Better print handling for bad UTF-8 characters

Author: mikera

convex-core/src/main/java/convex/core/data/AString.java

@@ -41,7 +41,7 @@ public boolean print(BlobBuilder sb, long limit) {
 			long avail=limit-sb.count();
 			if (avail>0) {
 				sb.append('"');
-				sb.append(slice(0,avail-1));
+				printEscaped(sb,0,avail-1);
 			}
 			return false;
 		}

convex-core/src/main/java/convex/core/data/StringShort.java

@@ -2,6 +2,7 @@
 
 import java.nio.charset.StandardCharsets;
 
+import convex.core.data.prim.CVMChar;
 import convex.core.data.util.BlobBuilder;
 import convex.core.exceptions.InvalidDataException;
 import convex.core.text.Text;
@@ -223,7 +224,15 @@ protected void printEscaped(BlobBuilder sb, long start, long end) {
 		if ((start<0)||(start>end)||(end>n)) throw new IllegalArgumentException(Errors.badRange(start, end));
 		for (long i=start; i<end; i++) {
 			byte b=data.byteAtUnchecked(i);
-			Text.writeEscapedByte(sb,b);
+			if (b>=0) {
+				// ASCII range, might be escape character
+				Text.writeEscapedByte(sb,b);
+			} else {
+				CVMChar ch=get(i);
+				if (ch==null) ch=CVMChar.BAD_CHARACTER;
+				sb.append(ch);
+				i+=CVMChar.utfLength(ch.getCodePoint())-1;
+			}
 		}
 		return;
 	}

convex-core/src/main/java/convex/core/data/prim/CVMChar.java

@@ -26,6 +26,8 @@ public final class CVMChar extends APrimitive implements Comparable<CVMChar> {
 	public static int MAX_CODEPOINT=0x10ffff; // 21 bits max Unicode value
 	public static CVMChar MAX_VALUE=create(MAX_CODEPOINT); // 21 bits max Unicode value
 
+	public static CVMChar BAD_CHARACTER=create(0xFFFD);
+	
 	private static final int CACHE_SIZE=256;
 
 	private static final CVMChar[] cache=new CVMChar[CACHE_SIZE];

convex-core/src/test/java/convex/core/lang/ReaderTest.java

@@ -275,6 +275,13 @@ public void testResolve() {
 		assertParseException(() -> Reader.read("@(foo)"));
 	}
 
+	@Test public void testUnprintablePrint() {
+		AString bad=Strings.create(Blob.fromHex("ff"));
+		AString pbad=RT.print(bad);
+		AString expected=Strings.create("\"\uFFFD\"");
+		assertEquals(expected,pbad);
+	}
+	
 	@Test
 	public void testTooManyClosingParens() {
 		// See #244