wc: more tests and fixes

My previous commits meant to bring our wc's output and behavior in line
with GNU's. There should be tests that check for these changes!

I found a stupid bug in my own changes, I was not adding 1 to the
indexes produced by .enumerate() when printing errors.

src/uu/wc/src/wc.rs

@@ -188,9 +188,9 @@ impl<'a> Inputs<'a> {
             },
         };
 
-        // The index of each yielded item must be tracked for error reporting.
+        // The 1-based index of each yielded item must be tracked for error reporting.
-        let mut with_idx = base.enumerate();
+        let mut with_idx = base.enumerate().map(|(i, v)| (i + 1, v));
-        let files0_from_path = settings.files0_from.as_ref().map(|i| i.as_borrowed());
+        let files0_from_path = settings.files0_from.as_ref().map(|p| p.as_borrowed());
 
         let iter = iter::from_fn(move || {
             let (idx, next) = with_idx.next()?;
@@ -303,23 +303,9 @@ fn is_stdin_small_file() -> bool {
     matches!(f.metadata(), Ok(meta) if meta.is_file() && meta.len() <= (10 << 20))
 }
 
-#[cfg(windows)]
+#[cfg(not(unix))]
-fn is_stdin_small_file() -> bool {
+// Windows presents a piped stdin as a "normal file" with a length equal to however many bytes
-    use std::os::windows::io::{AsRawHandle, FromRawHandle};
+// have been buffered at the time it's checked. To be safe, we must never assume it's a file.
-    // Safety: we'll rely on Rust to give us a valid RawHandle for stdin with which we can attempt
-    // to open a File, but only for the sake of fetching .metadata().  ManuallyDrop will ensure we
-    // don't do anything else to the FD if anything unexpected happens.
-    let f = std::mem::ManuallyDrop::new(unsafe {
-        let h = io::stdin().as_raw_handle();
-        if h.is_null() {
-            return false;
-        }
-        File::from_raw_handle(h)
-    });
-    matches!(f.metadata(), Ok(meta) if meta.is_file() && meta.len() <= (10 << 20))
-}
-
-#[cfg(not(any(unix, windows)))]
 fn is_stdin_small_file() -> bool {
     false
 }

tests/by-util/test_wc.rs

@@ -268,12 +268,16 @@ fn test_multiple_default() {
             "lorem_ipsum.txt",
             "moby_dick.txt",
             "alice_in_wonderland.txt",
+            "alice in wonderland.txt",
         ])
         .run()
-        .stdout_is(
+        .stdout_is(concat!(
-            "  13  109  772 lorem_ipsum.txt\n  18  204 1115 moby_dick.txt\n   5   57  302 \
+            "  13  109  772 lorem_ipsum.txt\n",
-             alice_in_wonderland.txt\n  36  370 2189 total\n",
+            "  18  204 1115 moby_dick.txt\n",
-        );
+            "   5   57  302 alice_in_wonderland.txt\n",
+            "   5   57  302 alice in wonderland.txt\n",
+            "  41  427 2491 total\n",
+        ));
 }
 
 /// Test for an empty file.
@@ -352,17 +356,24 @@ fn test_file_bytes_dictate_width() {
     new_ucmd!()
         .args(&["-lwc", "alice_in_wonderland.txt", "lorem_ipsum.txt"])
         .run()
-        .stdout_is(
+        .stdout_is(concat!(
-            "   5   57  302 alice_in_wonderland.txt\n  13  109  772 \
+            "   5   57  302 alice_in_wonderland.txt\n",
-                    lorem_ipsum.txt\n  18  166 1074 total\n",
+            "  13  109  772 lorem_ipsum.txt\n",
-        );
+            "  18  166 1074 total\n",
+        ));
 
