X-Git-Url: http://git.bitcoin.ninja/index.cgi?p=ldk-java;a=blobdiff_plain;f=java_strings.py;h=2fcf836d23b900bcf87bd3b7e5875c5ee0c52523;hp=f5df004d80d3dd709c54366165a712ceb461337e;hb=HEAD;hpb=40764e2a87c8cc70c5749a4d681f68842f975f59

diff --git a/java_strings.py b/java_strings.py
index f5df004d..7e35d175 100644
--- a/java_strings.py
+++ b/java_strings.py
@@ -17,6 +17,8 @@ class Consts:
             uint16_t = ['short'],
             uint32_t = ['int'],
             uint64_t = ['long'],
+            int64_t = ['long'],
+            double = ['double'],
         )
         self.java_type_map = dict(
             String = "String"
@@ -70,6 +72,7 @@ public class bindings {
 		// Fetching the LDK versions from C also checks that the header and binaries match
 		System.err.println("Loaded LDK-Java Bindings " + version.get_ldk_java_bindings_version() + " with LDK " + get_ldk_version() + " and LDK-C-Bindings " + get_ldk_c_bindings_version());
 	}
+	public static void run_statics() { /* Useful to force the statics to run */ }
 	static native void init(java.lang.Class c);
 	static native void init_class_cache();
 	static native String get_lib_version_string();
@@ -130,6 +133,38 @@ class CommonBase {
 	}
 }"""
 
+        self.txin_defn = """public class TxIn extends CommonBase {
+	/** The witness in this input, in serialized form */
+	public final byte[] witness;
+	/** The script_sig in this input */
+	public final byte[] script_sig;
+	/** The transaction output's sequence number */
+	public final int sequence;
+	/** The txid this input is spending */
+	public final byte[] previous_txid;
+	/** The output index within the spent transaction of the output this input is spending */
+	public final int previous_vout;
+
+	TxIn(java.lang.Object _dummy, long ptr) {
+		super(ptr);
+		this.witness = bindings.TxIn_get_witness(ptr);
+		this.script_sig = bindings.TxIn_get_script_sig(ptr);
+		this.sequence = bindings.TxIn_get_sequence(ptr);
+		this.previous_txid = bindings.TxIn_get_previous_txid(ptr);
+		this.previous_vout = bindings.TxIn_get_previous_vout(ptr);
+	}
+	/** Constructs a new TxIn, note that previous_txid must be exactly 32 bytes */
+	public TxIn(byte[] witness, byte[] script_sig, int sequence, byte[] previous_txid, int previous_vout) {
+		this(null, bindings.TxIn_new(witness, script_sig, sequence, previous_txid, previous_vout));
+	}
+
+	@Override @SuppressWarnings(\"deprecation\")
+	protected void finalize() throws Throwable {
+		super.finalize();
+		if (ptr != 0) { bindings.TxIn_free(ptr); }
+	}
+}"""
+
         self.scalar_defn = """public class BigEndianScalar extends CommonBase {
 	/** The bytes of the scalar value, in big endian */
 	public final byte[] scalar_bytes;
@@ -150,6 +185,34 @@ class CommonBase {
 	}
 }"""
 
+        self.witness_program_defn = """public class WitnessProgram extends CommonBase {
+	/** The witness program bytes themselves */
+	public final byte[] program;
+	/** The witness version */
+	public final WitnessVersion version;
+
+	WitnessProgram(java.lang.Object _dummy, long ptr) {
+		super(ptr);
+		this.program = bindings.WitnessProgram_get_program(ptr);
+		this.version = new WitnessVersion(bindings.WitnessProgram_get_version(ptr));
+	}
+	static byte check_args(byte[] program, WitnessVersion version) {
+		if (program.length < 2 || program.length > 40) throw new IllegalArgumentException();
+		if (version.getVal() == 0 && program.length != 20 && program.length != 32) throw new IllegalArgumentException();
