fix(parsing): Detect and move past the UTF8 byte order mark (BOM) at …

…the beginning of a data file (endless-sky#10121)

source/DataFile.cpp

@@ -109,7 +109,13 @@ void DataFile::LoadData(const string &data)
 	size_t lineNumber = 0;
 
 	size_t end = data.length();
-	for(size_t pos = 0; pos < end; )
+
+	size_t pos = 0;
+	// If the first character is the UTF8 byte order mark (BOM), skip it.
+	if(!Utf8::IsBOM(Utf8::DecodeCodePoint(data, pos)))
+		pos = 0;
+
+	while(pos < end)
 	{
 		++lineNumber;
 		size_t tokenPos = pos;

source/text/Utf8.cpp

@@ -23,6 +23,12 @@ this program. If not, see <https://www.gnu.org/licenses/>.
 
 using namespace std;
 
+namespace {
+	constexpr char32_t BOM = 0x0000FEFF;
+}
+
+
+
 namespace Utf8 {
 #if defined(_WIN32)
 	wstring ToUTF16(const string &input, bool isPath)
@@ -60,6 +66,14 @@ namespace Utf8 {
 
 
 
+	// Check if this character is the byte order mark (BOM) sequence.
+	bool IsBOM(char32_t c)
+	{
+		return c == BOM;
+	}
+
+
+
 	size_t NextCodePoint(const string &str, size_t pos)
 	{
 		if(pos >= str.length())

source/text/Utf8.h

@@ -25,6 +25,9 @@ namespace Utf8 {
 	std::string ToUTF8(const wchar_t *str);
 #endif
 
+	// Check if this character is the byte order mark (BOM) sequence.
+	bool IsBOM(char32_t c);
+
 	// Skip to the next unicode code point after pos in utf8.
 	// Return string::npos when there are no more code points.
 	std::size_t NextCodePoint(const std::string &str, std::size_t pos);