AK: Add SplitBehavior::KeepTrailingSeparator with tests

Author: https://github.com/demostanis
Commit: https://github.com/SerenityOS/serenity/commit/7c33f8f7df
Pull-request: https://github.com/SerenityOS/serenity/pull/15750
Reviewed-by: https://github.com/AtkinsSJ
Reviewed-by: https://github.com/davidot
Reviewed-by: https://github.com/linusg ✅

AK/String.cpp

@@ -110,12 +110,13 @@ Vector<String> String::split_limit(char separator, size_t limit, SplitBehavior s
     Vector<String> v;
     size_t substart = 0;
     bool keep_empty = has_flag(split_behavior, SplitBehavior::KeepEmpty);
+    bool keep_separator = has_flag(split_behavior, SplitBehavior::KeepTrailingSeparator);
     for (size_t i = 0; i < length() && (v.size() + 1) != limit; ++i) {
         char ch = characters()[i];
         if (ch == separator) {
             size_t sublen = i - substart;
             if (sublen != 0 || keep_empty)
-                v.append(substring(substart, sublen));
+                v.append(substring(substart, keep_separator ? sublen + 1 : sublen));
             substart = i + 1;
         }
     }
@@ -133,12 +134,13 @@ Vector<StringView> String::split_view(Function<bool(char)> separator, SplitBehav
     Vector<StringView> v;
     size_t substart = 0;
     bool keep_empty = has_flag(split_behavior, SplitBehavior::KeepEmpty);
+    bool keep_separator = has_flag(split_behavior, SplitBehavior::KeepTrailingSeparator);
     for (size_t i = 0; i < length(); ++i) {
         char ch = characters()[i];
         if (separator(ch)) {
             size_t sublen = i - substart;
             if (sublen != 0 || keep_empty)
-                v.append(substring_view(substart, sublen));
+                v.append(substring_view(substart, keep_separator ? sublen + 1 : sublen));
             substart = i + 1;
         }
     }

AK/StringUtils.h

@@ -40,10 +40,17 @@ enum class TrimWhitespace {
 };
 
 enum class SplitBehavior : unsigned {
+    // Neither keep empty substrings nor keep the trailing separator.
+    // This is the default behavior if unspecified.
     Nothing = 0,
+
     // If two separators follow each other without any characters
-    // in between, keep a "" in the resulting vector.
+    // in between, keep a "" in the resulting vector. (or only the
+    // separator if KeepTrailingSeparator is used)
     KeepEmpty = 1,
+
+    // Do not strip off the separator at the end of the string.
+    KeepTrailingSeparator = 2,
 };
 AK_ENUM_BITWISE_OPERATORS(SplitBehavior);
 

AK/StringView.cpp

@@ -272,12 +272,13 @@ Vector<StringView> StringView::split_view_if(Function<bool(char)> const& predica
     Vector<StringView> v;
     size_t substart = 0;
     bool keep_empty = has_flag(split_behavior, SplitBehavior::KeepEmpty);
+    bool keep_separator = has_flag(split_behavior, SplitBehavior::KeepTrailingSeparator);
     for (size_t i = 0; i < length(); ++i) {
         char ch = characters_without_null_termination()[i];
         if (predicate(ch)) {
             size_t sublen = i - substart;
             if (sublen != 0 || keep_empty)
-                v.append(substring_view(substart, sublen));
+                v.append(substring_view(substart, keep_separator ? sublen + 1 : sublen));
             substart = i + 1;
         }
     }

AK/StringView.h

@@ -170,11 +170,16 @@ public:
 
         auto maybe_separator_index = find(separator);
         bool keep_empty = has_flag(split_behavior, SplitBehavior::KeepEmpty);
+        bool keep_separator = has_flag(split_behavior, SplitBehavior::KeepTrailingSeparator);
         while (maybe_separator_index.has_value()) {
             auto separator_index = maybe_separator_index.value();
             auto part_with_separator = view.substring_view(0, separator_index + separator.length());
-            if (keep_empty || separator_index > 0)
+            if (keep_empty || separator_index > 0) {
-                callback(part_with_separator.substring_view(0, separator_index));
+                if (keep_separator)
+                    callback(part_with_separator);
+                else
+                    callback(part_with_separator.substring_view(0, separator_index));
+            }
             view = view.substring_view_starting_after_substring(part_with_separator);
             maybe_separator_index = view.find(separator);
         }

Tests/AK/TestStringView.cpp

@@ -163,6 +163,11 @@ TEST_CASE(split_view)
     EXPECT_EQ(test_string_view.split_view_if(predicate, SplitBehavior::KeepEmpty), Vector({ "a"sv, ""sv, "b"sv, "c"sv, "d"sv }));
     EXPECT_EQ(test_string_view.split_view_if(predicate), Vector({ "a"sv, "b"sv, "c"sv, "d"sv }));
     EXPECT_EQ(test_string_view.split_view_if(predicate, SplitBehavior::KeepEmpty), Vector({ "a"sv, ""sv, "b"sv, "c"sv, "d"sv }));
+
+    test_string_view = "a,,,b"sv;
+    EXPECT_EQ(test_string_view.split_view(","sv, SplitBehavior::KeepEmpty), Vector({ "a"sv, ""sv, ""sv, "b"sv }));
+    EXPECT_EQ(test_string_view.split_view(","sv, SplitBehavior::KeepTrailingSeparator), Vector({ "a,"sv, "b"sv }));
+    EXPECT_EQ(test_string_view.split_view(","sv, SplitBehavior::KeepTrailingSeparator | SplitBehavior::KeepEmpty), Vector({ "a,"sv, ","sv, ","sv, "b"sv }));
 }
 
 TEST_CASE(constexpr_stuff)