cat: Unrevert splice patch (#2020)

* cat: Unrevert splice patch

* cat: Add fifo test

* cat: Add tests for error cases

* cat: Add tests for character devices

* wc: Make sure we handle short splice writes

* cat: Fix tests for 1.40.0 compiler

* cat: Run rustfmt on test_cat.rs

* Run 'cargo +1.40.0 update'

Cargo.lock

@@ -1396,9 +1396,9 @@ dependencies = [
 
 [[package]]
 name = "syn"
-version = "1.0.68"
+version = "1.0.69"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3ce15dd3ed8aa2f8eeac4716d6ef5ab58b6b9256db41d7e1a0224c2788e8fd87"
+checksum = "48fe99c6bd8b1cc636890bcc071842de909d902c81ac7dab53ba33c421ab8ffb"
 dependencies = [
  "proc-macro2",
  "quote 1.0.9",
@@ -1618,7 +1618,8 @@ name = "uu_cat"
 version = "0.0.6"
 dependencies = [
  "clap",
- "quick-error",
+ "nix 0.20.0",
+ "thiserror",
  "unix_socket",
  "uucore",
  "uucore_procs",

src/uu/cat/Cargo.toml

@@ -16,13 +16,16 @@ path = "src/cat.rs"
 
 [dependencies]
 clap = "2.33"
-quick-error = "1.2.3"
+thiserror = "1.0"
 uucore = { version=">=0.0.8", package="uucore", path="../../uucore", features=["fs"] }
 uucore_procs = { version=">=0.0.5", package="uucore_procs", path="../../uucore_procs" }
 
 [target.'cfg(unix)'.dependencies]
 unix_socket = "0.5.0"
 
+[target.'cfg(any(target_os = "linux", target_os = "android"))'.dependencies]
+nix = "0.20"
+
 [[bin]]
 name = "cat"
 path = "src/main.rs"

src/uu/wc/src/count_bytes.rs

@@ -20,6 +20,21 @@ use nix::unistd::pipe;
 
 const BUF_SIZE: usize = 16384;
 
+/// 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(())
+}
+
 /// This is a Linux-specific function to count the number of bytes using the
 /// `splice` system call, which is faster than using `read`.
 #[inline]
@@ -39,7 +54,7 @@ fn count_bytes_using_splice(fd: RawFd) -> nix::Result<usize> {
             break;
         }
         byte_count += res;
-        splice(pipe_rd, None, null, None, res, SpliceFFlags::empty())?;
+        splice_exact(pipe_rd, null, res)?;
     }
 
     Ok(byte_count)

tests/by-util/test_cat.rs

@@ -1,7 +1,5 @@
-#[cfg(unix)]
-extern crate unix_socket;
-
 use crate::common::util::*;
+use std::io::Read;
 
 #[test]
 fn test_output_simple() {
@@ -11,6 +9,129 @@ fn test_output_simple() {
         .stdout_only("abcde\nfghij\nklmno\npqrst\nuvwxyz\n");
 }
 
