Make wrap_rows not wrap inside utf-8 multibyte sequences

est31 · est31 · commit 45a9bb310ee0 · 2015-06-17T22:19:56.000+02:00
Also count multibyte sequences as "one" character. Adds unittest for the bug reporter's case. Fixes #2796.
diff --git a/src/unittest/test_utilities.cpp b/src/unittest/test_utilities.cpp
@@ -242,6 +242,14 @@ void TestUtilities::testUTF8()
 void TestUtilities::testWrapRows()
 {
 	UASSERT(wrap_rows("12345678",4) == "1234\n5678");
+	// test that wrap_rows doesn't wrap inside multibyte sequences
+	const unsigned char s[] = {
+		0x2f, 0x68, 0x6f, 0x6d, 0x65, 0x2f, 0x72, 0x61, 0x70, 0x74, 0x6f,
+		0x72, 0x2f, 0xd1, 0x82, 0xd0, 0xb5, 0xd1, 0x81, 0xd1, 0x82, 0x2f,
+		0x6d, 0x69, 0x6e, 0x65, 0x74, 0x65, 0x73, 0x74, 0x2f, 0x62, 0x69,
+		0x6e, 0x2f, 0x2e, 0x2e, 0};
+	std::string str((char *)s);
+	UASSERT(utf8_to_wide(wrap_rows(str, 20)) != L"<invalid UTF-8 string>");
 }
 
 
diff --git a/src/util/string.h b/src/util/string.h
@@ -32,6 +32,11 @@ with this program; if not, write to the Free Software Foundation, Inc.,
 #define STRINGIFY(x) #x
 #define TOSTRING(x) STRINGIFY(x)
 
+// Checks whether a byte is an inner byte for an utf-8 multibyte sequence
+#define IS_UTF8_MULTB_INNER(x) (((unsigned char)x >= 0x80) && ((unsigned char)x <= 0xc0))
+// Checks whether a byte is a start byte for an utf-8 multibyte sequence
+#define IS_UTF8_MULTB_START(x) ((unsigned char)x >= 0xc0)
+
 typedef std::map<std::string, std::string> StringMap;
 
 struct FlagDesc {
@@ -411,7 +416,10 @@ inline bool string_allowed_blacklist(const std::string &str,
  *	every \p row_len characters whether it breaks a word or not.  It is
  *	intended to be used for, for example, showing paths in the GUI.
  *
- * @param from The string to be wrapped into rows.
+ * @note This function doesn't wrap inside utf-8 multibyte sequences and also
+ * 	counts multibyte sequences correcly as single characters.
+ *
+ * @param from The (utf-8) string to be wrapped into rows.
  * @param row_len The row length (in characters).
  * @return A new string with the wrapping applied.
  */
@@ -420,9 +428,12 @@ inline std::string wrap_rows(const std::string &from,
 {
 	std::string to;
 
+	size_t character_idx = 0;
 	for (size_t i = 0; i < from.size(); i++) {
-		if (i != 0 && i % row_len == 0)
+		if (character_idx > 0 && character_idx % row_len == 0)
 			to += '\n';
+		if (!IS_UTF8_MULTB_INNER(from[i]))
+			character_idx++;
 		to += from[i];
 	}
 

Original file line number	Diff line number	Diff line change
`@@ -242,6 +242,14 @@ void TestUtilities::testUTF8()`
`242`	`242`	`void TestUtilities::testWrapRows()`
`243`	`243`	`{`
`244`	`244`	`UASSERT(wrap_rows("12345678",4) == "1234\n5678");`
	`245`	`+ // test that wrap_rows doesn't wrap inside multibyte sequences`
	`246`	`+ const unsigned char s[] = {`
	`247`	`+ 0x2f, 0x68, 0x6f, 0x6d, 0x65, 0x2f, 0x72, 0x61, 0x70, 0x74, 0x6f,`
	`248`	`+ 0x72, 0x2f, 0xd1, 0x82, 0xd0, 0xb5, 0xd1, 0x81, 0xd1, 0x82, 0x2f,`
	`249`	`+ 0x6d, 0x69, 0x6e, 0x65, 0x74, 0x65, 0x73, 0x74, 0x2f, 0x62, 0x69,`
	`250`	`+ 0x6e, 0x2f, 0x2e, 0x2e, 0};`
	`251`	`+ std::string str((char *)s);`
	`252`	`+ UASSERT(utf8_to_wide(wrap_rows(str, 20)) != L"<invalid UTF-8 string>");`
`245`	`253`	`}`
`246`	`254`
`247`	`255`