+		return version.getVal();
+	}
+	public WitnessProgram(byte[] program, WitnessVersion version) {
+		super(bindings.WitnessProgram_new(check_args(program, version), program));
+		this.program = bindings.WitnessProgram_get_program(ptr);
+		this.version = new WitnessVersion(bindings.WitnessProgram_get_version(ptr));
+	}
+
+	@Override @SuppressWarnings(\"deprecation\")
+	protected void finalize() throws Throwable {
+		super.finalize();
+		if (ptr != 0) { bindings.WitnessProgram_free(ptr); }
+	}
+}"""
 
         self.c_file_pfx = """#include <jni.h>
 // On OSX jlong (ie long long) is not equivalent to int64_t, so we override here
@@ -429,25 +492,133 @@ typedef jlongArray int64_tArray;
 typedef jbyteArray int8_tArray;
 typedef jshortArray int16_tArray;
 
-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);
+static inline jstring str_ref_to_java(JNIEnv *env, const unsigned char* chars, size_t len) {
+	// Java uses "Modified UTF-8" rather than UTF-8. This requires special
+	// handling for codepoints above 0xFFFF, which get converted from four
+	// bytes to six. We don't know upfront how many codepoints in the string
+	// are above 0xFFFF, so we just allocate an extra 33% up front and waste a
+	// bit of space.
+	unsigned char* java_chars = MALLOC(len * 3 / 2 + 1, "str conv buf");
+	unsigned char* next_java_char = java_chars;
+	const unsigned char* next_in_char = chars;
+	const unsigned char* end = chars + len;
+	#define COPY_CHAR_TO_JAVA do { *next_java_char = *next_in_char; next_java_char++; next_in_char++; } while (0)
+
+	while (next_in_char < end) {
+		if (!*next_in_char) break;
+		if (!(*next_in_char & 0b10000000)) {
+			COPY_CHAR_TO_JAVA;
+		} else if ((*next_in_char & 0b11100000) == 0b11000000) {
+			if (next_in_char + 2 > end) { CHECK(false); break; } // bad string
+			COPY_CHAR_TO_JAVA;
+			COPY_CHAR_TO_JAVA;
+		} else if ((*next_in_char & 0b11110000) == 0b11100000) {
+			if (next_in_char + 3 > end) { CHECK(false); break; } // bad string
+			COPY_CHAR_TO_JAVA;
+			COPY_CHAR_TO_JAVA;
+			COPY_CHAR_TO_JAVA;
+		} else if ((*next_in_char & 0b11111000) == 0b11110000) {
+			if (next_in_char + 4 > end) { CHECK(false); break; } // bad string
+			uint32_t codepoint = 0;
+			codepoint |= (((uint32_t)*(next_in_char    )) & 0b00000111) << 18;
+			codepoint |= (((uint32_t)*(next_in_char + 1)) & 0b00111111) << 12;
+			codepoint |= (((uint32_t)*(next_in_char + 2)) & 0b00111111) << 6;
+			codepoint |= (((uint32_t)*(next_in_char + 3)) & 0b00111111) << 0;
+			codepoint -= 0x10000;
+			*next_java_char = 0b11101101;
+			next_java_char++;
+			*next_java_char = 0b10100000 | ((codepoint >> 16) & 0b00001111);
+			next_java_char++;
+			*next_java_char = 0b10000000 | ((codepoint >> 10) & 0b00111111);
+			next_java_char++;
+			*next_java_char = 0b11101101;
+			next_java_char++;
+			*next_java_char = 0b10110000 | ((codepoint >>  6) & 0b00001111);
+			next_java_char++;
+			*next_java_char = 0b10000000 | ((codepoint >>  0) & 0b00111111);
+			next_java_char++;
+			next_in_char += 4;
+		} else {
+			// Bad string
+			CHECK(false);
+			break;
+		}
+	}
+	*next_java_char = 0;
+	jstring ret = (*env)->NewStringUTF(env, java_chars);
+	FREE(java_chars);
 	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;
+	// Java uses "Modified UTF-8" rather than UTF-8. This requires special
+	// handling for codepoints above 0xFFFF, which we implement below.
