tee: remove output buffering

Maximkaaa · Maximkaaa · commit 01b6d2492200 · 2025-06-19T09:12:42.000+09:00
To comply with POSIX standard `tee` implementation must not buffer
its output, so we replace std::io::copy implementation that does
buffering with the custom one.
diff --git a/src/uu/tee/src/tee.rs b/src/uu/tee/src/tee.rs
@@ -7,7 +7,7 @@
 
 use clap::{Arg, ArgAction, Command, builder::PossibleValue};
 use std::fs::OpenOptions;
-use std::io::{Error, ErrorKind, Read, Result, Write, copy, stdin, stdout};
+use std::io::{Error, ErrorKind, Read, Result, Write, stdin, stdout};
 use std::path::PathBuf;
 use uucore::display::Quotable;
 use uucore::error::UResult;
@@ -190,6 +190,7 @@ fn tee(options: &Options) -> Result<()> {
         return Ok(());
     }
 
+    // We cannot use std::io::copy here as it doesn't flush the output buffer
     let res = match copy(input, &mut output) {
         // ErrorKind::Other is raised by MultiWriter when all writers
         // have exited, so that copy will abort. It's equivalent to
@@ -207,6 +208,46 @@ fn tee(options: &Options) -> Result<()> {
     }
 }
 
+/// Copies all bytes from the input buffer to the output buffer.
+///
+/// Returns the number of written bytes.
+fn copy(mut input: impl Read, mut output: impl Write) -> Result<usize> {
+    // The implementation for this function is adopted from the generic buffer copy implementation from
+    // the standard library:
+    // https://github.com/rust-lang/rust/blob/2feb91181882e525e698c4543063f4d0296fcf91/library/std/src/io/copy.rs#L271-L297
+
+    // Use buffer size from std implementation:
+    // https://github.com/rust-lang/rust/blob/2feb91181882e525e698c4543063f4d0296fcf91/library/std/src/sys/io/mod.rs#L44
+    // spell-checker:ignore espidf
+    const DEFAULT_BUF_SIZE: usize = if cfg!(target_os = "espidf") {
+        512
+    } else {
+        8 * 1024
+    };
+
+    let mut buffer = [0u8; DEFAULT_BUF_SIZE];
+    let mut len = 0;
+
+    loop {
+        let received = match input.read(&mut buffer) {
+            Ok(bytes_count) => bytes_count,
+            Err(e) if e.kind() == ErrorKind::Interrupted => continue,
+            Err(e) => return Err(e),
+        };
+
+        if received == 0 {
+            return Ok(len);
+        }
+
+        output.write_all(&buffer[0..received])?;
+
+        // We need to flush the buffer here to comply with POSIX requirement that
+        // `tee` does not buffer the input.
+        output.flush()?;
+        len += received;
+    }
+}
+
 /// Tries to open the indicated file and return it. Reports an error if that's not possible.
 /// If that error should lead to program termination, this function returns Some(Err()),
 /// otherwise it returns None.
diff --git a/tests/by-util/test_tee.rs b/tests/by-util/test_tee.rs
@@ -10,6 +10,8 @@ use uutests::{at_and_ucmd, new_ucmd, util_name};
 use regex::Regex;
 #[cfg(target_os = "linux")]
 use std::fmt::Write;
+use std::process::Stdio;
+use std::time::Duration;
 
 // tests for basic tee functionality.
 // inspired by:
@@ -134,6 +136,46 @@ fn test_readonly() {
     assert_eq!(at.read(writable_file), content_tee);
 }
 
+#[test]
+fn test_tee_output_not_buffered() {
+    // POSIX says: The tee utility shall not buffer output
+
+    // If the output is buffered, the test will hang, so we run it in
+    // a separate thread to stop execution by timeout.
+    let handle = std::thread::spawn(move || {
+        let content = "a";
+        let file_out = "tee_file_out";
+
+        let (at, mut ucmd) = at_and_ucmd!();
+        let mut child = ucmd
+            .arg(file_out)
+            .set_stdin(Stdio::piped())
+            .set_stdout(Stdio::piped())
+            .run_no_wait();
+
+        // We write to the input pipe, but do not close it. If the output is
+        // buffered, reading from output pipe will hang indefinitely, as we
+        // will never write anything else to it.
+        child.write_in(content.as_bytes());
+
+        let out = String::from_utf8(child.stdout_exact_bytes(1)).unwrap();
+        assert_eq!(&out, content);
+
+        // Writing to a file may take a couple hundreds nanoseconds
+        child.delay(1);
+        assert_eq!(at.read(file_out), content);
+    });
+
+    // Give some time for the `tee` to create an output file
+    std::thread::sleep(Duration::from_millis(10));
+
+    assert!(
+        handle.is_finished(),
+        "Nothing was received through output pipe"
+    );
+    handle.join().unwrap();
+}
+
 #[cfg(target_os = "linux")]
 mod linux_only {
     use uutests::util::{AtPath, CmdResult, TestScenario, UCommand};
