Add utf8_length in String class

core/io/pck_packer.cpp

@@ -178,11 +178,12 @@ Error PCKPacker::flush(bool p_verbose) {
 	}
 
 	for (int i = 0; i < files.size(); i++) {
-		int string_len = files[i].path.utf8().length();
+		CharString utf8_string = files[i].path.utf8();
+		int string_len = utf8_string.length();
 		int pad = _get_pad(4, string_len);
 
 		fhead->store_32(string_len + pad);
-		fhead->store_buffer((const uint8_t *)files[i].path.utf8().get_data(), string_len);
+		fhead->store_buffer((const uint8_t *)utf8_string.get_data(), string_len);
 		for (int j = 0; j < pad; j++) {
 			fhead->store_8(0);
 		}

core/io/plist.cpp

@@ -253,7 +253,7 @@ size_t PListNode::get_asn1_size(uint8_t p_len_octets) const {
 
 			for (const KeyValue<String, Ref<PListNode>> &E : data_dict) {
 				size += 1 + _asn1_size_len(p_len_octets); // Sequence.
-				size += 1 + _asn1_size_len(p_len_octets) + E.key.utf8().length(); //Key.
+				size += 1 + _asn1_size_len(p_len_octets) + E.key.utf8_length(); //Key.
 				size += 1 + _asn1_size_len(p_len_octets) + E.value->get_asn1_size(p_len_octets); // Value.
 			}
 			return size;

core/string/ustring.cpp

@@ -2243,43 +2243,56 @@ Error String::parse_utf8(const char *p_utf8, int p_len, bool p_skip_cr) {
 	}
 }
 
-CharString String::utf8() const {
-	int l = length();
-	if (!l) {
-		return CharString();
+uint32_t String::unicode_codepoint_as_utf8_length(char32_t c) {
+	uint32_t utf8_size = 0;
+	if (c <= 0x7f) { // 7 bits.
+		utf8_size += 1;
+	} else if (c <= 0x7ff) { // 11 bits
+		utf8_size += 2;
+	} else if (c <= 0xffff) { // 16 bits
+		utf8_size += 3;
+	} else if (c <= 0x001fffff) { // 21 bits
+		utf8_size += 4;
+	} else if (c <= 0x03ffffff) { // 26 bits
+		utf8_size += 5;
+		print_error(vformat("Invalid unicode codepoint (%x)", c));
+	} else if (c <= 0x7fffffff) { // 31 bits
+		utf8_size += 6;
+		print_error(vformat("Invalid unicode codepoint (%x)", c));
+	} else {
+		utf8_size += 1;
+		print_error(vformat("Invalid unicode codepoint (%x), cannot represent as UTF-8", c));
 	}
 
-	const char32_t *d = &operator[](0);
-	int fl = 0;
-	for (int i = 0; i < l; i++) {
-		uint32_t c = d[i];
-		if (c <= 0x7f) { // 7 bits.
-			fl += 1;
-		} else if (c <= 0x7ff) { // 11 bits
-			fl += 2;
-		} else if (c <= 0xffff) { // 16 bits
-			fl += 3;
-		} else if (c <= 0x001fffff) { // 21 bits
-			fl += 4;
-		} else if (c <= 0x03ffffff) { // 26 bits
-			fl += 5;
-			print_unicode_error(vformat("Invalid unicode codepoint (%x)", c));
-		} else if (c <= 0x7fffffff) { // 31 bits
-			fl += 6;
-			print_unicode_error(vformat("Invalid unicode codepoint (%x)", c));
-		} else {
-			fl += 1;
-			print_unicode_error(vformat("Invalid unicode codepoint (%x), cannot represent as UTF-8", c), true);
-		}
+	return utf8_size;
+}
+
+uint32_t String::utf8_length() const {
+	int l = length();
+	const char32_t *dst = ptr();
+
+	uint32_t utf8_size = 0;
+
+	for (int i = 0; i < l; ++i) {
+		char32_t c = dst[i];
+		utf8_size += unicode_codepoint_as_utf8_length(c);
 	}
 
+	return utf8_size;
+}
+
+CharString String::utf8() const {
+	int fl = utf8_length();
+
 	CharString utf8s;
 	if (fl == 0) {
 		return utf8s;
 	}
 
 	utf8s.resize(fl + 1);
 	uint8_t *cdst = (uint8_t *)utf8s.get_data();
+	const char32_t *d = &operator[](0);
+	int l = length();
 
 #define APPEND_CHAR(m_c) *(cdst++) = m_c
 

core/string/ustring.h

@@ -404,6 +404,8 @@ class String {
 
 	CharString ascii(bool p_allow_extended = false) const;
 	CharString utf8() const;
+	uint32_t utf8_length() const;
+	static uint32_t unicode_codepoint_as_utf8_length(char32_t c);
 	Error parse_utf8(const char *p_utf8, int p_len = -1, bool p_skip_cr = false);
 	static String utf8(const char *p_utf8, int p_len = -1);
 

modules/gltf/gltf_document.cpp

@@ -7167,7 +7167,7 @@ PackedByteArray GLTFDocument::_serialize_glb_buffer(Ref<GLTFState> p_state, Erro
 	const int32_t header_size = 12;
 	const int32_t chunk_header_size = 8;
 
