The
+/// function `splice()` is used to move data between two file descriptors
+/// without copying between kernel- and userspace. This results in a large
+/// speedup.
+///
+/// The `bool` in the result value indicates if we need to fall back to normal
+/// copying or not. False means we don't have to.
+#[inline]
+pub(super) fn write_fast_using_splice<R: Read>(
+    handle: &mut InputHandle<R>,
+    write_fd: RawFd,
+) -> CatResult<bool> {
+    let (pipe_rd, pipe_wr) = match pipe() {
+        Ok(r) => r,
+        Err(_) => {
+            // It is very rare that creating a pipe fails, but it can happen.
+            return Ok(true);
+        }
+    };
+
+    loop {
+        match splice(
+            handle.file_descriptor,
+            None,
+            pipe_wr,
+            None,
+            BUF_SIZE,
+            SpliceFFlags::empty(),
+        ) {
+            Ok(n) => {
+                if n == 0 {
+                    return Ok(false);
+                }
+                if splice_exact(pipe_rd, write_fd, n).is_err() {
+                    // If the first splice manages to copy to the intermediate
+                    // pipe, but the second splice to stdout fails for some reason
+                    // we can recover by copying the data that we have from the
+                    // intermediate pipe to stdout using normal read/write. Then
+                    // we tell the caller to fall back.
+                    copy_exact(pipe_rd, write_fd, n)?;
+                    return Ok(true);
+                }
+            }
+            Err(_) => {
+                return Ok(true);
+            }
+        }
+    }
+}
+
+/// Splice wrapper which handles short writes.
+#[inline]
+fn splice_exact(read_fd: RawFd, write_fd: RawFd, num_bytes: usize) -> nix::Result<()> {
+    let mut left = num_bytes;
+    loop {
+        let written = splice(read_fd, None, write_fd, None, left, SpliceFFlags::empty())?;
+        left -= written;
+        if left == 0 {
+            break;
+        }
+    }
+    Ok(())
+}
+
+/// Caller must ensure that `num_bytes <= BUF_SIZE`, otherwise this function
+/// will panic. The way we use this function in `write_fast_using_splice`
+/// above is safe because `splice` is set to write at most `BUF_SIZE` to the
+/// pipe.
+#[inline]
+fn copy_exact(read_fd: RawFd, write_fd: RawFd, num_bytes: usize) -> nix::Result<()> {
+    let mut left = num_bytes;
+    let mut buf = [0; BUF_SIZE];
+    loop {
+        let read = unistd::read(read_fd, &mut buf[..left])?;
+        let written = unistd::write(write_fd, &mut buf[..read])?;
+        left -= written;
+        if left == 0 {
+            break;
+        }
+    }
+    Ok(())
+}

src/uu/cp/src/cp.rs

@@ -155,7 +155,8 @@ pub enum OverwriteMode {
     NoClobber,
 }
 
-#[derive(Clone, Eq, PartialEq)]
+/// Possible arguments for `--reflink`.
+#[derive(Copy, Clone, Eq, PartialEq)]
 pub enum ReflinkMode {
     Always,
     Auto,
@@ -210,7 +211,6 @@ pub struct Options {
     overwrite: OverwriteMode,
     parents: bool,
     strip_trailing_slashes: bool,
-    reflink: bool,
     reflink_mode: ReflinkMode,
     preserve_attributes: Vec<Attribute>,
     recursive: bool,
@@ -633,12 +633,12 @@ impl Options {
             update: matches.is_present(OPT_UPDATE),
             verbose: matches.is_present(OPT_VERBOSE),
             strip_trailing_slashes: matches.is_present(OPT_STRIP_TRAILING_SLASHES),
-            reflink: matches.is_present(OPT_REFLINK),
             reflink_mode: {
                 if let Some(reflink) = matches.value_of(OPT_REFLINK) {
                     match reflink {
                         "always" => ReflinkMode::Always,
                         "auto" => ReflinkMode::Auto,
+                        "never" => ReflinkMode::Never,
                         value => {
                             return Err(Error::InvalidArgument(format!(
                                 "invalid argument '{}' for \'reflink\'",
@@ -1193,47 +1193,20 @@ fn copy_file(source: &Path, dest: &Path, options: &Options) -> CopyResult<()> {
     Ok(())
 }
 
-///Copy the file from `source` to `dest` either using the normal `fs::copy` or the
+/// Copy the file from `source` to `dest` either using the normal `fs::copy` or a
-///`FICLONE` ioctl if --reflink is specified and the filesystem supports it.
+/// copy-on-write scheme if --reflink is specified and the filesystem supports it.
 fn copy_helper(source: &Path, dest: &Path, options: &Options) -> CopyResult<()> {
-    if options.reflink {
+    if options.reflink_mode != ReflinkMode::Never {
-        #[cfg(not(target_os = "linux"))]
+        #[cfg(not(any(target_os = "linux", target_os = "macos")))]
-        return Err("--reflink is only supported on linux".to_string().into());
+        return Err("--reflink is only supported on linux and macOS"
+            .to_string()
+            .into());
+
+        #[cfg(target_os = "macos")]
+        copy_on_write_macos(source, dest, options.reflink_mode)?;
 
         #[cfg(target_os = "linux")]
-        {
+        copy_on_write_linux(source, dest, options.reflink_mode)?;
-            let src_file = File::open(source).unwrap().into_raw_fd();
-            let dst_file = OpenOptions::new()
-                .write(true)
-                .truncate(false)
-                .create(true)
-                .open(dest)
-                .unwrap()
-                .into_raw_fd();
-            match options.reflink_mode {
-                ReflinkMode::Always => unsafe {
-                    let result = ficlone(dst_file, src_file as *const i32);
-                    if result != 0 {
-                        return Err(format!(
-                            "failed to clone {:?} from {:?}: {}",
-                            source,
-                            dest,
-                            std::io::Error::last_os_error()
-                        )
-                        .into());
-                    } else {
-                        return Ok(());
-                    }
-                },
-                ReflinkMode::Auto => unsafe {
-                    let result = ficlone(dst_file, src_file as *const i32);
-                    if result != 0 {
-                        fs::copy(source, dest).context(&*context_for(source, dest))?;
-                    }
-                },
-                ReflinkMode::Never => {}
-            }
-        }
     } else if options.no_dereference && fs::symlink_metadata(&source)?.file_type().is_symlink() {
         // Here, we will copy the symlink itself (actually, just recreate it)
         let link = fs::read_link(&source)?;
@@ -1266,6 +1239,101 @@ fn copy_helper(source: &Path, dest: &Path, options: &Options) -> CopyResult<()>
     Ok(())
 }
 
+/// Copies `source` to `dest` using copy-on-write if possible.
+#[cfg(target_os = "linux")]
+fn copy_on_write_linux(source: &Path, dest: &Path, mode: ReflinkMode) -> CopyResult<()> {
+    debug_assert!(mode != ReflinkMode::Never);
+
+    let src_file = File::open(source).unwrap().into_raw_fd();
+    let dst_file = OpenOptions::new()
+        .write(true)
+        .truncate(false)
+        .create(true)
+        .open(dest)
+        .unwrap()
+        .into_raw_fd();
+    match mode {
+        ReflinkMode::Always => unsafe {
+            let result = ficlone(dst_file, src_file as *const i32);
+            if result != 0 {
+                return Err(format!(
+                    "failed to clone {:?} from {:?}: {}",
+                    source,
+                    dest,
+                    std::io::Error::last_os_error()
+                )
+                .into());
+            } else {
+                return Ok(());
+            }
+        },
+        ReflinkMode::Auto => unsafe {
+            let result = ficlone(dst_file, src_file as *const i32);
+            if result != 0 {
+                fs::copy(source, dest).context(&*context_for(source, dest))?;
+            }
+        },
+        ReflinkMode::Never => unreachable!(),
+    }
+
+    Ok(())
+}
+
+/// Copies `source` to `dest` using copy-on-write if possible.
+#[cfg(target_os = "macos")]
+fn copy_on_write_macos(source: &Path, dest: &Path, mode: ReflinkMode) -> CopyResult<()> {
+    debug_assert!(mode != ReflinkMode::Never);
+
+    // Extract paths in a form suitable to be passed to a syscall.
+    // The unwrap() is safe because they come from the command-line and so contain non nul
+    // character.
+    use std::os::unix::ffi::OsStrExt;
+    let src = CString::new(source.as_os_str().as_bytes()).unwrap();
+    let dst = CString::new(dest.as_os_str().as_bytes()).unwrap();
+
+    // clonefile(2) was introduced in macOS 10.12 so we cannot statically link against it
+    // for backward compatibility.
+    let clonefile = CString::new("clonefile").unwrap();
+    let raw_pfn = unsafe { libc::dlsym(libc::RTLD_NEXT, clonefile.as_ptr()) };
+
+    let mut error = 0;
+    if !raw_pfn.is_null() {
+        // Call clonefile(2).
+        // Safety: Casting a C function pointer to a rust function value is one of the few
+        // blessed uses of `transmute()`.
+        unsafe {
+            let pfn: extern "C" fn(
+                src: *const libc::c_char,
+                dst: *const libc::c_char,
+                flags: u32,
+            ) -> libc::c_int = std::mem::transmute(raw_pfn);
+            error = pfn(src.as_ptr(), dst.as_ptr(), 0);
+            if std::io::Error::last_os_error().kind() == std::io::ErrorKind::AlreadyExists {
+                // clonefile(2) fails if the destination exists.  Remove it and try again.  Do not
+                // bother to check if removal worked because we're going to try to clone again.
+                let _ = fs::remove_file(dest);
+                error = pfn(src.as_ptr(), dst.as_ptr(), 0);
+            }
+        }
+    }
+
+    if raw_pfn.is_null() || error != 0 {
+        // clonefile(2) is not supported or it error'ed out (possibly because the FS does not
+        // support COW).
+        match mode {
+            ReflinkMode::Always => {
+                return Err(
+                    format!("failed to clone {:?} from {:?}: {}", source, dest, error).into(),
+                )
+            }
+            ReflinkMode::Auto => fs::copy(source, dest).context(&*context_for(source, dest))?,
+            ReflinkMode::Never => unreachable!(),
+        };
+    }
+
+    Ok(())
+}
+
 /// Generate an error message if `target` is not the correct `target_type`
 pub fn verify_target_type(target: &Path, target_type: &TargetType) -> CopyResult<()> {
     match (target_type, target.is_dir()) {

src/uu/du/src/du.rs

@@ -500,7 +500,7 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
     };
 
     let strs = if matches.free.is_empty() {
-        vec!["./".to_owned()]
+        vec!["./".to_owned()] // TODO: gnu `du` doesn't use trailing "/" here
     } else {
         matches.free.clone()
     };

src/uu/install/src/install.rs

@@ -41,6 +41,7 @@ pub struct Behavior {
     compare: bool,
     strip: bool,
     strip_program: String,
+    create_leading: bool,
 }
 
 #[derive(Clone, Eq, PartialEq)]
@@ -70,7 +71,7 @@ static OPT_BACKUP: &str = "backup";
 static OPT_BACKUP_2: &str = "backup2";
 static OPT_DIRECTORY: &str = "directory";
 static OPT_IGNORED: &str = "ignored";
-static OPT_CREATED: &str = "created";
+static OPT_CREATE_LEADING: &str = "create-leading";
 static OPT_GROUP: &str = "group";
 static OPT_MODE: &str = "mode";
 static OPT_OWNER: &str = "owner";
@@ -133,9 +134,9 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
 
         .arg(
             // TODO implement flag
-            Arg::with_name(OPT_CREATED)
+            Arg::with_name(OPT_CREATE_LEADING)
                 .short("D")
-                .help("(unimplemented) create all leading components of DEST except the last, then copy SOURCE to DEST")
+                .help("create all leading components of DEST except the last, then copy SOURCE to DEST")
         )
         .arg(
             Arg::with_name(OPT_GROUP)
@@ -266,8 +267,6 @@ fn check_unimplemented<'a>(matches: &ArgMatches) -> Result<(), &'a str> {
         Err("--backup")
     } else if matches.is_present(OPT_BACKUP_2) {
         Err("-b")
-    } else if matches.is_present(OPT_CREATED) {
-        Err("-D")
     } else if matches.is_present(OPT_SUFFIX) {
         Err("--suffix, -S")
     } else if matches.is_present(OPT_TARGET_DIRECTORY) {
@@ -343,6 +342,7 @@ fn behavior(matches: &ArgMatches) -> Result<Behavior, i32> {
                 .value_of(OPT_STRIP_PROGRAM)
                 .unwrap_or(DEFAULT_STRIP_PROGRAM),
         ),
+        create_leading: matches.is_present(OPT_CREATE_LEADING),
     })
 }
 
@@ -410,12 +410,35 @@ fn standard(paths: Vec<String>, b: Behavior) -> i32 {
         .iter()
         .map(PathBuf::from)
         .collect::<Vec<_>>();
+
     let target = Path::new(paths.last().unwrap());
 
-    if (target.is_file() || is_new_file_path(target)) && sources.len() == 1 {
+    if sources.len() > 1 || (target.exists() && target.is_dir()) {
+        copy_files_into_dir(sources, &target.to_path_buf(), &b)
+    } else {
+        if let Some(parent) = target.parent() {
+            if !parent.exists() && b.create_leading {
+                if let Err(e) = fs::create_dir_all(parent) {
+                    show_error!("failed to create {}: {}", parent.display(), e);
+                    return 1;
+                }
+
+                if mode::chmod(&parent, b.mode()).is_err() {
+                    show_error!("failed to chmod {}", parent.display());
+                    return 1;
+                }
+            }
+        }
+
+        if target.is_file() || is_new_file_path(target) {
             copy_file_to_file(&sources[0], &target.to_path_buf(), &b)
         } else {
-        copy_files_into_dir(sources, &target.to_path_buf(), &b)
+            show_error!(
+                "invalid target {}: No such file or directory",
+                target.display()
+            );
+            1
+        }
     }
 }
 

src/uu/ls/BENCHMARKING.md

@@ -0,0 +1,34 @@
+# Benchmarking ls
+
+ls majorly involves fetching a lot of details (depending upon what details are requested, eg. time/date, inode details, etc) for each path using system calls. Ideally, any system call should be done only once for each of the paths - not adhering to this principle leads to a lot of system call overhead multiplying and bubbling up, especially for recursive ls, therefore it is important to always benchmark multiple scenarios.
+This is an overwiew over what was benchmarked, and if you make changes to `ls`, 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!
+
+## Simple recursive ls
+
+-   Get a large tree, for example linux kernel source tree.
+-   Benchmark simple recursive ls with hyperfine: `hyperfine --warmup 2 "target/release/coreutils ls -R tree > /dev/null"`.
+
+## Recursive ls with all and long options
+
+-   Same tree as above
+-   Benchmark recursive ls with -al -R options with hyperfine: `hyperfine --warmup 2 "target/release/coreutils ls -al -R tree > /dev/null"`.
+
+## Comparing with GNU ls
+
+Hyperfine accepts multiple commands to run and will compare them. To compare performance with GNU ls
+duplicate the string you passed to hyperfine but remove the `target/release/coreutils` bit from it.
+
+Example: `hyperfine --warmup 2 "target/release/coreutils ls -al -R tree > /dev/null"` becomes
+`hyperfine --warmup 2 "target/release/coreutils ls -al -R tree > /dev/null" "ls -al -R tree > /dev/null"`
+(This assumes GNU ls is installed as `ls`)
+
+This can also be used to compare with version of ls built before your changes to ensure your change does not regress this
+
+## Checking system call count
+
+- Another thing to look at would be system calls count using strace (on linux) or equivalent on other operating systems.
+- Example: `strace -c target/release/coreutils ls -al -R tree`

src/uu/ls/Cargo.toml

@@ -16,18 +16,19 @@ path = "src/ls.rs"
 
 [dependencies]
 clap = "2.33"
-lazy_static = "1.0.1"
 number_prefix = "0.4"
 term_grid = "0.1.5"
 termsize = "0.1.6"
 time = "0.1.40"
-unicode-width = "0.1.5"
 globset = "0.4.6"
+lscolors = { version="0.7.1", features=["ansi_term"] }
 uucore = { version=">=0.0.8", package="uucore", path="../../uucore", features=["entries", "fs"] }
 uucore_procs = { version=">=0.0.5", package="uucore_procs", path="../../uucore_procs" }
+once_cell = "1.7.2"
+atty = "0.2"
 
 [target.'cfg(unix)'.dependencies]
-atty = "0.2"
+lazy_static = "1.4.0"
 
 [[bin]]
 name = "ls"

src/uu/ls/src/ls.rs

@@ -7,41 +7,41 @@
 
 // spell-checker:ignore (ToDO) cpio svgz webm somegroup nlink rmvb xspf
 
+#[macro_use]
+extern crate uucore;
 #[cfg(unix)]
 #[macro_use]
 extern crate lazy_static;
-#[macro_use]
-extern crate uucore;
 
 mod quoting_style;
 mod version_cmp;
 
 use clap::{App, Arg};
 use globset::{self, Glob, GlobSet, GlobSetBuilder};
+use lscolors::LsColors;
 use number_prefix::NumberPrefix;
+use once_cell::unsync::OnceCell;
 use quoting_style::{escape_name, QuotingStyle};
-#[cfg(unix)]
-use std::collections::HashMap;
-use std::fs;
-use std::fs::{DirEntry, FileType, Metadata};
-#[cfg(unix)]
-use std::os::unix::fs::FileTypeExt;
-#[cfg(any(unix, target_os = "redox"))]
-use std::os::unix::fs::MetadataExt;
 #[cfg(windows)]
 use std::os::windows::fs::MetadataExt;
-use std::path::{Path, PathBuf};
+use std::{
+    cmp::Reverse,
+    fs::{self, DirEntry, FileType, Metadata},
+    io::{stdout, BufWriter, Stdout, Write},
+    path::{Path, PathBuf},
+    process::exit,
+    time::{SystemTime, UNIX_EPOCH},
+};
 #[cfg(unix)]
-use std::time::Duration;
+use std::{
-use std::time::{SystemTime, UNIX_EPOCH};
+    collections::HashMap,
-use std::{cmp::Reverse, process::exit};
+    os::unix::fs::{FileTypeExt, MetadataExt},
-
+    time::Duration,
+};
 use term_grid::{Cell, Direction, Filling, Grid, GridOptions};
 use time::{strftime, Timespec};
 #[cfg(unix)]
-use unicode_width::UnicodeWidthStr;
+use uucore::libc::{S_IXGRP, S_IXOTH, S_IXUSR};
-#[cfg(unix)]
-use uucore::libc::{mode_t, S_ISGID, S_ISUID, S_ISVTX, S_IWOTH, S_IXGRP, S_IXOTH, S_IXUSR};
 
 static VERSION: &str = env!("CARGO_PKG_VERSION");
 static ABOUT: &str = "
@@ -54,30 +54,6 @@ fn get_usage() -> String {
     format!("{0} [OPTION]... [FILE]...", executable!())
 }
 
-#[cfg(unix)]
-static DEFAULT_COLORS: &str = "rs=0:di=01;34:ln=01;36:mh=00:pi=40;33:so=01;35:do=01;35:bd=40;33;01:cd=40;33;01:or=40;31;01:mi=00:su=37;41:sg=30;43:ca=30;41:tw=30;42:ow=34;42:st=37;44:ex=01;32:*.tar=01;31:*.tgz=01;31:*.arc=01;31:*.arj=01;31:*.taz=01;31:*.lha=01;31:*.lz4=01;31:*.lzh=01;31:*.lzma=01;31:*.tlz=01;31:*.txz=01;31:*.tzo=01;31:*.t7z=01;31:*.zip=01;31:*.z=01;31:*.Z=01;31:*.dz=01;31:*.gz=01;31:*.lrz=01;31:*.lz=01;31:*.lzo=01;31:*.xz=01;31:*.bz2=01;31:*.bz=01;31:*.tbz=01;31:*.tbz2=01;31:*.tz=01;31:*.deb=01;31:*.rpm=01;31:*.jar=01;31:*.war=01;31:*.ear=01;31:*.sar=01;31:*.rar=01;31:*.alz=01;31:*.ace=01;31:*.zoo=01;31:*.cpio=01;31:*.7z=01;31:*.rz=01;31:*.cab=01;31:*.jpg=01;35:*.jpeg=01;35:*.gif=01;35:*.bmp=01;35:*.pbm=01;35:*.pgm=01;35:*.ppm=01;35:*.tga=01;35:*.xbm=01;35:*.xpm=01;35:*.tif=01;35:*.tiff=01;35:*.png=01;35:*.svg=01;35:*.svgz=01;35:*.mng=01;35:*.pcx=01;35:*.mov=01;35:*.mpg=01;35:*.mpeg=01;35:*.m2v=01;35:*.mkv=01;35:*.webm=01;35:*.ogm=01;35:*.mp4=01;35:*.m4v=01;35:*.mp4v=01;35:*.vob=01;35:*.qt=01;35:*.nuv=01;35:*.wmv=01;35:*.asf=01;35:*.rm=01;35:*.rmvb=01;35:*.flc=01;35:*.avi=01;35:*.fli=01;35:*.flv=01;35:*.gl=01;35:*.dl=01;35:*.xcf=01;35:*.xwd=01;35:*.yuv=01;35:*.cgm=01;35:*.emf=01;35:*.ogv=01;35:*.ogx=01;35:*.aac=00;36:*.au=00;36:*.flac=00;36:*.m4a=00;36:*.mid=00;36:*.midi=00;36:*.mka=00;36:*.mp3=00;36:*.mpc=00;36:*.ogg=00;36:*.ra=00;36:*.wav=00;36:*.oga=00;36:*.opus=00;36:*.spx=00;36:*.xspf=00;36:";
-
-#[cfg(unix)]
-lazy_static! {
-    static ref LS_COLORS: String =
-        std::env::var("LS_COLORS").unwrap_or_else(|_| DEFAULT_COLORS.to_string());
-    static ref COLOR_MAP: HashMap<&'static str, &'static str> = {
-        let codes = LS_COLORS.split(':');
-        let mut map = HashMap::new();
-        for c in codes {
-            let p: Vec<_> = c.splitn(2, '=').collect();
-            if p.len() == 2 {
-                map.insert(p[0], p[1]);
-            }
-        }
-        map
-    };
-    static ref RESET_CODE: &'static str = COLOR_MAP.get("rs").unwrap_or(&"0");
-    static ref LEFT_CODE: &'static str = COLOR_MAP.get("lc").unwrap_or(&"\x1b[");
-    static ref RIGHT_CODE: &'static str = COLOR_MAP.get("rc").unwrap_or(&"m");
-    static ref END_CODE: &'static str = COLOR_MAP.get("ec").unwrap_or(&"");
-}
-
 pub mod options {
     pub mod format {
         pub static ONELINE: &str = "1";
@@ -113,10 +89,8 @@ pub mod options {
         pub static C: &str = "quote-name";
     }
     pub static QUOTING_STYLE: &str = "quoting-style";
-
     pub mod indicator_style {
-        pub static NONE: &str = "none";
+        pub static SLASH: &str = "p";
-        pub static SLASH: &str = "slash";
         pub static FILE_TYPE: &str = "file-type";
         pub static CLASSIFY: &str = "classify";
     }
@@ -135,9 +109,6 @@ pub mod options {
     pub static TIME: &str = "time";
     pub static IGNORE_BACKUPS: &str = "ignore-backups";
     pub static DIRECTORY: &str = "directory";
-    pub static CLASSIFY: &str = "classify";
-    pub static FILE_TYPE: &str = "file-type";
-    pub static SLASH: &str = "p";
     pub static INODE: &str = "inode";
     pub static REVERSE: &str = "reverse";
     pub static RECURSIVE: &str = "recursive";
@@ -212,8 +183,7 @@ struct Config {
     time: Time,
     #[cfg(unix)]
     inode: bool,
-    #[cfg(unix)]
+    color: Option<LsColors>,
-    color: bool,
     long: LongFormat,
     width: Option<u16>,
     quoting_style: QuotingStyle,
@@ -337,8 +307,7 @@ impl Config {
             Time::Modification
         };
 
-        #[cfg(unix)]
+        let needs_color = match options.value_of(options::COLOR) {
-        let color = match options.value_of(options::COLOR) {
             None => options.is_present(options::COLOR),
             Some(val) => match val {
                 "" | "always" | "yes" | "force" => true,
@@ -347,6 +316,12 @@ impl Config {
             },
         };
 
+        let color = if needs_color {
+            Some(LsColors::from_env().unwrap_or_default())
+        } else {
+            None
+        };
+
         let size_format = if options.is_present(options::size::HUMAN_READABLE) {
             SizeFormat::Binary
         } else if options.is_present(options::size::SI) {
@@ -448,19 +423,11 @@ impl Config {
                 "slash" => IndicatorStyle::Slash,
                 &_ => IndicatorStyle::None,
             }
-        } else if options.is_present(options::indicator_style::NONE) {
+        } else if options.is_present(options::indicator_style::CLASSIFY) {
-            IndicatorStyle::None
-        } else if options.is_present(options::indicator_style::CLASSIFY)
-            || options.is_present(options::CLASSIFY)
-        {
             IndicatorStyle::Classify
-        } else if options.is_present(options::indicator_style::SLASH)
+        } else if options.is_present(options::indicator_style::SLASH) {
-            || options.is_present(options::SLASH)
-        {
             IndicatorStyle::Slash
-        } else if options.is_present(options::indicator_style::FILE_TYPE)
+        } else if options.is_present(options::indicator_style::FILE_TYPE) {
-            || options.is_present(options::FILE_TYPE)
-        {
             IndicatorStyle::FileType
         } else {
             IndicatorStyle::None
@@ -492,6 +459,10 @@ impl Config {
             }
         }
 
+        if files == Files::Normal {
+            ignore_patterns.add(Glob::new(".*").unwrap());
+        }
+
         let ignore_patterns = ignore_patterns.build().unwrap();
 
         let dereference = if options.is_present(options::dereference::ALL) {
@@ -520,7 +491,6 @@ impl Config {
             size_format,
             directory: options.is_present(options::DIRECTORY),
             time,
-            #[cfg(unix)]
             color,
             #[cfg(unix)]
             inode: options.is_present(options::INODE),
@@ -983,45 +953,45 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
                 .takes_value(true)
                 .possible_values(&["none", "slash", "file-type", "classify"])
                 .overrides_with_all(&[
-                    options::FILE_TYPE,
+                    options::indicator_style::FILE_TYPE,
-                    options::SLASH,
+                    options::indicator_style::SLASH,
-                    options::CLASSIFY,
+                    options::indicator_style::CLASSIFY,
                     options::INDICATOR_STYLE,
                 ]))
                 .arg(
-            Arg::with_name(options::CLASSIFY)
+            Arg::with_name(options::indicator_style::CLASSIFY)
                 .short("F")
-                .long(options::CLASSIFY)
+                .long(options::indicator_style::CLASSIFY)
                 .help("Append a character to each file name indicating the file type. Also, for \
                        regular files that are executable, append '*'. The file type indicators are \
                        '/' for directories, '@' for symbolic links, '|' for FIFOs, '=' for sockets, \
                        '>' for doors, and nothing for regular files.")
                 .overrides_with_all(&[
-                    options::FILE_TYPE,
+                    options::indicator_style::FILE_TYPE,
-                    options::SLASH,
+                    options::indicator_style::SLASH,
-                    options::CLASSIFY,
+                    options::indicator_style::CLASSIFY,
                     options::INDICATOR_STYLE,
                 ])
         )
         .arg(
-            Arg::with_name(options::FILE_TYPE)
+            Arg::with_name(options::indicator_style::FILE_TYPE)
-                .long(options::FILE_TYPE)
+                .long(options::indicator_style::FILE_TYPE)
                 .help("Same as --classify, but do not append '*'")
                 .overrides_with_all(&[
-                    options::FILE_TYPE,
+                    options::indicator_style::FILE_TYPE,
-                    options::SLASH,
+                    options::indicator_style::SLASH,
-                    options::CLASSIFY,
+                    options::indicator_style::CLASSIFY,
                     options::INDICATOR_STYLE,
                 ]))
         .arg(
-            Arg::with_name(options::SLASH)
+            Arg::with_name(options::indicator_style::SLASH)
-                .short(options::SLASH)
+                .short(options::indicator_style::SLASH)
                 .help("Append / indicator to directories."
                 )
                 .overrides_with_all(&[
-                    options::FILE_TYPE,
+                    options::indicator_style::FILE_TYPE,
-                    options::SLASH,
+                    options::indicator_style::SLASH,
-                    options::CLASSIFY,
+                    options::indicator_style::CLASSIFY,
                     options::INDICATOR_STYLE,
                 ]))
 
@@ -1038,12 +1008,82 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
     list(locs, Config::from(matches))
 }
 
