1 files changed, 196 insertions, 0 deletions
diff --git a/src/u8.hh b/src/u8.hh
new file mode 100644
index 0000000..d673caa
--- /dev/null
+++ b/src/u8.hh
@@ -0,0 +1,196 @@
+#ifndef U8_HH
+#define U8_HH
+
+#include "u.hh"  // IWYU pragma: export
+
+#include <cstdint>  // IWYU pragma: export
+#include <expected>
+#include <iterator>
+#include <type_traits>
+#include <utility>
+
+namespace u8 {
+
+template <std::forward_iterator T>
+  requires std::is_same_v<std::iter_value_t<T>, uint8_t>
+std::expected<uint32_t, u::ReadError> read(T& start, T const& end) {
+  if (start == end)
+    return std::unexpected(u::ReadError::End);
+  uint32_t u;
+  switch (*start >> 4) {
+    case 0xf:
+      // 11110uvv 10vvwwww 10xxxxyy 10yyzzzz
+      if (std::distance(start, end) < 4) {
+        return std::unexpected(u::ReadError::Incomplete);
+      }
+      u = (*start & 0x07) << 18;
+      std::advance(start, 1);
+      if ((*start & 0xc0) != 0x80) {
+        std::advance(start, 3);
+        return std::unexpected(u::ReadError::Invalid);
+      }
+      u |= (*start & 0x3f) << 12;
+      std::advance(start, 1);
+      if ((*start & 0xc0) != 0x80) {
+        std::advance(start, 2);
+        return std::unexpected(u::ReadError::Invalid);
+      }
+      u |= (*start & 0x3f) << 6;
+      std::advance(start, 1);
+      if ((*start & 0xc0) != 0x80) {
+        std::advance(start, 1);
+        return std::unexpected(u::ReadError::Invalid);
+      }
+      u |= *start & 0x3f;
+      if (u < 0x10000 || u > 0x10ffff) {
+        std::advance(start, 1);
+        return std::unexpected(u::ReadError::Invalid);
+      }
+      break;
+    case 0xe:
+      // 1110wwww 10xxxxyy 10yyzzzz
+      if (std::distance(start, end) < 3) {
+        return std::unexpected(u::ReadError::Incomplete);
+      }
+      u = (*start & 0x0f) << 12;
+      std::advance(start, 1);
+      if ((*start & 0xc0) != 0x80) {
+        std::advance(start, 2);
+        return std::unexpected(u::ReadError::Invalid);
+      }
+      u |= (*start & 0x3f) << 6;
+      std::advance(start, 1);
+      if ((*start & 0xc0) != 0x80) {
+        std::advance(start, 1);
+        return std::unexpected(u::ReadError::Invalid);
+      }
+      u |= *start & 0x3f;
+      if (u < 0x800 || (u >= 0xd800 && u <= 0xdfff)) {
+        std::advance(start, 1);
+        return std::unexpected(u::ReadError::Invalid);
+      }
+      break;
+    case 0xd:
+    case 0xc:
+      // 110xxxyy 10yyzzzz
+      if (std::distance(start, end) < 2) {
+        return std::unexpected(u::ReadError::Incomplete);
+      }
+      u = (*start & 0x1f) << 6;
+      std::advance(start, 1);
+      if ((*start & 0xc0) != 0x80) {
+        std::advance(start, 1);
+        return std::unexpected(u::ReadError::Invalid);
+      }
+      u |= *start & 0x3f;
+      if (u < 0x80) {
+        std::advance(start, 1);
+        return std::unexpected(u::ReadError::Invalid);
+      }
+      break;
+    case 0xb:
+    case 0xa:
+    case 0x9:
+    case 0x8:
+      std::advance(start, 1);
+      return std::unexpected(u::ReadError::Invalid);
+    default:
+      // 0yyyzzzz
+      u = *start;
+      break;
+  }
+  std::advance(start, 1);
+  return u;
+}
+
+template <std::forward_iterator T>
+  requires std::is_same_v<std::iter_value_t<T>, uint8_t>
+std::expected<uint32_t, u::ReadErrorReplace> read_replace(T& start,
+                                                          T const& end,
+                                                          bool eof) {
+  auto const tmp = start;
+  auto ret = read(start, end);
+  if (ret.has_value())
+    return *ret;
+  switch (ret.error()) {
+    case u::ReadError::Incomplete:
+      if (eof)
+        break;
+      return std::unexpected(u::ReadErrorReplace::Incomplete);
+    case u::ReadError::End:
+      return std::unexpected(u::ReadErrorReplace::End);
+    case u::ReadError::Invalid:
+      break;
+  }
+  start = tmp + 1;
+  return 0xfffd;
+}
+
+template <std::forward_iterator T>
+  requires std::is_same_v<std::iter_value_t<T>, uint8_t>
+bool write(T& start, T const& end, uint32_t code) {
+  if (code < 0x80) {
+    if (start == end)
+      return false;
+    *start = static_cast<uint8_t>(code);
+  } else if (code < 0x800) {
+    if (std::distance(start, end) < 2)
+      return false;
+    *start = 0xc0 | static_cast<uint8_t>(code >> 6);
+    std::advance(start, 1);
+    *start = 0x80 | static_cast<uint8_t>(code & 0x3f);
+  } else if (code < 0x10000) {
+    if (std::distance(start, end) < 3)
+      return false;
+    *start = 0xe0 | static_cast<uint8_t>(code >> 12);
+    std::advance(start, 1);
+    *start = 0x80 | static_cast<uint8_t>((code >> 6) & 0x3f);
+    std::advance(start, 1);
+    *start = 0x80 | static_cast<uint8_t>(code & 0x3f);
+  } else {
+    if (std::distance(start, end) < 4)
+      return false;
+    *start = 0xf0 | static_cast<uint8_t>(code >> 18);
+    std::advance(start, 1);
+    *start = 0x80 | static_cast<uint8_t>((code >> 12) & 0x3f);
+    std::advance(start, 1);
+    *start = 0x80 | static_cast<uint8_t>((code >> 6) & 0x3f);
+    std::advance(start, 1);
+    *start = 0x80 | static_cast<uint8_t>(code & 0x3f);
+  }
+  std::advance(start, 1);
+  return true;
+}
+
+template <std::forward_iterator T>
+  requires std::is_same_v<std::iter_value_t<T>, uint8_t>
+bool skip(T& start, T const& end) {
+  if (start == end)
+    return false;
+  switch (*start >> 4) {
+    case 0xf:
+      if (std::distance(start, end) < 4)
+        return false;
+      std::advance(start, 4);
+      break;
+    case 0xe:
+      if (std::distance(start, end) < 3)
+        return false;
+      std::advance(start, 3);
+      break;
+    case 0xc:
+    case 0xd:
+      if (std::distance(start, end) < 2)
+        return false;
+      std::advance(start, 2);
+      break;
+    default:
+      std::advance(start, 1);
+      break;
+  }
+  return true;
+}
+
+}  // namespace u8
+
+#endif  // U8_HH