+#[test]
+fn test_no_options() {
+    for fixture in &["empty.txt", "alpha.txt", "nonewline.txt"] {
+        // Give fixture through command line file argument
+        new_ucmd!()
+            .args(&[fixture])
+            .succeeds()
+            .stdout_is_fixture(fixture);
+        // Give fixture through stdin
+        new_ucmd!()
+            .pipe_in_fixture(fixture)
+            .succeeds()
+            .stdout_is_fixture(fixture);
+    }
+}
+
+#[test]
+fn test_no_options_big_input() {
+    for &n in &[
+        0,
+        1,
+        42,
+        16 * 1024 - 7,
+        16 * 1024 - 1,
+        16 * 1024,
+        16 * 1024 + 1,
+        16 * 1024 + 3,
+        32 * 1024,
+        64 * 1024,
+        80 * 1024,
+        96 * 1024,
+        112 * 1024,
+        128 * 1024,
+    ] {
+        let data = vec_of_size(n);
+        let data2 = data.clone();
+        assert_eq!(data.len(), data2.len());
+        new_ucmd!().pipe_in(data).succeeds().stdout_is_bytes(&data2);
+    }
+}
+
+#[test]
+#[cfg(unix)]
+fn test_fifo_symlink() {
+    use std::fs::OpenOptions;
+    use std::io::Write;
+    use std::thread;
+
+    let s = TestScenario::new(util_name!());
+    s.fixtures.mkdir("dir");
+    s.fixtures.mkfifo("dir/pipe");
+    assert!(s.fixtures.is_fifo("dir/pipe"));
+
+    // Make cat read the pipe through a symlink
+    s.fixtures.symlink_file("dir/pipe", "sympipe");
+    let proc = s.ucmd().args(&["sympipe"]).run_no_wait();
+
+    let data = vec_of_size(128 * 1024);
+    let data2 = data.clone();
+
+    let pipe_path = s.fixtures.plus("dir/pipe");
+    let thread = thread::spawn(move || {
+        let mut pipe = OpenOptions::new()
+            .write(true)
+            .create(false)
+            .open(pipe_path)
+            .unwrap();
+        pipe.write_all(&data).unwrap();
+    });
+
+    let output = proc.wait_with_output().unwrap();
+    assert_eq!(&output.stdout, &data2);
+    thread.join().unwrap();
+}
+
+#[test]
+fn test_directory() {
+    let s = TestScenario::new(util_name!());
+    s.fixtures.mkdir("test_directory");
+    s.ucmd()
+        .args(&["test_directory"])
+        .fails()
+        .stderr_is("cat: test_directory: Is a directory");
+}
+
+#[test]
+fn test_directory_and_file() {
+    let s = TestScenario::new(util_name!());
+    s.fixtures.mkdir("test_directory2");
+    for fixture in &["empty.txt", "alpha.txt", "nonewline.txt"] {
+        s.ucmd()
+            .args(&["test_directory2", fixture])
+            .fails()
+            .stderr_is("cat: test_directory2: Is a directory")
+            .stdout_is_fixture(fixture);
+    }
+}
+
+#[test]
+fn test_three_directories_and_file_and_stdin() {
+    let s = TestScenario::new(util_name!());
+    s.fixtures.mkdir("test_directory3");
+    s.fixtures.mkdir("test_directory3/test_directory4");
+    s.fixtures.mkdir("test_directory3/test_directory5");
+    s.ucmd()
+        .args(&[
+            "test_directory3/test_directory4",
+            "alpha.txt",
+            "-",
+            "filewhichdoesnotexist.txt",
+            "nonewline.txt",
+            "test_directory3/test_directory5",
+            "test_directory3/../test_directory3/test_directory5",
+            "test_directory3",
+        ])
+        .pipe_in("stdout bytes")
+        .fails()
+        .stderr_is_fixture("three_directories_and_file_and_stdin.stderr.expected")
+        .stdout_is(
+            "abcde\nfghij\nklmno\npqrst\nuvwxyz\nstdout bytestext without a trailing newline",
+        );
+}
+
 #[test]
 fn test_output_multi_files_print_all_chars() {
     new_ucmd!()
@@ -149,13 +270,64 @@ fn test_squeeze_blank_before_numbering() {
     }
 }
 