+	unsigned char* newchars = MALLOC(str_len, "String chars");
+	unsigned char* next_newchar = newchars;
+	uint64_t utf8_len = 0;
+
+	const unsigned char* jchars = (*env)->GetStringUTFChars(env, str, NULL);
+	const unsigned char* next_char = jchars;
+	const unsigned char* end = jchars + str_len;
+
+	#define COPY_CHAR_FROM_JAVA do { *next_newchar = *next_char; next_newchar++; next_char++; utf8_len++; } while (0)
+
+	while (next_char < end) {
+		if (!(*next_char & 0b10000000)) {
+			CHECK(*next_char != 0); // Bad Modified UTF-8 string, but we'll just cut here
+			COPY_CHAR_FROM_JAVA;
+		} else if ((*next_char & 0b11100000) == 0b11000000) {
+			if (next_char + 2 > end) { CHECK(false); break; } // bad string
+			uint16_t codepoint = 0;
+			codepoint |= (((uint16_t)(*next_char & 0x1f)) << 6);
+			codepoint |= *(next_char + 1) & 0x3f;
+			if (codepoint == 0) {
+				// We should really never get null codepoints, but java allows them.
+				// Just skip it.
+				next_char += 2;
+			} else {
+				COPY_CHAR_FROM_JAVA;
+				COPY_CHAR_FROM_JAVA;
+			}
+		} else if ((*next_char & 0b11110000) == 0b11100000) {
+			if (next_char + 3 > end) { CHECK(false); break; } // bad string
+			if (*next_char == 0b11101101 && (*(next_char + 1) & 0b11110000) == 0b10100000) {
+				// Surrogate code unit shoul indicate we have a codepoint above
+				// 0xFFFF, which is where Modified UTF-8 and UTF-8 diverge.
+				if (next_char + 6 > end) { CHECK(false); break; } // bad string
+				CHECK(*(next_char + 3) == 0b11101101);
+				CHECK((*(next_char + 4) & 0b11110000) == 0b10110000);
+				// Calculate the codepoint per https://docs.oracle.com/javase/1.5.0/docs/guide/jni/spec/types.html#wp16542
+				uint32_t codepoint = 0x10000;
+				codepoint += ((((uint32_t)*(next_char + 1)) & 0x0f) << 16);
+				codepoint += ((((uint32_t)*(next_char + 2)) & 0x3f) << 10);
+				codepoint += ((((uint32_t)*(next_char + 4)) & 0x0f) <<  6);
+				codepoint +=  (((uint32_t)*(next_char + 5)) & 0x3f);
+				*next_newchar = 0b11110000 | ((codepoint >> 18) &    0b111);
+				next_newchar++;
+				*next_newchar = 0b10000000 | ((codepoint >> 12) & 0b111111);
+				next_newchar++;
+				*next_newchar = 0b10000000 | ((codepoint >>  6) & 0b111111);
+				next_newchar++;
+				*next_newchar = 0b10000000 | ( codepoint        & 0b111111);
+				next_newchar++;
+				next_char += 6;
+				utf8_len += 4;
+			} else {
+				COPY_CHAR_FROM_JAVA;
+				COPY_CHAR_FROM_JAVA;
+				COPY_CHAR_FROM_JAVA;
+			}
+		} else {
+			// Bad string
+			CHECK(false);
+			break;
+		}
+	}
 	(*env)->ReleaseStringUTFChars(env, str, jchars);
 	LDKStr res = {
 		.chars = newchars,
-		.len = str_len,
+		.len = utf8_len,
 		.chars_is_owned = true
 	};
 	return res;
@@ -485,7 +656,7 @@ import javax.annotation.Nullable;
         self.usize_c_ty = "int64_t"
         self.usize_native_ty = "long"
         self.native_zero_ptr = "0"
-        self.result_c_ty = "jclass"
+        self.unitary_enum_c_ty = "jclass"
         self.ptr_arr = "jobjectArray"