+/// Represents a Path along with it's associated data
+/// Any data that will be reused several times makes sense to be added to this structure
+/// Caching data here helps eliminate redundant syscalls to fetch same information
+struct PathData {
+    // Result<MetaData> got from symlink_metadata() or metadata() based on config
+    md: OnceCell<Option<Metadata>>,
+    ft: OnceCell<Option<FileType>>,
+    // Name of the file - will be empty for . or ..
+    file_name: String,
+    // PathBuf that all above data corresponds to
+    p_buf: PathBuf,
+    must_dereference: bool,
+}
+
+impl PathData {
+    fn new(
+        p_buf: PathBuf,
+        file_type: Option<std::io::Result<FileType>>,
+        config: &Config,
+        command_line: bool,
+    ) -> Self {
+        let name = p_buf
+            .file_name()
+            .map_or(String::new(), |s| s.to_string_lossy().into_owned());
+        let must_dereference = match &config.dereference {
+            Dereference::All => true,
+            Dereference::Args => command_line,
+            Dereference::DirArgs => {
+                if command_line {
+                    if let Ok(md) = p_buf.metadata() {
+                        md.is_dir()
+                    } else {
+                        false
+                    }
+                } else {
+                    false
+                }
+            }
+            Dereference::None => false,
+        };
+        let ft = match file_type {
+            Some(ft) => OnceCell::from(ft.ok()),
+            None => OnceCell::new(),
+        };
+
+        Self {
+            md: OnceCell::new(),
+            ft,
+            file_name: name,
+            p_buf,
+            must_dereference,
+        }
+    }
+
+    fn md(&self) -> Option<&Metadata> {
+        self.md
+            .get_or_init(|| get_metadata(&self.p_buf, self.must_dereference).ok())
+            .as_ref()
+    }
+
+    fn file_type(&self) -> Option<&FileType> {
+        self.ft
+            .get_or_init(|| self.md().map(|md| md.file_type()))
+            .as_ref()
+    }
+}
+
 fn list(locs: Vec<String>, config: Config) -> i32 {
     let number_of_locs = locs.len();
 
-    let mut files = Vec::<PathBuf>::new();
+    let mut files = Vec::<PathData>::new();
-    let mut dirs = Vec::<PathBuf>::new();
+    let mut dirs = Vec::<PathData>::new();
     let mut has_failed = false;
+
+    let mut out = BufWriter::new(stdout());
+
     for loc in locs {
         let p = PathBuf::from(&loc);
         if !p.exists() {
@@ -1054,38 +1094,30 @@ fn list(locs: Vec<String>, config: Config) -> i32 {
             continue;
         }
 
-        let show_dir_contents = if !config.directory {
+        let path_data = PathData::new(p, None, &config, true);
-            match config.dereference {
+
-                Dereference::None => {
+        let show_dir_contents = if let Some(ft) = path_data.file_type() {
-                    if let Ok(md) = p.symlink_metadata() {
+            !config.directory && ft.is_dir()
-                        md.is_dir()
         } else {
-                        show_error!("'{}': {}", &loc, "No such file or directory");
             has_failed = true;
-                        continue;
-                    }
-                }
-                _ => p.is_dir(),
-            }
-        } else {
             false
         };
 
         if show_dir_contents {
-            dirs.push(p);
+            dirs.push(path_data);
         } else {
-            files.push(p);
+            files.push(path_data);
         }
     }
     sort_entries(&mut files, &config);
-    display_items(&files, None, &config, true);
+    display_items(&files, None, &config, &mut out);
 
     sort_entries(&mut dirs, &config);
     for dir in dirs {
         if number_of_locs > 1 {
-            println!("\n{}:", dir.to_string_lossy());
+            let _ = writeln!(out, "\n{}:", dir.p_buf.display());
         }
-        enter_directory(&dir, &config);
+        enter_directory(&dir, &config, &mut out);
     }
     if has_failed {
         1
@@ -1094,22 +1126,21 @@ fn list(locs: Vec<String>, config: Config) -> i32 {
     }
 }
 
-fn sort_entries(entries: &mut Vec<PathBuf>, config: &Config) {
+fn sort_entries(entries: &mut Vec<PathData>, config: &Config) {
     match config.sort {
         Sort::Time => entries.sort_by_key(|k| {
             Reverse(
-                get_metadata(k, false)
+                k.md()
-                    .ok()
                     .and_then(|md| get_system_time(&md, config))
                     .unwrap_or(UNIX_EPOCH),
             )
         }),
         Sort::Size => {
-            entries.sort_by_key(|k| Reverse(get_metadata(k, false).map(|md| md.len()).unwrap_or(0)))
+            entries.sort_by_key(|k| Reverse(k.md().as_ref().map(|md| md.len()).unwrap_or(0)))
         }
         // The default sort in GNU ls is case insensitive
-        Sort::Name => entries.sort_by_key(|k| k.to_string_lossy().to_lowercase()),
+        Sort::Name => entries.sort_by_cached_key(|k| k.file_name.to_lowercase()),
-        Sort::Version => entries.sort_by(|a, b| version_cmp::version_cmp(a, b)),
+        Sort::Version => entries.sort_by(|k, j| version_cmp::version_cmp(&k.p_buf, &j.p_buf)),
         Sort::None => {}
     }
 
@@ -1122,48 +1153,53 @@ fn sort_entries(entries: &mut Vec<PathBuf>, config: &Config) {
 fn is_hidden(file_path: &DirEntry) -> bool {
     let metadata = fs::metadata(file_path.path()).unwrap();
     let attr = metadata.file_attributes();
-    ((attr & 0x2) > 0) || file_path.file_name().to_string_lossy().starts_with('.')
+    (attr & 0x2) > 0
-}
-
-#[cfg(unix)]
-fn is_hidden(file_path: &DirEntry) -> bool {
-    file_path.file_name().to_string_lossy().starts_with('.')
 }
 
 fn should_display(entry: &DirEntry, config: &Config) -> bool {
     let ffi_name = entry.file_name();
 
+    // For unix, the hidden files are already included in the ignore pattern
+    #[cfg(windows)]
+    {
         if config.files == Files::Normal && is_hidden(entry) {
             return false;
         }
-
-    if config.ignore_patterns.is_match(&ffi_name) {
-        return false;
     }
-    true
+
+    !config.ignore_patterns.is_match(&ffi_name)
 }
 
-fn enter_directory(dir: &Path, config: &Config) {
+fn enter_directory(dir: &PathData, config: &Config, out: &mut BufWriter<Stdout>) {
-    let mut entries: Vec<_> = safe_unwrap!(fs::read_dir(dir).and_then(Iterator::collect));
+    let mut entries: Vec<_> = if config.files == Files::All {
-
+        vec![
-    entries.retain(|e| should_display(e, config));
+            PathData::new(dir.p_buf.join("."), None, config, false),
-
+            PathData::new(dir.p_buf.join(".."), None, config, false),
-    let mut entries: Vec<_> = entries.iter().map(DirEntry::path).collect();
+        ]
-    sort_entries(&mut entries, config);
-
-    if config.files == Files::All {
-        let mut display_entries = entries.clone();
-        display_entries.insert(0, dir.join(".."));
-        display_entries.insert(0, dir.join("."));
-        display_items(&display_entries, Some(dir), config, false);
     } else {
-        display_items(&entries, Some(dir), config, false);
+        vec![]
-    }
+    };
+
+    let mut temp: Vec<_> = safe_unwrap!(fs::read_dir(&dir.p_buf))
+        .map(|res| safe_unwrap!(res))
+        .filter(|e| should_display(e, config))
+        .map(|e| PathData::new(DirEntry::path(&e), Some(e.file_type()), config, false))
+        .collect();
+
+    sort_entries(&mut temp, config);
+
+    entries.append(&mut temp);
+
+    display_items(&entries, Some(&dir.p_buf), config, out);
 
     if config.recursive {
-        for e in entries.iter().filter(|p| p.is_dir()) {
+        for e in entries
-            println!("\n{}:", e.to_string_lossy());
+            .iter()
-            enter_directory(&e, config);
+            .skip(if config.files == Files::All { 2 } else { 0 })
+            .filter(|p| p.file_type().map(|ft| ft.is_dir()).unwrap_or(false))
+        {
+            let _ = writeln!(out, "\n{}:", e.p_buf.display());
+            enter_directory(&e, config, out);
         }
     }
 }
@@ -1176,8 +1212,8 @@ fn get_metadata(entry: &Path, dereference: bool) -> std::io::Result<Metadata> {
     }
 }
 
-fn display_dir_entry_size(entry: &Path, config: &Config) -> (usize, usize) {
+fn display_dir_entry_size(entry: &PathData, config: &Config) -> (usize, usize) {
-    if let Ok(md) = get_metadata(entry, false) {
+    if let Some(md) = entry.md() {
         (
             display_symlink_count(&md).len(),
             display_file_size(&md, config).len(),
@@ -1191,7 +1227,12 @@ fn pad_left(string: String, count: usize) -> String {
     format!("{:>width$}", string, width = count)
 }
 
-fn display_items(items: &[PathBuf], strip: Option<&Path>, config: &Config, command_line: bool) {
+fn display_items(
+    items: &[PathData],
+    strip: Option<&Path>,
+    config: &Config,
+    out: &mut BufWriter<Stdout>,
+) {
     if config.format == Format::Long {
         let (mut max_links, mut max_size) = (1, 1);
         for item in items {
@@ -1200,58 +1241,59 @@ fn display_items(items: &[PathBuf], strip: Option<&Path>, config: &Config, comma
             max_size = size.max(max_size);
         }
         for item in items {
-            display_item_long(item, strip, max_links, max_size, config, command_line);
+            display_item_long(item, strip, max_links, max_size, config, out);
         }
     } else {
-        let names = items.iter().filter_map(|i| {
+        let names = items
-            let md = get_metadata(i, false);
+            .iter()
-            match md {
+            .filter_map(|i| display_file_name(&i, strip, config));
-                Err(e) => {
-                    let filename = get_file_name(i, strip);
-                    show_error!("'{}': {}", filename, e);
-                    None
-                }
-                Ok(md) => Some(display_file_name(&i, strip, &md, config)),
-            }
-        });
 
         match (&config.format, config.width) {
-            (Format::Columns, Some(width)) => display_grid(names, width, Direction::TopToBottom),
+            (Format::Columns, Some(width)) => {
-            (Format::Across, Some(width)) => display_grid(names, width, Direction::LeftToRight),
+                display_grid(names, width, Direction::TopToBottom, out)
+            }
+            (Format::Across, Some(width)) => {
+                display_grid(names, width, Direction::LeftToRight, out)
+            }
             (Format::Commas, width_opt) => {
                 let term_width = width_opt.unwrap_or(1);
                 let mut current_col = 0;
                 let mut names = names;
                 if let Some(name) = names.next() {
-                    print!("{}", name.contents);
+                    let _ = write!(out, "{}", name.contents);
                     current_col = name.width as u16 + 2;
                 }
                 for name in names {
                     let name_width = name.width as u16;
                     if current_col + name_width + 1 > term_width {
                         current_col = name_width + 2;
-                        print!(",\n{}", name.contents);
+                        let _ = write!(out, ",\n{}", name.contents);
                     } else {
                         current_col += name_width + 2;
-                        print!(", {}", name.contents);
+                        let _ = write!(out, ", {}", name.contents);
                     }
                 }
                 // Current col is never zero again if names have been printed.
                 // So we print a newline.
                 if current_col > 0 {
-                    println!();
+                    let _ = writeln!(out,);
                 }
             }
             _ => {
                 for name in names {
-                    println!("{}", name.contents);
+                    let _ = writeln!(out, "{}", name.contents);
                 }
             }
         }
     }
 }
 
-fn display_grid(names: impl Iterator<Item = Cell>, width: u16, direction: Direction) {
+fn display_grid(
+    names: impl Iterator<Item = Cell>,
+    width: u16,
+    direction: Direction,
+    out: &mut BufWriter<Stdout>,
+) {
     let mut grid = Grid::new(GridOptions {
         filling: Filling::Spaces(2),
         direction,
@@ -1262,56 +1304,44 @@ fn display_grid(names: impl Iterator<Item = Cell>, width: u16, direction: Direct
     }
 
     match grid.fit_into_width(width as usize) {
-        Some(output) => print!("{}", output),
+        Some(output) => {
+            let _ = write!(out, "{}", output);
+        }
         // Width is too small for the grid, so we fit it in one column
-        None => print!("{}", grid.fit_into_columns(1)),
+        None => {
+            let _ = write!(out, "{}", grid.fit_into_columns(1));
+        }
     }
 }
 
 use uucore::fs::display_permissions;
 
 fn display_item_long(
-    item: &Path,
+    item: &PathData,
     strip: Option<&Path>,
     max_links: usize,
     max_size: usize,
     config: &Config,
-    command_line: bool,
+    out: &mut BufWriter<Stdout>,
 ) {
-    let dereference = match &config.dereference {
+    let md = match item.md() {
-        Dereference::All => true,
+        None => {
-        Dereference::Args => command_line,
+            let filename = get_file_name(&item.p_buf, strip);
-        Dereference::DirArgs => {
+            show_error!("could not show file: {}", filename);
-            if command_line {
-                if let Ok(md) = item.metadata() {
-                    md.is_dir()
-                } else {
-                    false
-                }
-            } else {
-                false
-            }
-        }
-        Dereference::None => false,
-    };
-
-    let md = match get_metadata(item, dereference) {
-        Err(e) => {
-            let filename = get_file_name(&item, strip);
-            show_error!("{}: {}", filename, e);
             return;
         }
-        Ok(md) => md,
+        Some(md) => md,
     };
 
     #[cfg(unix)]
     {
         if config.inode {
-            print!("{} ", get_inode(&md));
+            let _ = write!(out, "{} ", get_inode(&md));
         }
     }
 
-    print!(
+    let _ = write!(
+        out,
         "{}{} {}",
         display_file_type(md.file_type()),
         display_permissions(&md),
@@ -1319,24 +1349,28 @@ fn display_item_long(
     );
 
     if config.long.owner {
-        print!(" {}", display_uname(&md, config));
+        let _ = write!(out, " {}", display_uname(&md, config));
     }
 
     if config.long.group {
-        print!(" {}", display_group(&md, config));
+        let _ = write!(out, " {}", display_group(&md, config));
     }
 
     // Author is only different from owner on GNU/Hurd, so we reuse
     // the owner, since GNU/Hurd is not currently supported by Rust.
     if config.long.author {
-        print!(" {}", display_uname(&md, config));
+        let _ = write!(out, " {}", display_uname(&md, config));
     }
 
-    println!(
+    let _ = writeln!(
+        out,
         " {} {} {}",
         pad_left(display_file_size(&md, config), max_size),
         display_date(&md, config),
-        display_file_name(&item, strip, &md, config).contents,
+        // unwrap is fine because it fails when metadata is not available
+        // but we already know that it is because it's checked at the
+        // start of the function.
+        display_file_name(&item, strip, config).unwrap().contents,
     );
 }
 
@@ -1348,23 +1382,51 @@ fn get_inode(metadata: &Metadata) -> String {
 // Currently getpwuid is `linux` target only. If it's broken out into
 // a posix-compliant attribute this can be updated...
 #[cfg(unix)]
+use std::sync::Mutex;
+#[cfg(unix)]
 use uucore::entries;
 
+#[cfg(unix)]
+fn cached_uid2usr(uid: u32) -> String {
+    lazy_static! {
+        static ref UID_CACHE: Mutex<HashMap<u32, String>> = Mutex::new(HashMap::new());
+    }
+
+    let mut uid_cache = UID_CACHE.lock().unwrap();
+    uid_cache
+        .entry(uid)
+        .or_insert_with(|| entries::uid2usr(uid).unwrap_or_else(|_| uid.to_string()))
+        .clone()
+}
+
 #[cfg(unix)]
 fn display_uname(metadata: &Metadata, config: &Config) -> String {
     if config.long.numeric_uid_gid {
         metadata.uid().to_string()
     } else {
-        entries::uid2usr(metadata.uid()).unwrap_or_else(|_| metadata.uid().to_string())
+        cached_uid2usr(metadata.uid())
     }
 }
 
+#[cfg(unix)]
+fn cached_gid2grp(gid: u32) -> String {
+    lazy_static! {
+        static ref GID_CACHE: Mutex<HashMap<u32, String>> = Mutex::new(HashMap::new());
+    }
+
+    let mut gid_cache = GID_CACHE.lock().unwrap();
+    gid_cache
+        .entry(gid)
+        .or_insert_with(|| entries::gid2grp(gid).unwrap_or_else(|_| gid.to_string()))
+        .clone()
+}
+
 #[cfg(unix)]
 fn display_group(metadata: &Metadata, config: &Config) -> String {
     if config.long.numeric_uid_gid {
         metadata.gid().to_string()
     } else {
-        entries::gid2grp(metadata.gid()).unwrap_or_else(|_| metadata.gid().to_string())
+        cached_gid2grp(metadata.gid())
     }
 }
 
@@ -1374,7 +1436,6 @@ fn display_uname(_metadata: &Metadata, _config: &Config) -> String {
 }
 
 #[cfg(not(unix))]
-#[allow(unused_variables)]
 fn display_group(_metadata: &Metadata, _config: &Config) -> String {
     "somegroup".to_string()
 }
@@ -1449,13 +1510,13 @@ fn display_file_size(metadata: &Metadata, config: &Config) -> String {
     }
 }
 
-fn display_file_type(file_type: FileType) -> String {
+fn display_file_type(file_type: FileType) -> char {
     if file_type.is_dir() {
-        "d".to_string()
+        'd'
     } else if file_type.is_symlink() {
-        "l".to_string()
+        'l'
     } else {
-        "-".to_string()
+        '-'
     }
 }
 
@@ -1470,140 +1531,58 @@ fn get_file_name(name: &Path, strip: Option<&Path>) -> String {
     name.to_string_lossy().into_owned()
 }
 
-#[cfg(not(unix))]
+#[cfg(unix)]
-fn display_file_name(
+fn file_is_executable(md: &Metadata) -> bool {
-    path: &Path,
+    // Mode always returns u32, but the flags might not be, based on the platform
-    strip: Option<&Path>,
+    // e.g. linux has u32, mac has u16.
-    metadata: &Metadata,
+    // S_IXUSR -> user has execute permission
-    config: &Config,
+    // S_IXGRP -> group has execute persmission
-) -> Cell {
+    // S_IXOTH -> other users have execute permission
-    let mut name = escape_name(get_file_name(path, strip), &config.quoting_style);
+    md.mode() & ((S_IXUSR | S_IXGRP | S_IXOTH) as u32) != 0
-    let file_type = metadata.file_type();
+}
+
+#[allow(clippy::clippy::collapsible_else_if)]
+fn classify_file(path: &PathData) -> Option<char> {
+    let file_type = path.file_type()?;
 
-    match config.indicator_style {
-        IndicatorStyle::Classify | IndicatorStyle::FileType => {
     if file_type.is_dir() {
-                name.push('/');
+        Some('/')
-            }
-            if file_type.is_symlink() {
-                name.push('@');
-            }
-        }
-        IndicatorStyle::Slash => {
-            if file_type.is_dir() {
-                name.push('/');
-            }
-        }
-        _ => (),
-    };
-
-    if config.format == Format::Long && metadata.file_type().is_symlink() {
-        if let Ok(target) = path.read_link() {
-            // We don't bother updating width here because it's not used for long listings
-            let target_name = target.to_string_lossy().to_string();
-            name.push_str(" -> ");
-            name.push_str(&target_name);
-        }
-    }
-
-    name.into()
-}
-
-#[cfg(unix)]
-fn color_name(name: String, typ: &str) -> String {
-    let mut typ = typ;
-    if !COLOR_MAP.contains_key(typ) {
-        if typ == "or" {
-            typ = "ln";
-        } else if typ == "mi" {
-            typ = "fi";
-        }
-    };
-    if let Some(code) = COLOR_MAP.get(typ) {
-        format!(
-            "{}{}{}{}{}{}{}{}",
-            *LEFT_CODE, code, *RIGHT_CODE, name, *END_CODE, *LEFT_CODE, *RESET_CODE, *RIGHT_CODE,
-        )
-    } else {
-        name
-    }
-}
-
-#[cfg(unix)]
-macro_rules! has {
-    ($mode:expr, $perm:expr) => {
-        $mode & ($perm as mode_t) != 0
-    };
-}
-
-#[cfg(unix)]
-#[allow(clippy::cognitive_complexity)]
-fn display_file_name(
-    path: &Path,
-    strip: Option<&Path>,
-    metadata: &Metadata,
-    config: &Config,
-) -> Cell {
-    let mut name = escape_name(get_file_name(path, strip), &config.quoting_style);
-    if config.format != Format::Long && config.inode {
-        name = get_inode(metadata) + " " + &name;
-    }
-    let mut width = UnicodeWidthStr::width(&*name);
-
-    let ext;
-    if config.color || config.indicator_style != IndicatorStyle::None {
-        let file_type = metadata.file_type();
-
-        let (code, sym) = if file_type.is_dir() {
-            ("di", Some('/'))
     } else if file_type.is_symlink() {
-            if path.exists() {
+        Some('@')
-                ("ln", Some('@'))
     } else {
-                ("or", Some('@'))
+        #[cfg(unix)]
-            }
+        {
-        } else if file_type.is_socket() {
+            if file_type.is_socket() {
-            ("so", Some('='))
+                Some('=')
             } else if file_type.is_fifo() {
-            ("pi", Some('|'))
+                Some('|')
-        } else if file_type.is_block_device() {
+            } else if file_type.is_file() && file_is_executable(path.md()?) {
-            ("bd", None)
-        } else if file_type.is_char_device() {
-            ("cd", None)
-        } else if file_type.is_file() {
-            let mode = metadata.mode() as mode_t;
-            let sym = if has!(mode, S_IXUSR | S_IXGRP | S_IXOTH) {
                 Some('*')
             } else {
                 None
-            };
-            if has!(mode, S_ISUID) {
-                ("su", sym)
-            } else if has!(mode, S_ISGID) {
-                ("sg", sym)
-            } else if has!(mode, S_ISVTX) && has!(mode, S_IWOTH) {
-                ("tw", sym)
-            } else if has!(mode, S_ISVTX) {
-                ("st", sym)
-            } else if has!(mode, S_IWOTH) {
-                ("ow", sym)
-            } else if has!(mode, S_IXUSR | S_IXGRP | S_IXOTH) {
-                ("ex", sym)
-            } else if metadata.nlink() > 1 {
-                ("mh", sym)
-            } else if let Some(e) = path.extension() {
-                ext = format!("*.{}", e.to_string_lossy());
-                (ext.as_str(), None)
-            } else {
-                ("fi", None)
             }
-        } else {
+        }
-            ("", None)
+        #[cfg(not(unix))]
-        };
+        None
+    }
+}
 
-        if config.color {
+fn display_file_name(path: &PathData, strip: Option<&Path>, config: &Config) -> Option<Cell> {
-            name = color_name(name, code);
+    let mut name = escape_name(get_file_name(&path.p_buf, strip), &config.quoting_style);
+
+    #[cfg(unix)]
+    {
+        if config.format != Format::Long && config.inode {
+            name = get_inode(path.md()?) + " " + &name;
         }
+    }
+
+    if let Some(ls_colors) = &config.color {
+        name = color_name(&ls_colors, &path.p_buf, name, path.md()?);
+    }
+
+    if config.indicator_style != IndicatorStyle::None {
+        let sym = classify_file(path);
 
         let char_opt = match config.indicator_style {
             IndicatorStyle::Classify => sym,
@@ -1626,23 +1605,23 @@ fn display_file_name(
 
         if let Some(c) = char_opt {
             name.push(c);
-            width += 1;
         }
     }
 
-    if config.format == Format::Long && metadata.file_type().is_symlink() {
+    if config.format == Format::Long && path.file_type()?.is_symlink() {
-        if let Ok(target) = path.read_link() {
+        if let Ok(target) = path.p_buf.read_link() {
-            // We don't bother updating width here because it's not used for long listings
-            let code = if target.exists() { "fi" } else { "mi" };
-            let target_name = color_name(target.to_string_lossy().to_string(), code);
             name.push_str(" -> ");
-            name.push_str(&target_name);
+            name.push_str(&target.to_string_lossy());
         }
     }
 
-    Cell {
+    Some(name.into())
-        contents: name,
+}
-        width,
+
+fn color_name(ls_colors: &LsColors, path: &Path, name: String, md: &Metadata) -> String {
+    match ls_colors.style_for_path_with_metadata(path, Some(&md)) {
+        Some(style) => style.to_ansi_term_style().paint(name).to_string(),
+        None => name,
     }
 }
 

src/uu/ls/src/quoting_style.rs

@@ -1,6 +1,6 @@
 use std::char::from_digit;
 
-const SPECIAL_SHELL_CHARS: &str = "~`#$&*()\\|[]{};'\"<>?! ";
+const SPECIAL_SHELL_CHARS: &str = "~`#$&*()|[]{};\\'\"<>?! ";
 
 pub(super) enum QuotingStyle {
     Shell {
@@ -27,12 +27,10 @@ pub(super) enum Quotes {
 // This implementation is heavily inspired by the std::char::EscapeDefault implementation
 // in the Rust standard library. This custom implementation is needed because the
 // characters \a, \b, \e, \f & \v are not recognized by Rust.
-#[derive(Clone, Debug)]
 struct EscapedChar {
     state: EscapeState,
 }
 
-#[derive(Clone, Debug)]
 enum EscapeState {
     Done,
     Char(char),
@@ -41,14 +39,12 @@ enum EscapeState {
     Octal(EscapeOctal),
 }
 
-#[derive(Clone, Debug)]
 struct EscapeOctal {
     c: char,
     state: EscapeOctalState,
     idx: usize,
 }
 
-#[derive(Clone, Debug)]
 enum EscapeOctalState {
     Done,
     Backslash,
@@ -135,7 +131,6 @@ impl EscapedChar {
             '\x0B' => Backslash('v'),
             '\x0C' => Backslash('f'),
             '\r' => Backslash('r'),
-            '\\' => Backslash('\\'),
             '\x00'..='\x1F' | '\x7F' => Octal(EscapeOctal::from(c)),
             '\'' => match quotes {
                 Quotes::Single => Backslash('\''),
@@ -511,6 +506,23 @@ mod tests {
             ],
         );
 
+        // A control character followed by a special shell character
+        check_names(
+            "one\n&two",
+            vec![
+                ("one?&two", "literal"),
+                ("one\n&two", "literal-show"),
+                ("one\\n&two", "escape"),
+                ("\"one\\n&two\"", "c"),
+                ("'one?&two'", "shell"),
+                ("'one\n&two'", "shell-show"),
+                ("'one?&two'", "shell-always"),
+                ("'one\n&two'", "shell-always-show"),
+                ("'one'$'\\n''&two'", "shell-escape"),
+                ("'one'$'\\n''&two'", "shell-escape-always"),
+            ],
+        );
+
         // The first 16 control characters. NUL is also included, even though it is of
         // no importance for file names.
         check_names(
@@ -627,4 +639,22 @@ mod tests {
             ],
         );
     }
+
+    #[test]
+    fn test_backslash() {
+        // Escaped in C-style, but not in Shell-style escaping
+        check_names(
+            "one\\two",
+            vec![
+                ("one\\two", "literal"),
+                ("one\\two", "literal-show"),
+                ("one\\\\two", "escape"),
+                ("\"one\\\\two\"", "c"),
+                ("'one\\two'", "shell"),
+                ("\'one\\two\'", "shell-always"),
+                ("'one\\two'", "shell-escape"),
+                ("'one\\two'", "shell-escape-always"),
+            ],
+        );
+    }
 }

src/uu/mknod/src/mknod.rs

@@ -7,8 +7,6 @@
 
 // spell-checker:ignore (ToDO) parsemode makedev sysmacros makenod newmode perror IFBLK IFCHR IFIFO
 
-mod parsemode;
-
 #[macro_use]
 extern crate uucore;
 
