Incorrect String Conversion in JNI to C Bindings

[zpv]

What

Incorrect handling of string conversion between Java strings and C strings in the JNI bindings. The conversion does not correctly handle characters that are encoded into 4-bytes in UTF-8 (e.g. some emojis).

ldk-garbagecollected/java_strings.py

Lines 465 to 487 in 3e33cfd

	static inline jstring str_ref_to_java(JNIEnv *env, const char* chars, size_t len) {
	// Sadly we need to create a temporary because Java can't accept a char* without a 0-terminator
	char* conv_buf = MALLOC(len + 1, "str conv buf");
	memcpy(conv_buf, chars, len);
	conv_buf[len] = 0;
	jstring ret = (*env)->NewStringUTF(env, conv_buf);
	FREE(conv_buf);
	return ret;
	}
	static inline LDKStr java_to_owned_str(JNIEnv *env, jstring str) {
	uint64_t str_len = (*env)->GetStringUTFLength(env, str);
	char* newchars = MALLOC(str_len + 1, "String chars");
	const char* jchars = (*env)->GetStringUTFChars(env, str, NULL);
	memcpy(newchars, jchars, str_len);
	newchars[str_len] = 0;
	(*env)->ReleaseStringUTFChars(env, str, jchars);
	LDKStr res = {
	.chars = newchars,
	.len = str_len,
	.chars_is_owned = true
	};
	return res;
	}

• Current conversion functions GetStringUTFChars and NewStringUTF expect Modified UTF-8 (MUTF-8) format and not regular UTF-8.

• Java's MUTF-8 to C's UTF-8 encoding is causing data corruption for certain UTF-8 characters.

Possible Fix?

https://stackoverflow.com/a/32215302

• Use GetStringChars() to get the original UTF-16 encoded characters, and then create your own UTF-8 byte array from that. The conversion from UTF-16 to UTF-8 is a very simply algorithm to implement by hand, or you can use any pre-existing implementation provided by your platform or 3rd party libraries.

• Have your JNI code call back into Java to invoke the String.getBytes(String charsetName) method to encode the jstring object to a UTF-8 byte array, eg:

[TheBlueMatt]

I'm fine with either proposed solution, the second is easier to reason about but a chunk more code and criss-crossing different parts of the codebase, if the first solution is as simple as random stackoverflow commenters think that's probably simpler.