From fc4547b412e28164af1bf8981234c6af959ccc0b Mon Sep 17 00:00:00 2001
From: Joel Klinghed <the_jk@spawned.biz>
Date: Tue, 13 Jun 2023 10:07:16 +0200
Subject: WIP

---
 utf/src/utf16.cc | 67 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
 utf/src/utf32.cc | 43 +++++++++++++++++++++++++++++++++++
 utf/src/utf8.cc  | 68 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 178 insertions(+)
 create mode 100644 utf/src/utf16.cc
 create mode 100644 utf/src/utf32.cc
 create mode 100644 utf/src/utf8.cc

(limited to 'utf/src')

diff --git a/utf/src/utf16.cc b/utf/src/utf16.cc
new file mode 100644
index 0000000..43595bf
--- /dev/null
+++ b/utf/src/utf16.cc
@@ -0,0 +1,67 @@
+#include "utf16.hh"
+
+#include "utf_error.hh"
+
+namespace utf {
+
+namespace {
+
+inline bool is_high_surrogate(uint16_t c) {
+  return c >= 0xd800 && c <= 0xdbff;
+}
+
+inline bool is_low_surrogate(uint16_t c) {
+  return c >= 0xdc00 && c <= 0xdfff;
+}
+
+}  // namespace
+
+uint32_t read16be(std::string_view data, std::size_t& offset) {
+  if (offset > data.size() || data.size() - offset < 2)
+    return NEED_MORE;
+  uint16_t c = static_cast<uint16_t>(data[offset]) << 8
+               | static_cast<uint16_t>(data[offset + 1] & 0xff);
+  if (is_high_surrogate(c)) {
+    if (data.size() - offset < 4)
+      return NEED_MORE;
+    uint16_t d = static_cast<uint16_t>(data[offset + 2]) << 8
+                 | static_cast<uint16_t>(data[offset + 3] & 0xff);
+    if (is_low_surrogate(d)) {
+      offset += 4;
+      return 0x10000
+          + (static_cast<uint32_t>(c & 0x3ff) << 10
+             | (d & 0x3ff));
+    }
+    return INVALID;
+  } else if (is_low_surrogate(c)) {
+    return INVALID;
+  }
+  offset += 2;
+  return c;
+}
+
+uint32_t read16le(std::string_view data, std::size_t& offset) {
+  if (offset > data.size() || data.size() - offset < 2)
+    return NEED_MORE;
+  uint16_t c = static_cast<uint16_t>(data[offset + 1]) << 8
+               | static_cast<uint16_t>(data[offset] & 0xff);
+  if (is_high_surrogate(c)) {
+    if (data.size() - offset < 4)
+      return NEED_MORE;
+    uint16_t d = static_cast<uint16_t>(data[offset + 3]) << 8
+                 | static_cast<uint16_t>(data[offset + 2] & 0xff);
+    if (is_low_surrogate(d)) {
+      offset += 4;
+      return 0x10000
+          + (static_cast<uint32_t>(c & 0x3ff) << 10
+             | (d & 0x3ff));
+    }
+    return INVALID;
+  } else if (is_low_surrogate(c)) {
+    return INVALID;
+  }
+  offset += 2;
+  return c;
+}
+
+}  // namespace utf
diff --git a/utf/src/utf32.cc b/utf/src/utf32.cc
new file mode 100644
index 0000000..cfa29b6
--- /dev/null
+++ b/utf/src/utf32.cc
@@ -0,0 +1,43 @@
+#include "utf32.hh"
+
+#include "utf_error.hh"
+
+namespace utf {
+
+namespace {
+
+inline bool valid_codepoint(uint32_t c) {
+  return (c < 0xd800) || (c > 0xdfff && c <= 0x10ffff);
+}
+
+}  // namespace
+
+uint32_t read32be(std::string_view data, std::size_t& offset) {
+  if (offset > data.size() || data.size() - offset < 4)
+    return NEED_MORE;
+  uint32_t c = static_cast<uint32_t>(data[offset]) << 24
+               | static_cast<uint32_t>(data[offset + 1] & 0xff) << 16
+               | static_cast<uint32_t>(data[offset + 2] & 0xff) << 8
+               | static_cast<uint32_t>(data[offset + 3] & 0xff);
+  if (valid_codepoint(c)) {
+    offset += 4;
+    return c;
+  }
+  return INVALID;
+}
+
+uint32_t read32le(std::string_view data, std::size_t& offset) {
+  if (offset > data.size() || data.size() - offset < 4)
+    return NEED_MORE;
+  uint32_t c = static_cast<uint32_t>(data[offset + 3]) << 24
+               | static_cast<uint32_t>(data[offset + 2] & 0xff) << 16
+               | static_cast<uint32_t>(data[offset + 1] & 0xff) << 8
+               | static_cast<uint32_t>(data[offset] & 0xff);
+  if (valid_codepoint(c)) {
+    offset += 4;
+    return c;
+  }
+  return INVALID;
+}
+
+}  // namespace utf
diff --git a/utf/src/utf8.cc b/utf/src/utf8.cc
new file mode 100644
index 0000000..54b0296
--- /dev/null
+++ b/utf/src/utf8.cc
@@ -0,0 +1,68 @@
+#include "utf8.hh"
+
+#include "utf_error.hh"
+
+namespace utf {
+
+namespace {
+
+inline bool valid_codepoint(uint32_t c) {
+  return (c < 0xd800) || (c > 0xdfff && c <= 0x10ffff);
+}
+
+}  // namespace
+
+uint32_t read8(std::string_view data, std::size_t& offset) {
+  if (offset >= data.size())
+    return NEED_MORE;
+  uint32_t ret;
+  uint8_t size;
+  switch (static_cast<uint8_t>(data[offset]) >> 4) {
+    case 15:
+      if (data[offset] & 0x08)
+        return INVALID;
+      ret = static_cast<uint32_t>(data[offset] & 0x07) << 18;
+      size = 4;
+      break;
+    case 14:
+      ret = static_cast<uint32_t>(data[offset] & 0x0f) << 12;
+      size = 3;
+      break;
+    case 13:
+    case 12:
+      ret = static_cast<uint32_t>(data[offset] & 0x1f) << 6;
+      size = 2;
+      break;
+    default:
+      if (data[offset] & 0x80)
+        return INVALID;
+      return data[offset++];
+  }
+  if (data.size() - offset < size)
+    return NEED_MORE;
+  for (uint8_t i = 1; i < size; ++i) {
+    if ((data[offset + i] & 0xc0) != 0x80)
+      return INVALID;
+    ret |= static_cast<uint32_t>(data[offset + i] & 0x3f) << (size - i - 1) * 6;
+  }
+  if (!valid_codepoint(ret))
+    return INVALID;
+  switch (size) {
+    case 4:
+      if (ret < 0x10000)
+        return INVALID;
+      break;
+    case 3:
+      if (ret < 0x800)
+        return INVALID;
+      break;
+    case 2:
+      if (ret < 0x80)
+        return INVALID;
+      break;
+  }
+  offset += size;
+  return ret;
+}
+
+}  // namespace utf
-- 
cgit v1.2.3-70-g09d2