@@ -98,7 +96,7 @@ for details about the options it supports.",
     let mut last_umask: mode_t = 0;
     let mut newmode: mode_t = MODE_RW_UGO;
     if matches.opt_present("mode") {
-        match parsemode::parse_mode(matches.opt_str("mode")) {
+        match uucore::mode::parse_mode(matches.opt_str("mode")) {
             Ok(parsed) => {
                 if parsed > 0o777 {
                     show_info!("mode must specify only file permission bits");

src/uu/mknod/src/parsemode.rs

@@ -1,58 +0,0 @@
-// spell-checker:ignore (ToDO) fperm
-
-use libc::{mode_t, S_IRGRP, S_IROTH, S_IRUSR, S_IWGRP, S_IWOTH, S_IWUSR};
-
-use uucore::mode;
-
-pub fn parse_mode(mode: Option<String>) -> Result<mode_t, String> {
-    let fperm = S_IRUSR | S_IWUSR | S_IRGRP | S_IWGRP | S_IROTH | S_IWOTH;
-    if let Some(mode) = mode {
-        let arr: &[char] = &['0', '1', '2', '3', '4', '5', '6', '7', '8', '9'];
-        let result = if mode.contains(arr) {
-            mode::parse_numeric(fperm as u32, mode.as_str())
-        } else {
-            mode::parse_symbolic(fperm as u32, mode.as_str(), true)
-        };
-        result.map(|mode| mode as mode_t)
-    } else {
-        Ok(fperm)
-    }
-}
-
-#[cfg(test)]
-mod test {
-    /// Test if the program is running under WSL
-    // ref: <https://github.com/microsoft/WSL/issues/4555> @@ <https://archive.is/dP0bz>
-    // ToDO: test on WSL2 which likely doesn't need special handling; plan change to `is_wsl_1()` if WSL2 is less needy
-    pub fn is_wsl() -> bool {
-        #[cfg(target_os = "linux")]
-        {
-            if let Ok(b) = std::fs::read("/proc/sys/kernel/osrelease") {
-                if let Ok(s) = std::str::from_utf8(&b) {
-                    let a = s.to_ascii_lowercase();
-                    return a.contains("microsoft") || a.contains("wsl");
-                }
-            }
-        }
-        false
-    }
-
-    #[test]
-    fn symbolic_modes() {
-        assert_eq!(super::parse_mode(Some("u+x".to_owned())).unwrap(), 0o766);
-        assert_eq!(
-            super::parse_mode(Some("+x".to_owned())).unwrap(),
-            if !is_wsl() { 0o777 } else { 0o776 }
-        );
-        assert_eq!(super::parse_mode(Some("a-w".to_owned())).unwrap(), 0o444);
-        assert_eq!(super::parse_mode(Some("g-r".to_owned())).unwrap(), 0o626);
-    }
-
-    #[test]
-    fn numeric_modes() {
-        assert_eq!(super::parse_mode(Some("644".to_owned())).unwrap(), 0o644);
-        assert_eq!(super::parse_mode(Some("+100".to_owned())).unwrap(), 0o766);
-        assert_eq!(super::parse_mode(Some("-4".to_owned())).unwrap(), 0o662);
-        assert_eq!(super::parse_mode(None).unwrap(), 0o666);
-    }
-}

src/uu/printf/src/cli.rs

@@ -18,18 +18,15 @@ pub fn err_msg(msg: &str) {
 
 // by default stdout only flushes
 // to console when a newline is passed.
-#[allow(unused_must_use)]
 pub fn flush_char(c: char) {
     print!("{}", c);
-    stdout().flush();
+    let _ = stdout().flush();
 }
-#[allow(unused_must_use)]
 pub fn flush_str(s: &str) {
     print!("{}", s);
-    stdout().flush();
+    let _ = stdout().flush();
 }
-#[allow(unused_must_use)]
 pub fn flush_bytes(bslice: &[u8]) {
-    stdout().write(bslice);
+    let _ = stdout().write(bslice);
-    stdout().flush();
+    let _ = stdout().flush();
 }

src/uu/printf/src/printf.rs

@@ -1,5 +1,4 @@
 #![allow(dead_code)]
-
 // spell-checker:ignore (change!) each's
 // spell-checker:ignore (ToDO) LONGHELP FORMATSTRING templating parameterizing formatstr
 
@@ -9,7 +8,6 @@ mod tokenize;
 
 static NAME: &str = "printf";
 static VERSION: &str = env!("CARGO_PKG_VERSION");
-static SHORT_USAGE: &str = "printf: usage: printf [-v var] format [arguments]";
 static LONGHELP_LEAD: &str = "printf
 
  USAGE: printf FORMATSTRING [ARGUMENT]...

src/uu/printf/src/tokenize/num_format/formatters/base_conv/mod.rs

@@ -28,8 +28,7 @@ pub fn arrnum_int_mult(arr_num: &[u8], basenum: u8, base_ten_int_fact: u8) -> Ve
             }
         }
     }
-    #[allow(clippy::map_clone)]
+    let ret: Vec<u8> = ret_rev.into_iter().rev().collect();
-    let ret: Vec<u8> = ret_rev.iter().rev().map(|x| *x).collect();
     ret
 }
 
@@ -102,70 +101,6 @@ pub fn arrnum_int_div_step(
         remainder: rem_out,
     }
 }
-// pub struct ArrFloat {
-// pub leading_zeros: u8,
-// pub values: Vec<u8>,
-// pub basenum: u8
-// }
-//
-// pub struct ArrFloatDivOut {
-// pub quotient: u8,
-// pub remainder: ArrFloat
-// }
-//
-// pub fn arrfloat_int_div(
-// arrfloat_in : &ArrFloat,
-// base_ten_int_divisor : u8,
-// precision : u16
-// ) -> DivOut  {
-//
-// let mut remainder = ArrFloat {
-// basenum: arrfloat_in.basenum,
-// leading_zeros: arrfloat_in.leading_zeroes,
-// values: Vec<u8>::new()
-// }
-// let mut quotient = 0;
-//
-// let mut bufferval : u16 = 0;
-// let base : u16 = arrfloat_in.basenum as u16;
-// let divisor : u16 = base_ten_int_divisor as u16;
-//
-// let mut it_f = arrfloat_in.values.iter();
-// let mut position = 0 + arrfloat_in.leading_zeroes as u16;
-// let mut at_end = false;
-// while position< precision {
-// let next_digit = match it_f.next() {
-// Some(c) => {}
-// None => { 0 }
-// }
-// match u_cur {
-// Some(u) => {
-// bufferval += u.clone() as u16;
-// if bufferval > divisor {
-// while bufferval >= divisor {
-// quotient+=1;
-// bufferval -= divisor;
-// }
-// if bufferval == 0 {
-// rem_out.position +=1;
-// } else {
-// rem_out.replace = Some(bufferval as u8);
-// }
-// break;
-// } else {
-// bufferval *= base;
-// }
-// },
-// None => {
-// break;
-// }
-// }
-// u_cur = it_f.next().clone();
-// rem_out.position+=1;
-// }
-// ArrFloatDivOut { quotient: quotient, remainder: remainder }
-// }
-//
 pub fn arrnum_int_add(arrnum: &[u8], basenum: u8, base_ten_int_term: u8) -> Vec<u8> {
     let mut carry: u16 = u16::from(base_ten_int_term);
     let mut rem: u16;
@@ -193,8 +128,7 @@ pub fn arrnum_int_add(arrnum: &[u8], basenum: u8, base_ten_int_term: u8) -> Vec<
             }
         }
     }
-    #[allow(clippy::map_clone)]
+    let ret: Vec<u8> = ret_rev.into_iter().rev().collect();
-    let ret: Vec<u8> = ret_rev.iter().rev().map(|x| *x).collect();
     ret
 }
 
@@ -219,8 +153,7 @@ pub fn unsigned_to_arrnum(src: u16) -> Vec<u8> {
 }
 
 // temporary needs-improvement-function
-#[allow(unused_variables)]
+pub fn base_conv_float(src: &[u8], radix_src: u8, _radix_dest: u8) -> f64 {
-pub fn base_conv_float(src: &[u8], radix_src: u8, radix_dest: u8) -> f64 {
     // it would require a lot of addl code
     // to implement this for arbitrary string input.
     // until then, the below operates as an outline
@@ -267,7 +200,6 @@ pub fn arrnum_to_str(src: &[u8], radix_def_dest: &dyn RadixDef) -> String {
     str_out
 }
 
-#[allow(unused_variables)]
 pub fn base_conv_str(
     src: &str,
     radix_def_src: &dyn RadixDef,

src/uu/printf/src/tokenize/num_format/formatters/cninetyninehexfloatf.rs

@@ -43,45 +43,15 @@ impl Formatter for CninetyNineHexFloatf {
 // c99 hex has unique requirements of all floating point subs in pretty much every part of building a primitive, from prefix and suffix to need for base conversion (in all other cases if you don't have decimal you must have decimal, here it's the other way around)
 
 // on the todo list is to have a trait for get_primitive that is implemented by each float formatter and can override a default. when that happens we can take the parts of get_primitive_dec specific to dec and spin them out to their own functions that can be overridden.
-#[allow(unused_variables)]
-#[allow(unused_assignments)]
 fn get_primitive_hex(
     inprefix: &InPrefix,
-    str_in: &str,
+    _str_in: &str,
-    analysis: &FloatAnalysis,
+    _analysis: &FloatAnalysis,
-    last_dec_place: usize,
+    _last_dec_place: usize,
     capitalized: bool,
 ) -> FormatPrimitive {
     let prefix = Some(String::from(if inprefix.sign == -1 { "-0x" } else { "0x" }));
 
-    // assign the digits before and after the decimal points
-    // to separate slices. If no digits after decimal point,
-    // assign 0
-    let (mut first_segment_raw, second_segment_raw) = match analysis.decimal_pos {
-        Some(pos) => (&str_in[..pos], &str_in[pos + 1..]),
-        None => (str_in, "0"),
-    };
-    if first_segment_raw.is_empty() {
-        first_segment_raw = "0";
-    }
-    // convert to string, hexifying if input is in dec.
-    // let (first_segment, second_segment) =
-    // match inprefix.radix_in {
-    // Base::Ten => {
-    // (to_hex(first_segment_raw, true),
-    // to_hex(second_segment_raw, false))
-    // }
-    // _ => {
-    // (String::from(first_segment_raw),
-    // String::from(second_segment_raw))
-    // }
-    // };
-    //
-    //
-    // f.pre_decimal = Some(first_segment);
-    // f.post_decimal = Some(second_segment);
-    //
-
     // TODO actual conversion, make sure to get back mantissa.
     // for hex to hex, it's really just a matter of moving the
     // decimal point and calculating the mantissa by its initial

src/uu/printf/src/tokenize/num_format/formatters/decf.rs

@@ -22,12 +22,11 @@ fn get_len_fprim(fprim: &FormatPrimitive) -> usize {
     len
 }
 
-pub struct Decf {
+pub struct Decf;
-    as_num: f64,
+
-}
 impl Decf {
     pub fn new() -> Decf {
-        Decf { as_num: 0.0 }
+        Decf
     }
 }
 impl Formatter for Decf {

src/uu/printf/src/tokenize/num_format/formatters/floatf.rs

@@ -5,12 +5,10 @@ use super::super::format_field::FormatField;
 use super::super::formatter::{FormatPrimitive, Formatter, InPrefix};
 use super::float_common::{get_primitive_dec, primitive_to_str_common, FloatAnalysis};
 
-pub struct Floatf {
+pub struct Floatf;
-    as_num: f64,
-}
 impl Floatf {
     pub fn new() -> Floatf {
-        Floatf { as_num: 0.0 }
+        Floatf
     }
 }
 impl Formatter for Floatf {

src/uu/printf/src/tokenize/num_format/formatters/intf.rs

@@ -11,7 +11,7 @@ use std::i64;
 use std::u64;
 
 pub struct Intf {
-    a: u32,
+    _a: u32,
 }
 
 // see the Intf::analyze() function below
@@ -24,7 +24,7 @@ struct IntAnalysis {
 
 impl Intf {
     pub fn new() -> Intf {
-        Intf { a: 0 }
+        Intf { _a: 0 }
     }
     // take a ref to argument string, and basic information
     // about prefix (offset, radix, sign), and analyze string

src/uu/printf/src/tokenize/num_format/formatters/scif.rs

@@ -5,12 +5,11 @@ use super::super::format_field::FormatField;
 use super::super::formatter::{FormatPrimitive, Formatter, InPrefix};
 use super::float_common::{get_primitive_dec, primitive_to_str_common, FloatAnalysis};
 
-pub struct Scif {
+pub struct Scif;
-    as_num: f64,
+
-}
 impl Scif {
     pub fn new() -> Scif {
-        Scif { as_num: 0.0 }
+        Scif
     }
 }
 impl Formatter for Scif {

src/uu/printf/src/tokenize/unescaped_text.rs

@@ -242,18 +242,16 @@ impl UnescapedText {
         }
     }
 }
-#[allow(unused_variables)]
 impl token::Tokenizer for UnescapedText {
     fn from_it(
         it: &mut PutBackN<Chars>,
-        args: &mut Peekable<Iter<String>>,
+        _: &mut Peekable<Iter<String>>,
     ) -> Option<Box<dyn token::Token>> {
         UnescapedText::from_it_core(it, false)
     }
 }
-#[allow(unused_variables)]
 impl token::Token for UnescapedText {
-    fn print(&self, pf_args_it: &mut Peekable<Iter<String>>) {
+    fn print(&self, _: &mut Peekable<Iter<String>>) {
         cli::flush_bytes(&self.0[..]);
     }
 }

src/uu/sort/Cargo.toml

@@ -23,7 +23,8 @@ clap = "2.33"
 fnv = "1.0.7"
 itertools = "0.10.0"
 semver = "0.9.0"
-smallvec = { version = "1.6.1", features = ["serde"] }
+smallvec = "1.6.1"
+unicode-width = "0.1.8"
 uucore = { version=">=0.0.8", package="uucore", path="../../uucore", features=["fs"] }
 uucore_procs = { version=">=0.0.5", package="uucore_procs", path="../../uucore_procs" }
 tempdir = "0.3.7"

src/uu/sort/src/numeric_str_cmp.rs

@@ -138,7 +138,15 @@ impl NumInfo {
                     sign: if had_digit { sign } else { Sign::Positive },
                     exponent: 0,
                 },
-                0..0,
+                if had_digit {
+                    // In this case there were only zeroes.
+                    // For debug output to work properly, we have to claim to match the end of the number.
+                    num.len()..num.len()
+                } else {
+                    // This was no number at all.
+                    // For debug output to work properly, we have to claim to match the start of the number.
+                    0..0
+                },
             )
         }
     }

src/uu/sort/src/sort.rs

@@ -29,7 +29,6 @@ use rayon::prelude::*;
 use semver::Version;
 use serde::{Deserializer, Deserialize, Serialize};
 use smallvec::SmallVec;
-use std::borrow::Cow;
 use std::cmp::Ordering;
 use std::collections::BinaryHeap;
 use std::env;
@@ -38,9 +37,10 @@ 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::ops::Range;
-use std::path::{Path, PathBuf};
+use std::path::Path;
-use uucore::fs::is_stdin_interactive; // for Iterator::dedup();
+use unicode_width::UnicodeWidthStr;
+use uucore::fs::is_stdin_interactive; // for Iterator::dedup()
 
 static NAME: &str = "sort";
 static ABOUT: &str = "Display sorted concatenation of all FILE(s).";
@@ -66,6 +66,7 @@ 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_DEBUG: &str = "debug";
 static OPT_IGNORE_CASE: &str = "ignore-case";
 static OPT_IGNORE_BLANKS: &str = "ignore-blanks";
 static OPT_IGNORE_NONPRINTING: &str = "ignore-nonprinting";
@@ -94,7 +95,7 @@ static DEFAULT_TMPDIR: &str = r"/tmp";
 // 16GB buffer for Vec<Line> before we dump to disk, never used
 static DEFAULT_BUF_SIZE: usize = 16000000000;
 
-#[derive(Eq, Ord, PartialEq, PartialOrd, Clone)]
+#[derive(Eq, Ord, PartialEq, PartialOrd, Clone, Copy)]
 enum SortMode {
     Numeric,
     HumanNumeric,
@@ -106,6 +107,7 @@ enum SortMode {
 #[derive(Clone)]
 struct GlobalSettings {
     mode: SortMode,
+    debug: bool,
     ignore_blanks: bool,
     ignore_case: bool,
     dictionary_order: bool,
@@ -154,6 +156,7 @@ impl Default for GlobalSettings {
     fn default() -> GlobalSettings {
         GlobalSettings {
             mode: SortMode::Default,
+            debug: false,
             ignore_blanks: false,
             ignore_case: false,
             dictionary_order: false,
@@ -190,7 +193,7 @@ struct KeySettings {
 impl From<&GlobalSettings> for KeySettings {
     fn from(settings: &GlobalSettings) -> Self {
         Self {
-            mode: settings.mode.clone(),
+            mode: settings.mode,
             ignore_blanks: settings.ignore_blanks,
             ignore_case: settings.ignore_case,
             ignore_non_printing: settings.ignore_non_printing,
@@ -236,7 +239,7 @@ impl SelectionRange {
 #[derive(Debug, Serialize, Deserialize, Clone)]
 enum NumCache {
     #[serde(deserialize_with="bailout_parse_f64")]
-    AsF64(f64),
+    AsF64(GeneralF64ParseResult),
     WithInfo(NumInfo),
     None,
 }
@@ -250,7 +253,7 @@ fn bailout_parse_f64<'de, D>(d: D) -> Result<f64, D::Error> where D: Deserialize
 }
 
 impl NumCache {
-    fn as_f64(&self) -> f64 {
+    fn as_f64(&self) -> GeneralF64ParseResult {
         match self {
             NumCache::AsF64(n) => *n,
             _ => unreachable!(),
@@ -309,18 +312,13 @@ impl Line {
             .selectors
             .iter()
             .map(|selector| {
-                let mut range =
+                let range = selector.get_selection(&line, fields.as_deref());
-                    if let Some(range) = selector.get_field_selection(&line, fields.as_deref()) {
+                let mut range = if let Some(transformed) =
-                        if let Some(transformed) =
                     transform(&line[range.to_owned()], &selector.settings)
                 {
                     SelectionRange::String(transformed)
                 } else {
-                            SelectionRange::ByIndex(range.start().to_owned()..range.end() + 1)
+                    SelectionRange::ByIndex(range)
-                        }
-                    } else {
-                        // If there is no match, match the empty string.
-                        SelectionRange::ByIndex(0..0)
                 };
                 let num_cache = if selector.settings.mode == SortMode::Numeric
                     || selector.settings.mode == SortMode::HumanNumeric
@@ -336,7 +334,8 @@ impl Line {
                     range.shorten(num_range);
                     NumCache::WithInfo(info)
                 } else if selector.settings.mode == SortMode::GeneralNumeric && settings.buffer_size == DEFAULT_BUF_SIZE {
-                    NumCache::AsF64(permissive_f64_parse(get_leading_gen(range.get_str(&line))))
+                    let str = range.get_str(&line);
+                    NumCache::AsF64(general_f64_parse(&str[get_leading_gen(str)]))
                 } else {
                     NumCache::None
                 };
@@ -345,6 +344,129 @@ impl Line {
             .collect();
         Self { line, selections }
     }
+
+    /// Writes indicators for the selections this line matched. The original line content is NOT expected
+    /// to be already printed.
+    fn print_debug(
+        &self,
+        settings: &GlobalSettings,
+        writer: &mut dyn Write,
+    ) -> std::io::Result<()> {
+        // We do not consider this function performance critical, as debug output is only useful for small files,
+        // which are not a performance problem in any case. Therefore there aren't any special performance
+        // optimizations here.
+
+        let line = self.line.replace('\t', ">");
+        writeln!(writer, "{}", line)?;
+
+        let fields = tokenize(&self.line, settings.separator);
+        for selector in settings.selectors.iter() {
+            let mut selection = selector.get_selection(&self.line, Some(&fields));
+            match selector.settings.mode {
+                SortMode::Numeric | SortMode::HumanNumeric => {
+                    // find out which range is used for numeric comparisons
+                    let (_, num_range) = NumInfo::parse(
+                        &self.line[selection.clone()],
+                        NumInfoParseSettings {
+                            accept_si_units: selector.settings.mode == SortMode::HumanNumeric,
+                            thousands_separator: Some(THOUSANDS_SEP),
+                            decimal_pt: Some(DECIMAL_PT),
+                        },
+                    );
+                    let initial_selection = selection.clone();
+
+                    // Shorten selection to num_range.
+                    selection.start += num_range.start;
+                    selection.end = selection.start + num_range.len();
+
+                    // include a trailing si unit
+                    if selector.settings.mode == SortMode::HumanNumeric
+                        && self.line[selection.end..initial_selection.end]
+                            .starts_with(&['k', 'K', 'M', 'G', 'T', 'P', 'E', 'Z', 'Y'][..])
+                    {
+                        selection.end += 1;
+                    }
+
+                    // include leading zeroes, a leading minus or a leading decimal point
+                    while self.line[initial_selection.start..selection.start]
+                        .ends_with(&['-', '0', '.'][..])
+                    {
+                        selection.start -= 1;
+                    }
+                }
+                SortMode::GeneralNumeric => {
+                    let initial_selection = &self.line[selection.clone()];
+
+                    let leading = get_leading_gen(initial_selection);
+
+                    // Shorten selection to leading.
+                    selection.start += leading.start;
+                    selection.end = selection.start + leading.len();
+                }
+                SortMode::Month => {
+                    let initial_selection = &self.line[selection.clone()];
+
+                    let month = if month_parse(initial_selection) == Month::Unknown {
+                        // We failed to parse a month, which is equivalent to matching nothing.
+                        0..0
+                    } else {
+                        // We parsed a month. Match the three first non-whitespace characters, which must be the month we parsed.
+                        let mut chars = initial_selection
+                            .char_indices()
+                            .skip_while(|(_, c)| c.is_whitespace());
+                        chars.next().unwrap().0
+                            ..chars.nth(2).map_or(initial_selection.len(), |(idx, _)| idx)
+                    };
+
+                    // Shorten selection to month.
+                    selection.start += month.start;
+                    selection.end = selection.start + month.len();
+                }
+                _ => {}
+            }
+
+            write!(
+                writer,
+                "{}",
+                " ".repeat(UnicodeWidthStr::width(&line[..selection.start]))
+            )?;
+
+            // TODO: Once our minimum supported rust version is at least 1.47, use selection.is_empty() instead.
+            #[allow(clippy::len_zero)]
+            {
+                if selection.len() == 0 {
+                    writeln!(writer, "^ no match for key")?;
+                } else {
+                    writeln!(
+                        writer,
+                        "{}",
+                        "_".repeat(UnicodeWidthStr::width(&line[selection]))
+                    )?;
+                }
+            }
+        }
+        if !(settings.random
+            || settings.stable
+            || settings.unique
+            || !(settings.dictionary_order
+                || settings.ignore_blanks
+                || settings.ignore_case
+                || settings.ignore_non_printing
+                || settings.mode != SortMode::Default))
+        {
+            // A last resort comparator is in use, underline the whole line.
+            if self.line.is_empty() {
+                writeln!(writer, "^ no match for key")?;
+            } else {
+                writeln!(
+                    writer,
+                    "{}",
+                    "_".repeat(UnicodeWidthStr::width(line.as_str()))
+                )?;
+            }
+        }
+        Ok(())
+    }
 }
 
 /// Transform this line. Returns None if there's no need to transform.
@@ -407,20 +529,18 @@ fn tokenize_default(line: &str) -> Vec<Field> {
 
 /// 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 tokens = vec![];
-    let mut previous_was_separator = false;
+    let separator_indices =
-    for (idx, char) in line.char_indices() {
+        line.char_indices()
-        if previous_was_separator {
+            .filter_map(|(i, c)| if c == separator { Some(i) } else { None });
-            tokens.push(idx..0);
+    let mut start = 0;
+    for sep_idx in separator_indices {
+        tokens.push(start..sep_idx);
+        start = sep_idx + 1;
     }
-        if char == separator {
+    if start < line.len() {
-            tokens.last_mut().unwrap().end = idx;
+        tokens.push(start..line.len());
-            previous_was_separator = true;
-        } else {
-            previous_was_separator = false;
     }
-    }
-    tokens.last_mut().unwrap().end = line.len();
     tokens
 }
 
@@ -464,6 +584,28 @@ impl KeyPosition {
                         crash!(1, "invalid option for key: `{}`", c)
                     }
                 }
+                // All numeric sorts and month sort conflict with dictionary_order and ignore_non_printing.
+                // Instad of reporting an error, let them overwrite each other.
+
+                // FIXME: This should only override if the overridden flag is a global flag.
+                // If conflicting flags are attached to the key, GNU sort crashes and we should probably too.
+                match option {
+                    'h' | 'n' | 'g' | 'M' => {
+                        settings.dictionary_order = false;
+                        settings.ignore_non_printing = false;
+                    }
+                    'd' | 'i' => {
+                        settings.mode = match settings.mode {
+                            SortMode::Numeric
+                            | SortMode::HumanNumeric
+                            | SortMode::GeneralNumeric
+                            | SortMode::Month => SortMode::Default,
+                            // Only SortMode::Default and SortMode::Version work with dictionary_order and ignore_non_printing
+                            m @ SortMode::Default | m @ SortMode::Version => m,
+                        }
+                    }
+                    _ => {}
+                }
             }
             // Strip away option characters from the original value so we can parse it later
             *value_with_options = &value_with_options[..options_start];
@@ -506,13 +648,16 @@ impl FieldSelector {
 
     /// Look up the slice that corresponds to this selector for the given line.
     /// If needs_fields returned false, fields may be None.
-    fn get_field_selection<'a>(
+    fn get_selection<'a>(&self, line: &'a str, tokens: Option<&[Field]>) -> Range<usize> {
-        &self,
+        enum Resolution {
-        line: &'a str,
+            // The start index of the resolved character, inclusive
-        tokens: Option<&[Field]>,
+            StartOfChar(usize),
-    ) -> Option<RangeInclusive<usize>> {
+            // The end index of the resolved character, exclusive.
-        enum ResolutionErr {
+            // This is only returned if the character index is 0.
+            EndOfChar(usize),
+            // The resolved character would be in front of the first character
             TooLow,
+            // The resolved character would be after the last character
             TooHigh,
         }
 
@@ -521,15 +666,15 @@ impl FieldSelector {
             line: &str,
             tokens: Option<&[Field]>,
             position: &KeyPosition,
-        ) -> Result<usize, ResolutionErr> {
+        ) -> Resolution {
             if tokens.map_or(false, |fields| fields.len() < position.field) {
-                Err(ResolutionErr::TooHigh)
+                Resolution::TooHigh
             } else if position.char == 0 {
                 let end = tokens.unwrap()[position.field - 1].end;
                 if end == 0 {
-                    Err(ResolutionErr::TooLow)
+                    Resolution::TooLow
                 } else {
-                    Ok(end - 1)
+                    Resolution::EndOfChar(end)
                 }
             } else {
                 let mut idx = if position.field == 1 {
@@ -538,38 +683,52 @@ impl FieldSelector {
                     0
                 } else {
                     tokens.unwrap()[position.field - 1].start
-                } + position.char
+                };
-                    - 1;
+                idx += line[idx..]
+                    .char_indices()
+                    .nth(position.char - 1)
+                    .map_or(line.len(), |(idx, _)| idx);
                 if idx >= line.len() {
-                    Err(ResolutionErr::TooHigh)
+                    Resolution::TooHigh
                 } else {
                     if position.ignore_blanks {
-                        if let Some(not_whitespace) =
+                        if let Some((not_whitespace, _)) =
-                            line[idx..].chars().position(|c| !c.is_whitespace())
+                            line[idx..].char_indices().find(|(_, c)| !c.is_whitespace())
                         {
                             idx += not_whitespace;
                         } else {
-                            return Err(ResolutionErr::TooHigh);
+                            return Resolution::TooHigh;
                         }
                     }
-                    Ok(idx)
+                    Resolution::StartOfChar(idx)
                 }
             }
         }
 
-        if let Ok(from) = resolve_index(line, tokens, &self.from) {
+        match resolve_index(line, tokens, &self.from) {
+            Resolution::StartOfChar(from) => {
                 let to = self.to.as_ref().map(|to| resolve_index(line, tokens, &to));
+
                 match to {
-                Some(Ok(to)) => Some(from..=to),
+                    Some(Resolution::StartOfChar(mut to)) => {
+                        to += line[to..].chars().next().map_or(1, |c| c.len_utf8());
+                        from..to
+                    }
+                    Some(Resolution::EndOfChar(to)) => 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),
+                    None | Some(Resolution::TooHigh) => from..line.len(),
                     // 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,
+                    Some(Resolution::TooLow) => 0..0,
                 }
-        } else {
+            }
-            None
+            Resolution::TooLow | Resolution::EndOfChar(_) => {
+                unreachable!("This should only happen if the field start index is 0, but that should already have caused an error.")
+            }
+            // While for comparisons it's only important that this is an empty slice,
+            // to produce accurate debug output we need to match an empty slice at the end of the line.
+            Resolution::TooHigh => line.len()..line.len(),
         }
     }
 }
@@ -597,7 +756,7 @@ 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 == self.cmp(other)
     }
 }
 