+/// This tests reading from Unix character devices
 #[test]
-#[cfg(foo)]
+#[cfg(unix)]
+fn test_dev_random() {
+    let mut buf = [0; 2048];
+    let mut proc = new_ucmd!().args(&["/dev/random"]).run_no_wait();
+    let mut proc_stdout = proc.stdout.take().unwrap();
+    proc_stdout.read_exact(&mut buf).unwrap();
+
+    let num_zeroes = buf.iter().fold(0, |mut acc, &n| {
+        if n == 0 {
+            acc += 1;
+        }
+        acc
+    });
+    // The probability of more than 512 zero bytes is essentially zero if the
+    // output is truly random.
+    assert!(num_zeroes < 512);
+    proc.kill().unwrap();
+}
+
+/// Reading from /dev/full should return an infinite amount of zero bytes.
+/// Wikipedia says there is support on Linux, FreeBSD, and NetBSD.
+#[test]
+#[cfg(any(target_os = "linux", target_os = "freebsd", target_os = "netbsd"))]
+fn test_dev_full() {
+    let mut buf = [0; 2048];
+    let mut proc = new_ucmd!().args(&["/dev/full"]).run_no_wait();
+    let mut proc_stdout = proc.stdout.take().unwrap();
+    let expected = [0; 2048];
+    proc_stdout.read_exact(&mut buf).unwrap();
+    assert_eq!(&buf[..], &expected[..]);
+    proc.kill().unwrap();
+}
+
+#[test]
+#[cfg(any(target_os = "linux", target_os = "freebsd", target_os = "netbsd"))]
+fn test_dev_full_show_all() {
+    let mut buf = [0; 2048];
+    let mut proc = new_ucmd!().args(&["-A", "/dev/full"]).run_no_wait();
+    let mut proc_stdout = proc.stdout.take().unwrap();
+    proc_stdout.read_exact(&mut buf).unwrap();
+
+    let expected: Vec<u8> = (0..buf.len())
+        .map(|n| if n & 1 == 0 { b'^' } else { b'@' })
+        .collect();
+
+    assert_eq!(&buf[..], &expected[..]);
+    proc.kill().unwrap();
+}
+
+#[test]
+#[cfg(unix)]
 fn test_domain_socket() {
-    use self::tempdir::TempDir;
-    use self::unix_socket::UnixListener;
     use std::io::prelude::*;
     use std::thread;
+    use tempdir::TempDir;
+    use unix_socket::UnixListener;
 
     let dir = TempDir::new("unix_socket").expect("failed to create dir");
     let socket_path = dir.path().join("sock");

tests/by-util/test_wc.rs

@@ -1,5 +1,33 @@
 use crate::common::util::*;
 
+#[test]
+fn test_count_bytes_large_stdin() {
+    for &n in &[
+        0,
+        1,
+        42,
+        16 * 1024 - 7,
+        16 * 1024 - 1,
+        16 * 1024,
+        16 * 1024 + 1,
+        16 * 1024 + 3,
+        32 * 1024,
+        64 * 1024,
+        80 * 1024,
+        96 * 1024,
+        112 * 1024,
+        128 * 1024,
+    ] {
+        let data = vec_of_size(n);
+        let expected = format!("{}\n", n);
+        new_ucmd!()
+            .args(&["-c"])
+            .pipe_in(data)
+            .succeeds()
+            .stdout_is_bytes(&expected.as_bytes());
+    }
+}
+
 #[test]
 fn test_stdin_default() {
     new_ucmd!()

tests/common/util.rs

@@ -222,6 +222,12 @@ impl CmdResult {
         self
     }
 
+    /// 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)
+    }
+
     /// asserts that
     /// 1.  the command resulted in stdout stream output that equals the
     ///     passed in value
@@ -804,3 +810,12 @@ pub fn read_size(child: &mut Child, size: usize) -> String {
         .unwrap();
     String::from_utf8(output).unwrap()
 }
+
+pub fn vec_of_size(n: usize) -> Vec<u8> {
+    let mut result = Vec::new();
+    for _ in 0..n {
+        result.push('a' as u8);
+    }
+    assert_eq!(result.len(), n);
+    result
+}

tests/fixtures/cat/three_directories_and_file_and_stdin.stderr.expected

@@ -0,0 +1,5 @@
+cat: test_directory3/test_directory4: Is a directory
+cat: filewhichdoesnotexist.txt: No such file or directory (os error 2)
+cat: test_directory3/test_directory5: Is a directory
+cat: test_directory3/../test_directory3/test_directory5: Is a directory
+cat: test_directory3: Is a directory