Support UTF-8 by StringRefId::toDebugString

apps/openmw_test_suite/esm/testrefid.cpp

@@ -287,7 +287,11 @@ namespace ESM
             { RefId(), "Empty{}" },
             { RefId::stringRefId("foo"), "\"foo\"" },
             { RefId::stringRefId("BAR"), "\"BAR\"" },
-            { RefId::stringRefId(std::string({ 'a', 0, -1, '\n', '\t' })), "\"a\\x0\\xFF\\xA\\x9\"" },
+            { RefId::stringRefId(std::string({ 'a', 0, -1, '\n', '\t' })), "\"a\\x0\\xff\\xa\\x9\"" },
+            { RefId::stringRefId("Логово дракона"), "\"Логово дракона\"" },
+            { RefId::stringRefId("\xd0\x9b"), "\"Л\"" },
+            { RefId::stringRefId("\xff\x9b"), "\"\\xff\\x9b\"" },
+            { RefId::stringRefId("\xd0\xd0"), "\"\\xd0\\xd0\"" },
             { RefId::formIdRefId({ .mIndex = 42, .mContentFile = 0 }), "FormId:0x2a" },
             { RefId::formIdRefId({ .mIndex = 0xffffff, .mContentFile = std::numeric_limits<std::int32_t>::min() }),
                 "FormId:0xff80000000ffffff" },

components/esm/stringrefid.cpp

@@ -1,13 +1,17 @@
 #include "stringrefid.hpp"
+#include "serializerefid.hpp"
 
+#include <charconv>
 #include <iomanip>
 #include <mutex>
 #include <ostream>
 #include <sstream>
+#include <system_error>
 #include <unordered_set>
 
 #include "components/misc/guarded.hpp"
 #include "components/misc/strings/algorithm.hpp"
+#include "components/misc/utf8stream.hpp"
 
 namespace ESM
 {
@@ -26,6 +30,18 @@ namespace ESM
                 it = locked->emplace(id).first;
             return &*it;
         }
+
+        void addHex(unsigned char value, std::string& result)
+        {
+            const std::size_t size = 2 + getHexIntegralSize(value);
+            const std::size_t shift = result.size();
+            result.resize(shift + size);
+            result[shift] = '\\';
+            result[shift + 1] = 'x';
+            const auto [end, ec] = std::to_chars(result.data() + shift + 2, result.data() + result.size(), value, 16);
+            if (ec != std::errc())
+                throw std::system_error(std::make_error_code(ec));
+        }
     }
 
     StringRefId::StringRefId()
@@ -60,20 +76,43 @@ namespace ESM
 
     std::ostream& operator<<(std::ostream& stream, StringRefId value)
     {
-        stream << '"';
-        for (char c : *value.mValue)
-            if (std::isprint(c) && c != '\t' && c != '\n' && c != '\r')
-                stream << c;
-            else
-                stream << "\\x" << std::hex << std::uppercase << static_cast<unsigned>(static_cast<unsigned char>(c));
-        return stream << '"';
+        return stream << value.toDebugString();
     }
 
     std::string StringRefId::toDebugString() const
     {
-        std::ostringstream stream;
-        stream << *this;
-        return stream.str();
+        std::string result;
+        result.reserve(2 + mValue->size());
+        result.push_back('"');
+        const unsigned char* ptr = reinterpret_cast<const unsigned char*>(mValue->data());
+        const unsigned char* const end = reinterpret_cast<const unsigned char*>(mValue->data() + mValue->size());
+        while (ptr != end)
+        {
+            if (Utf8Stream::isAscii(*ptr))
+            {
+                if (std::isprint(*ptr) && *ptr != '\t' && *ptr != '\n' && *ptr != '\r')
+                    result.push_back(*ptr);
+                else
+                    addHex(*ptr, result);
+                ++ptr;
+                continue;
+            }
+            const auto [octets, first] = Utf8Stream::getOctetCount(*ptr);
+            const auto [chr, next] = Utf8Stream::decode(ptr + 1, end, first, octets);
+            if (chr == Utf8Stream::sBadChar())
+            {
+                while (ptr != std::min(end, ptr + octets))
+                {
+                    addHex(*ptr, result);
+                    ++ptr;
+                }
+                continue;
+            }
+            result.append(ptr, next);
+            ptr = next;
+        }
+        result.push_back('"');
+        return result;
     }
 
     bool StringRefId::startsWith(std::string_view prefix) const

components/misc/utf8stream.hpp

@@ -1,6 +1,7 @@
 #ifndef MISC_UTF8ITER_HPP
 #define MISC_UTF8ITER_HPP
 
+#include <cstdint>
 #include <cstring>
 #include <string>
 #include <string_view>
@@ -63,9 +64,11 @@ public:
         return val;
     }
 
+    static bool isAscii(unsigned char value) { return (value & 0x80) == 0; }
+
     static std::pair<UnicodeChar, Point> decode(Point cur, Point end)
     {
-        if ((*cur & 0x80) == 0)
+        if (isAscii(*cur))
         {
             UnicodeChar chr = *cur++;
 
@@ -75,8 +78,13 @@ public:
         int octets;
         UnicodeChar chr;
 
-        std::tie(octets, chr) = octet_count(*cur++);
+        std::tie(octets, chr) = getOctetCount(*cur++);
 
+        return decode(cur, end, chr, octets);
+    }
+
+    static std::pair<UnicodeChar, Point> decode(Point cur, Point end, UnicodeChar chr, std::size_t octets)
+    {
         if (octets > 5)
             return std::make_pair(sBadChar(), cur);
 
@@ -161,10 +169,9 @@ public:
         return out;
     }
 
-private:
-    static std::pair<int, UnicodeChar> octet_count(unsigned char octet)
+    static std::pair<std::size_t, UnicodeChar> getOctetCount(unsigned char octet)
     {
-        int octets;
+        std::size_t octets;
 
         unsigned char mark = 0xC0;
         unsigned char mask = 0xE0;
@@ -181,6 +188,7 @@ private:
         return std::make_pair(octets, octet & ~mask);
     }
 
+private:
     void next() { std::tie(val, nxt) = decode(nxt, end); }
 
     Point cur;