@@ -628,8 +787,8 @@ impl<'a> FileMerger<'a> {
 }
 
 impl<'a> Iterator for FileMerger<'a> {
-    type Item = String;
+    type Item = Line;
-    fn next(&mut self) -> Option<String> {
+    fn next(&mut self) -> Option<Line> {
         match self.heap.pop() {
             Some(mut current) => {
                 match current.lines.next() {
@@ -639,12 +798,12 @@ impl<'a> Iterator for FileMerger<'a> {
                             Line::new(next_line, &self.settings),
                         );
                         self.heap.push(current);
-                        Some(ret.line)
+                        Some(ret)
                     }
                     _ => {
                         // Don't put it back in the heap (it's empty/erroring)
                         // but its first line is still valid.
-                        Some(current.current_line.line)
+                        Some(current.current_line)
                     }
                 }
             }
@@ -708,7 +867,8 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
             Arg::with_name(OPT_DICTIONARY_ORDER)
                 .short("d")
                 .long(OPT_DICTIONARY_ORDER)
-                .help("consider only blanks and alphanumeric characters"),
+                .help("consider only blanks and alphanumeric characters")
+                .conflicts_with_all(&[OPT_NUMERIC_SORT, OPT_GENERAL_NUMERIC_SORT, OPT_HUMAN_NUMERIC_SORT, OPT_MONTH_SORT]),
         )
         .arg(
             Arg::with_name(OPT_MERGE)
@@ -736,9 +896,10 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
         )
         .arg(
             Arg::with_name(OPT_IGNORE_NONPRINTING)
-                .short("-i")
+                .short("i")
                 .long(OPT_IGNORE_NONPRINTING)
-                .help("ignore nonprinting characters"),
+                .help("ignore nonprinting characters")
+                .conflicts_with_all(&[OPT_NUMERIC_SORT, OPT_GENERAL_NUMERIC_SORT, OPT_HUMAN_NUMERIC_SORT, OPT_MONTH_SORT]),
         )
         .arg(
             Arg::with_name(OPT_IGNORE_BLANKS)
@@ -829,9 +990,16 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
                 .value_name("NUL_FILES")
                 .multiple(true),
         )
+        .arg(
+            Arg::with_name(OPT_DEBUG)
+                .long(OPT_DEBUG)
+                .help("underline the parts of the line that are actually used for sorting"),
+        )
         .arg(Arg::with_name(ARG_FILES).multiple(true).takes_value(true))
         .get_matches_from(args);
 
+    settings.debug = matches.is_present(OPT_DEBUG);
+
     // 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
@@ -964,6 +1132,13 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
                 1,
                 &mut key_settings,
             );
+            if from.char == 0 {
+                crash!(
+                    1,
+                    "invalid character index 0 in `{}` for the start position of a field",
+                    key
+                )
+            }
             let to = from_to
                 .next()
                 .map(|to| KeyPosition::parse(to, 0, &mut key_settings));
@@ -1042,7 +1217,10 @@ fn exec(files: Vec<String>, settings: GlobalSettings) -> i32 {
 
     if settings.merge {
         if settings.unique {
-            print_sorted(file_merger.dedup(), &settings)
+            print_sorted(
+                file_merger.dedup_by(|a, b| compare_by(a, b, settings) == Ordering::Equal),
+                &settings,
+            )
         } else {
             print_sorted(file_merger, &settings)
         }
@@ -1050,12 +1228,11 @@ fn exec(files: Vec<String>, settings: GlobalSettings) -> i32 {
         print_sorted(
             lines
                 .into_iter()
-                .dedup_by(|a, b| compare_by(a, b, &settings) == Ordering::Equal)
+                .dedup_by(|a, b| compare_by(a, b, settings) == Ordering::Equal),
-                .map(|line| line.line),
             &settings,
         )
     } else {
-        print_sorted(lines.into_iter().map(|line| line.line), &settings)
+        print_sorted(lines.into_iter(), &settings)
     }
 
     0
@@ -1163,107 +1340,80 @@ fn default_compare(a: &str, b: &str) -> Ordering {
     a.cmp(b)
 }
 
-/// This function does the initial detection of numeric lines for FP compares.
+// This function cleans up the initial comparison done by leading_num_common for a general numeric compare.
-// 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 = "";
-
-    // check whether char is numeric, whitespace or decimal point or thousand separator
-    for (idx, c) in a.char_indices() {
-        if !c.is_numeric()
-            && !c.is_whitespace()
-            && !c.eq(&THOUSANDS_SEP)
-            && !c.eq(&DECIMAL_PT)
-            // check for e notation
-            && !c.eq(&'e')
-            && !c.eq(&'E')
-            // check whether first char is + or - 
-            && !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];
-            break;
-        }
-        // If line is not a number line, return the line as is
-        s = &a;
-    }
-    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) -> &str {
+fn get_leading_gen(input: &str) -> Range<usize> {
+    let trimmed = input.trim_start();
+    let leading_whitespace_len = input.len() - trimmed.len();
+    for allowed_prefix in &["inf", "-inf", "nan"] {
+        if trimmed.is_char_boundary(allowed_prefix.len())
+            && trimmed[..allowed_prefix.len()].eq_ignore_ascii_case(allowed_prefix)
+        {
+            return leading_whitespace_len..(leading_whitespace_len + allowed_prefix.len());
+        }
+    }
     // Make this iter peekable to see if next char is numeric
-    let raw_leading_num = leading_num_common(a);
+    let mut char_indices = trimmed.char_indices().peekable();
-    let mut p_iter = raw_leading_num.chars().peekable();
+
-    let mut result = "";
+    let first = char_indices.peek();
-    // Cleanup raw stripped strings
+
-    while let Some(c) = p_iter.next() {
+    if first.map_or(false, |&(_, c)| c == NEGATIVE || c == POSITIVE) {
-        let next_char_numeric = p_iter.peek().unwrap_or(&'\0').is_numeric();
+        char_indices.next();
-        // 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
-        // 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 {
-            result = a.trim().trim_start_matches('+');
-            break;
     }
-        // If no further processing needed to be done, return the line as-is to be sorted
+
-        result = a;
+    let mut had_e_notation = false;
+    let mut had_decimal_pt = false;
+    while let Some((idx, c)) = char_indices.next() {
+        if c.is_ascii_digit() {
+            continue;
         }
-    result
+        if c == DECIMAL_PT && !had_decimal_pt {
+            had_decimal_pt = true;
+            continue;
+        }
+        let next_char_numeric = char_indices
+            .peek()
+            .map_or(false, |(_, c)| c.is_ascii_digit());
+        if (c == 'e' || c == 'E') && !had_e_notation && next_char_numeric {
+            had_e_notation = true;
+            continue;
+        }
+        return leading_whitespace_len..(leading_whitespace_len + idx);
+    }
+    leading_whitespace_len..input.len()
 }
 
-#[inline(always)]
+#[derive(Copy, Clone, PartialEq, PartialOrd)]
-fn remove_trailing_dec<'a, S: Into<Cow<'a, str>>>(input: S) -> Cow<'a, str> {
+enum GeneralF64ParseResult {
-    let input = input.into();
+    Invalid,
-    if let Some(s) = input.find(DECIMAL_PT) {
+    NaN,
-        let (leading, trailing) = input.split_at(s);
+    NegInfinity,
-        let output = [leading, ".", trailing.replace(DECIMAL_PT, "").as_str()].concat();
+    Number(f64),
-        Cow::Owned(output)
+    Infinity,
-    } 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 {
+fn general_f64_parse(a: &str) -> GeneralF64ParseResult {
-    // GNU sort treats "NaN" as non-number in numeric, so it needs special care.
+    // The actual behavior here relies on Rust's implementation of parsing floating points.
-    // *Keep this trim before parse* despite what POSIX may say about -b and -n
+    // For example "nan", "inf" (ignoring the case) and "infinity" are only parsed to floats starting from 1.53.
-    // because GNU and BSD both seem to require it to match their behavior
+    // TODO: Once our minimum supported Rust version is 1.53 or above, we should add tests for those cases.
-    //
+    match a.parse::<f64>() {
-    // Remove any trailing decimals, ie 4568..890... becomes 4568.890
+        Ok(a) if a.is_nan() => GeneralF64ParseResult::NaN,
-    // Then, we trim whitespace and parse
+        Ok(a) if a == std::f64::NEG_INFINITY => GeneralF64ParseResult::NegInfinity,
-    match remove_trailing_dec(a).trim().parse::<f64>() {
+        Ok(a) if a == std::f64::INFINITY => GeneralF64ParseResult::Infinity,
-        Ok(val) if val.is_nan() => std::f64::NEG_INFINITY,
+        Ok(a) => GeneralF64ParseResult::Number(a),
-        Ok(val) => val,
+        Err(_) => GeneralF64ParseResult::Invalid,
-        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.
 /// We explicitly need to convert to f64 in this case.
-fn general_numeric_compare(a: f64, b: f64) -> Ordering {
+fn general_numeric_compare(a: GeneralF64ParseResult, b: GeneralF64ParseResult) -> Ordering {
-    #![allow(clippy::comparison_chain)]
+    a.partial_cmp(&b).unwrap()
-    // f64::cmp isn't implemented (due to NaN issues); implement directly instead
-    if a > b {
-        Ordering::Greater
-    } else if a < b {
-        Ordering::Less
-    } else {
-        Ordering::Equal
-    }
 }
 
 fn get_rand_string() -> String {
@@ -1290,7 +1440,7 @@ fn random_shuffle(a: &str, b: &str, x: String) -> Ordering {
     da.cmp(&db)
 }
 
-#[derive(Eq, Ord, PartialEq, PartialOrd)]
+#[derive(Eq, Ord, PartialEq, PartialOrd, Clone, Copy)]
 enum Month {
     Unknown,
     January,
@@ -1309,30 +1459,32 @@ enum Month {
 
 /// Parse the beginning string into a Month, returning Month::Unknown on errors.
 fn month_parse(line: &str) -> Month {
-    // GNU splits at any 3 letter match "JUNNNN" is JUN
+    let line = line.trim();
-    let pattern = if line.trim().len().ge(&3) {
-        // Split a 3 and get first element of tuple ".0"
-        line.trim().split_at(3).0
-    } else {
-        ""
-    };
 
-    let result = match pattern.to_uppercase().as_ref() {
+    const MONTHS: [(&str, Month); 12] = [
-        "JAN" => Month::January,
+        ("JAN", Month::January),
-        "FEB" => Month::February,
+        ("FEB", Month::February),
-        "MAR" => Month::March,
+        ("MAR", Month::March),
-        "APR" => Month::April,
+        ("APR", Month::April),
-        "MAY" => Month::May,
+        ("MAY", Month::May),
-        "JUN" => Month::June,
+        ("JUN", Month::June),
-        "JUL" => Month::July,
+        ("JUL", Month::July),
-        "AUG" => Month::August,
+        ("AUG", Month::August),
-        "SEP" => Month::September,
+        ("SEP", Month::September),
-        "OCT" => Month::October,
+        ("OCT", Month::October),
-        "NOV" => Month::November,
+        ("NOV", Month::November),
-        "DEC" => Month::December,
+        ("DEC", Month::December),
-        _ => Month::Unknown,
+    ];
-    };
+
-    result
+    for (month_str, month) in &MONTHS {
+        if line.is_char_boundary(month_str.len())
+            && line[..month_str.len()].eq_ignore_ascii_case(month_str)
+        {
+            return *month;
+        }
+    }
+
+    Month::Unknown
 }
 
 fn month_compare(a: &str, b: &str) -> Ordering {
@@ -1390,7 +1542,7 @@ fn remove_nonprinting_chars(s: &str) -> String {
         .collect::<String>()
 }
 
-fn print_sorted<T: Iterator<Item = String>>(iter: T, settings: &GlobalSettings) {
+fn print_sorted<T: Iterator<Item = Line>>(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>,
@@ -1401,15 +1553,19 @@ fn print_sorted<T: Iterator<Item = String>>(iter: T, settings: &GlobalSettings)
         },
         None => Box::new(BufWriter::new(stdout())) as Box<dyn Write>,
     };
-    if settings.zero_terminated {
+    if settings.zero_terminated && !settings.debug {
         for line in iter {
-            crash_if_err!(1, file.write_all(line.as_bytes()));
+            crash_if_err!(1, file.write_all(line.line.as_bytes()));
             crash_if_err!(1, file.write_all("\0".as_bytes()));
         }
     } else {
         for line in iter {
-            crash_if_err!(1, file.write_all(line.as_bytes()));
+            if !settings.debug {
+                crash_if_err!(1, file.write_all(line.line.as_bytes()));
                 crash_if_err!(1, file.write_all("\n".as_bytes()));
+            } else {
+                crash_if_err!(1, line.print_debug(settings, &mut file));
+            }
         }
     }
     crash_if_err!(1, file.flush());
@@ -1504,4 +1660,14 @@ mod tests {
             vec![0..0, 1..1, 2..2, 3..9, 10..18,]
         );
     }
+
+    #[test]
+    fn test_tokenize_fields_trailing_custom_separator() {
+        let line = "a";
+        assert_eq!(tokenize(line, Some('a')), vec![0..0]);
+        let line = "aa";
+        assert_eq!(tokenize(line, Some('a')), vec![0..0, 1..1]);
+        let line = "..a..a";
+        assert_eq!(tokenize(line, Some('a')), vec![0..2, 3..5]);
+    }
 }

src/uu/tac/src/tac.rs

@@ -91,10 +91,15 @@ fn tac(filenames: Vec<String>, before: bool, _: bool, separator: &str) -> i32 {
         } else {
             let path = Path::new(filename);
             if path.is_dir() || path.metadata().is_err() {
+                if path.is_dir() {
+                    show_error!("dir: read error: Invalid argument");
+                } else {
                     show_error!(
                         "failed to open '{}' for reading: No such file or directory",
                         filename
                     );
+                }
+                exit_code = 1;
                 continue;
             }
             match File::open(path) {

src/uu/tail/src/tail.rs

@@ -117,6 +117,7 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
         .arg(
             Arg::with_name(options::SLEEP_INT)
                 .short("s")
+                .takes_value(true)
                 .long(options::SLEEP_INT)
                 .help("Number or seconds to sleep between polling the file when running with -f"),
         )

src/uu/truncate/src/truncate.rs

@@ -192,7 +192,8 @@ fn truncate(
 }
 
 fn parse_size(size: &str) -> (u64, TruncateMode) {
-    let mode = match size.chars().next().unwrap() {
+    let clean_size = size.replace(" ", "");
+    let mode = match clean_size.chars().next().unwrap() {
         '+' => TruncateMode::Extend,
         '-' => TruncateMode::Reduce,
         '<' => TruncateMode::AtMost,
@@ -203,9 +204,9 @@ fn parse_size(size: &str) -> (u64, TruncateMode) {
     };
     let bytes = {
         let mut slice = if mode == TruncateMode::Reference {
-            size
+            &clean_size
         } else {
-            &size[1..]
+            &clean_size[1..]
         };
         if slice.chars().last().unwrap().is_alphabetic() {
             slice = &slice[..slice.len() - 1];
@@ -220,11 +221,11 @@ fn parse_size(size: &str) -> (u64, TruncateMode) {
         Ok(num) => num,
         Err(e) => crash!(1, "'{}' is not a valid number: {}", size, e),
     };
-    if size.chars().last().unwrap().is_alphabetic() {
+    if clean_size.chars().last().unwrap().is_alphabetic() {
-        number *= match size.chars().last().unwrap().to_ascii_uppercase() {
+        number *= match clean_size.chars().last().unwrap().to_ascii_uppercase() {
-            'B' => match size
+            'B' => match clean_size
                 .chars()
-                .nth(size.len() - 2)
+                .nth(clean_size.len() - 2)
                 .unwrap()
                 .to_ascii_uppercase()
             {

src/uu/uniq/src/uniq.rs

@@ -61,34 +61,43 @@ impl Uniq {
         reader: &mut BufReader<R>,
         writer: &mut BufWriter<W>,
     ) {
-        let mut lines: Vec<String> = vec![];
         let mut first_line_printed = false;
-        let delimiters = self.delimiters;
+        let mut group_count = 1;
         let line_terminator = self.get_line_terminator();
-        // Don't print any delimiting lines before, after or between groups if delimiting method is 'none'
+        let mut lines = reader.split(line_terminator).map(get_line_string);
-        let no_delimiters = delimiters == Delimiters::None;
+        let mut line = match lines.next() {
-        // The 'prepend' and 'both' delimit methods will cause output to start with delimiter line
+            Some(l) => l,
-        let prepend_delimiter = delimiters == Delimiters::Prepend || delimiters == Delimiters::Both;
+            None => return,
-        // The 'append' and 'both' delimit methods will cause output to end with delimiter line
+        };
-        let append_delimiter = delimiters == Delimiters::Append || delimiters == Delimiters::Both;
 
-        for line in reader.split(line_terminator).map(get_line_string) {
+        // compare current `line` with consecutive lines (`next_line`) of the input
-            if !lines.is_empty() && self.cmp_keys(&lines[0], &line) {
+        // and if needed, print `line` based on the command line options provided
-                // Print delimiter if delimit method is not 'none' and any line has been output
+        for next_line in lines {
-                // before or if we need to start output with delimiter
+            if self.cmp_keys(&line, &next_line) {
-                let print_delimiter = !no_delimiters && (prepend_delimiter || first_line_printed);
+                if (group_count == 1 && !self.repeats_only)
-                first_line_printed |= self.print_lines(writer, &lines, print_delimiter);
+                    || (group_count > 1 && !self.uniques_only)
-                lines.truncate(0);
+                {
+                    self.print_line(writer, &line, group_count, first_line_printed);
+                    first_line_printed = true;
                 }
-            lines.push(line);
+                line = next_line;
+                group_count = 1;
+            } else {
+                if self.all_repeated {
+                    self.print_line(writer, &line, group_count, first_line_printed);
+                    first_line_printed = true;
+                    line = next_line;
                 }
-        if !lines.is_empty() {
+                group_count += 1;
-            // Print delimiter if delimit method is not 'none' and any line has been output
-            // before or if we need to start output with delimiter
-            let print_delimiter = !no_delimiters && (prepend_delimiter || first_line_printed);
-            first_line_printed |= self.print_lines(writer, &lines, print_delimiter);
             }
-        if append_delimiter && first_line_printed {
+        }
+        if (group_count == 1 && !self.repeats_only) || (group_count > 1 && !self.uniques_only) {
+            self.print_line(writer, &line, group_count, first_line_printed);
+            first_line_printed = true;
+        }
+        if (self.delimiters == Delimiters::Append || self.delimiters == Delimiters::Both)
+            && first_line_printed
+        {
             crash_if_err!(1, writer.write_all(&[line_terminator]));
         }
     }
@@ -163,27 +172,17 @@ impl Uniq {
         }
     }
 
-    fn print_lines<W: Write>(
+    fn should_print_delimiter(&self, group_count: usize, first_line_printed: bool) -> bool {
-        &self,
+        // if no delimiter option is selected then no other checks needed
-        writer: &mut BufWriter<W>,
+        self.delimiters != Delimiters::None
-        lines: &[String],
+            // print delimiter only before the first line of a group, not between lines of a group
-        print_delimiter: bool,
+            && group_count == 1
-    ) -> bool {
+            // if at least one line has been output before current group then print delimiter
-        let mut first_line_printed = false;
+            && (first_line_printed
-        let mut count = if self.all_repeated { 1 } else { lines.len() };
+                // or if we need to prepend delimiter then print it even at the start of the output
-        if lines.len() == 1 && !self.repeats_only || lines.len() > 1 && !self.uniques_only {
+                || self.delimiters == Delimiters::Prepend
-            self.print_line(writer, &lines[0], count, print_delimiter);
+                // the 'both' delimit mode should prepend and append delimiters
-            first_line_printed = true;
+                || self.delimiters == Delimiters::Both)
-            count += 1;
-        }
-        if self.all_repeated {
-            for line in lines[1..].iter() {
-                self.print_line(writer, line, count, print_delimiter && !first_line_printed);
-                first_line_printed = true;
-                count += 1;
-            }
-        }
-        first_line_printed
     }
 
     fn print_line<W: Write>(
@@ -191,11 +190,11 @@ impl Uniq {
         writer: &mut BufWriter<W>,
         line: &str,
         count: usize,
-        print_delimiter: bool,
+        first_line_printed: bool,
     ) {
         let line_terminator = self.get_line_terminator();
 
-        if print_delimiter {
+        if self.should_print_delimiter(count, first_line_printed) {
             crash_if_err!(1, writer.write_all(&[line_terminator]));
         }
 

src/uucore/src/lib/features/mode.rs

@@ -7,6 +7,8 @@
 
 // spell-checker:ignore (vars) fperm srwx
 
+use libc::{mode_t, S_IRGRP, S_IROTH, S_IRUSR, S_IWGRP, S_IWOTH, S_IWUSR};
+
 pub fn parse_numeric(fperm: u32, mut mode: &str) -> Result<u32, String> {
     let (op, pos) = parse_op(mode, Some('='))?;
     mode = mode[pos..].trim().trim_start_matches('0');
@@ -129,3 +131,41 @@ fn parse_change(mode: &str, fperm: u32, considering_dir: bool) -> (u32, usize) {
     }
     (srwx, pos)
 }
+
+pub fn parse_mode(mode: Option<String>) -> Result<mode_t, String> {
+    let fperm = S_IRUSR | S_IWUSR | S_IRGRP | S_IWGRP | S_IROTH | S_IWOTH;
+    if let Some(mode) = mode {
+        let arr: &[char] = &['0', '1', '2', '3', '4', '5', '6', '7', '8', '9'];
+        let result = if mode.contains(arr) {
+            parse_numeric(fperm as u32, mode.as_str())
+        } else {
+            parse_symbolic(fperm as u32, mode.as_str(), true)
+        };
+        result.map(|mode| mode as mode_t)
+    } else {
+        Ok(fperm)
+    }
+}
+
+#[cfg(test)]
+mod test {
+
+    #[test]
+    fn symbolic_modes() {
+        assert_eq!(super::parse_mode(Some("u+x".to_owned())).unwrap(), 0o766);
+        assert_eq!(
+            super::parse_mode(Some("+x".to_owned())).unwrap(),
+            if !crate::os::is_wsl_1() { 0o777 } else { 0o776 }
+        );
+        assert_eq!(super::parse_mode(Some("a-w".to_owned())).unwrap(), 0o444);
+        assert_eq!(super::parse_mode(Some("g-r".to_owned())).unwrap(), 0o626);
+    }
+
+    #[test]
+    fn numeric_modes() {
+        assert_eq!(super::parse_mode(Some("644".to_owned())).unwrap(), 0o644);
+        assert_eq!(super::parse_mode(Some("+100".to_owned())).unwrap(), 0o766);
+        assert_eq!(super::parse_mode(Some("-4".to_owned())).unwrap(), 0o662);
+        assert_eq!(super::parse_mode(None).unwrap(), 0o666);
+    }
+}

src/uucore/src/lib/lib.rs

@@ -26,6 +26,7 @@ mod mods; // core cross-platform modules
 
 // * cross-platform modules
 pub use crate::mods::coreopts;
+pub use crate::mods::os;
 pub use crate::mods::panic;
 pub use crate::mods::ranges;
 

src/uucore/src/lib/mods.rs

@@ -1,5 +1,6 @@
 // mods ~ cross-platforms modules (core/bundler file)
 
 pub mod coreopts;
+pub mod os;
 pub mod panic;
 pub mod ranges;

src/uucore/src/lib/mods/os.rs

@@ -0,0 +1,30 @@
+/// Test if the program is running under WSL
+// ref: <https://github.com/microsoft/WSL/issues/4555> @@ <https://archive.is/dP0bz>
+pub fn is_wsl_1() -> bool {
+    #[cfg(target_os = "linux")]
+    {
+        if is_wsl_2() {
+            return false;
+        }
+        if let Ok(b) = std::fs::read("/proc/sys/kernel/osrelease") {
+            if let Ok(s) = std::str::from_utf8(&b) {
+                let a = s.to_ascii_lowercase();
+                return a.contains("microsoft") || a.contains("wsl");
+            }
+        }
+    }
+    false
+}
+
+pub fn is_wsl_2() -> bool {
+    #[cfg(target_os = "linux")]
+    {
+        if let Ok(b) = std::fs::read("/proc/sys/kernel/osrelease") {
+            if let Ok(s) = std::str::from_utf8(&b) {
+                let a = s.to_ascii_lowercase();
+                return a.contains("wsl2");
+            }
+        }
+    }
+    false
+}

tests/by-util/test_cat.rs

@@ -1,4 +1,7 @@
 use crate::common::util::*;
+#[cfg(unix)]
+use std::fs::OpenOptions;
+#[cfg(unix)]
 use std::io::Read;
 
 #[test]
@@ -26,7 +29,7 @@ fn test_no_options() {
 }
 
 #[test]
-#[cfg(unix)]
+#[cfg(any(target_vendor = "apple", target_os = "linux", target_os = "android"))]
 fn test_no_options_big_input() {
     for &n in &[
         0,
@@ -54,7 +57,6 @@ fn test_no_options_big_input() {
 #[test]
 #[cfg(unix)]
 fn test_fifo_symlink() {
-    use std::fs::OpenOptions;
     use std::io::Write;
     use std::thread;
 
@@ -85,6 +87,74 @@ fn test_fifo_symlink() {
     thread.join().unwrap();
 }
 
+#[test]
+#[cfg(unix)]
+fn test_piped_to_regular_file() {
+    use std::fs::read_to_string;
+
+    for &append in &[true, false] {
+        let s = TestScenario::new(util_name!());
+        let file_path = s.fixtures.plus("file.txt");
+
+        {
+            let file = OpenOptions::new()
+                .create_new(true)
+                .write(true)
+                .append(append)
+                .open(&file_path)
+                .unwrap();
+
+            s.ucmd()
+                .set_stdout(file)
+                .pipe_in_fixture("alpha.txt")
+                .succeeds();
+        }
+        let contents = read_to_string(&file_path).unwrap();
+        assert_eq!(contents, "abcde\nfghij\nklmno\npqrst\nuvwxyz\n");
+    }
+}
+
+#[test]
+#[cfg(unix)]
+fn test_piped_to_dev_null() {
+    for &append in &[true, false] {
+        let s = TestScenario::new(util_name!());
+        {
+            let dev_null = OpenOptions::new()
+                .write(true)
+                .append(append)
+                .open("/dev/null")
+                .unwrap();
+
+            s.ucmd()
+                .set_stdout(dev_null)
+                .pipe_in_fixture("alpha.txt")
+                .succeeds();
+        }
+    }
+}
+
+#[test]
+#[cfg(any(target_os = "linux", target_os = "freebsd", target_os = "netbsd"))]
+fn test_piped_to_dev_full() {
+    for &append in &[true, false] {
+        let s = TestScenario::new(util_name!());
+        {
+            let dev_full = OpenOptions::new()
+                .write(true)
+                .append(append)
+                .open("/dev/full")
+                .unwrap();
+
+            s.ucmd()
+                .set_stdout(dev_full)
+                .pipe_in_fixture("alpha.txt")
+                .fails()
+                .stderr_contains(&"No space left on device".to_owned());
+        }
+    }
+}
+
 #[test]
 fn test_directory() {
     let s = TestScenario::new(util_name!());
@@ -327,6 +397,7 @@ fn test_dev_full_show_all() {
 #[cfg(unix)]
 fn test_domain_socket() {
     use std::io::prelude::*;
+    use std::sync::{Arc, Barrier};
     use std::thread;
     use tempdir::TempDir;
     use unix_socket::UnixListener;
@@ -335,17 +406,23 @@ fn test_domain_socket() {
     let socket_path = dir.path().join("sock");
     let listener = UnixListener::bind(&socket_path).expect("failed to create socket");
 
+    // use a barrier to ensure we don't run cat before the listener is setup
+    let barrier = Arc::new(Barrier::new(2));
+    let barrier2 = Arc::clone(&barrier);
+
     let thread = thread::spawn(move || {
         let mut stream = listener.accept().expect("failed to accept connection").0;
+        barrier2.wait();
         stream
             .write_all(b"a\tb")
             .expect("failed to write test data");
     });
 
-    new_ucmd!()
+    let child = new_ucmd!().args(&[socket_path]).run_no_wait();
-        .args(&[socket_path])
+    barrier.wait();
-        .succeeds()
+    let stdout = &child.wait_with_output().unwrap().stdout.clone();
-        .stdout_only("a\tb");
+    let output = String::from_utf8_lossy(&stdout);
+    assert_eq!("a\tb", output);
 
     thread.join().unwrap();
 }

tests/by-util/test_chgrp.rs

@@ -104,7 +104,7 @@ fn test_reference() {
     // skip for root or MS-WSL
     // * MS-WSL is bugged (as of 2019-12-25), allowing non-root accounts su-level privileges for `chgrp`
     // * for MS-WSL, succeeds and stdout == 'group of /etc retained as root'
-    if !(get_effective_gid() == 0 || is_wsl()) {
+    if !(get_effective_gid() == 0 || uucore::os::is_wsl_1()) {
         new_ucmd!()
             .arg("-v")
             .arg("--reference=/etc/passwd")

tests/by-util/test_chmod.rs

@@ -47,7 +47,7 @@ fn run_single_test(test: &TestCase, at: AtPath, mut ucmd: UCommand) {
         ucmd.arg(arg);
     }
     let r = ucmd.run();
-    if !r.success {
+    if !r.succeeded() {
         println!("{}", r.stderr_str());
         panic!("{:?}: failed", ucmd.raw);
     }

tests/by-util/test_chroot.rs

@@ -4,14 +4,11 @@ use crate::common::util::*;
 fn test_missing_operand() {
     let result = new_ucmd!().run();
 