     // . is a directory, so minimum_width should get set to 7
     #[cfg(not(windows))]
-    const STDOUT: &str = "      0       0       0 emptyfile.txt\n      0       0       0 \
+    const STDOUT: &str = concat!(
-                          .\n      0       0       0 total\n";
+        "      0       0       0 emptyfile.txt\n",
+        "      0       0       0 .\n",
+        "      0       0       0 total\n",
+    );
     #[cfg(windows)]
-    const STDOUT: &str = "      0       0       0 emptyfile.txt\n      0       0       0 total\n";
+    const STDOUT: &str = concat!(
+        "      0       0       0 emptyfile.txt\n",
+        "      0       0       0 total\n",
+    );
     new_ucmd!()
         .args(&["-lwc", "emptyfile.txt", "."])
         .run()
@@ -392,12 +403,10 @@ fn test_read_from_directory_error() {
 /// Test that getting counts from nonexistent file is an error.
 #[test]
 fn test_read_from_nonexistent_file() {
-    const MSG: &str = "bogusfile: No such file or directory";
     new_ucmd!()
         .args(&["bogusfile"])
         .fails()
-        .stderr_contains(MSG)
+        .stderr_only("wc: bogusfile: No such file or directory\n");
-        .stdout_is("");
 }
 
 #[test]
@@ -421,15 +430,30 @@ fn test_files0_disabled_files_argument() {
 
 #[test]
 fn test_files0_from() {
+    // file
     new_ucmd!()
         .args(&["--files0-from=files0_list.txt"])
         .run()
+        .success()
         .stdout_is(concat!(
             "  13  109  772 lorem_ipsum.txt\n",
             "  18  204 1115 moby_dick.txt\n",
             "   5   57  302 alice_in_wonderland.txt\n",
             "  36  370 2189 total\n",
         ));
+
+    // stream
+    new_ucmd!()
+        .args(&["--files0-from=-"])
+        .pipe_in_fixture("files0_list.txt")
+        .run()
+        .success()
+        .stdout_is(concat!(
+            "13 109 772 lorem_ipsum.txt\n",
+            "18 204 1115 moby_dick.txt\n",
+            "5 57 302 alice_in_wonderland.txt\n",
+            "36 370 2189 total\n",
+        ));
 }
 