-	int32_t padding = (chunk_header_size + json.utf8().length()) % 4;
+	int32_t padding = (chunk_header_size + json.utf8_length()) % 4;
 	json += String(" ").repeat(padding);
 
 	CharString cs = json.utf8();

platform/linuxbsd/x11/display_server_x11.cpp

@@ -1920,7 +1920,8 @@ void DisplayServerX11::window_set_title(const String &p_title, WindowID p_window
 	Atom _net_wm_name = XInternAtom(x11_display, "_NET_WM_NAME", false);
 	Atom utf8_string = XInternAtom(x11_display, "UTF8_STRING", false);
 	if (_net_wm_name != None && utf8_string != None) {
-		XChangeProperty(x11_display, wd.x11_window, _net_wm_name, utf8_string, 8, PropModeReplace, (unsigned char *)p_title.utf8().get_data(), p_title.utf8().length());
+		CharString utf8_title = p_title.utf8();
+		XChangeProperty(x11_display, wd.x11_window, _net_wm_name, utf8_string, 8, PropModeReplace, (unsigned char *)utf8_title.get_data(), utf8_title.length());
 	}
 }
 

platform/windows/windows_utils.cpp

@@ -161,32 +161,37 @@ Error WindowsUtils::copy_and_rename_pdb(const String &p_dll_path) {
 		return ERR_SKIP;
 	}
 
-	String new_pdb_base_name = p_dll_path.get_file().get_basename() + "_";
+	String new_pdb_base_name = p_dll_path.get_file().get_basename();
 
 	// Checking the available space for the updated string
 	// and trying to shorten it if there is not much space.
 	{
-		// e.g. 999.pdb
-		const uint8_t suffix_size = String::num_characters((int64_t)max_pdb_names - 1) + 4;
-		// e.g. ~lib_ + 1 for the \0
-		const uint8_t min_base_size = 5 + 1;
-		int original_path_size = pdb_info.path.utf8().length();
-		CharString utf8_name = new_pdb_base_name.utf8();
-		int new_expected_buffer_size = utf8_name.length() + suffix_size;
+		// e.g. _999.pdb
+		const uint8_t suffix_size = String::num_characters((int64_t)max_pdb_names - 1) + 5;
+		int original_path_size = pdb_info.path.utf8_length();
+		int new_expected_buffer_size = new_pdb_base_name.utf8_length() + suffix_size;
 
 		// Since we have limited space inside the DLL to patch the path to the PDB,
 		// it is necessary to limit the size based on the number of bytes occupied by the string.
 		if (new_expected_buffer_size > original_path_size) {
+			// e.g. ~lib + 1 for the \0
+			const uint8_t min_base_size = 4 + 1;
 			ERR_FAIL_COND_V_MSG(original_path_size < min_base_size + suffix_size, FAILED, vformat("The original PDB path size in bytes is too small: '%s'. Expected size: %d or more bytes, but available %d.", pdb_info.path, min_base_size + suffix_size, original_path_size));
 
-			utf8_name.resize(original_path_size - suffix_size + 1); // +1 for the \0
-			utf8_name[utf8_name.size() - 1] = '\0';
-			new_pdb_base_name.parse_utf8(utf8_name);
-			new_pdb_base_name[new_pdb_base_name.length() - 1] = '_'; // Restore the last '_'
+			int too_much = new_expected_buffer_size - original_path_size;
+			const char32_t *ptr = new_pdb_base_name.ptr();
+			int32_t pdb_utf32_size = new_pdb_base_name.length();
+			while (too_much > 0) {
+				too_much -= String::unicode_codepoint_as_utf8_length(ptr[pdb_utf32_size]);
+				--pdb_utf32_size;
+			}
+			new_pdb_base_name.resize(pdb_utf32_size);
 			WARN_PRINT(vformat("The original path size of '%s' in bytes was too small to fit the new name, so it was shortened to '%s%d.pdb'.", pdb_info.path, new_pdb_base_name, max_pdb_names - 1));
 		}
 	}
 
+	new_pdb_base_name += '_';
+
 	// Delete old PDB files.
 	for (const String &file : DirAccess::get_files_at(dll_base_dir)) {
 		if (file.begins_with(new_pdb_base_name) && file.ends_with(".pdb")) {
@@ -222,7 +227,7 @@ Error WindowsUtils::copy_and_rename_pdb(const String &p_dll_path) {
 		Ref<FileAccess> file = FileAccess::open(p_dll_path, FileAccess::READ_WRITE, &err);
 		ERR_FAIL_COND_V_MSG(err != OK, err, vformat("Failed to open '%s' to patch the PDB path.", p_dll_path));
 
-		int original_path_size = pdb_info.path.utf8().length();
+		int original_path_size = pdb_info.path.utf8_length();
 		// Double-check file bounds.
 		ERR_FAIL_UNSIGNED_INDEX_V_MSG(pdb_info.address + original_path_size, file->get_length(), FAILED, vformat("Failed to write a new PDB path. Probably '%s' has been changed.", p_dll_path));