-    assert_eq!(
+    assert!(result
-        true,
+        .stderr_str()
-        result
+        .starts_with("error: The following required arguments were not provided"));
-            .stderr
-            .starts_with("error: The following required arguments were not provided")
-    );
 
-    assert_eq!(true, result.stderr.contains("<newroot>"));
+    assert!(result.stderr_str().contains("<newroot>"));
 }
 
 #[test]
@@ -20,14 +17,11 @@ fn test_enter_chroot_fails() {
 
     at.mkdir("jail");
 
-    let result = ucmd.arg("jail").run();
+    let result = ucmd.arg("jail").fails();
 
-    assert_eq!(
+    assert!(result
-        true,
+        .stderr_str()
-        result.stderr.starts_with(
+        .starts_with("chroot: error: cannot chroot to jail: Operation not permitted (os error 1)"));
-            "chroot: error: cannot chroot to jail: Operation not permitted (os error 1)"
-        )
-    )
 }
 
 #[test]
@@ -47,19 +41,18 @@ fn test_invalid_user_spec() {
 
     at.mkdir("a");
 
-    let result = ucmd.arg("a").arg("--userspec=ARABA:").run();
+    let result = ucmd.arg("a").arg("--userspec=ARABA:").fails();
 
-    assert_eq!(
+    assert!(result
-        true,
+        .stderr_str()
-        result.stderr.starts_with("chroot: error: invalid userspec")
+        .starts_with("chroot: error: invalid userspec"));
-    );
 }
 
 #[test]
 fn test_preference_of_userspec() {
     let scene = TestScenario::new(util_name!());
     let result = scene.cmd("whoami").run();
-    if is_ci() && result.stderr.contains("No such user/group") {
+    if is_ci() && result.stderr_str().contains("No such user/group") {
         // In the CI, some server are failing to return whoami.
         // As seems to be a configuration issue, ignoring it
         return;
@@ -73,7 +66,7 @@ fn test_preference_of_userspec() {
     println!("result.stdout = {}", result.stdout_str());
     println!("result.stderr = {}", result.stderr_str());
 
-    if is_ci() && result.stderr.contains("cannot find name for user ID") {
+    if is_ci() && result.stderr_str().contains("cannot find name for user ID") {
         // In the CI, some server are failing to return id.
         // As seems to be a configuration issue, ignoring it
         return;

tests/by-util/test_cp.rs

@@ -42,13 +42,9 @@ static TEST_MOUNT_OTHER_FILESYSTEM_FILE: &str = "mount/DO_NOT_copy_me.txt";
 fn test_cp_cp() {
     let (at, mut ucmd) = at_and_ucmd!();
     // Invoke our binary to make the copy.
-    let result = ucmd
+    ucmd.arg(TEST_HELLO_WORLD_SOURCE)
-        .arg(TEST_HELLO_WORLD_SOURCE)
         .arg(TEST_HELLO_WORLD_DEST)
-        .run();
+        .succeeds();
-
-    // Check that the exit code represents a successful copy.
-    assert!(result.success);
 
     // Check the content of the destination file that was copied.
     assert_eq!(at.read(TEST_HELLO_WORLD_DEST), "Hello, World!\n");
@@ -57,12 +53,9 @@ fn test_cp_cp() {
 #[test]
 fn test_cp_existing_target() {
     let (at, mut ucmd) = at_and_ucmd!();
-    let result = ucmd
+    ucmd.arg(TEST_HELLO_WORLD_SOURCE)
-        .arg(TEST_HELLO_WORLD_SOURCE)
         .arg(TEST_EXISTING_FILE)
-        .run();
+        .succeeds();
-
-    assert!(result.success);
 
     // Check the content of the destination file
     assert_eq!(at.read(TEST_EXISTING_FILE), "Hello, World!\n");
@@ -74,52 +67,41 @@ fn test_cp_existing_target() {
 #[test]
 fn test_cp_duplicate_files() {
     let (at, mut ucmd) = at_and_ucmd!();
-    let result = ucmd
+    ucmd.arg(TEST_HELLO_WORLD_SOURCE)
-        .arg(TEST_HELLO_WORLD_SOURCE)
         .arg(TEST_HELLO_WORLD_SOURCE)
         .arg(TEST_COPY_TO_FOLDER)
-        .run();
+        .succeeds()
-
+        .stderr_contains("specified more than once");
-    assert!(result.success);
-    assert!(result.stderr.contains("specified more than once"));
     assert_eq!(at.read(TEST_COPY_TO_FOLDER_FILE), "Hello, World!\n");
 }
 
 #[test]
 fn test_cp_multiple_files_target_is_file() {
-    let (_, mut ucmd) = at_and_ucmd!();
+    new_ucmd!()
-    let result = ucmd
         .arg(TEST_HELLO_WORLD_SOURCE)
         .arg(TEST_HELLO_WORLD_SOURCE)
         .arg(TEST_EXISTING_FILE)
-        .run();
+        .fails()
-
+        .stderr_contains("not a directory");
-    assert!(!result.success);
-    assert!(result.stderr.contains("not a directory"));
 }
 
 #[test]
 fn test_cp_directory_not_recursive() {
-    let (_, mut ucmd) = at_and_ucmd!();
+    new_ucmd!()
-    let result = ucmd
         .arg(TEST_COPY_TO_FOLDER)
         .arg(TEST_HELLO_WORLD_DEST)
-        .run();
+        .fails()
-
+        .stderr_contains("omitting directory");
-    assert!(!result.success);
-    assert!(result.stderr.contains("omitting directory"));
 }
 
 #[test]
 fn test_cp_multiple_files() {
     let (at, mut ucmd) = at_and_ucmd!();
-    let result = ucmd
+    ucmd.arg(TEST_HELLO_WORLD_SOURCE)
-        .arg(TEST_HELLO_WORLD_SOURCE)
         .arg(TEST_HOW_ARE_YOU_SOURCE)
         .arg(TEST_COPY_TO_FOLDER)
-        .run();
+        .succeeds();
 
-    assert!(result.success);
     assert_eq!(at.read(TEST_COPY_TO_FOLDER_FILE), "Hello, World!\n");
     assert_eq!(at.read(TEST_HOW_ARE_YOU_DEST), "How are you?\n");
 }
@@ -129,14 +111,11 @@ fn test_cp_multiple_files() {
 #[cfg(not(macos))]
 fn test_cp_recurse() {
     let (at, mut ucmd) = at_and_ucmd!();
-
+    ucmd.arg("-r")
-    let result = ucmd
-        .arg("-r")
         .arg(TEST_COPY_FROM_FOLDER)
         .arg(TEST_COPY_TO_FOLDER_NEW)
-        .run();
+        .succeeds();
 
-    assert!(result.success);
     // Check the content of the destination file that was copied.
     assert_eq!(at.read(TEST_COPY_TO_FOLDER_NEW_FILE), "Hello, World!\n");
 }
@@ -144,14 +123,10 @@ fn test_cp_recurse() {
 #[test]
 fn test_cp_with_dirs_t() {
     let (at, mut ucmd) = at_and_ucmd!();
-
+    ucmd.arg("-t")
-    //using -t option
-    let result_to_dir_t = ucmd
-        .arg("-t")
         .arg(TEST_COPY_TO_FOLDER)
         .arg(TEST_HELLO_WORLD_SOURCE)
-        .run();
+        .succeeds();
-    assert!(result_to_dir_t.success);
     assert_eq!(at.read(TEST_COPY_TO_FOLDER_FILE), "Hello, World!\n");
 }
 
@@ -162,63 +137,52 @@ fn test_cp_with_dirs() {
     let scene = TestScenario::new(util_name!());
     let at = &scene.fixtures;
 
-    //using -t option
+    scene
-    let result_to_dir = scene
         .ucmd()
         .arg(TEST_HELLO_WORLD_SOURCE)
         .arg(TEST_COPY_TO_FOLDER)
-        .run();
+        .succeeds();
-    assert!(result_to_dir.success);
     assert_eq!(at.read(TEST_COPY_TO_FOLDER_FILE), "Hello, World!\n");
 
-    let result_from_dir = scene
+    scene
         .ucmd()
         .arg(TEST_COPY_FROM_FOLDER_FILE)
         .arg(TEST_HELLO_WORLD_DEST)
-        .run();
+        .succeeds();
-    assert!(result_from_dir.success);
     assert_eq!(at.read(TEST_HELLO_WORLD_DEST), "Hello, World!\n");
 }
 
 #[test]
 fn test_cp_arg_target_directory() {
     let (at, mut ucmd) = at_and_ucmd!();
-    let result = ucmd
+    ucmd.arg(TEST_HELLO_WORLD_SOURCE)
-        .arg(TEST_HELLO_WORLD_SOURCE)
         .arg("-t")
         .arg(TEST_COPY_TO_FOLDER)
-        .run();
+        .succeeds();
 
-    assert!(result.success);
     assert_eq!(at.read(TEST_COPY_TO_FOLDER_FILE), "Hello, World!\n");
 }
 
 #[test]
 fn test_cp_arg_no_target_directory() {
-    let (_, mut ucmd) = at_and_ucmd!();
+    new_ucmd!()
-    let result = ucmd
         .arg(TEST_HELLO_WORLD_SOURCE)
         .arg("-v")
         .arg("-T")
         .arg(TEST_COPY_TO_FOLDER)
-        .run();
+        .fails()
-
+        .stderr_contains("cannot overwrite directory");
-    assert!(!result.success);
-    assert!(result.stderr.contains("cannot overwrite directory"));
 }
 
 #[test]
 fn test_cp_arg_interactive() {
-    let (_, mut ucmd) = at_and_ucmd!();
+    new_ucmd!()
-    let result = ucmd
         .arg(TEST_HELLO_WORLD_SOURCE)
         .arg(TEST_HOW_ARE_YOU_SOURCE)
         .arg("-i")
         .pipe_in("N\n")
-        .run();
+        .succeeds()
-
+        .stderr_contains("Not overwriting");
-    assert!(result.success);
-    assert!(result.stderr.contains("Not overwriting"));
 }
 
 #[test]
@@ -227,39 +191,33 @@ fn test_cp_arg_link() {
     use std::os::linux::fs::MetadataExt;
 
     let (at, mut ucmd) = at_and_ucmd!();
-    let result = ucmd
+    ucmd.arg(TEST_HELLO_WORLD_SOURCE)
-        .arg(TEST_HELLO_WORLD_SOURCE)
         .arg("--link")
         .arg(TEST_HELLO_WORLD_DEST)
-        .run();
+        .succeeds();
 
-    assert!(result.success);
     assert_eq!(at.metadata(TEST_HELLO_WORLD_SOURCE).st_nlink(), 2);
 }
 
 #[test]
 fn test_cp_arg_symlink() {
     let (at, mut ucmd) = at_and_ucmd!();
-    let result = ucmd
+    ucmd.arg(TEST_HELLO_WORLD_SOURCE)
-        .arg(TEST_HELLO_WORLD_SOURCE)
         .arg("--symbolic-link")
         .arg(TEST_HELLO_WORLD_DEST)
-        .run();
+        .succeeds();
 
-    assert!(result.success);
     assert!(at.is_symlink(TEST_HELLO_WORLD_DEST));
 }
 
 #[test]
 fn test_cp_arg_no_clobber() {
     let (at, mut ucmd) = at_and_ucmd!();
-    let result = ucmd
+    ucmd.arg(TEST_HELLO_WORLD_SOURCE)
-        .arg(TEST_HELLO_WORLD_SOURCE)
         .arg("--no-clobber")
         .arg(TEST_HOW_ARE_YOU_SOURCE)
-        .run();
+        .succeeds();
 
-    assert!(result.success);
     assert_eq!(at.read(TEST_HOW_ARE_YOU_SOURCE), "How are you?\n");
 }
 
@@ -267,34 +225,31 @@ fn test_cp_arg_no_clobber() {
 fn test_cp_arg_no_clobber_twice() {
     let scene = TestScenario::new(util_name!());
     let at = &scene.fixtures;
+
     at.touch("source.txt");
-    let result = scene
+    scene
         .ucmd()
         .arg("--no-clobber")
         .arg("source.txt")
         .arg("dest.txt")
-        .run();
+        .succeeds()
+        .no_stderr();
 
-    println!("stderr = {:?}", result.stderr_str());
-    println!("stdout = {:?}", result.stdout_str());
-    assert!(result.success);
-    assert!(result.stderr.is_empty());
     assert_eq!(at.read("source.txt"), "");
 
     at.append("source.txt", "some-content");
-    let result = scene
+    scene
         .ucmd()
         .arg("--no-clobber")
         .arg("source.txt")
         .arg("dest.txt")
-        .run();
+        .succeeds()
+        .stdout_does_not_contain("Not overwriting");
 
-    assert!(result.success);
     assert_eq!(at.read("source.txt"), "some-content");
     // Should be empty as the "no-clobber" should keep
     // the previous version
     assert_eq!(at.read("dest.txt"), "");
-    assert!(!result.stderr.contains("Not overwriting"));
 }
 
 #[test]
@@ -311,16 +266,11 @@ fn test_cp_arg_force() {
     permissions.set_readonly(true);
     set_permissions(at.plus(TEST_HELLO_WORLD_DEST), permissions).unwrap();
 
-    let result = ucmd
+    ucmd.arg(TEST_HELLO_WORLD_SOURCE)
-        .arg(TEST_HELLO_WORLD_SOURCE)
         .arg("--force")
         .arg(TEST_HELLO_WORLD_DEST)
-        .run();
+        .succeeds();
 
-    println!("{:?}", result.stderr_str());
-    println!("{:?}", result.stdout_str());
-
-    assert!(result.success);
     assert_eq!(at.read(TEST_HELLO_WORLD_DEST), "Hello, World!\n");
 }
 
@@ -342,13 +292,11 @@ fn test_cp_arg_remove_destination() {
     permissions.set_readonly(true);
     set_permissions(at.plus(TEST_HELLO_WORLD_DEST), permissions).unwrap();
 
-    let result = ucmd
+    ucmd.arg(TEST_HELLO_WORLD_SOURCE)
-        .arg(TEST_HELLO_WORLD_SOURCE)
         .arg("--remove-destination")
         .arg(TEST_HELLO_WORLD_DEST)
-        .run();
+        .succeeds();
 
-    assert!(result.success);
     assert_eq!(at.read(TEST_HELLO_WORLD_DEST), "Hello, World!\n");
 }
 
@@ -356,13 +304,11 @@ fn test_cp_arg_remove_destination() {
 fn test_cp_arg_backup() {
     let (at, mut ucmd) = at_and_ucmd!();
 
-    let result = ucmd
+    ucmd.arg(TEST_HELLO_WORLD_SOURCE)
-        .arg(TEST_HELLO_WORLD_SOURCE)
         .arg("--backup")
         .arg(TEST_HOW_ARE_YOU_SOURCE)
-        .run();
+        .succeeds();
 
-    assert!(result.success);
     assert_eq!(at.read(TEST_HOW_ARE_YOU_SOURCE), "Hello, World!\n");
     assert_eq!(
         at.read(&*format!("{}~", TEST_HOW_ARE_YOU_SOURCE)),
@@ -374,14 +320,12 @@ fn test_cp_arg_backup() {
 fn test_cp_arg_suffix() {
     let (at, mut ucmd) = at_and_ucmd!();
 
-    let result = ucmd
+    ucmd.arg(TEST_HELLO_WORLD_SOURCE)
-        .arg(TEST_HELLO_WORLD_SOURCE)
         .arg("--suffix")
         .arg(".bak")
         .arg(TEST_HOW_ARE_YOU_SOURCE)
-        .run();
+        .succeeds();
 
-    assert!(result.success);
     assert_eq!(at.read(TEST_HOW_ARE_YOU_SOURCE), "Hello, World!\n");
     assert_eq!(
         at.read(&*format!("{}.bak", TEST_HOW_ARE_YOU_SOURCE)),
@@ -391,9 +335,8 @@ fn test_cp_arg_suffix() {
 
 #[test]
 fn test_cp_deref_conflicting_options() {
-    let (_at, mut ucmd) = at_and_ucmd!();
+    new_ucmd!()
-
+        .arg("-LP")
-    ucmd.arg("-LP")
         .arg(TEST_COPY_TO_FOLDER)
         .arg(TEST_HELLO_WORLD_SOURCE)
         .fails();
@@ -401,8 +344,7 @@ fn test_cp_deref_conflicting_options() {
 
 #[test]
 fn test_cp_deref() {
-    let scene = TestScenario::new(util_name!());
+    let (at, mut ucmd) = at_and_ucmd!();
-    let at = &scene.fixtures;
 
     #[cfg(not(windows))]
     let _r = fs::symlink(
@@ -415,16 +357,12 @@ fn test_cp_deref() {
         at.subdir.join(TEST_HELLO_WORLD_SOURCE_SYMLINK),
     );
     //using -L option
-    let result = scene
+    ucmd.arg("-L")
-        .ucmd()
-        .arg("-L")
         .arg(TEST_HELLO_WORLD_SOURCE)
         .arg(TEST_HELLO_WORLD_SOURCE_SYMLINK)
         .arg(TEST_COPY_TO_FOLDER)
-        .run();
+        .succeeds();
 
-    // Check that the exit code represents a successful copy.
-    assert!(result.success);
     let path_to_new_symlink = at
         .subdir
         .join(TEST_COPY_TO_FOLDER)
@@ -444,8 +382,7 @@ fn test_cp_deref() {
 }
 #[test]
 fn test_cp_no_deref() {
-    let scene = TestScenario::new(util_name!());
+    let (at, mut ucmd) = at_and_ucmd!();
-    let at = &scene.fixtures;
 
     #[cfg(not(windows))]
     let _r = fs::symlink(
@@ -458,16 +395,12 @@ fn test_cp_no_deref() {
         at.subdir.join(TEST_HELLO_WORLD_SOURCE_SYMLINK),
     );
     //using -P option
-    let result = scene
+    ucmd.arg("-P")
-        .ucmd()
-        .arg("-P")
         .arg(TEST_HELLO_WORLD_SOURCE)
         .arg(TEST_HELLO_WORLD_SOURCE_SYMLINK)
         .arg(TEST_COPY_TO_FOLDER)
-        .run();
+        .succeeds();
 
-    // Check that the exit code represents a successful copy.
-    assert!(result.success);
     let path_to_new_symlink = at
         .subdir
         .join(TEST_COPY_TO_FOLDER)
@@ -490,14 +423,10 @@ fn test_cp_strip_trailing_slashes() {
     let (at, mut ucmd) = at_and_ucmd!();
 
     //using --strip-trailing-slashes option
-    let result = ucmd
+    ucmd.arg("--strip-trailing-slashes")
-        .arg("--strip-trailing-slashes")
         .arg(format!("{}/", TEST_HELLO_WORLD_SOURCE))
         .arg(TEST_HELLO_WORLD_DEST)
-        .run();
+        .succeeds();
-
-    // Check that the exit code represents a successful copy.
-    assert!(result.success);
 
     // Check the content of the destination file that was copied.
     assert_eq!(at.read(TEST_HELLO_WORLD_DEST), "Hello, World!\n");
@@ -507,14 +436,11 @@ fn test_cp_strip_trailing_slashes() {
 fn test_cp_parents() {
     let (at, mut ucmd) = at_and_ucmd!();
 
-    let result = ucmd
+    ucmd.arg("--parents")
-        .arg("--parents")
         .arg(TEST_COPY_FROM_FOLDER_FILE)
         .arg(TEST_COPY_TO_FOLDER)
-        .run();
+        .succeeds();
 
-    assert!(result.success);
-    // Check the content of the destination file that was copied.
     assert_eq!(
         at.read(&format!(
             "{}/{}",
@@ -528,14 +454,12 @@ fn test_cp_parents() {
 fn test_cp_parents_multiple_files() {
     let (at, mut ucmd) = at_and_ucmd!();
 
-    let result = ucmd
+    ucmd.arg("--parents")
-        .arg("--parents")
         .arg(TEST_COPY_FROM_FOLDER_FILE)
         .arg(TEST_HOW_ARE_YOU_SOURCE)
         .arg(TEST_COPY_TO_FOLDER)
-        .run();
+        .succeeds();
 
-    assert!(result.success);
     assert_eq!(
         at.read(&format!(
             "{}/{}",
@@ -554,20 +478,12 @@ fn test_cp_parents_multiple_files() {
 
 #[test]
 fn test_cp_parents_dest_not_directory() {
-    let (_, mut ucmd) = at_and_ucmd!();
+    new_ucmd!()
-
-    let result = ucmd
         .arg("--parents")
         .arg(TEST_COPY_FROM_FOLDER_FILE)
         .arg(TEST_HELLO_WORLD_DEST)
-        .run();
+        .fails()
-    println!("{:?}", result);
+        .stderr_contains("with --parents, the destination must be a directory");
-
-    // Check that we did not succeed in copying.
-    assert!(!result.success);
-    assert!(result
-        .stderr
-        .contains("with --parents, the destination must be a directory"));
 }
 
 #[test]
@@ -594,18 +510,14 @@ fn test_cp_deref_folder_to_folder() {
     assert!(env::set_current_dir(&cwd).is_ok());
 
     //using -P -R option
-    let result = scene
+    scene
         .ucmd()
         .arg("-L")
         .arg("-R")
         .arg("-v")
         .arg(TEST_COPY_FROM_FOLDER)
         .arg(TEST_COPY_TO_FOLDER_NEW)
-        .run();
+        .succeeds();
-    println!("cp output {}", result.stdout_str());
-
-    // Check that the exit code represents a successful copy.
-    assert!(result.success);
 
     #[cfg(not(windows))]
     {
@@ -698,18 +610,14 @@ fn test_cp_no_deref_folder_to_folder() {
     assert!(env::set_current_dir(&cwd).is_ok());
 
     //using -P -R option
-    let result = scene
+    scene
         .ucmd()
         .arg("-P")
         .arg("-R")
         .arg("-v")
         .arg(TEST_COPY_FROM_FOLDER)
         .arg(TEST_COPY_TO_FOLDER_NEW)
-        .run();
+        .succeeds();
-    println!("cp output {}", result.stdout_str());
-
-    // Check that the exit code represents a successful copy.
-    assert!(result.success);
 
     #[cfg(not(windows))]
     {
@@ -791,13 +699,11 @@ fn test_cp_archive() {
         previous,
     )
     .unwrap();
-    let result = ucmd
+    ucmd.arg(TEST_HELLO_WORLD_SOURCE)
-        .arg(TEST_HELLO_WORLD_SOURCE)
         .arg("--archive")
         .arg(TEST_HOW_ARE_YOU_SOURCE)
-        .run();
+        .succeeds();
 
-    assert!(result.success);
     assert_eq!(at.read(TEST_HOW_ARE_YOU_SOURCE), "Hello, World!\n");
 
     let metadata = std_fs::metadata(at.subdir.join(TEST_HELLO_WORLD_SOURCE)).unwrap();
@@ -807,11 +713,10 @@ fn test_cp_archive() {
     let creation2 = metadata2.modified().unwrap();
 
     let scene2 = TestScenario::new("ls");
-    let result = scene2.cmd("ls").arg("-al").arg(at.subdir).run();
+    let result = scene2.cmd("ls").arg("-al").arg(at.subdir).succeeds();
 
     println!("ls dest {}", result.stdout_str());
     assert_eq!(creation, creation2);
-    assert!(result.success);
 }
 
 #[test]
@@ -850,11 +755,10 @@ fn test_cp_archive_recursive() {
     // Back to the initial cwd (breaks the other tests)
     assert!(env::set_current_dir(&cwd).is_ok());
 
-    let resultg = ucmd
+    ucmd.arg("--archive")
-        .arg("--archive")
         .arg(TEST_COPY_TO_FOLDER)
         .arg(TEST_COPY_TO_FOLDER_NEW)
-        .run();
+        .fails(); // fails for now
 
     let scene2 = TestScenario::new("ls");
     let result = scene2
@@ -865,7 +769,6 @@ fn test_cp_archive_recursive() {
 
     println!("ls dest {}", result.stdout_str());
 
-    let scene2 = TestScenario::new("ls");
     let result = scene2
         .cmd("ls")
         .arg("-al")
@@ -910,9 +813,6 @@ fn test_cp_archive_recursive() {
             .join("2.link")
             .to_string_lossy()
     ));
-
-    // fails for now
-    assert!(resultg.success);
 }
 
 #[test]
@@ -928,13 +828,11 @@ fn test_cp_preserve_timestamps() {
         previous,
     )
     .unwrap();
-    let result = ucmd
+    ucmd.arg(TEST_HELLO_WORLD_SOURCE)
-        .arg(TEST_HELLO_WORLD_SOURCE)
         .arg("--preserve=timestamps")
         .arg(TEST_HOW_ARE_YOU_SOURCE)
-        .run();
+        .succeeds();
 
-    assert!(result.success);
     assert_eq!(at.read(TEST_HOW_ARE_YOU_SOURCE), "Hello, World!\n");
 
     let metadata = std_fs::metadata(at.subdir.join(TEST_HELLO_WORLD_SOURCE)).unwrap();
@@ -948,7 +846,6 @@ fn test_cp_preserve_timestamps() {
 
     println!("ls dest {}", result.stdout_str());
     assert_eq!(creation, creation2);
-    assert!(result.success);
 }
 
 #[test]
@@ -966,13 +863,11 @@ fn test_cp_dont_preserve_timestamps() {
     .unwrap();
     sleep(Duration::from_secs(3));
 
-    let result = ucmd
+    ucmd.arg(TEST_HELLO_WORLD_SOURCE)
-        .arg(TEST_HELLO_WORLD_SOURCE)
         .arg("--no-preserve=timestamps")
         .arg(TEST_HOW_ARE_YOU_SOURCE)
-        .run();
+        .succeeds();
 
-    assert!(result.success);
     assert_eq!(at.read(TEST_HOW_ARE_YOU_SOURCE), "Hello, World!\n");
 
     let metadata = std_fs::metadata(at.subdir.join(TEST_HELLO_WORLD_SOURCE)).unwrap();
@@ -992,7 +887,6 @@ fn test_cp_dont_preserve_timestamps() {
     // Some margins with time check
     assert!(res.as_secs() > 3595);
     assert!(res.as_secs() < 3605);
-    assert!(result.success);
 }
 
 #[test]
@@ -1017,7 +911,7 @@ fn test_cp_one_file_system() {
     let scene = TestScenario::new(util_name!());
 
     // Test must be run as root (or with `sudo -E`)
-    if scene.cmd("whoami").run().stdout != "root\n" {
+    if scene.cmd("whoami").run().stdout_str() != "root\n" {
         return;
     }
 
@@ -1042,17 +936,16 @@ fn test_cp_one_file_system() {
     at_src.touch(TEST_MOUNT_OTHER_FILESYSTEM_FILE);
 
     // Begin testing -x flag
-    let result = scene
+    scene
         .ucmd()
         .arg("-rx")
         .arg(TEST_MOUNT_COPY_FROM_FOLDER)
         .arg(TEST_COPY_TO_FOLDER_NEW)
-        .run();
+        .succeeds();
 
     // Ditch the mount before the asserts
     scene.cmd("umount").arg(mountpoint_path).succeeds();
 
