split: fix for GNU Tests regression + tests

src/uu/split/src/split.rs

@@ -76,8 +76,16 @@ fn handle_obsolete(args: &[String]) -> (Vec<String>, Option<String>) {
     let filtered_args = args
         .iter()
         .filter_map(|slice| {
-            if slice.starts_with('-') && !slice.starts_with("--") {
+            if slice.starts_with('-')
+                && !slice.starts_with("--")
+                && !slice.starts_with("-a")
+                && !slice.starts_with("-b")
+                && !slice.starts_with("-C")
+                && !slice.starts_with("-l")
+                && !slice.starts_with("-n")
+            {
                 // start of the short option string
+                // that can have obsolete lines option value in it
                 // extract numeric part and filter it out
                 let mut obs_lines_extracted: Vec<char> = vec![];
                 let filtered_slice: Vec<char> = slice
@@ -102,15 +110,14 @@ fn handle_obsolete(args: &[String]) -> (Vec<String>, Option<String>) {
                         // there were some short options in front of or after obsolete lines value
                         // i.e. '-xd100' or '-100de' or similar, which after extraction of obsolete lines value
                         // would look like '-xd' or '-de' or similar
-                        // preserve it
                         Some(filtered_slice.iter().collect())
                     } else {
                         None
                     }
                 }
             } else {
-                // not a short option
-                // preserve it
+                // either not a short option
+                // or a short option that cannot have obsolete lines value in it
                 Some(slice.to_owned())
             }
         })

tests/by-util/test_split.rs

@@ -170,6 +170,22 @@ fn test_split_str_prefixed_chunks_by_bytes() {
     assert_eq!(glob.collate(), at.read_bytes(name));
 }
 
+// Test short bytes option concatenated with value
+#[test]
+fn test_split_by_bytes_short_concatenated_with_value() {
+    let (at, mut ucmd) = at_and_ucmd!();
+    let name = "split_by_bytes_short_concatenated_with_value";
+    RandomFile::new(&at, name).add_bytes(10000);
+    ucmd.args(&["-b1000", name]).succeeds();
+
+    let glob = Glob::new(&at, ".", r"x[[:alpha:]][[:alpha:]]$");
+    assert_eq!(glob.count(), 10);
+    for filename in glob.collect() {
+        assert_eq!(glob.directory.metadata(&filename).len(), 1000);
+    }
+    assert_eq!(glob.collate(), at.read_bytes(name));
+}
+
 // This is designed to test what happens when the desired part size is not a
 // multiple of the buffer size and we hopefully don't overshoot the desired part
 // size.
@@ -326,6 +342,19 @@ fn test_split_lines_number() {
         .stderr_only("split: invalid number of lines: 'file'\n");
 }
 
+// Test short lines option with value concatenated
+#[test]
+fn test_split_lines_short_concatenated_with_value() {
+    let (at, mut ucmd) = at_and_ucmd!();
+    let name = "split_num_prefixed_chunks_by_lines";
+    RandomFile::new(&at, name).add_lines(10000);
+    ucmd.args(&["-l1000", name]).succeeds();
+
+    let glob = Glob::new(&at, ".", r"x[[:alpha:]][[:alpha:]]$");
+    assert_eq!(glob.count(), 10);
+    assert_eq!(glob.collate(), at.read_bytes(name));
+}
+
 /// Test for obsolete lines option standalone
 #[test]
 fn test_split_obs_lines_standalone() {
@@ -692,6 +721,19 @@ fn test_invalid_suffix_length() {
         .stderr_contains("invalid suffix length: 'xyz'");
 }
 
+// Test short suffix length option with value concatenated
+#[test]
+fn test_split_suffix_length_short_concatenated_with_value() {
+    let (at, mut ucmd) = at_and_ucmd!();
+    let name = "split_num_prefixed_chunks_by_lines";
+    RandomFile::new(&at, name).add_lines(10000);
+    ucmd.args(&["-a4", name]).succeeds();
+
+    let glob = Glob::new(&at, ".", r"x[[:alpha:]][[:alpha:]][[:alpha:]][[:alpha:]]$");
+    assert_eq!(glob.count(), 10);
+    assert_eq!(glob.collate(), at.read_bytes(name));
+}
+
 #[test]
 fn test_include_newlines() {
     let (at, mut ucmd) = at_and_ucmd!();
@@ -710,6 +752,19 @@ fn test_include_newlines() {
     assert_eq!(s, "5\n");
 }
 
+// Test short number of chunks option concatenated with value
+#[test]
+fn test_split_number_chunks_short_concatenated_with_value() {
+    let (at, mut ucmd) = at_and_ucmd!();
+    ucmd.args(&["-n3", "threebytes.txt"])
+        .succeeds()
+        .no_stdout()
+        .no_stderr();
+    assert_eq!(at.read("xaa"), "a");
+    assert_eq!(at.read("xab"), "b");
+    assert_eq!(at.read("xac"), "c");
+}
+
 #[test]
 fn test_allow_empty_files() {
     let (at, mut ucmd) = at_and_ucmd!();
@@ -784,6 +839,16 @@ fn test_line_bytes() {
     assert_eq!(at.read("xad"), "ee\n");
 }
 
+#[test]
+fn test_line_bytes_concatenated_with_value() {
+    let (at, mut ucmd) = at_and_ucmd!();
+    ucmd.args(&["-C8", "letters.txt"]).succeeds();
+    assert_eq!(at.read("xaa"), "aaaaaaaa");
+    assert_eq!(at.read("xab"), "a\nbbbb\n");
+    assert_eq!(at.read("xac"), "cccc\ndd\n");
+    assert_eq!(at.read("xad"), "ee\n");
+}
+
 #[test]
 fn test_line_bytes_no_final_newline() {
     let (at, mut ucmd) = at_and_ucmd!();