Create a string_view like thing for utf8 strings

utf8_string_view provides length and character iteration over a std::string_view containing utf8 data.

2 files changed, 78 insertions, 0 deletions

diff --git a/lib/unicode.h b/lib/unicode.h
index dde1d3d..2945650 100644
--- a/lib/unicode.h
+++ b/lib/unicode.h
@@ -14,4 +14,69 @@ uint32_t get_codepoint(const char *);
 
 #ifdef __cplusplus
 }
+
+#	include <string_view>
+#	include <algorithm>
+
+struct utf8_string_view {
+	struct iter {
+		constexpr explicit iter(const char * p) : pos {p} { }
+
+		[[nodiscard]] auto
+		operator!=(const iter & other) const
+		{
+			return pos != other.pos;
+		}
+
+		auto &
+		operator++()
+		{
+			pos = next_char(pos);
+			return *this;
+		}
+
+		[[nodiscard]] auto
+		operator*() const
+		{
+			return get_codepoint(pos);
+		}
+
+	private:
+		const char * pos;
+	};
+
+	// cppcheck-suppress noExplicitConstructor; NOLINTNEXTLINE(hicpp-explicit-conversions)
+	template<typename... Args> constexpr utf8_string_view(Args &&... args) : str {std::forward<Args>(args)...} { }
+
+	[[nodiscard]] auto
+	begin() const
+	{
+		return iter {str.cbegin()};
+	}
+
+	[[nodiscard]] auto
+	end() const
+	{
+		return iter {str.cend()};
+	}
+
+	[[nodiscard]] size_t length() const;
+
+private:
+	std::string_view str;
+};
+template<> struct std::iterator_traits<utf8_string_view::iter> {
+	using difference_type = size_t;
+	using value_type = uint32_t;
+	using pointer = void;
+	using reference = void;
+	using iterator_category = std::forward_iterator_tag;
+};
+
+[[nodiscard]] inline size_t
+utf8_string_view::length() const
+{
+	return std::distance(begin(), end());
+}
+
 #endif
diff --git a/test/test-text.cpp b/test/test-text.cpp
index 6762b09..0df6885 100644
--- a/test/test-text.cpp
+++ b/test/test-text.cpp
@@ -9,8 +9,21 @@
 #include <glm/glm.hpp>
 #include <span>
 #include <ui/font.h>
+#include <unicode.h>
 #include <vector>
 
+BOOST_AUTO_TEST_CASE(utf8_string_view_iter)
+{
+	static constexpr utf8_string_view text {"Some UTF-8 €£²¹ text."};
+	static constexpr std::array codepoints {
+			83, 111, 109, 101, 32, 85, 84, 70, 45, 56, 32, 8364, 163, 178, 185, 32, 116, 101, 120, 116, 46};
+	BOOST_CHECK_EQUAL(std::count_if(text.begin(), text.end(), isspace), 3);
+	BOOST_CHECK_EQUAL(text.length(), 21);
+	std::vector<uint32_t> codepointsOut;
+	std::copy(text.begin(), text.end(), std::back_inserter(codepointsOut));
+	BOOST_CHECK_EQUAL_COLLECTIONS(codepoints.begin(), codepoints.end(), codepointsOut.begin(), codepointsOut.end());
+}
+
 struct FontTest : public Font {
 	FontTest() : Font {"/usr/share/fonts/corefonts/arial.ttf", 48} { }
 };