-    assert!(result.success);
     assert!(!at_dst.file_exists(TEST_MOUNT_OTHER_FILESYSTEM_FILE));
     // Check if the other files were copied from the source folder hirerarchy
     for entry in WalkDir::new(at_src.as_string()) {
@@ -1072,3 +965,59 @@ fn test_cp_one_file_system() {
         }
     }
 }
+
+#[test]
+#[cfg(any(target_os = "linux", target_os = "macos"))]
+fn test_cp_reflink_always() {
+    let (at, mut ucmd) = at_and_ucmd!();
+    let result = ucmd
+        .arg("--reflink=always")
+        .arg(TEST_HELLO_WORLD_SOURCE)
+        .arg(TEST_EXISTING_FILE)
+        .run();
+
+    if result.succeeded() {
+        // Check the content of the destination file
+        assert_eq!(at.read(TEST_EXISTING_FILE), "Hello, World!\n");
+    } else {
+        // Older Linux versions do not support cloning.
+    }
+}
+
+#[test]
+#[cfg(any(target_os = "linux", target_os = "macos"))]
+fn test_cp_reflink_auto() {
+    let (at, mut ucmd) = at_and_ucmd!();
+    ucmd.arg("--reflink=auto")
+        .arg(TEST_HELLO_WORLD_SOURCE)
+        .arg(TEST_EXISTING_FILE)
+        .succeeds();
+
+    // Check the content of the destination file
+    assert_eq!(at.read(TEST_EXISTING_FILE), "Hello, World!\n");
+}
+
+#[test]
+#[cfg(any(target_os = "linux", target_os = "macos"))]
+fn test_cp_reflink_never() {
+    let (at, mut ucmd) = at_and_ucmd!();
+    ucmd.arg("--reflink=never")
+        .arg(TEST_HELLO_WORLD_SOURCE)
+        .arg(TEST_EXISTING_FILE)
+        .succeeds();
+
+    // Check the content of the destination file
+    assert_eq!(at.read(TEST_EXISTING_FILE), "Hello, World!\n");
+}
+
+#[test]
+#[cfg(any(target_os = "linux", target_os = "macos"))]
+fn test_cp_reflink_bad() {
+    let (_, mut ucmd) = at_and_ucmd!();
+    let _result = ucmd
+        .arg("--reflink=bad")
+        .arg(TEST_HELLO_WORLD_SOURCE)
+        .arg(TEST_EXISTING_FILE)
+        .fails()
+        .stderr_contains("invalid argument");
+}

tests/by-util/test_date.rs

@@ -7,174 +7,147 @@ use rust_users::*;
 
 #[test]
 fn test_date_email() {
-    let (_, mut ucmd) = at_and_ucmd!();
+    new_ucmd!().arg("--rfc-email").succeeds();
-    let result = ucmd.arg("--rfc-email").run();
-    assert!(result.success);
 }
 
 #[test]
 fn test_date_email2() {
-    let (_, mut ucmd) = at_and_ucmd!();
+    new_ucmd!().arg("-R").succeeds();
-    let result = ucmd.arg("-R").run();
-    assert!(result.success);
 }
 
 #[test]
 fn test_date_rfc_3339() {
     let scene = TestScenario::new(util_name!());
 
-    let mut result = scene.ucmd().arg("--rfc-3339=ns").succeeds();
+    let rfc_regexp = concat!(
+        r#"(\d+)-(0[1-9]|1[012])-(0[1-9]|[12]\d|3[01])\s([01]\d|2[0-3]):"#,
+        r#"([0-5]\d):([0-5]\d|60)(\.\d+)?(([Zz])|([\+|\-]([01]\d|2[0-3])))"#
+    );
+    let re = Regex::new(rfc_regexp).unwrap();
 
     // Check that the output matches the regexp
-    let rfc_regexp = r"(\d+)-(0[1-9]|1[012])-(0[1-9]|[12]\d|3[01])\s([01]\d|2[0-3]):([0-5]\d):([0-5]\d|60)(\.\d+)?(([Zz])|([\+|\-]([01]\d|2[0-3])))";
+    scene
-    let re = Regex::new(rfc_regexp).unwrap();
+        .ucmd()
-    assert!(re.is_match(&result.stdout_str().trim()));
+        .arg("--rfc-3339=ns")
+        .succeeds()
+        .stdout_matches(&re);
 
-    result = scene.ucmd().arg("--rfc-3339=seconds").succeeds();
+    scene
-
+        .ucmd()
-    // Check that the output matches the regexp
+        .arg("--rfc-3339=seconds")
-    let re = Regex::new(rfc_regexp).unwrap();
+        .succeeds()
-    assert!(re.is_match(&result.stdout_str().trim()));
+        .stdout_matches(&re);
 }
 
 #[test]
 fn test_date_rfc_8601() {
-    let (_, mut ucmd) = at_and_ucmd!();
+    new_ucmd!().arg("--iso-8601=ns").succeeds();
-    let result = ucmd.arg("--iso-8601=ns").run();
-    assert!(result.success);
 }
 
 #[test]
 fn test_date_rfc_8601_second() {
-    let (_, mut ucmd) = at_and_ucmd!();
+    new_ucmd!().arg("--iso-8601=second").succeeds();
-    let result = ucmd.arg("--iso-8601=second").run();
-    assert!(result.success);
 }
 
 #[test]
 fn test_date_utc() {
-    let (_, mut ucmd) = at_and_ucmd!();
+    new_ucmd!().arg("--utc").succeeds();
-    let result = ucmd.arg("--utc").run();
-    assert!(result.success);
 }
 
 #[test]
 fn test_date_universal() {
-    let (_, mut ucmd) = at_and_ucmd!();
+    new_ucmd!().arg("--universal").succeeds();
-    let result = ucmd.arg("--universal").run();
-    assert!(result.success);
 }
 
 #[test]
 fn test_date_format_y() {
     let scene = TestScenario::new(util_name!());
 
-    let mut result = scene.ucmd().arg("+%Y").succeeds();
-
-    assert!(result.success);
     let mut re = Regex::new(r"^\d{4}$").unwrap();
-    assert!(re.is_match(&result.stdout_str().trim()));
+    scene.ucmd().arg("+%Y").succeeds().stdout_matches(&re);
 
-    result = scene.ucmd().arg("+%y").succeeds();
-
-    assert!(result.success);
     re = Regex::new(r"^\d{2}$").unwrap();
-    assert!(re.is_match(&result.stdout_str().trim()));
+    scene.ucmd().arg("+%y").succeeds().stdout_matches(&re);
 }
 
 #[test]
 fn test_date_format_m() {
     let scene = TestScenario::new(util_name!());
 
-    let mut result = scene.ucmd().arg("+%b").succeeds();
-
-    assert!(result.success);
     let mut re = Regex::new(r"\S+").unwrap();
-    assert!(re.is_match(&result.stdout_str().trim()));
+    scene.ucmd().arg("+%b").succeeds().stdout_matches(&re);
 
-    result = scene.ucmd().arg("+%m").succeeds();
-
-    assert!(result.success);
     re = Regex::new(r"^\d{2}$").unwrap();
-    assert!(re.is_match(&result.stdout_str().trim()));
+    scene.ucmd().arg("+%m").succeeds().stdout_matches(&re);
 }
 
 #[test]
 fn test_date_format_day() {
     let scene = TestScenario::new(util_name!());
 
-    let mut result = scene.ucmd().arg("+%a").succeeds();
-
-    assert!(result.success);
     let mut re = Regex::new(r"\S+").unwrap();
-    assert!(re.is_match(&result.stdout_str().trim()));
+    scene.ucmd().arg("+%a").succeeds().stdout_matches(&re);
-
-    result = scene.ucmd().arg("+%A").succeeds();
-
-    assert!(result.success);
 
     re = Regex::new(r"\S+").unwrap();
-    assert!(re.is_match(&result.stdout_str().trim()));
+    scene.ucmd().arg("+%A").succeeds().stdout_matches(&re);
 
-    result = scene.ucmd().arg("+%u").succeeds();
-
-    assert!(result.success);
     re = Regex::new(r"^\d{1}$").unwrap();
-    assert!(re.is_match(&result.stdout_str().trim()));
+    scene.ucmd().arg("+%u").succeeds().stdout_matches(&re);
 }
 
 #[test]
 fn test_date_format_full_day() {
-    let scene = TestScenario::new(util_name!());
-
-    let result = scene.ucmd().arg("+'%a %Y-%m-%d'").succeeds();
-
-    assert!(result.success);
     let re = Regex::new(r"\S+ \d{4}-\d{2}-\d{2}").unwrap();
-    assert!(re.is_match(&result.stdout_str().trim()));
+    new_ucmd!()
+        .arg("+'%a %Y-%m-%d'")
+        .succeeds()
+        .stdout_matches(&re);
 }
 
 #[test]
 #[cfg(all(unix, not(target_os = "macos")))]
 fn test_date_set_valid() {
     if get_effective_uid() == 0 {
-        let (_, mut ucmd) = at_and_ucmd!();
+        new_ucmd!()
-        let result = ucmd
             .arg("--set")
             .arg("2020-03-12 13:30:00+08:00")
-            .succeeds();
+            .succeeds()
-        result.no_stdout().no_stderr();
+            .no_stdout()
+            .no_stderr();
     }
 }
 
 #[test]
 #[cfg(any(windows, all(unix, not(target_os = "macos"))))]
 fn test_date_set_invalid() {
-    let (_, mut ucmd) = at_and_ucmd!();
+    let result = new_ucmd!().arg("--set").arg("123abcd").fails();
-    let result = ucmd.arg("--set").arg("123abcd").fails();
+    result.no_stdout();
-    let result = result.no_stdout();
+    assert!(result.stderr_str().starts_with("date: invalid date "));
-    assert!(result.stderr.starts_with("date: invalid date "));
 }
 
 #[test]
 #[cfg(all(unix, not(target_os = "macos")))]
 fn test_date_set_permissions_error() {
-    if !(get_effective_uid() == 0 || is_wsl()) {
+    if !(get_effective_uid() == 0 || uucore::os::is_wsl_1()) {
-        let (_, mut ucmd) = at_and_ucmd!();
+        let result = new_ucmd!()
-        let result = ucmd.arg("--set").arg("2020-03-11 21:45:00+08:00").fails();
+            .arg("--set")
-        let result = result.no_stdout();
+            .arg("2020-03-11 21:45:00+08:00")
-        assert!(result.stderr.starts_with("date: cannot set date: "));
+            .fails();
+        result.no_stdout();
+        assert!(result.stderr_str().starts_with("date: cannot set date: "));
     }
 }
 
 #[test]
 #[cfg(target_os = "macos")]
 fn test_date_set_mac_unavailable() {
-    let (_, mut ucmd) = at_and_ucmd!();
+    let result = new_ucmd!()
-    let result = ucmd.arg("--set").arg("2020-03-11 21:45:00+08:00").fails();
+        .arg("--set")
-    let result = result.no_stdout();
+        .arg("2020-03-11 21:45:00+08:00")
+        .fails();
+    result.no_stdout();
     assert!(result
-        .stderr
+        .stderr_str()
         .starts_with("date: setting the date is not supported by macOS"));
 }
 
@@ -183,13 +156,12 @@ fn test_date_set_mac_unavailable() {
 /// TODO: expected to fail currently; change to succeeds() when required.
 fn test_date_set_valid_2() {
     if get_effective_uid() == 0 {
-        let (_, mut ucmd) = at_and_ucmd!();
+        let result = new_ucmd!()
-        let result = ucmd
             .arg("--set")
             .arg("Sat 20 Mar 2021 14:53:01 AWST")
             .fails();
-        let result = result.no_stdout();
+        result.no_stdout();
-        assert!(result.stderr.starts_with("date: invalid date "));
+        assert!(result.stderr_str().starts_with("date: invalid date "));
     }
 }
 
@@ -198,13 +170,12 @@ fn test_date_set_valid_2() {
 /// TODO: expected to fail currently; change to succeeds() when required.
 fn test_date_set_valid_3() {
     if get_effective_uid() == 0 {
-        let (_, mut ucmd) = at_and_ucmd!();
+        let result = new_ucmd!()
-        let result = ucmd
             .arg("--set")
             .arg("Sat 20 Mar 2021 14:53:01") // Local timezone
             .fails();
-        let result = result.no_stdout();
+        result.no_stdout();
-        assert!(result.stderr.starts_with("date: invalid date "));
+        assert!(result.stderr_str().starts_with("date: invalid date "));
     }
 }
 
@@ -213,12 +184,11 @@ fn test_date_set_valid_3() {
 /// TODO: expected to fail currently; change to succeeds() when required.
 fn test_date_set_valid_4() {
     if get_effective_uid() == 0 {
-        let (_, mut ucmd) = at_and_ucmd!();
+        let result = new_ucmd!()
-        let result = ucmd
             .arg("--set")
             .arg("2020-03-11 21:45:00") // Local timezone
             .fails();
-        let result = result.no_stdout();
+        result.no_stdout();
-        assert!(result.stderr.starts_with("date: invalid date "));
+        assert!(result.stderr_str().starts_with("date: invalid date "));
     }
 }

tests/by-util/test_df.rs

@@ -2,30 +2,22 @@ use crate::common::util::*;
 
 #[test]
 fn test_df_compatible_no_size_arg() {
-    let (_, mut ucmd) = at_and_ucmd!();
+    new_ucmd!().arg("-a").succeeds();
-    let result = ucmd.arg("-a").run();
-    assert!(result.success);
 }
 
 #[test]
 fn test_df_compatible() {
-    let (_, mut ucmd) = at_and_ucmd!();
+    new_ucmd!().arg("-ah").succeeds();
-    let result = ucmd.arg("-ah").run();
-    assert!(result.success);
 }
 
 #[test]
 fn test_df_compatible_type() {
-    let (_, mut ucmd) = at_and_ucmd!();
+    new_ucmd!().arg("-aT").succeeds();
-    let result = ucmd.arg("-aT").run();
-    assert!(result.success);
 }
 
 #[test]
 fn test_df_compatible_si() {
-    let (_, mut ucmd) = at_and_ucmd!();
+    new_ucmd!().arg("-aH").succeeds();
-    let result = ucmd.arg("-aH").run();
-    assert!(result.success);
 }
 
 // ToDO: more tests...

tests/by-util/test_du.rs

@@ -10,7 +10,7 @@ fn test_du_basics() {
     new_ucmd!().succeeds().no_stderr();
 }
 #[cfg(target_vendor = "apple")]
-fn _du_basics(s: String) {
+fn _du_basics(s: &str) {
     let answer = "32\t./subdir
 8\t./subdir/deeper
 24\t./subdir/links
@@ -30,11 +30,18 @@ fn _du_basics(s: &str) {
 
 #[test]
 fn test_du_basics_subdir() {
-    let (_at, mut ucmd) = at_and_ucmd!();
+    let scene = TestScenario::new(util_name!());
 
-    let result = ucmd.arg(SUB_DIR).run();
+    let result = scene.ucmd().arg(SUB_DIR).succeeds();
-    assert!(result.success);
+
-    assert_eq!(result.stderr, "");
+    #[cfg(target_os = "linux")]
+    {
+        let result_reference = scene.cmd("du").arg(SUB_DIR).run();
+        if result_reference.succeeded() {
+            assert_eq!(result.stdout_str(), result_reference.stdout_str());
+            return;
+        }
+    }
     _du_basics_subdir(result.stdout_str());
 }
 
@@ -49,7 +56,7 @@ fn _du_basics_subdir(s: &str) {
 #[cfg(all(not(target_vendor = "apple"), not(target_os = "windows")))]
 fn _du_basics_subdir(s: &str) {
     // MS-WSL linux has altered expected output
-    if !is_wsl() {
+    if !uucore::os::is_wsl_1() {
         assert_eq!(s, "8\tsubdir/deeper\n");
     } else {
         assert_eq!(s, "0\tsubdir/deeper\n");
@@ -58,26 +65,29 @@ fn _du_basics_subdir(s: &str) {
 
 #[test]
 fn test_du_basics_bad_name() {
-    let (_at, mut ucmd) = at_and_ucmd!();
+    new_ucmd!()
-
+        .arg("bad_name")
-    let result = ucmd.arg("bad_name").run();
+        .succeeds() // TODO: replace with ".fails()" once `du` is fixed
-    assert_eq!(result.stdout_str(), "");
+        .stderr_only("du: error: bad_name: No such file or directory\n");
-    assert_eq!(
-        result.stderr,
-        "du: error: bad_name: No such file or directory\n"
-    );
 }
 
 #[test]
 fn test_du_soft_link() {
-    let ts = TestScenario::new("du");
+    let scene = TestScenario::new(util_name!());
+    let at = &scene.fixtures;
 
-    let link = ts.ccmd("ln").arg("-s").arg(SUB_FILE).arg(SUB_LINK).run();
+    at.symlink_file(SUB_FILE, SUB_LINK);
-    assert!(link.success);
 
-    let result = ts.ucmd().arg(SUB_DIR_LINKS).run();
+    let result = scene.ucmd().arg(SUB_DIR_LINKS).succeeds();
-    assert!(result.success);
+
-    assert_eq!(result.stderr, "");
+    #[cfg(target_os = "linux")]
+    {
+        let result_reference = scene.cmd("du").arg(SUB_DIR_LINKS).run();
+        if result_reference.succeeded() {
+            assert_eq!(result.stdout_str(), result_reference.stdout_str());
+            return;
+        }
+    }
     _du_soft_link(result.stdout_str());
 }
 
@@ -93,7 +103,7 @@ fn _du_soft_link(s: &str) {
 #[cfg(all(not(target_vendor = "apple"), not(target_os = "windows")))]
 fn _du_soft_link(s: &str) {
     // MS-WSL linux has altered expected output
-    if !is_wsl() {
+    if !uucore::os::is_wsl_1() {
         assert_eq!(s, "16\tsubdir/links\n");
     } else {
         assert_eq!(s, "8\tsubdir/links\n");
@@ -102,14 +112,23 @@ fn _du_soft_link(s: &str) {
 
 #[test]
 fn test_du_hard_link() {
-    let ts = TestScenario::new("du");
+    let scene = TestScenario::new(util_name!());
 
-    let link = ts.ccmd("ln").arg(SUB_FILE).arg(SUB_LINK).run();
+    let result_ln = scene.cmd("ln").arg(SUB_FILE).arg(SUB_LINK).run();
-    assert!(link.success);
+    if !result_ln.succeeded() {
+        scene.ccmd("ln").arg(SUB_FILE).arg(SUB_LINK).succeeds();
+    }
 
-    let result = ts.ucmd().arg(SUB_DIR_LINKS).run();
+    let result = scene.ucmd().arg(SUB_DIR_LINKS).succeeds();
-    assert!(result.success);
+
-    assert_eq!(result.stderr, "");
+    #[cfg(target_os = "linux")]
+    {
+        let result_reference = scene.cmd("du").arg(SUB_DIR_LINKS).run();
+        if result_reference.succeeded() {
+            assert_eq!(result.stdout_str(), result_reference.stdout_str());
+            return;
+        }
+    }
     // We do not double count hard links as the inodes are identical
     _du_hard_link(result.stdout_str());
 }
@@ -125,7 +144,7 @@ fn _du_hard_link(s: &str) {
 #[cfg(all(not(target_vendor = "apple"), not(target_os = "windows")))]
 fn _du_hard_link(s: &str) {
     // MS-WSL linux has altered expected output
-    if !is_wsl() {
+    if !uucore::os::is_wsl_1() {
         assert_eq!(s, "16\tsubdir/links\n");
     } else {
         assert_eq!(s, "8\tsubdir/links\n");
@@ -134,11 +153,23 @@ fn _du_hard_link(s: &str) {
 
 #[test]
 fn test_du_d_flag() {
-    let ts = TestScenario::new("du");
+    let scene = TestScenario::new(util_name!());
 
-    let result = ts.ucmd().arg("-d").arg("1").run();
+    let result = scene.ucmd().arg("-d1").succeeds();
-    assert!(result.success);
+
-    assert_eq!(result.stderr, "");
+    #[cfg(target_os = "linux")]
+    {
+        let result_reference = scene.cmd("du").arg("-d1").run();
+        if result_reference.succeeded() {
+            assert_eq!(
+                // TODO: gnu `du` doesn't use trailing "/" here
+                // result.stdout_str(), result_reference.stdout_str()
+                result.stdout_str().trim_end_matches("/\n"),
+                result_reference.stdout_str().trim_end_matches("\n")
+            );
+            return;
+        }
+    }
     _du_d_flag(result.stdout_str());
 }
 
@@ -153,7 +184,7 @@ fn _du_d_flag(s: &str) {
 #[cfg(all(not(target_vendor = "apple"), not(target_os = "windows")))]
 fn _du_d_flag(s: &str) {
     // MS-WSL linux has altered expected output
-    if !is_wsl() {
+    if !uucore::os::is_wsl_1() {
         assert_eq!(s, "28\t./subdir\n36\t./\n");
     } else {
         assert_eq!(s, "8\t./subdir\n8\t./\n");
@@ -162,9 +193,7 @@ fn _du_d_flag(s: &str) {
 
 #[test]
 fn test_du_h_flag_empty_file() {
-    let ts = TestScenario::new("du");
+    new_ucmd!()
-
-    ts.ucmd()
         .arg("-h")
         .arg("empty.txt")
         .succeeds()
@@ -174,54 +203,61 @@ fn test_du_h_flag_empty_file() {
 #[cfg(feature = "touch")]
 #[test]
 fn test_du_time() {
-    let ts = TestScenario::new("du");
+    let scene = TestScenario::new(util_name!());
 
-    let touch = ts
+    scene
         .ccmd("touch")
         .arg("-a")
         .arg("-m")
         .arg("-t")
         .arg("201505150000")
         .arg("date_test")
-        .run();
+        .succeeds();
-    assert!(touch.success);
 
-    let result = ts.ucmd().arg("--time").arg("date_test").run();
+    scene
-
+        .ucmd()
-    // cleanup by removing test file
+        .arg("--time")
-    ts.cmd("rm").arg("date_test").run();
+        .arg("date_test")
-
+        .succeeds()
-    assert!(result.success);
+        .stdout_only("0\t2015-05-15 00:00\tdate_test\n");
-    assert_eq!(result.stderr, "");
-    assert_eq!(result.stdout, "0\t2015-05-15 00:00\tdate_test\n");
 }
 
 #[cfg(not(target_os = "windows"))]
 #[cfg(feature = "chmod")]
 #[test]
 fn test_du_no_permission() {
-    let ts = TestScenario::new("du");
+    let scene = TestScenario::new(util_name!());
+    let at = &scene.fixtures;
 
-    let chmod = ts.ccmd("chmod").arg("-r").arg(SUB_DIR_LINKS).run();
+    at.mkdir_all(SUB_DIR_LINKS);
-    println!("chmod output: {:?}", chmod);
-    assert!(chmod.success);
-    let result = ts.ucmd().arg(SUB_DIR_LINKS).run();
 
-    ts.ccmd("chmod").arg("+r").arg(SUB_DIR_LINKS).run();
+    scene.ccmd("chmod").arg("-r").arg(SUB_DIR_LINKS).succeeds();
 
-    assert!(result.success);
+    let result = scene.ucmd().arg(SUB_DIR_LINKS).run(); // TODO: replace with ".fails()" once `du` is fixed
-    assert_eq!(
+    result.stderr_contains(
-        result.stderr,
+        "du: cannot read directory ‘subdir/links‘: Permission denied (os error 13)",
-        "du: cannot read directory ‘subdir/links‘: Permission denied (os error 13)\n"
     );
-    _du_no_permission(result.stdout);
+
+    #[cfg(target_os = "linux")]
+    {
+        let result_reference = scene.cmd("du").arg(SUB_DIR_LINKS).fails();
+        if result_reference
+            .stderr_str()
+            .contains("du: cannot read directory 'subdir/links': Permission denied")
+        {
+            assert_eq!(result.stdout_str(), result_reference.stdout_str());
+            return;
+        }
+    }
+
+    _du_no_permission(result.stdout_str());
 }
 
 #[cfg(target_vendor = "apple")]
-fn _du_no_permission(s: String) {
+fn _du_no_permission(s: &str) {
     assert_eq!(s, "0\tsubdir/links\n");
 }
 #[cfg(all(not(target_vendor = "apple"), not(target_os = "windows")))]
-fn _du_no_permission(s: String) {
+fn _du_no_permission(s: &str) {
     assert_eq!(s, "4\tsubdir/links\n");
 }

tests/by-util/test_env.rs

@@ -140,8 +140,11 @@ fn test_unset_variable() {
 
 #[test]
 fn test_fail_null_with_program() {
-    let out = new_ucmd!().arg("--null").arg("cd").fails().stderr;
+    new_ucmd!()
-    assert!(out.contains("cannot specify --null (-0) with command"));
+        .arg("--null")
+        .arg("cd")
+        .fails()
+        .stderr_contains("cannot specify --null (-0) with command");
 }
 
 #[cfg(not(windows))]

tests/by-util/test_fmt.rs

@@ -29,7 +29,7 @@ fn test_fmt_w_too_big() {
         .run();
     //.stdout_is_fixture("call_graph.expected");
     assert_eq!(
-        result.stderr.trim(),
+        result.stderr_str().trim(),
         "fmt: error: invalid width: '2501': Numerical result out of range"
     );
 }

tests/by-util/test_groups.rs

@@ -10,7 +10,7 @@ fn test_groups() {
         // As seems to be a configuration issue, ignoring it
         return;
     }
-    assert!(result.success);
+    result.success();
     assert!(!result.stdout_str().trim().is_empty());
 }
 
@@ -30,16 +30,12 @@ fn test_groups_arg() {
 
     println!("result.stdout = {}", result.stdout_str());
     println!("result.stderr = {}", result.stderr_str());
-    assert!(result.success);
+    result.success();
     assert!(!result.stdout_str().is_empty());
     let username = result.stdout_str().trim();
 
     // call groups with the user name to check that we
     // are getting something
-    let (_, mut ucmd) = at_and_ucmd!();
+    new_ucmd!().arg(username).succeeds();
-    let result = ucmd.arg(username).run();
-    println!("result.stdout = {}", result.stdout_str());
-    println!("result.stderr = {}", result.stderr_str());
-    assert!(result.success);
     assert!(!result.stdout_str().is_empty());
 }

tests/by-util/test_head.rs

@@ -156,14 +156,10 @@ fn test_negative_zero_bytes() {
 }
 #[test]
 fn test_no_such_file_or_directory() {
-    let result = new_ucmd!().arg("no_such_file.toml").run();
+    new_ucmd!()
-
+        .arg("no_such_file.toml")
-    assert_eq!(
+        .fails()
-        true,
+        .stderr_contains("cannot open 'no_such_file.toml' for reading: No such file or directory");
-        result
-            .stderr
-            .contains("cannot open 'no_such_file.toml' for reading: No such file or directory")
-    )
 }
 
 // there was a bug not caught by previous tests

tests/by-util/test_install.rs

@@ -11,12 +11,10 @@ use std::thread::sleep;
 fn test_install_help() {
     let (_, mut ucmd) = at_and_ucmd!();
 
-    assert!(ucmd
+    ucmd.arg("--help")
-        .arg("--help")
         .succeeds()
         .no_stderr()
-        .stdout
+        .stdout_contains("FLAGS:");
-        .contains("FLAGS:"));
 }
 
 #[test]
@@ -59,13 +57,11 @@ fn test_install_failing_not_dir() {
     at.touch(file1);
     at.touch(file2);
     at.touch(file3);
-    assert!(ucmd
+    ucmd.arg(file1)
-        .arg(file1)
         .arg(file2)
         .arg(file3)
         .fails()
-        .stderr
+        .stderr_contains("not a directory");
-        .contains("not a directory"));
 }
 
 #[test]
@@ -77,13 +73,11 @@ fn test_install_unimplemented_arg() {
 
     at.touch(file);
     at.mkdir(dir);
-    assert!(ucmd
+    ucmd.arg(context_arg)
-        .arg(context_arg)
         .arg(file)
         .arg(dir)
         .fails()
-        .stderr
+        .stderr_contains("Unimplemented");
-        .contains("Unimplemented"));
 
     assert!(!at.file_exists(&format!("{}/{}", dir, file)));
 }
@@ -231,13 +225,11 @@ fn test_install_mode_failing() {
 
     at.touch(file);
     at.mkdir(dir);
-    assert!(ucmd
+    ucmd.arg(file)
-        .arg(file)
         .arg(dir)
         .arg(mode_arg)
         .fails()
-        .stderr
+        .stderr_contains("Invalid mode string: invalid digit found in string");
-        .contains("Invalid mode string: invalid digit found in string"));
 
     let dest_file = &format!("{}/{}", dir, file);
     assert!(at.file_exists(file));
@@ -336,7 +328,7 @@ fn test_install_target_new_file_with_owner() {
         .arg(format!("{}/{}", dir, file))
         .run();
 
-    if is_ci() && result.stderr.contains("error: no such user:") {
+    if is_ci() && result.stderr_str().contains("error: no such user:") {
         // In the CI, some server are failing to return the user id.
         // As seems to be a configuration issue, ignoring it
         return;
@@ -359,7 +351,7 @@ fn test_install_target_new_file_failing_nonexistent_parent() {
     ucmd.arg(file1)
         .arg(format!("{}/{}", dir, file2))
         .fails()
-        .stderr_contains(&"not a directory");
+        .stderr_contains(&"No such file or directory");
 }
 
 #[test]
@@ -619,33 +611,64 @@ fn test_install_and_strip_with_program() {
 #[test]
 #[cfg(not(windows))]
 fn test_install_and_strip_with_invalid_program() {
-    let scene = TestScenario::new(util_name!());
+    new_ucmd!()
-
-    let stderr = scene
-        .ucmd()
         .arg("-s")
         .arg("--strip-program")
         .arg("/bin/date")
         .arg(strip_source_file())
         .arg(STRIP_TARGET_FILE)
         .fails()
-        .stderr;
+        .stderr_contains("strip program failed");
-    assert!(stderr.contains("strip program failed"));
 }
 
