Utilities: Allow uniq to work on lines of arbitrary length

Calls to `read_line` are replaced with `read_line_with_resize`
and `swap`s of StringViews, which assume a consistent location
of the underlying ByteBuffers, are replaced. A test file has
been added for uniq, which includes a test case for long lines.

Tests/Utilities/CMakeLists.txt

@@ -1,6 +1,7 @@
 set(TEST_SOURCES
     TestSed.cpp
     TestPatch.cpp
+    TestUniq.cpp
 )
 
 foreach(source IN LISTS TEST_SOURCES)

Tests/Utilities/TestUniq.cpp

@@ -0,0 +1,52 @@
+/*
+ * Copyright (c) 2023, Rodrigo Tobar <rtobarc@gmail.com>.
+ * Copyright (c) 2024, Daniel Gaston <tfd@tuta.io>.
+ *
+ * SPDX-License-Identifier: BSD-2-Clause
+ */
+
+#include <AK/ScopeGuard.h>
+#include <AK/StringView.h>
+#include <LibCore/Command.h>
+#include <LibCore/File.h>
+#include <LibTest/Macros.h>
+#include <LibTest/TestCase.h>
+
+static void run_uniq(Vector<char const*>&& arguments, StringView standard_input, StringView expected_stdout)
+{
+    MUST(arguments.try_insert(0, "uniq"));
+    MUST(arguments.try_append(nullptr));
+    auto uniq = MUST(Core::Command::create("uniq"sv, arguments.data()));
+    MUST(uniq->write(standard_input));
+    auto [stdout, stderr] = MUST(uniq->read_all());
+    auto status = MUST(uniq->status());
+    if (status != Core::Command::ProcessResult::DoneWithZeroExitCode) {
+        FAIL(ByteString::formatted("uniq didn't exit cleanly: status: {}, stdout: {}, stderr: {}", static_cast<int>(status), StringView { stdout.bytes() }, StringView { stderr.bytes() }));
+    }
+    EXPECT_EQ(StringView { expected_stdout.bytes() }, StringView { stdout.bytes() });
+}
+
+TEST_CASE(two_duplicate_lines)
+{
+    run_uniq({}, "AAA\nAAA\n"sv, "AAA\n"sv);
+}
+
+TEST_CASE(two_unique_lines)
+{
+    run_uniq({}, "AAA\nAaA\n"sv, "AAA\nAaA\n"sv);
+}
+
+TEST_CASE(long_line)
+{
+    auto input = Array<u8, 4096> {};
+    auto expected_output = Array<u8, 2048> {};
+    // Create two lines of 2047 A's and a newline.
+    input.fill('A');
+    input[2047] = '\n';
+    input[4095] = '\n';
+
+    expected_output.fill('A');
+    expected_output[2047] = '\n';
+
+    run_uniq({}, StringView { input }, StringView { expected_output });
+}

Userland/Utilities/uniq.cpp

@@ -1,5 +1,6 @@
 /*
  * Copyright (c) 2020, Matthew L. Curry <matthew.curry@gmail.com>
+ * Copyright (c) 2024, Daniel Gaston <tfd@tuta.io>.
  *
  * SPDX-License-Identifier: BSD-2-Clause
  */
@@ -86,14 +87,12 @@ ErrorOr<int> serenity_main(Main::Arguments arguments)
     ByteBuffer previous_buf = TRY(ByteBuffer::create_uninitialized(1024));
     ByteBuffer current_buf = TRY(ByteBuffer::create_uninitialized(1024));
 
-    StringView previous = TRY(infile->read_line(previous_buf));
+    StringView previous = TRY(infile->read_line_with_resize(previous_buf));
     StringView previous_to_compare = skip(previous, skip_chars, skip_fields);
 
     while (TRY(infile->can_read_line())) {
-        // FIXME: The buffer does not automatically resize,
-        // and this will return EMSGSIZE if the read line
-        // is more than 1024 bytes.
-        StringView current = TRY(infile->read_line(current_buf));
+
+        StringView current = TRY(infile->read_line_with_resize(current_buf));
 
         StringView current_to_compare = skip(current, skip_chars, skip_fields);
         bool lines_equal = ignore_case ? current_to_compare.equals_ignoring_ascii_case(previous_to_compare) : current_to_compare == previous_to_compare;
@@ -103,9 +102,13 @@ ErrorOr<int> serenity_main(Main::Arguments arguments)
         } else {
             count++;
         }
-        swap(current_to_compare, previous_to_compare);
+
         swap(current_buf, previous_buf);
-        swap(current, previous);
+        // The StringViews cannot be swapped since read_line_with_resize
+        // potentially changes the location of the buffers due to reallocation.
+        // Instead create a new StringView of what was most recently read in.
+        previous = StringView { previous_buf.span().trim(current.length()) };
+        previous_to_compare = skip(previous, skip_chars, skip_fields);
     }
 
     TRY(write_line_content(previous, count, duplicates_only, print_count, *outfile));