Userland: Add partial support for complex specifications to tr

Userland/Utilities/tr.cpp

@@ -7,8 +7,86 @@
 #include <AK/Optional.h>
 #include <AK/String.h>
 #include <LibCore/ArgsParser.h>
+#include <ctype.h>
 #include <stdio.h>
 
+static bool is_octal(int c)
+{
+    return c == '0' || c == '1' || c == '2' || c == '3' || c == '4' || c == '5' || c == '6' || c == '7';
+}
+
+static void generate_character_class(Function<int(int)> oracle, StringBuilder& out)
+{
+    for (int i = 0; i < 128; i++) {
+        if (oracle(i))
+            out.append(static_cast<char>(i));
+    }
+}
+
+static String build_set(StringView specification)
+{
+    StringBuilder out;
+    GenericLexer lexer(specification);
+
+    while (!lexer.is_eof()) {
+        if (lexer.consume_specific("[:alnum:]"sv))
+            generate_character_class(isalnum, out);
+        else if (lexer.consume_specific("[:blank:]"sv))
+            generate_character_class(isblank, out);
+        else if (lexer.consume_specific("[:digit:]"sv))
+            generate_character_class(isdigit, out);
+        else if (lexer.consume_specific("[:lower:]"sv))
+            generate_character_class(islower, out);
+        else if (lexer.consume_specific("[:punct:]"sv))
+            generate_character_class(ispunct, out);
+        else if (lexer.consume_specific("[:upper:]"sv))
+            generate_character_class(isupper, out);
+        else if (lexer.consume_specific("[:alpha:]"sv))
+            generate_character_class(isalpha, out);
+        else if (lexer.consume_specific("[:cntrl:]"sv))
+            generate_character_class(iscntrl, out);
+        else if (lexer.consume_specific("[:graph:]"sv))
+            generate_character_class(isgraph, out);
+        else if (lexer.consume_specific("[:print:]"sv))
+            generate_character_class(isprint, out);
+        else if (lexer.consume_specific("[:space:]"sv))
+            generate_character_class(isspace, out);
+        else if (lexer.consume_specific("[:xdigit:]"sv))
+            generate_character_class(isxdigit, out);
+        else if (lexer.consume_specific("\\\\"sv))
+            out.append('\\');
+        else if (lexer.consume_specific("\\a"sv))
+            out.append('\a');
+        else if (lexer.consume_specific("\\b"sv))
+            out.append('\b');
+        else if (lexer.consume_specific("\\f"sv))
+            out.append('\f');
+        else if (lexer.consume_specific("\\n"sv))
+            out.append('\n');
+        else if (lexer.consume_specific("\\r"sv))
+            out.append('\r');
+        else if (lexer.consume_specific("\\t"sv))
+            out.append('\t');
+        else if (lexer.consume_specific("\\v"sv))
+            out.append('\v');
+        else if (lexer.next_is('\\') && is_octal(lexer.peek(1))) {
+            lexer.consume_specific('\\');
+            int max_left_over = 3;
+            auto octal_digits = lexer.consume_while([&](char i) -> bool {
+                return is_octal(i) && max_left_over--;
+            });
+
+            int value = 0;
+            for (char ch : octal_digits)
+                value = value * 8 + (ch - '0');
+            out.append(static_cast<char>(value));
+        } else
+            out.append(lexer.consume(1));
+    }
+
+    return out.to_string();
+}
+
 int main(int argc, char** argv)
 {
     bool complement_flag = false;
@@ -28,19 +106,14 @@ int main(int argc, char** argv)
         return 1;
     }
 
-    String from_complement;
+    String from_str = build_set(from_chars);
-    StringView from_str;
     if (complement_flag) {
-        auto original_set = StringView(from_chars);
         StringBuilder complement_set;
-        for (int i = 0; i < 256; i++) {
+        for (int ch = 0; ch < 256; ch++) {
-            if (!original_set.contains(i))
+            if (!from_str.contains(static_cast<char>(ch)))
-                complement_set.append(static_cast<int>(i));
+                complement_set.append(static_cast<char>(ch));
         }
-        from_complement = complement_set.to_string();
+        from_str = complement_set.to_string();
-        from_str = from_complement;
-    } else {
-        from_str = from_chars;
     }
 
     if (delete_flag) {
@@ -52,7 +125,7 @@ int main(int argc, char** argv)
                 putchar(ch);
         }
     } else {
-        auto to_str = AK::StringView(to_chars);
+        auto to_str = build_set(to_chars);
 
         for (;;) {
             char ch = fgetc(stdin);