 #[test]
 #[cfg(not(windows))]
 fn test_install_and_strip_with_non_existent_program() {
-    let scene = TestScenario::new(util_name!());
+    new_ucmd!()
-
-    let stderr = scene
-        .ucmd()
         .arg("-s")
         .arg("--strip-program")
         .arg("/usr/bin/non_existent_program")
         .arg(strip_source_file())
         .arg(STRIP_TARGET_FILE)
         .fails()
-        .stderr;
+        .stderr_contains("No such file or directory");
-    assert!(stderr.contains("No such file or directory"));
+}
+
+#[test]
+fn test_install_creating_leading_dirs() {
+    let scene = TestScenario::new(util_name!());
+    let at = &scene.fixtures;
+
+    let source = "create_leading_test_file";
+    let target = "dir1/dir2/dir3/test_file";
+
+    at.touch(source);
+
+    scene
+        .ucmd()
+        .arg("-D")
+        .arg(source)
+        .arg(at.plus(target))
+        .succeeds()
+        .no_stderr();
+}
+
+#[test]
+#[cfg(not(windows))]
+fn test_install_creating_leading_dir_fails_on_long_name() {
+    let scene = TestScenario::new(util_name!());
+    let at = &scene.fixtures;
+
+    let source = "create_leading_test_file";
+    let target = format!("{}/test_file", "d".repeat(libc::PATH_MAX as usize + 1));
+
+    at.touch(source);
+
+    scene
+        .ucmd()
+        .arg("-D")
+        .arg(source)
+        .arg(at.plus(target.as_str()))
+        .fails()
+        .stderr_contains("failed to create");
 }

tests/by-util/test_ln.rs

@@ -299,13 +299,11 @@ fn test_symlink_overwrite_dir_fail() {
     at.touch(path_a);
     at.mkdir(path_b);
 
-    assert!(
+    assert!(!ucmd
-        ucmd.args(&["-s", "-T", path_a, path_b])
+        .args(&["-s", "-T", path_a, path_b])
         .fails()
-            .stderr
+        .stderr_str()
-            .len()
+        .is_empty());
-            > 0
-    );
 }
 
 #[test]
@@ -358,7 +356,11 @@ fn test_symlink_target_only() {
 
     at.mkdir(dir);
 
-    assert!(ucmd.args(&["-s", "-t", dir]).fails().stderr.len() > 0);
+    assert!(!ucmd
+        .args(&["-s", "-t", dir])
+        .fails()
+        .stderr_str()
+        .is_empty());
 }
 
 #[test]

tests/by-util/test_logname.rs

@@ -9,7 +9,7 @@ fn test_normal() {
     for (key, value) in env::vars() {
         println!("{}: {}", key, value);
     }
-    if (is_ci() || is_wsl()) && result.stderr_str().contains("error: no login name") {
+    if (is_ci() || uucore::os::is_wsl_1()) && result.stderr_str().contains("error: no login name") {
         // ToDO: investigate WSL failure
         // In the CI, some server are failing to return logname.
         // As seems to be a configuration issue, ignoring it

tests/by-util/test_ls.rs

@@ -103,6 +103,20 @@ fn test_ls_width() {
             .succeeds()
             .stdout_only("test-width-1\ntest-width-2\ntest-width-3\ntest-width-4\n");
     }
+
+    scene
+        .ucmd()
+        .arg("-w=bad")
+        .fails()
+        .stderr_contains("invalid line width");
+
+    for option in &["-w 1a", "-w=1a", "--width=1a", "--width 1a"] {
+        scene
+            .ucmd()
+            .args(&option.split(" ").collect::<Vec<_>>())
+            .fails()
+            .stderr_only("ls: error: invalid line width: ‘1a’");
+    }
 }
 
 #[test]
@@ -436,6 +450,39 @@ fn test_ls_deref() {
     assert!(!re.is_match(result.stdout_str().trim()));
 }
 
+#[test]
+fn test_ls_sort_none() {
+    let scene = TestScenario::new(util_name!());
+    let at = &scene.fixtures;
+
+    at.touch("test-3");
+    at.touch("test-1");
+    at.touch("test-2");
+
+    // Order is not specified so we just check that it doesn't
+    // give any errors.
+    scene.ucmd().arg("--sort=none").succeeds();
+    scene.ucmd().arg("-U").succeeds();
+}
+
+#[test]
+fn test_ls_sort_name() {
+    let scene = TestScenario::new(util_name!());
+    let at = &scene.fixtures;
+
+    at.touch("test-3");
+    at.touch("test-1");
+    at.touch("test-2");
+
+    let sep = if cfg!(unix) { "\n" } else { "  " };
+
+    scene
+        .ucmd()
+        .arg("--sort=name")
+        .succeeds()
+        .stdout_is(["test-1", "test-2", "test-3\n"].join(sep));
+}
+
 #[test]
 fn test_ls_order_size() {
     let scene = TestScenario::new(util_name!());
@@ -464,6 +511,18 @@ fn test_ls_order_size() {
     result.stdout_only("test-1\ntest-2\ntest-3\ntest-4\n");
     #[cfg(windows)]
     result.stdout_only("test-1  test-2  test-3  test-4\n");
+
+    let result = scene.ucmd().arg("--sort=size").succeeds();
+    #[cfg(not(windows))]
+    result.stdout_only("test-4\ntest-3\ntest-2\ntest-1\n");
+    #[cfg(windows)]
+    result.stdout_only("test-4  test-3  test-2  test-1\n");
+
+    let result = scene.ucmd().arg("--sort=size").arg("-r").succeeds();
+    #[cfg(not(windows))]
+    result.stdout_only("test-1\ntest-2\ntest-3\ntest-4\n");
+    #[cfg(windows)]
+    result.stdout_only("test-1  test-2  test-3  test-4\n");
 }
 
 #[test]
@@ -472,13 +531,16 @@ fn test_ls_long_ctime() {
     let at = &scene.fixtures;
 
     at.touch("test-long-ctime-1");
-    let result = scene.ucmd().arg("-lc").succeeds();
+
+    for arg in &["-c", "--time=ctime", "--time=status"] {
+        let result = scene.ucmd().arg("-l").arg(arg).succeeds();
 
         // Should show the time on Unix, but question marks on windows.
         #[cfg(unix)]
         result.stdout_contains(":");
         #[cfg(not(unix))]
         result.stdout_contains("???");
+    }
 }
 
 #[test]
@@ -519,15 +581,28 @@ fn test_ls_order_time() {
     #[cfg(windows)]
     result.stdout_only("test-4  test-3  test-2  test-1\n");
 
+    let result = scene.ucmd().arg("--sort=time").succeeds();
+    #[cfg(not(windows))]
+    result.stdout_only("test-4\ntest-3\ntest-2\ntest-1\n");
+    #[cfg(windows)]
+    result.stdout_only("test-4  test-3  test-2  test-1\n");
+
     let result = scene.ucmd().arg("-tr").succeeds();
     #[cfg(not(windows))]
     result.stdout_only("test-1\ntest-2\ntest-3\ntest-4\n");
     #[cfg(windows)]
     result.stdout_only("test-1  test-2  test-3  test-4\n");
 
+    let result = scene.ucmd().arg("--sort=time").arg("-r").succeeds();
+    #[cfg(not(windows))]
+    result.stdout_only("test-1\ntest-2\ntest-3\ntest-4\n");
+    #[cfg(windows)]
+    result.stdout_only("test-1  test-2  test-3  test-4\n");
+
     // 3 was accessed last in the read
     // So the order should be 2 3 4 1
-    let result = scene.ucmd().arg("-tu").succeeds();
+    for arg in &["-u", "--time=atime", "--time=access", "--time=use"] {
+        let result = scene.ucmd().arg("-t").arg(arg).succeeds();
         let file3_access = at.open("test-3").metadata().unwrap().accessed().unwrap();
         let file4_access = at.open("test-4").metadata().unwrap().accessed().unwrap();
 
@@ -547,6 +622,7 @@ fn test_ls_order_time() {
                 result.stdout_only("test-4  test-3  test-2  test-1\n");
             }
         }
+    }
 
     // test-2 had the last ctime change when the permissions were set
     // So the order should be 2 4 3 1
@@ -621,20 +697,27 @@ fn test_ls_recursive() {
     result.stdout_contains(&"a\\b:\nb");
 }
 
-#[cfg(unix)]
 #[test]
-fn test_ls_ls_color() {
+fn test_ls_color() {
     let scene = TestScenario::new(util_name!());
     let at = &scene.fixtures;
     at.mkdir("a");
-    at.mkdir("a/nested_dir");
+    let nested_dir = Path::new("a")
+        .join("nested_dir")
+        .to_string_lossy()
+        .to_string();
+    at.mkdir(&nested_dir);
     at.mkdir("z");
-    at.touch(&at.plus_as_string("a/nested_file"));
+    let nested_file = Path::new("a")
+        .join("nested_file")
+        .to_string_lossy()
+        .to_string();
+    at.touch(&nested_file);
     at.touch("test-color");
 
-    let a_with_colors = "\x1b[01;34ma\x1b[0m";
+    let a_with_colors = "\x1b[1;34ma\x1b[0m";
-    let z_with_colors = "\x1b[01;34mz\x1b[0m";
+    let z_with_colors = "\x1b[1;34mz\x1b[0m";
-    let nested_dir_with_colors = "\x1b[01;34mnested_dir\x1b[0m";
+    let nested_dir_with_colors = "\x1b[1;34mnested_dir\x1b[0m";
 
     // Color is disabled by default
     let result = scene.ucmd().succeeds();
@@ -670,14 +753,6 @@ fn test_ls_ls_color() {
         .succeeds()
         .stdout_contains(nested_dir_with_colors);
 
-    // Color has no effect
-    scene
-        .ucmd()
-        .arg("--color=always")
-        .arg("a/nested_file")
-        .succeeds()
-        .stdout_contains("a/nested_file\n");
-
     // No output
     scene
         .ucmd()
@@ -817,7 +892,7 @@ fn test_ls_indicator_style() {
     let options = vec!["classify", "file-type", "slash"];
     for opt in options {
         // Verify that classify and file-type both contain indicators for symlinks.
-        let result = scene
+        scene
             .ucmd()
             .arg(format!("--indicator-style={}", opt))
             .succeeds()
@@ -827,7 +902,7 @@ fn test_ls_indicator_style() {
     // Same test as above, but with the alternate flags.
     let options = vec!["--classify", "--file-type", "-p"];
     for opt in options {
-        let result = scene
+        scene
             .ucmd()
             .arg(format!("{}", opt))
             .succeeds()
@@ -838,7 +913,7 @@ fn test_ls_indicator_style() {
     let options = vec!["classify", "file-type"];
     for opt in options {
         // Verify that classify and file-type both contain indicators for symlinks.
-        let result = scene
+        scene
             .ucmd()
             .arg(format!("--indicator-style={}", opt))
             .succeeds()
@@ -962,7 +1037,7 @@ fn test_ls_hidden_windows() {
 
     let result = scene.ucmd().succeeds();
     assert!(!result.stdout_str().contains(file));
-    let result = scene.ucmd().arg("-a").succeeds().stdout_contains(file);
+    scene.ucmd().arg("-a").succeeds().stdout_contains(file);
 }
 
 #[test]
@@ -1052,9 +1127,11 @@ fn test_ls_quoting_style() {
     at.touch("one");
 
     // It seems that windows doesn't allow \n in filenames.
+    // And it also doesn't like \, of course.
     #[cfg(unix)]
     {
         at.touch("one\ntwo");
+        at.touch("one\\two");
         // Default is shell-escape
         scene
             .ucmd()
@@ -1116,6 +1193,42 @@ fn test_ls_quoting_style() {
                 .succeeds()
                 .stdout_only(format!("{}\n", correct));
         }
+
+        for (arg, correct) in &[
+            ("--quoting-style=literal", "one\\two"),
+            ("-N", "one\\two"),
+            ("--quoting-style=c", "\"one\\\\two\""),
+            ("-Q", "\"one\\\\two\""),
+            ("--quote-name", "\"one\\\\two\""),
+            ("--quoting-style=escape", "one\\\\two"),
+            ("-b", "one\\\\two"),
+            ("--quoting-style=shell-escape", "'one\\two'"),
+            ("--quoting-style=shell-escape-always", "'one\\two'"),
+            ("--quoting-style=shell", "'one\\two'"),
+            ("--quoting-style=shell-always", "'one\\two'"),
+        ] {
+            scene
+                .ucmd()
+                .arg(arg)
+                .arg("one\\two")
+                .succeeds()
+                .stdout_only(format!("{}\n", correct));
+        }
+
+        // Tests for a character that forces quotation in shell-style escaping
+        // after a character in a dollar expression
+        at.touch("one\n&two");
+        for (arg, correct) in &[
+            ("--quoting-style=shell-escape", "'one'$'\\n''&two'"),
+            ("--quoting-style=shell-escape-always", "'one'$'\\n''&two'"),
+        ] {
+            scene
+                .ucmd()
+                .arg(arg)
+                .arg("one\n&two")
+                .succeeds()
+                .stdout_only(format!("{}\n", correct));
+        }
     }
 
     scene
@@ -1316,6 +1429,43 @@ fn test_ls_ignore_hide() {
         .stdout_is("CONTRIBUTING.md\nREADME.md\nREADMECAREFULLY.md\nsome_other_file\n");
 }
 
+#[test]
+fn test_ls_ignore_backups() {
+    let scene = TestScenario::new(util_name!());
+    let at = &scene.fixtures;
+
+    at.touch("somefile");
+    at.touch("somebackup~");
+    at.touch(".somehiddenfile");
+    at.touch(".somehiddenbackup~");
+
+    scene.ucmd().arg("-B").succeeds().stdout_is("somefile\n");
+    scene
+        .ucmd()
+        .arg("--ignore-backups")
+        .succeeds()
+        .stdout_is("somefile\n");
+
+    scene
+        .ucmd()
+        .arg("-aB")
+        .succeeds()
+        .stdout_contains(".somehiddenfile")
+        .stdout_contains("somefile")
+        .stdout_does_not_contain("somebackup")
+        .stdout_does_not_contain(".somehiddenbackup~");
+
+    scene
+        .ucmd()
+        .arg("-a")
+        .arg("--ignore-backups")
+        .succeeds()
+        .stdout_contains(".somehiddenfile")
+        .stdout_contains("somefile")
+        .stdout_does_not_contain("somebackup")
+        .stdout_does_not_contain(".somehiddenbackup~");
+}
+
 #[test]
 fn test_ls_directory() {
     let scene = TestScenario::new(util_name!());

tests/by-util/test_mkfifo.rs

@@ -19,8 +19,7 @@ fn test_create_one_fifo_with_invalid_mode() {
         .arg("-m")
         .arg("invalid")
         .fails()
-        .stderr
+        .stderr_contains("invalid mode");
-        .contains("invalid mode");
 }
 
 #[test]

tests/by-util/test_more.rs

@@ -2,18 +2,15 @@ use crate::common::util::*;
 
 #[test]
 fn test_more_no_arg() {
-    let (_, mut ucmd) = at_and_ucmd!();
+    // stderr = more: Reading from stdin isn't supported yet.
-    let result = ucmd.run();
+    new_ucmd!().fails();
-    assert!(!result.success);
 }
 
 #[test]
 fn test_more_dir_arg() {
-    let (_, mut ucmd) = at_and_ucmd!();
+    let result = new_ucmd!().arg(".").run();
-    ucmd.arg(".");
+    result.failure();
-    let result = ucmd.run();
-    assert!(!result.success);
     const EXPECTED_ERROR_MESSAGE: &str =
         "more: '.' is a directory.\nTry 'more --help' for more information.";
-    assert_eq!(result.stderr.trim(), EXPECTED_ERROR_MESSAGE);
+    assert_eq!(result.stderr_str().trim(), EXPECTED_ERROR_MESSAGE);
 }

tests/by-util/test_mv.rs

@@ -476,16 +476,9 @@ fn test_mv_overwrite_nonempty_dir() {
     // GNU:     "mv: cannot move ‘a’ to ‘b’: Directory not empty"
 
     // Verbose output for the move should not be shown on failure
-    assert!(
+    let result = ucmd.arg("-vT").arg(dir_a).arg(dir_b).fails();
-        ucmd.arg("-vT")
+    result.no_stdout();
-            .arg(dir_a)
+    assert!(!result.stderr_str().is_empty());
-            .arg(dir_b)
-            .fails()
-            .no_stdout()
-            .stderr
-            .len()
-            > 0
-    );
 
     assert!(at.dir_exists(dir_a));
     assert!(at.dir_exists(dir_b));
@@ -526,15 +519,15 @@ fn test_mv_errors() {
 
     // $ mv -T -t a b
     // mv: cannot combine --target-directory (-t) and --no-target-directory (-T)
-    let result = scene
+    scene
         .ucmd()
         .arg("-T")
         .arg("-t")
         .arg(dir)
         .arg(file_a)
         .arg(file_b)
-        .fails();
+        .fails()
-    assert!(result.stderr.contains("cannot be used with"));
+        .stderr_contains("cannot be used with");
 
     // $ at.touch file && at.mkdir dir
     // $ mv -T file dir
@@ -553,7 +546,13 @@ fn test_mv_errors() {
     // $ at.mkdir dir && at.touch file
     // $ mv dir file
     // err == mv: cannot overwrite non-directory ‘file’ with directory ‘dir’
-    assert!(scene.ucmd().arg(dir).arg(file_a).fails().stderr.len() > 0);
+    assert!(!scene
+        .ucmd()
+        .arg(dir)
+        .arg(file_a)
+        .fails()
+        .stderr_str()
+        .is_empty());
 }
 
 #[test]

tests/by-util/test_nice.rs

@@ -16,7 +16,7 @@ fn test_negative_adjustment() {
 
     let res = new_ucmd!().args(&["-n", "-1", "true"]).run();
     assert!(res
-        .stderr
+        .stderr_str()
         .starts_with("nice: warning: setpriority: Permission denied"));
 }
 

tests/by-util/test_shuf.rs

@@ -9,35 +9,28 @@ fn test_output_is_random_permutation() {
         .collect::<Vec<String>>()
         .join("\n");
 
-    let result = new_ucmd!()
+    let result = new_ucmd!().pipe_in(input.as_bytes()).succeeds();
-        .pipe_in(input.as_bytes())
+    result.no_stderr();
-        .succeeds()
-        .no_stderr()
-        .stdout
-        .clone();
 
     let mut result_seq: Vec<i32> = result
+        .stdout_str()
         .split("\n")
         .filter(|x| !x.is_empty())
         .map(|x| x.parse().unwrap())
         .collect();
     result_seq.sort();
-    assert_ne!(result, input, "Output is not randomised");
+    assert_ne!(result.stdout_str(), input, "Output is not randomised");
     assert_eq!(result_seq, input_seq, "Output is not a permutation");
 }
 
 #[test]
 fn test_zero_termination() {
     let input_seq = vec![1, 2, 3, 4, 5, 6, 7, 8, 9, 10];
-    let result = new_ucmd!()
+    let result = new_ucmd!().arg("-z").arg("-i1-10").succeeds();
-        .arg("-z")
+    result.no_stderr();
-        .arg("-i1-10")
-        .succeeds()
-        .no_stderr()
-        .stdout
-        .clone();
 
     let mut result_seq: Vec<i32> = result
+        .stdout_str()
         .split("\0")
         .filter(|x| !x.is_empty())
         .map(|x| x.parse().unwrap())
@@ -57,12 +50,11 @@ fn test_echo() {
                 .map(|x| x.to_string())
                 .collect::<Vec<String>>(),
         )
-        .succeeds()
+        .succeeds();
-        .no_stderr()
+    result.no_stderr();
-        .stdout
-        .clone();
 
     let mut result_seq: Vec<i32> = result
+        .stdout_str()
         .split("\n")
         .filter(|x| !x.is_empty())
         .map(|x| x.parse().unwrap())
@@ -84,12 +76,11 @@ fn test_head_count() {
     let result = new_ucmd!()
         .args(&["-n", &repeat_limit.to_string()])
         .pipe_in(input.as_bytes())
-        .succeeds()
+        .succeeds();
-        .no_stderr()
+    result.no_stderr();
-        .stdout
-        .clone();
 
     let mut result_seq: Vec<i32> = result
+        .stdout_str()
         .split("\n")
         .filter(|x| !x.is_empty())
         .map(|x| x.parse().unwrap())
@@ -99,7 +90,7 @@ fn test_head_count() {
     assert!(
         result_seq.iter().all(|x| input_seq.contains(x)),
         "Output includes element not from input: {}",
-        result
+        result.stdout_str()
     )
 }
 
@@ -117,12 +108,11 @@ fn test_repeat() {
         .arg("-r")
         .args(&["-n", &repeat_limit.to_string()])
         .pipe_in(input.as_bytes())
-        .succeeds()
+        .succeeds();
-        .no_stderr()
+    result.no_stderr();
-        .stdout
-        .clone();
 
     let result_seq: Vec<i32> = result
+        .stdout_str()
         .split("\n")
         .filter(|x| !x.is_empty())
         .map(|x| x.parse().unwrap())
@@ -146,14 +136,11 @@ fn test_repeat() {
 fn test_file_input() {
     let expected_seq = vec![11, 12, 13, 14, 15, 16, 17, 18, 19, 20];
 
-    let result = new_ucmd!()
+    let result = new_ucmd!().arg("file_input.txt").succeeds();
-        .arg("file_input.txt")
+    result.no_stderr();
-        .succeeds()
-        .no_stderr()
-        .stdout
-        .clone();
 
     let mut result_seq: Vec<i32> = result
+        .stdout_str()
         .split("\n")
         .filter(|x| !x.is_empty())
         .map(|x| x.parse().unwrap())
@@ -164,52 +151,50 @@ fn test_file_input() {
 
 #[test]
 fn test_shuf_echo_and_input_range_not_allowed() {
-    let result = new_ucmd!().args(&["-e", "0", "-i", "0-2"]).run();
+    new_ucmd!()
-
+        .args(&["-e", "0", "-i", "0-2"])
-    assert!(!result.success);
+        .fails()
-    assert!(result
+        .stderr_contains(
-        .stderr
+            "The argument '--input-range <LO-HI>' cannot be used with '--echo <ARG>...'",
-        .contains("The argument '--input-range <LO-HI>' cannot be used with '--echo <ARG>...'"));
+        );
 }
 
 #[test]
 fn test_shuf_input_range_and_file_not_allowed() {
-    let result = new_ucmd!().args(&["-i", "0-9", "file"]).run();
+    new_ucmd!()
-
+        .args(&["-i", "0-9", "file"])
-    assert!(!result.success);
+        .fails()
-    assert!(result
+        .stderr_contains("The argument '<file>' cannot be used with '--input-range <LO-HI>'");
-        .stderr
-        .contains("The argument '<file>' cannot be used with '--input-range <LO-HI>'"));
 }
 
 #[test]
 fn test_shuf_invalid_input_range_one() {
-    let result = new_ucmd!().args(&["-i", "0"]).run();
+    new_ucmd!()
-
+        .args(&["-i", "0"])
-    assert!(!result.success);
+        .fails()
-    assert!(result.stderr.contains("invalid input range"));
+        .stderr_contains("invalid input range");
 }
 
 #[test]
 fn test_shuf_invalid_input_range_two() {
-    let result = new_ucmd!().args(&["-i", "a-9"]).run();
+    new_ucmd!()
-
+        .args(&["-i", "a-9"])
-    assert!(!result.success);
+        .fails()
-    assert!(result.stderr.contains("invalid input range: 'a'"));
+        .stderr_contains("invalid input range: 'a'");
 }
 
 #[test]
 fn test_shuf_invalid_input_range_three() {
-    let result = new_ucmd!().args(&["-i", "0-b"]).run();
+    new_ucmd!()
-
+        .args(&["-i", "0-b"])
-    assert!(!result.success);
+        .fails()
-    assert!(result.stderr.contains("invalid input range: 'b'"));
+        .stderr_contains("invalid input range: 'b'");
 }
 
 #[test]
 fn test_shuf_invalid_input_line_count() {
-    let result = new_ucmd!().args(&["-n", "a"]).run();
+    new_ucmd!()
-
+        .args(&["-n", "a"])
-    assert!(!result.success);
+        .fails()
-    assert!(result.stderr.contains("invalid line count: 'a'"));
+        .stderr_contains("invalid line count: 'a'");
 }

tests/by-util/test_sort.rs

@@ -2,10 +2,17 @@ use crate::common::util::*;
 
 fn test_helper(file_name: &str, args: &str) {
     new_ucmd!()
-        .arg(args)
         .arg(format!("{}.txt", file_name))
+        .args(&args.split(' ').collect::<Vec<&str>>())
         .succeeds()
         .stdout_is_fixture(format!("{}.expected", file_name));
+
+    new_ucmd!()
+        .arg(format!("{}.txt", file_name))
+        .arg("--debug")
+        .args(&args.split(' ').collect::<Vec<&str>>())
+        .succeeds()
+        .stdout_is_fixture(format!("{}.expected.debug", file_name));
 }
 
 // FYI, the initialization size of our Line struct is 96 bytes.
@@ -65,11 +72,7 @@ fn test_extsort_as64_bailout() {
 
 #[test]
 fn test_multiple_decimals_general() {
-    new_ucmd!()
+    test_helper("multiple_decimals_general", "-g")
-        .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]
@@ -99,7 +102,7 @@ fn test_check_zero_terminated_success() {
 #[test]
 fn test_random_shuffle_len() {
     // check whether output is the same length as the input
-    const FILE: &'static str = "default_unsorted_ints.expected";
+    const FILE: &str = "default_unsorted_ints.expected";
     let (at, _ucmd) = at_and_ucmd!();
     let result = new_ucmd!().arg("-R").arg(FILE).run().stdout_move_str();
     let expected = at.read(FILE);
@@ -111,7 +114,7 @@ fn test_random_shuffle_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";
+    const FILE: &str = "default_unsorted_ints.expected";
     let (at, _ucmd) = at_and_ucmd!();
     let result = new_ucmd!().arg("-R").arg(FILE).run().stdout_move_str();
     let expected = at.read(FILE);
@@ -126,7 +129,22 @@ 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";
+    const FILE: &str = "default_unsorted_ints.expected";
+    let (at, _ucmd) = at_and_ucmd!();
+    let result = new_ucmd!().arg("-R").arg(FILE).run().stdout_move_str();
+    let expected = at.read(FILE);
+    let unexpected = new_ucmd!().arg("-R").arg(FILE).run().stdout_move_str();
+
+    assert_ne!(result, expected);
+    assert_ne!(result, unexpected);
+}
+
+#[test]
+fn test_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: &str = "default_unsorted_ints.expected";
     let (at, _ucmd) = at_and_ucmd!();
     let result = new_ucmd!().arg("-R").arg(FILE).run().stdout_move_str();
     let expected = at.read(FILE);
@@ -228,6 +246,11 @@ fn test_non_printing_chars() {
     }
 }
 
+#[test]
+fn test_exponents_positive_general_fixed() {
+    test_helper("exponents_general", "-g");
+}
+
 #[test]
 fn test_exponents_positive_numeric() {
     test_helper("exponents-positive-numeric", "-n");
@@ -344,62 +367,32 @@ fn test_numeric_unique_ints2() {
 
 #[test]
 fn test_keys_open_ended() {
-    let input = "aa bb cc\ndd aa ff\ngg aa cc\n";
+    test_helper("keys_open_ended", "-k 2.3");
-    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";
+    test_helper("keys_closed_range", "-k 2.2,2.2");
-    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";
+    test_helper("keys_multiple_ranges", "-k 2,2 -k 3,3");
-    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";
+    test_helper("keys_no_field_match", "-k 4,4");
-    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";
+    test_helper("keys_no_char_match", "-k 1.2");
-    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";
+    test_helper("keys_custom_separator", "-k 2.2,2.2 -t x");
-    new_ucmd!()
-        .args(&["-k", "2.2,2.2", "-t", "x"])
-        .pipe_in(input)
-        .succeeds()
-        .stdout_only("ddxaaxff\nggxaaxcc\naaxbbxcc\n");
 }
 
 #[test]
@@ -426,6 +419,13 @@ fn test_keys_invalid_field_zero() {
         .stderr_only("sort: error: field index was 0");
 }
 
