Utilities: Rewrite sort(1) to be more posixy

Userland/Utilities/sort.cpp

@@ -1,5 +1,6 @@
 /*
  * Copyright (c) 2018-2020, Andreas Kling <kling@serenityos.org>
+ * Copyright (c) 2022, Peter Elliott <pelliott@serenityos.org>
  *
  * SPDX-License-Identifier: BSD-2-Clause
  */
@@ -7,39 +8,119 @@
 #include <AK/DeprecatedString.h>
 #include <AK/QuickSort.h>
 #include <AK/Vector.h>
+#include <LibCore/ArgsParser.h>
+#include <LibCore/Stream.h>
 #include <LibCore/System.h>
 #include <LibMain/Main.h>
-#include <errno.h>
+#include <ctype.h>
-#include <stdio.h>
+
-#include <stdlib.h>
+struct Line {
-#include <string.h>
+    StringView key;
-#include <unistd.h>
+    long int numeric_key;
+    DeprecatedString line;
+    bool numeric;
+
+    bool operator<(Line const& other) const
+    {
+        if (numeric)
+            return numeric_key < other.numeric_key;
+
+        return key < other.key;
+    }
+
+    bool operator==(Line const& other) const
+    {
+        if (numeric)
+            return numeric_key == other.numeric_key;
+
+        return key == other.key;
+    }
+
+private:
+};
+
+template<>
+struct AK::Traits<Line> : public GenericTraits<Line> {
+    static unsigned hash(Line l)
+    {
+        if (l.numeric)
+            return l.numeric_key;
+
+        return l.key.hash();
+    }
+};
+
+struct Options {
+    size_t key_field { 0 };
+    bool unique { false };
+    bool numeric { false };
+    StringView separator { "\0", 1 };
+    Vector<DeprecatedString> files;
+};
+
+static ErrorOr<void> load_file(Options options, StringView filename, Vector<Line>& lines, HashTable<Line>& seen)
+{
+    auto file = TRY(Core::Stream::BufferedFile::create(
+        TRY(Core::Stream::File::open_file_or_standard_stream(filename, Core::Stream::OpenMode::Read))));
+
+    // FIXME: Unlimited line length
+    auto buffer = TRY(ByteBuffer::create_uninitialized(4096));
+    while (TRY(file->can_read_line())) {
+        DeprecatedString line = TRY(file->read_line(buffer));
+
+        StringView key = line;
+        if (options.key_field != 0) {
+            auto split = (options.separator[0])
+                ? line.split_view(options.separator[0])
+                : line.split_view(isspace);
+            if (options.key_field - 1 >= split.size()) {
+                key = ""sv;
+            } else {
+                key = split[options.key_field - 1];
+            }
+        }
+
+        Line l = { key, key.to_int().value_or(0), line, options.numeric };
+
+        if (!options.unique || !seen.contains(l)) {
+            lines.append(l);
+            if (options.unique)
+                seen.set(l);
+        }
+    }
+
+    return {};
+}
 
 ErrorOr<int> serenity_main([[maybe_unused]] Main::Arguments arguments)
 {
-    TRY(Core::System::pledge("stdio"sv));
+    TRY(Core::System::pledge("stdio rpath"));
 
-    Vector<DeprecatedString> lines;
+    Options options;
 
-    for (;;) {
+    Core::ArgsParser args_parser;
-        char* buffer = nullptr;
+    args_parser.add_option(options.key_field, "The field to sort by", "key-field", 'k', "keydef");
-        ssize_t buflen = 0;
+    args_parser.add_option(options.unique, "Don't emit duplicate lines", "unique", 'u');
-        size_t n;
+    args_parser.add_option(options.numeric, "treat the key field as a number", "numeric", 'n');
-        errno = 0;
+    args_parser.add_option(options.separator, "The separator to split fields by", "sep", 't', "char");
-        buflen = getline(&buffer, &n, stdin);
+    args_parser.add_positional_argument(options.files, "Files to sort", "file", Core::ArgsParser::Required::No);
-        if (buflen == -1 && errno != 0) {
+    args_parser.parse(arguments);
-            perror("getline");
+
-            exit(1);
+    Vector<Line> lines;
+    HashTable<Line> seen;
+
+    if (options.files.size() == 0) {
+        TRY(load_file(options, "-"sv, lines, seen));
+    } else {
+        for (auto& file : options.files) {
+            TRY(load_file(options, file, lines, seen));
         }
-        if (buflen == -1)
-            break;
-        lines.append({ buffer, AK::ShouldChomp::Chomp });
     }
 
     quick_sort(lines);
 
     for (auto& line : lines) {
-        outln("{}", line);
+        outln("{}", line.line);
     }
 
     return 0;