 #[test]
@@ -450,7 +474,7 @@ fn test_files0_from_with_stdin_in_file() {
         .stdout_is(concat!(
             "     13     109     772 lorem_ipsum.txt\n",
             "     18     204    1115 moby_dick.txt\n",
-            "      5      57     302 -\n",
+            "      5      57     302 -\n", // alice_in_wonderland.txt
             "     36     370    2189 total\n",
         ));
 }
@@ -531,3 +555,152 @@ fn test_total_only() {
         .run()
         .stdout_is("31 313 1887\n");
 }
+
+#[test]
+fn test_zero_length_files() {
+    // A trailing zero is ignored, but otherwise empty file names are an error...
+    const LIST: &str = "\0moby_dick.txt\0\0alice_in_wonderland.txt\0\0lorem_ipsum.txt\0";
+
+    // Try with and without the last \0
+    for l in [LIST.len(), LIST.len() - 1] {
+        new_ucmd!()
+            .args(&["--files0-from=-"])
+            .pipe_in(&LIST[..l])
+            .run()
+            .failure()
+            .stdout_is(concat!(
+                "18 204 1115 moby_dick.txt\n",
+                "5 57 302 alice_in_wonderland.txt\n",
+                "13 109 772 lorem_ipsum.txt\n",
+                "36 370 2189 total\n",
+            ))
+            .stderr_is(concat!(
+                "wc: -:1: invalid zero-length file name\n",
+                "wc: -:3: invalid zero-length file name\n",
+                "wc: -:5: invalid zero-length file name\n",
+            ));
+    }
+
+    // But, just as important, a zero-length file name may still be at the end...
+    new_ucmd!()
+        .args(&["--files0-from=-"])
+        .pipe_in(
+            LIST.as_bytes()
+                .iter()
+                .chain(b"\0")
+                .copied()
+                .collect::<Vec<_>>(),
+        )
+        .run()
+        .failure()
+        .stdout_is(concat!(
+            "18 204 1115 moby_dick.txt\n",
+            "5 57 302 alice_in_wonderland.txt\n",
+            "13 109 772 lorem_ipsum.txt\n",
+            "36 370 2189 total\n",
+        ))
+        .stderr_is(concat!(
+            "wc: -:1: invalid zero-length file name\n",
+            "wc: -:3: invalid zero-length file name\n",
+            "wc: -:5: invalid zero-length file name\n",
+            "wc: -:7: invalid zero-length file name\n",
+        ));
+}
+
+#[test]
+fn test_files0_errors_quoting() {
+    new_ucmd!()
+        .args(&["--files0-from=files0 with nonexistent.txt"])
+        .run()
+        .failure()
+        .stderr_is(concat!(
+            "wc: this_file_does_not_exist.txt: No such file or directory\n",
+            "wc: 'files0 with nonexistent.txt':2: invalid zero-length file name\n",
+            "wc: 'this file does not exist.txt': No such file or directory\n",
+            "wc: \"this files doesn't exist either.txt\": No such file or directory\n",
+        ))
+        .stdout_is("0 0 0 total\n");
+}
+
+#[test]
+fn test_files0_progressive_stream() {
+    use std::process::Stdio;
+    // You should be able to run wc and have a back-and-forth exchange with wc...
+    let mut child = new_ucmd!()
+        .args(&["--files0-from=-"])
+        .set_stdin(Stdio::piped())
+        .set_stdout(Stdio::piped())
+        .set_stderr(Stdio::piped())
+        .run_no_wait();
+
+    macro_rules! chk {
+        ($fn:ident, $exp:literal) => {
+            assert_eq!(child.$fn($exp.len()), $exp.as_bytes());
+        };
+    }
+
+    // File in, count out...
+    child.write_in("moby_dick.txt\0");
+    chk!(stdout_exact_bytes, "18 204 1115 moby_dick.txt\n");
+    child.write_in("lorem_ipsum.txt\0");
+    chk!(stdout_exact_bytes, "13 109 772 lorem_ipsum.txt\n");
+
+    // Introduce an error!
+    child.write_in("\0");
+    chk!(
+        stderr_exact_bytes,
+        "wc: -:3: invalid zero-length file name\n"
+    );
+
+    // wc is quick to forgive, let's move on...
+    child.write_in("alice_in_wonderland.txt\0");
+    chk!(stdout_exact_bytes, "5 57 302 alice_in_wonderland.txt\n");
+
+    // Fin.
+    child
+        .wait()
+        .expect("wc should finish")
+        .failure()
+        .stdout_only("36 370 2189 total\n");
+}
+
+#[test]
+fn files0_from_dir() {
+    // On Unix, `read(open("."))` fails. On Windows, `open(".")` fails. Thus, the errors happen in
+    // different contexts.
+    #[cfg(not(windows))]
+    macro_rules! dir_err {
+        ($p:literal) => {
+            concat!("wc: ", $p, ": read error: Is a directory\n")
+        };
+    }
+    #[cfg(windows)]
+    macro_rules! dir_err {
+        ($p:literal) => {
+            concat!("wc: cannot open ", $p, " for reading: Permission denied\n")
+        };
+    }
+
+    new_ucmd!()
+        .args(&["--files0-from=dir with spaces"])
+        .fails()
+        .stderr_only(dir_err!("'dir with spaces'"));
+
+    // Those contexts have different rules about quoting in errors...
+    #[cfg(windows)]
+    const DOT_ERR: &str = dir_err!("'.'");
+    #[cfg(not(windows))]
+    const DOT_ERR: &str = dir_err!(".");
+    new_ucmd!()
+        .args(&["--files0-from=."])
+        .fails()
+        .stderr_only(DOT_ERR);
+
+    // That also means you cannot `< . wc --files0-from=-` on Windows.
+    #[cfg(not(windows))]
+    new_ucmd!()
+        .args(&["--files0-from=-"])
+        .set_stdin(std::fs::File::open(".").unwrap())
+        .fails()
+        .stderr_only(dir_err!("-"));
+}

tests/fixtures/wc/alice in wonderland.txt

@@ -0,0 +1,5 @@
+Alice was beginning to get very tired of sitting by
+her sister on the bank, and of having nothing to do: once or twice
+she had peeped into the book her sister was reading, but it had no
+pictures or conversations in it, "and what is the use of a book,"
+thought Alice "without pictures or conversation?"