+#[test]
+fn test_keys_invalid_char_zero() {
+    new_ucmd!().args(&["-k", "1.0"]).fails().stderr_only(
+        "sort: error: invalid character index 0 in `1.0` for the start position of a field",
+    );
+}
+
 #[test]
 fn test_keys_with_options() {
     let input = "aa 3 cc\ndd 1 ff\ngg 2 cc\n";
@@ -591,3 +591,39 @@ fn test_check_silent() {
         .fails()
         .stdout_is("");
 }
+
+#[test]
+fn test_dictionary_and_nonprinting_conflicts() {
+    let conflicting_args = ["n", "h", "g", "M"];
+    for restricted_arg in &["d", "i"] {
+        for conflicting_arg in &conflicting_args {
+            new_ucmd!()
+                .arg(&format!("-{}{}", restricted_arg, conflicting_arg))
+                .fails();
+        }
+        for conflicting_arg in &conflicting_args {
+            new_ucmd!()
+                .args(&[
+                    format!("-{}", restricted_arg).as_str(),
+                    "-k",
+                    &format!("1,1{}", conflicting_arg),
+                ])
+                .succeeds();
+        }
+        for conflicting_arg in &conflicting_args {
+            // FIXME: this should ideally fail.
+            new_ucmd!()
+                .args(&["-k", &format!("1{},1{}", restricted_arg, conflicting_arg)])
+                .succeeds();
+        }
+    }
+}
+
+#[test]
+fn test_trailing_separator() {
+    new_ucmd!()
+        .args(&["-t", "x", "-k", "1,1"])
+        .pipe_in("aax\naaa\n")
+        .succeeds()
+        .stdout_is("aax\naaa\n");
+}

tests/by-util/test_stat.rs

@@ -337,5 +337,5 @@ fn expected_result(args: &[&str]) -> String {
         .env("LANGUAGE", "C")
         .args(args)
         .run()
-        .stdout
+        .stdout_move_str()
 }

tests/by-util/test_tac.rs

@@ -52,18 +52,19 @@ fn test_single_non_newline_separator_before() {
 
 #[test]
 fn test_invalid_input() {
-    let (_, mut ucmd) = at_and_ucmd!();
+    let scene = TestScenario::new(util_name!());
+    let at = &scene.fixtures;
 
-    ucmd.arg("b")
+    scene
-        .run()
+        .ucmd()
-        .stderr
+        .arg("b")
-        .contains("tac: error: failed to open 'b' for reading");
+        .fails()
-
+        .stderr_contains("failed to open 'b' for reading: No such file or directory");
-    let (at, mut ucmd) = at_and_ucmd!();
 
     at.mkdir("a");
-    ucmd.arg("a")
+    scene
-        .run()
+        .ucmd()
-        .stderr
+        .arg("a")
-        .contains("tac: error: failed to read 'a'");
+        .fails()
+        .stderr_contains("dir: read error: Invalid argument");
 }

tests/by-util/test_tail.rs

@@ -343,3 +343,8 @@ fn test_negative_indexing() {
     assert_eq!(positive_lines_index.stdout(), negative_lines_index.stdout());
     assert_eq!(positive_bytes_index.stdout(), negative_bytes_index.stdout());
 }
+
+#[test]
+fn test_sleep_interval() {
+    new_ucmd!().arg("-s").arg("10").arg(FOOBAR_TXT).succeeds();
+}

tests/by-util/test_touch.rs

@@ -367,7 +367,58 @@ fn test_touch_mtime_dst_succeeds() {
 
     let target_time = str_to_filetime("%Y%m%d%H%M", "202103140300");
     let (_, mtime) = get_file_times(&at, file);
-    eprintln!("target_time: {:?}", target_time);
-    eprintln!("mtime: {:?}", mtime);
     assert!(target_time == mtime);
 }
+
+// is_dst_switch_hour returns true if timespec ts is just before the switch
+// to Daylight Saving Time.
+// For example, in EST (UTC-5), Timespec { sec: 1583647200, nsec: 0 }
+// for March 8 2020 01:00:00 AM
+// is just before the switch because on that day clock jumps by 1 hour,
+// so 1 minute after 01:59:00 is 03:00:00.
+fn is_dst_switch_hour(ts: time::Timespec) -> bool {
+    let ts_after = ts + time::Duration::hours(1);
+    let tm = time::at(ts);
+    let tm_after = time::at(ts_after);
+    tm_after.tm_hour == tm.tm_hour + 2
+}
+
+// get_dstswitch_hour returns date string for which touch -m -t fails.
+// For example, in EST (UTC-5), that will be "202003080200" so
+// touch -m -t 202003080200 somefile
+// fails (that date/time does not exist).
+// In other locales it will be a different date/time, and in some locales
+// it doesn't exist at all, in which case this function will return None.
+fn get_dstswitch_hour() -> Option<String> {
+    let now = time::now();
+    // Start from January 1, 2020, 00:00.
+    let mut tm = time::strptime("20200101-0000", "%Y%m%d-%H%M").unwrap();
+    tm.tm_isdst = -1;
+    tm.tm_utcoff = now.tm_utcoff;
+    let mut ts = tm.to_timespec();
+    // Loop through all hours in year 2020 until we find the hour just
+    // before the switch to DST.
+    for _i in 0..(366 * 24) {
+        if is_dst_switch_hour(ts) {
+            let mut tm = time::at(ts);
+            tm.tm_hour = tm.tm_hour + 1;
+            let s = time::strftime("%Y%m%d%H%M", &tm).unwrap().to_string();
+            return Some(s);
+        }
+        ts = ts + time::Duration::hours(1);
+    }
+    None
+}
+
+#[test]
+fn test_touch_mtime_dst_fails() {
+    let (_at, mut ucmd) = at_and_ucmd!();
+    let file = "test_touch_set_mtime_dst_fails";
+
+    match get_dstswitch_hour() {
+        Some(s) => {
+            ucmd.args(&["-m", "-t", &s, file]).fails();
+        }
+        None => (),
+    }
+}

tests/by-util/test_tr.rs

@@ -120,19 +120,15 @@ fn test_truncate_with_set1_shorter_than_set2() {
 #[test]
 fn missing_args_fails() {
     let (_, mut ucmd) = at_and_ucmd!();
-    let result = ucmd.run();
+    ucmd.fails().stderr_contains("missing operand");
-
-    assert!(!result.success);
-    assert!(result.stderr.contains("missing operand"));
 }
 
 #[test]
 fn missing_required_second_arg_fails() {
     let (_, mut ucmd) = at_and_ucmd!();
-    let result = ucmd.args(&["foo"]).run();
+    ucmd.args(&["foo"])
-
+        .fails()
-    assert!(!result.success);
+        .stderr_contains("missing operand after");
-    assert!(result.stderr.contains("missing operand after"));
 }
 
 #[test]

tests/by-util/test_truncate.rs

@@ -53,6 +53,16 @@ fn test_decrease_file_size() {
     assert!(file.seek(SeekFrom::Current(0)).unwrap() == 6);
 }
 
+#[test]
+fn test_space_in_size() {
+    let (at, mut ucmd) = at_and_ucmd!();
+    let mut file = at.make_file(TFILE2);
+    file.write_all(b"1234567890").unwrap();
+    ucmd.args(&["--size", " 4", TFILE2]).succeeds();
+    file.seek(SeekFrom::End(0)).unwrap();
+    assert!(file.seek(SeekFrom::Current(0)).unwrap() == 4);
+}
+
 #[test]
 fn test_failed() {
     new_ucmd!().fails();

tests/by-util/test_tsort.rs

@@ -18,33 +18,35 @@ fn test_sort_self_loop() {
 
 #[test]
 fn test_no_such_file() {
-    let result = new_ucmd!().arg("invalid_file_txt").run();
+    new_ucmd!()
-
+        .arg("invalid_file_txt")
-    assert_eq!(true, result.stderr.contains("No such file or directory"));
+        .fails()
+        .stderr_contains("No such file or directory");
 }
 
 #[test]
 fn test_version_flag() {
-    let version_short = new_ucmd!().arg("-V").run();
+    let version_short = new_ucmd!().arg("-V").succeeds();
-    let version_long = new_ucmd!().arg("--version").run();
+    let version_long = new_ucmd!().arg("--version").succeeds();
 
-    assert_eq!(version_short.stdout(), version_long.stdout());
+    assert_eq!(version_short.stdout_str(), version_long.stdout_str());
 }
 
 #[test]
 fn test_help_flag() {
-    let help_short = new_ucmd!().arg("-h").run();
+    let help_short = new_ucmd!().arg("-h").succeeds();
-    let help_long = new_ucmd!().arg("--help").run();
+    let help_long = new_ucmd!().arg("--help").succeeds();
 
-    assert_eq!(help_short.stdout(), help_long.stdout());
+    assert_eq!(help_short.stdout_str(), help_long.stdout_str());
 }
 
 #[test]
 fn test_multiple_arguments() {
-    let result = new_ucmd!()
+    new_ucmd!()
         .arg("call_graph.txt")
-        .arg("invalid_file.txt")
+        .arg("invalid_file")
-        .run();
+        .fails()
-
+        .stderr_contains(
-    assert_eq!(true, result.stderr.contains("error: Found argument 'invalid_file.txt' which wasn't expected, or isn't valid in this context"))
+            "Found argument 'invalid_file' which wasn't expected, or isn't valid in this context",
+        );
 }

tests/by-util/test_who.rs

@@ -23,7 +23,7 @@ fn test_heading() {
     for opt in vec!["-H"] {
         // allow whitespace variation
         // * minor whitespace differences occur between platform built-in outputs; specifically number of TABs between "TIME" and "COMMENT" may be variant
-        let actual = new_ucmd!().arg(opt).run().stdout;
+        let actual = new_ucmd!().arg(opt).run().stdout_move_str();
         let expect = expected_result(opt);
         println!("actual: {:?}", actual);
         println!("expect: {:?}", expect);
@@ -80,5 +80,5 @@ fn expected_result(arg: &str) -> String {
         .env("LANGUAGE", "C")
         .args(&[arg])
         .run()
-        .stdout
+        .stdout_move_str()
 }

tests/common/util.rs

@@ -2,6 +2,7 @@
 
 #[cfg(not(windows))]
 use libc;
+use pretty_assertions::assert_eq;
 use std::env;
 #[cfg(not(windows))]
 use std::ffi::CString;
@@ -42,22 +43,6 @@ pub fn is_ci() -> bool {
         .eq_ignore_ascii_case("true")
 }
 
-/// Test if the program is running under WSL
-// ref: <https://github.com/microsoft/WSL/issues/4555> @@ <https://archive.is/dP0bz>
-// ToDO: test on WSL2 which likely doesn't need special handling; plan change to `is_wsl_1()` if WSL2 is less needy
-pub fn is_wsl() -> bool {
-    #[cfg(target_os = "linux")]
-    {
-        if let Ok(b) = std::fs::read("/proc/sys/kernel/osrelease") {
-            if let Ok(s) = std::str::from_utf8(&b) {
-                let a = s.to_ascii_lowercase();
-                return a.contains("microsoft") || a.contains("wsl");
-            }
-        }
-    }
-    false
-}
-
 /// Read a test scenario fixture, returning its bytes
 fn read_scenario_fixture<S: AsRef<OsStr>>(tmpd: &Option<Rc<TempDir>>, file_rel_path: S) -> Vec<u8> {
     let tmpdir_path = tmpd.as_ref().unwrap().as_ref().path();
@@ -74,11 +59,11 @@ pub struct CmdResult {
     code: Option<i32>,
     /// zero-exit from running the Command?
     /// see [`success`]
-    pub success: bool,
+    success: bool,
     /// captured standard output after running the Command
-    pub stdout: String,
+    stdout: String,
     /// captured standard error after running the Command
-    pub stderr: String,
+    stderr: String,
 }
 
 impl CmdResult {
@@ -237,7 +222,7 @@ impl CmdResult {
     /// like stdout_is(...), but expects the contents of the file at the provided relative path
     pub fn stdout_is_fixture<T: AsRef<OsStr>>(&self, file_rel_path: T) -> &CmdResult {
         let contents = read_scenario_fixture(&self.tmpd, file_rel_path);
-        self.stdout_is_bytes(contents)
+        self.stdout_is(String::from_utf8(contents).unwrap())
     }
 
     /// asserts that the command resulted in stderr stream output that equals the
@@ -261,7 +246,7 @@ impl CmdResult {
     /// Like stdout_is_fixture, but for stderr
     pub fn stderr_is_fixture<T: AsRef<OsStr>>(&self, file_rel_path: T) -> &CmdResult {
         let contents = read_scenario_fixture(&self.tmpd, file_rel_path);
-        self.stderr_is_bytes(contents)
+        self.stderr_is(String::from_utf8(contents).unwrap())
     }
 
     /// asserts that
@@ -329,14 +314,14 @@ impl CmdResult {
     }
 
     pub fn stdout_matches(&self, regex: &regex::Regex) -> &CmdResult {
-        if !regex.is_match(self.stdout_str()) {
+        if !regex.is_match(self.stdout_str().trim()) {
             panic!("Stdout does not match regex:\n{}", self.stdout_str())
         }
         self
     }
 
     pub fn stdout_does_not_match(&self, regex: &regex::Regex) -> &CmdResult {
-        if regex.is_match(self.stdout_str()) {
+        if regex.is_match(self.stdout_str().trim()) {
             panic!("Stdout matches regex:\n{}", self.stdout_str())
         }
         self
@@ -635,7 +620,7 @@ impl TestScenario {
             },
             util_name: String::from(util_name),
             fixtures: AtPath::new(tmpd.as_ref().path()),
-            tmpd: tmpd,
+            tmpd,
         };
         let mut fixture_path_builder = env::current_dir().unwrap();
         fixture_path_builder.push(TESTS_DIR);
@@ -696,8 +681,11 @@ pub struct UCommand {
     comm_string: String,
     tmpd: Option<Rc<TempDir>>,
     has_run: bool,
-    stdin: Option<Vec<u8>>,
     ignore_stdin_write_error: bool,
+    stdin: Option<Stdio>,
+    stdout: Option<Stdio>,
+    stderr: Option<Stdio>,
+    bytes_into_stdin: Option<Vec<u8>>,
 }
 
 impl UCommand {
@@ -726,8 +714,11 @@ impl UCommand {
                 cmd
             },
             comm_string: String::from(arg.as_ref().to_str().unwrap()),
-            stdin: None,
             ignore_stdin_write_error: false,
+            bytes_into_stdin: None,
+            stdin: None,
+            stdout: None,
+            stderr: None,
         }
     }
 
@@ -738,6 +729,21 @@ impl UCommand {
         ucmd
     }
 
+    pub fn set_stdin<T: Into<Stdio>>(&mut self, stdin: T) -> &mut UCommand {
+        self.stdin = Some(stdin.into());
+        self
+    }
+
+    pub fn set_stdout<T: Into<Stdio>>(&mut self, stdout: T) -> &mut UCommand {
+        self.stdout = Some(stdout.into());
+        self
+    }
+
+    pub fn set_stderr<T: Into<Stdio>>(&mut self, stderr: T) -> &mut UCommand {
+        self.stderr = Some(stderr.into());
+        self
+    }
+
     /// Add a parameter to the invocation. Path arguments are treated relative
     /// to the test environment directory.
     pub fn arg<S: AsRef<OsStr>>(&mut self, arg: S) -> &mut UCommand {
@@ -767,10 +773,10 @@ impl UCommand {
 
     /// provides stdinput to feed in to the command when spawned
     pub fn pipe_in<T: Into<Vec<u8>>>(&mut self, input: T) -> &mut UCommand {
-        if self.stdin.is_some() {
+        if self.bytes_into_stdin.is_some() {
             panic!("{}", MULTIPLE_STDIN_MEANINGLESS);
         }
-        self.stdin = Some(input.into());
+        self.bytes_into_stdin = Some(input.into());
         self
     }
 
@@ -784,7 +790,7 @@ impl UCommand {
     /// This is typically useful to test non-standard workflows
     /// like feeding something to a command that does not read it
     pub fn ignore_stdin_write_error(&mut self) -> &mut UCommand {
-        if self.stdin.is_none() {
+        if self.bytes_into_stdin.is_none() {
             panic!("{}", NO_STDIN_MEANINGLESS);
         }
         self.ignore_stdin_write_error = true;
@@ -813,13 +819,13 @@ impl UCommand {
         log_info("run", &self.comm_string);
         let mut child = self
             .raw
-            .stdin(Stdio::piped())
+            .stdin(self.stdin.take().unwrap_or_else(|| Stdio::piped()))
-            .stdout(Stdio::piped())
+            .stdout(self.stdout.take().unwrap_or_else(|| Stdio::piped()))
-            .stderr(Stdio::piped())
+            .stderr(self.stderr.take().unwrap_or_else(|| Stdio::piped()))
             .spawn()
             .unwrap();
 
-        if let Some(ref input) = self.stdin {
+        if let Some(ref input) = self.bytes_into_stdin {
             let write_result = child
                 .stdin
                 .take()

tests/fixtures/sort/default_unsorted_ints.expected.debug

@@ -0,0 +1,200 @@
+1
+_
+10
+__
+100
+___
+11
+__
+12
+__
+13
+__
+14
+__
+15
+__
+16
+__
+17
+__
+18
+__
+19
+__
+2
+_
+20
+__
+21
+__
+22
+__
+23
+__
+24
+__
+25
+__
+26
+__
+27
+__
+28
+__
+29
+__
+3
+_
+30
+__
+31
+__
+32
+__
+33
+__
+34
+__
+35
+__
+36
+__
+37
+__
+38
+__
+39
+__
+4
+_
+40
+__
+41
+__
+42
+__
+43
+__
+44
+__
+45
+__
+46
+__
+47
+__
+48
+__
+49
+__
+5
+_
+50
+__
+51
+__
+52
+__
+53
+__
+54
+__
+55
+__
+56
+__
+57
+__
+58
+__
+59
+__
+6
+_
+60
+__
+61
+__
+62
+__
+63
+__
+64
+__
+65
+__
+66
+__
+67
+__
+68
+__
+69
+__
+7
+_
+70
+__
+71
+__
+72
+__
+73
+__
+74
+__
+75
+__
+76
+__
+77
+__
+78
+__
+79
+__
+8
+_
+80
+__
+81
+__
+82
+__
+83
+__
+84
+__
+85
+__
+86
+__
+87
+__
+88
+__
+89
+__
+9
+_
+90
+__
+91
+__
+92
+__
+93
+__
+94
+__
+95
+__
+96
+__
+97
+__
+98
+__
+99
+__

tests/fixtures/sort/dictionary_order.expected.debug

@@ -0,0 +1,9 @@
+bbb
+___
+___
+./bbc
+_____
+_____
+bbd
+___
+___

tests/fixtures/sort/exponents-positive-numeric.expected.debug

@@ -0,0 +1,36 @@
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
++100000
+^ no match for key
+_______
+10E
+__
+___
+50e10
+__
+_____
+100E6
+___
+_____
+1000EDKLD
+____
+_________
+10000K78
+_____
+________

tests/fixtures/sort/exponents_general.expected

@@ -0,0 +1,19 @@
+
+
+
+
+
+
+
+
+5.5.5.5
+10E
+1000EDKLD
+10000K78
++100000
++100000
+100E6
+100E6
+10e10e10e10
+50e10
+50e10

tests/fixtures/sort/exponents_general.expected.debug

@@ -0,0 +1,57 @@
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+5.5.5.5
+___
+_______
+10E
+__
+___
+1000EDKLD
+____
+_________
+10000K78
+_____
+________
++100000
+_______
+_______
++100000
+_______
+_______
+100E6
+_____
+_____
+100E6
+_____
+_____
+10e10e10e10
+_____
+___________
+50e10
+_____
+_____
+50e10
+_____
+_____

tests/fixtures/sort/exponents_general.txt

@@ -0,0 +1,19 @@
+100E6
+
+50e10
++100000
+
+10000K78
+10E
+
+
+1000EDKLD
+
+
+100E6
+
+50e10
++100000
+
+10e10e10e10
+5.5.5.5

tests/fixtures/sort/human-numeric-whitespace.expected.debug

@@ -0,0 +1,33 @@
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+456K
+____
+____
+4568K
+_____
+_____
+>>>456M
+   ____
+_______
+              6.2G
+              ____
+__________________

tests/fixtures/sort/human_block_sizes.expected.debug

@@ -0,0 +1,33 @@
+844K
+____
+____
+981K
+____
+____
+11M
+___
+___
+13M
+___
+___
+14M
+___
+___
+16M
+___
+___
+18M
+___
+___
+19M
+___
+___
+20M
+___
+___
+981T
+____
+____
+20P
+___
+___

tests/fixtures/sort/ignore_case.expected.debug

@@ -0,0 +1,21 @@
+aaa
+___
+___
+BBB
+___
+___
+ccc
+___
+___
+DDD
+___
+___
+eee
+___
+___
+FFF
+___
+___
+ggg
+___
+___

tests/fixtures/sort/keys_closed_range.expected

@@ -0,0 +1,6 @@
+dd aa ff
+gg aa cc
+aa bb cc
+èè éé èè
+👩‍🔬 👩‍🔬 👩‍🔬
+💣💣 💣💣 💣💣

tests/fixtures/sort/keys_closed_range.expected.debug

@@ -0,0 +1,18 @@
+dd aa ff
+   _
+________
+gg aa cc
+   _
+________
+aa bb cc
+   _
+________
+èè éé èè
+   _
+________
+👩‍🔬 👩‍🔬 👩‍🔬
+     __
+______________
+💣💣 💣💣 💣💣
+     __
+______________

tests/fixtures/sort/keys_closed_range.txt

@@ -0,0 +1,6 @@
+aa bb cc
+dd aa ff
+gg aa cc
+èè éé èè
+💣💣 💣💣 💣💣
+👩‍🔬 👩‍🔬 👩‍🔬

tests/fixtures/sort/keys_custom_separator.expected

@@ -0,0 +1,3 @@
+ddxaaxff
+ggxaaxcc
+aaxbbxcc

tests/fixtures/sort/keys_custom_separator.expected.debug

@@ -0,0 +1,9 @@
+ddxaaxff
+    _
+________
+ggxaaxcc
+    _
+________
+aaxbbxcc
+    _
+________

tests/fixtures/sort/keys_custom_separator.txt

@@ -0,0 +1,3 @@
+aaxbbxcc
+ddxaaxff
+ggxaaxcc

tests/fixtures/sort/keys_multiple_ranges.expected

@@ -0,0 +1,6 @@
+gg aa cc
+dd aa ff
+aa bb cc
+èè éé èè
+👩‍🔬 👩‍🔬 👩‍🔬
+💣💣 💣💣 💣💣

tests/fixtures/sort/keys_multiple_ranges.expected.debug

@@ -0,0 +1,24 @@
+gg aa cc
+  ___
+     ___
+________
+dd aa ff
+  ___
+     ___
+________
+aa bb cc
+  ___
+     ___
+________
+èè éé èè
+  ___
+     ___
+________
+👩‍🔬 👩‍🔬 👩‍🔬
+    _____
+         _____
+______________
+💣💣 💣💣 💣💣
+    _____
+         _____
+______________

tests/fixtures/sort/keys_multiple_ranges.txt

@@ -0,0 +1,6 @@
+aa bb cc
+dd aa ff
+gg aa cc
+èè éé èè
+💣💣 💣💣 💣💣
+👩‍🔬 👩‍🔬 👩‍🔬

tests/fixtures/sort/keys_no_char_match.expected

@@ -0,0 +1,3 @@
+c
+ba
+aaa

tests/fixtures/sort/keys_no_char_match.expected.debug

@@ -0,0 +1,9 @@
+c
+ ^ no match for key
+_
+ba
+ _
+__
+aaa
+ __
+___

tests/fixtures/sort/keys_no_char_match.txt

@@ -0,0 +1,3 @@
+aaa
+ba
+c

tests/fixtures/sort/keys_no_field_match.expected

@@ -0,0 +1,6 @@
+aa bb cc
+dd aa ff
+gg aa cc
+èè éé èè
+👩‍🔬 👩‍🔬 👩‍🔬
+💣💣 💣💣 💣💣

tests/fixtures/sort/keys_no_field_match.expected.debug

@@ -0,0 +1,18 @@
+aa bb cc
+        ^ no match for key
+________
+dd aa ff
+        ^ no match for key
+________
+gg aa cc
+        ^ no match for key
+________
+èè éé èè
+        ^ no match for key
+________
+👩‍🔬 👩‍🔬 👩‍🔬
+              ^ no match for key
+______________
+💣💣 💣💣 💣💣
+              ^ no match for key
+______________

tests/fixtures/sort/keys_no_field_match.txt

@@ -0,0 +1,6 @@
+aa bb cc
+dd aa ff
+gg aa cc
+èè éé èè
+💣💣 💣💣 💣💣
+👩‍🔬 👩‍🔬 👩‍🔬

tests/fixtures/sort/keys_open_ended.expected

@@ -0,0 +1,6 @@
+gg aa cc
+dd aa ff
+aa bb cc
+èè éé èè
+👩‍🔬 👩‍🔬 👩‍🔬
+💣💣 💣💣 💣💣

tests/fixtures/sort/keys_open_ended.expected.debug

@@ -0,0 +1,18 @@
+gg aa cc
+    ____
+________
+dd aa ff
+    ____
+________
+aa bb cc
+    ____
+________
+èè éé èè
+    ____
+________
+👩‍🔬 👩‍🔬 👩‍🔬
+       _______
+______________
+💣💣 💣💣 💣💣
+       _______
+______________

tests/fixtures/sort/keys_open_ended.txt

@@ -0,0 +1,6 @@
+aa bb cc
+dd aa ff
+gg aa cc
+èè éé èè
+💣💣 💣💣 💣💣
+👩‍🔬 👩‍🔬 👩‍🔬

tests/fixtures/sort/mixed_floats_ints_chars_numeric.expected.debug

@@ -0,0 +1,90 @@
+-2028789030
+___________
+___________
+-896689
+_______
+_______
+-8.90880
+________
+________
+-1
+__
+__
+-.05
+____
+____
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+000
+___
+___
+CARAvan
+^ no match for key
+_______
+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
+_______________
+_______________

tests/fixtures/sort/mixed_floats_ints_chars_numeric_stable.expected.debug

@@ -0,0 +1,60 @@
+-2028789030
+___________
+-896689
+_______
+-8.90880
+________
+-1
+__
+-.05
+____
+
+^ no match for key
+
+^ no match for key
+
+^ no match for key
+
+^ no match for key
+
+^ no match for key
+CARAvan
+^ no match for key
+
+^ no match for key
+
+^ no match for key
+
+^ no match for key
+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
+_______________

tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique.expected.debug

@@ -0,0 +1,40 @@
+-2028789030
+___________
+-896689
+_______
+-8.90880
+________
+-1
+__
+-.05
+____
+
+^ no match for key
+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
+_______________

tests/fixtures/sort/mixed_floats_ints_chars_numeric_unique_reverse.expected.debug

@@ -0,0 +1,40 @@
+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
+_
+
+^ no match for key
+-.05
+____
+-1
+__
+-8.90880
+________
+-896689
+_______
+-2028789030
+___________

tests/fixtures/sort/month_default.expected.debug

@@ -0,0 +1,30 @@
+N/A Ut enim ad minim veniam, quis
+^ no match for key
+_________________________________
+Jan Lorem ipsum dolor sit amet
+___
+______________________________
+mar laboris nisi ut aliquip ex ea
+___
+_________________________________
+May sed do eiusmod tempor incididunt
+___
+____________________________________
+JUN nostrud exercitation ullamco
+___
+________________________________
+Jul 1 should remain 2,1,3
+___
+_________________________
+Jul 2 these three lines
+___
+_______________________
+Jul 3 if --stable is provided
+___
+_____________________________
+Oct ut labore et dolore magna aliqua
+___
+____________________________________
+Dec consectetur adipiscing elit
+___
+_______________________________

tests/fixtures/sort/month_stable.expected.debug

@@ -0,0 +1,20 @@
+N/A Ut enim ad minim veniam, quis
+^ no match for key
+Jan Lorem ipsum dolor sit amet
+___
+mar laboris nisi ut aliquip ex ea
+___
+May sed do eiusmod tempor incididunt
+___
+JUN nostrud exercitation ullamco
+___
+Jul 2 these three lines
+___
+Jul 1 should remain 2,1,3
+___
+Jul 3 if --stable is provided
+___
+Oct ut labore et dolore magna aliqua
+___
+Dec consectetur adipiscing elit
+___

tests/fixtures/sort/months-dedup.expected.debug

@@ -0,0 +1,12 @@
+
+^ no match for key
+JAN
+___
+apr
+___
+MAY
+___
+JUNNNN
+___
+AUG 
+___

tests/fixtures/sort/months-whitespace.expected.debug

@@ -0,0 +1,24 @@
+
+^ no match for key
+^ no match for key
+
+^ no match for key
+^ no match for key
+JAN
+___
+___
+  FEb
+  ___
+_____
+ apr
+ ___
+____
+ apr
+ ___
+____
+>>>JUNNNN
+   ___
+_________
+AUG 
+___
+____