Add UTF-8, UTF-16 and Modified UTF-8 support

1 files changed, 86 insertions, 0 deletions

diff --git a/src/u16.hh b/src/u16.hh
new file mode 100644
index 0000000..6894a84
--- /dev/null
+++ b/src/u16.hh
@@ -0,0 +1,86 @@
+#ifndef U16_HH
+#define U16_HH
+
+#include <cstdint>
+#include <expected>
+#include <iterator>
+#include <type_traits>
+
+#include "u.hh"
+
+namespace u16 {
+
+template<std::forward_iterator T>
+  requires std::is_same_v<std::iter_value_t<T>, uint16_t>
+std::expected<uint32_t, u::ReadError> read(T& start, const T& end) {
+  if (start == end) return std::unexpected(u::ReadError::End);
+  uint16_t u = *start;
+  if (u >= 0xd800 && u <= 0xdbff) {
+    if (std::distance(start, end) < 2) {
+      return std::unexpected(u::ReadError::Incomplete);
+    }
+    std::advance(start, 1);
+    if (*start >= 0xdc00 && *start <= 0xdfff) {
+      uint16_t v = *start;
+      std::advance(start, 1);
+      return 0x10000 + (((u - 0xd800) << 10) | (v - 0xdc00));
+    }
+    return std::unexpected(u::ReadError::Invalid);
+  }
+  std::advance(start, 1);
+  if (u >= 0xdc00 && u <= 0xdfff) {
+    return std::unexpected(u::ReadError::Invalid);
+  }
+  return u;
+}
+
+template<std::forward_iterator T>
+  requires std::is_same_v<std::iter_value_t<T>, uint16_t>
+std::expected<uint32_t, u::ReadErrorReplace> read_replace(T& start,
+                                                          const T& end) {
+  auto ret = read(start, end);
+  if (ret.has_value())
+    return *ret;
+  switch (ret.error()) {
+    case u::ReadError::Incomplete:
+      return std::unexpected(u::ReadErrorReplace::Incomplete);
+    case u::ReadError::End:
+      return std::unexpected(u::ReadErrorReplace::End);
+    case u::ReadError::Invalid:
+      return 0xfffd;
+  }
+}
+
+template<std::forward_iterator T>
+  requires std::is_same_v<std::iter_value_t<T>, uint16_t>
+bool write(T& start, const T& end, uint32_t code) {
+  if (code < 0x10000) {
+    if (start == end) return false;
+    *start = static_cast<uint16_t>(code);
+  } else {
+    if (std::distance(start, end) < 2) return false;
+    code -= 0x10000;
+    *start = static_cast<uint16_t>(0xd800 + (code >> 10));
+    std::advance(start, 1);
+    *start = static_cast<uint16_t>(0xdc00 + (code & 0x3ff));
+  }
+  std::advance(start, 1);
+  return true;
+}
+
+template<std::forward_iterator T>
+  requires std::is_same_v<std::iter_value_t<T>, uint16_t>
+bool skip(T& start, const T& end) {
+  if (start == end) return false;
+  if (*start >= 0xd800 && *start <= 0xdbff) {
+    if (std::distance(start, end) < 2) return false;
+    std::advance(start, 2);
+    return true;
+  }
+  std::advance(start, 1);
+  return true;
+}
+
+}  // namespace u16
+
+#endif  // U16_HH