Edits for better handling of slightly tricky unprintable UTF-8

Author: mikera

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

@@ -228,10 +228,18 @@ protected void printEscaped(BlobBuilder sb, long start, long end) {
 				// 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;
+				int cp=charAt(i);
+				if (cp<0) {
+					sb.append(CVMChar.BAD_CHARACTER);
+					i+=1; // skip one byte? or should we error correct?
+				} else {
+					// need to copy exactly one UTF character
+					int len=CVMChar.utfLength(cp);
+					for (int j=0; j<len; j++) {
+						sb.append(byteAt(i+j));
+					}
+					i+=len-1;
+				}
 			}
 		}
 		return;

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

@@ -2,6 +2,7 @@
 
 import static convex.test.Assertions.*;
 import static org.junit.jupiter.api.Assertions.assertEquals;
+import static org.junit.jupiter.api.Assertions.assertNotEquals;
 import static org.junit.jupiter.api.Assertions.assertNull;
 import static org.junit.jupiter.api.Assertions.assertSame;
 import static org.junit.jupiter.api.Assertions.assertThrows;
@@ -276,10 +277,15 @@ public void testResolve() {
 	}
 
 	@Test public void testUnprintablePrint() {
-		AString bad=Strings.create(Blob.fromHex("ff"));
+		AString bad=Strings.create(Blob.fromHex("ff")); // bad UTF-8
 		AString pbad=RT.print(bad);
 		AString expected=Strings.create("\"\uFFFD\"");
 		assertEquals(expected,pbad);
+		
+		String ps=pbad.toString();
+		assertNotEquals(bad,Reader.read(ps)); // not reproducing the bad UTF-8
+		assertEquals(pbad,Reader.read(ps).print()); // printed version should round trip
+		doReadPrintTest(ps);
 	}
 
 	@Test