         self.is_arr_some_check = ("", " != NULL")
         self.get_native_arr_len_call = ("(*env)->GetArrayLength(env, ", ")")
@@ -565,7 +736,7 @@ import javax.annotation.Nullable;
         else:
             return "(*env)->Release" + ty_info.java_ty.strip("[]").title() + "ArrayElements(env, " + arr_name + ", " + dest_name + ", 0)"
 
-    def map_hu_array_elems(self, arr_name, conv_name, arr_ty, elem_ty):
+    def map_hu_array_elems(self, arr_name, conv_name, arr_ty, elem_ty, is_nullable):
         if elem_ty.java_ty == "long" and elem_ty.java_hu_ty != "long":
             return arr_name + " != null ? Arrays.stream(" + arr_name + ").mapToLong(" + conv_name + " -> " + elem_ty.from_hu_conv[0] + ").toArray() : null"
         elif elem_ty.java_ty == "long":
@@ -691,7 +862,7 @@ import javax.annotation.Nullable;
             out_c = out_c + "\t\tcase %d: return %s;\n" % (ord_v, var)
             ord_v = ord_v + 1
         out_java_enum = out_java_enum + "\t; static native void init();\n"
-        out_java_enum = out_java_enum + "\tstatic { init(); }\n"
+        out_java_enum = out_java_enum + "\tstatic { org.ldk.impl.bindings.run_statics(); init(); }\n"
         out_java_enum = out_java_enum + "}"
         out_java = out_java + "\tstatic { " + struct_name + ".values(); /* Force enum statics to run */ }\n"
         out_c += "\t}\n"
@@ -1219,7 +1390,7 @@ import javax.annotation.Nullable;
         out_opaque_struct_human += self.hu_struct_file_prefix
         out_opaque_struct_human += "\n/**\n * " + struct_doc_comment.replace("\n", "\n * ") + "\n */\n"
         out_opaque_struct_human += "@SuppressWarnings(\"unchecked\") // We correctly assign various generic arrays\n"
-        hu_name = struct_name.replace("LDKC2Tuple", "TwoTuple").replace("LDKC3Tuple", "ThreeTuple").replace("LDK", "")
+        hu_name = struct_name.replace("LDKC2Tuple", "TwoTuple").replace("LDKC4Tuple", "FourTuple").replace("LDKC3Tuple", "ThreeTuple").replace("LDK", "")
         out_opaque_struct_human += ("public class " + hu_name + " extends CommonBase")
         if struct_name.startswith("LDKLocked") or struct_name.startswith("LDKReadOnly"):
             out_opaque_struct_human += (" implements AutoCloseable")
@@ -1247,6 +1418,9 @@ import javax.annotation.Nullable;
         java_hu_struct += "\tprotected void finalize() throws Throwable {\n"
         java_hu_struct += "\t\tif (ptr != 0) { bindings." + struct_name.replace("LDK","") + "_free(ptr); } super.finalize();\n"
         java_hu_struct += "\t}\n\n"
+        java_hu_struct += "\tprotected void force_free() {\n" # Used by NioPeerHandler
+        java_hu_struct += "\t\tif (ptr != 0) { bindings." + struct_name.replace("LDK","") + "_free(ptr); ptr = 0; }\n"
+        java_hu_struct += "\t}\n\n"
         java_hu_struct += "\tstatic " + human_ty + " constr_from_ptr(long ptr) {\n"
         java_hu_struct += "\t\tif (bindings." + struct_name.replace("LDK", "") + "_is_ok(ptr)) {\n"
         java_hu_struct += "\t\t\treturn new " + human_ty + "_OK(null, ptr);\n"