Plumbing for extensible encode reading

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

@@ -19,7 +19,11 @@ public abstract class AEncoder<T> {
 	 */
 	public abstract Blob encode(T a);
 
-	public abstract T decode(Blob encoding) throws BadFormatException;
+	public T decode(Blob encoding) throws BadFormatException {
+		return read(encoding,0);
+	}
+
+	public abstract T read(Blob encoding, int offset) throws BadFormatException;
 
 	/**
 	 * Reads a value from a Blob of data

convex-core/src/main/java/convex/core/data/CAD3Encoder.java

@@ -1,20 +1,159 @@
 package convex.core.data;
 
+import convex.core.data.prim.AByteFlag;
+import convex.core.data.prim.ANumeric;
+import convex.core.data.prim.CVMBigInteger;
+import convex.core.data.prim.CVMChar;
+import convex.core.data.prim.CVMDouble;
+import convex.core.data.prim.CVMLong;
 import convex.core.exceptions.BadFormatException;
+import convex.core.util.ErrorMessages;
 
 /**
- * Base Encoder for CAD3 data / stores
+ * Base Encoder for CAD3 data / stores.
+ * 
+ * Does NOT directly decode custom CVM value types. Use the derived CVMEncoder if you need that behaviour.
  */
 public class CAD3Encoder extends AEncoder<ACell> {
 
 	public Blob encode(ACell a) {
 		return Cells.encode(a);
 	}
 
+	@Override
 	public ACell decode(Blob encoding) throws BadFormatException {
-		return Format.read(encoding);
+		if (encoding.count()<1) throw new BadFormatException("Empty encoding");
+		return read(encoding,0);
+	}
+
+	@Override
+	public ACell read(Blob encoding, int offset) throws BadFormatException {
+		byte tag = encoding.byteAt(offset);
+		ACell result= read(tag,encoding,offset);
+		return result;
+	}
+
+	protected ACell read(byte tag, Blob encoding, int offset) throws BadFormatException {	
+		switch (tag>>4) {
+		case 0: // 0x00-0x0F : Only null is valid
+			if (tag==Tag.NULL) return null;
+			break;
+
+		case 1: // 0x10-0x1F : Numeric values 
+			return readNumeric(tag,encoding,offset);
+
+		case 2: // 0x20-0x2F : Addresses and references 
+			if (tag == Tag.ADDRESS) return Address.read(encoding,offset);
+			// Note: 0x20 reference is invalid as a top level encoding
+			break;
+
+		case 3: // 0x30-0x3F : BAsic string / blob-like objects
+			return readBasicObject(tag, encoding, offset);
+
+		case 4: case 5: case 6: case 7: // 0x40-0x7F currently reserved
+			break;
+
+		case 8: // 0x80-0x8F : BAsic string / blob-like objects
+			return readDataStructure(tag, encoding, offset);
+
+		case 9: // 0x90-0x9F : Crypto / signature objects
+			return readSignedData(tag,encoding, offset);
+
+		case 10: // 0xA0-0xAF : Sparse records. A is for Airy.
+			return readSparseRecord(tag,encoding,offset);
+
+		case 11: // 0xB0-0xBF : Byte flags including booleans
+			return AByteFlag.read(tag);
+
+		case 12: // 0xC0-0xCF : Coded data objects
+			return readCodedData(tag,encoding, offset);
+
+		case 13: // 0xD0-0xDF : Dense records
+			return readDenseRecord(tag,encoding, offset);
+
+		case 14: // 0xE0-0xEF : Extension values
+			return readExtension(tag,encoding, offset);
+
+		case 15: // 0xF0-0xFF : Reserved / failure  Values
+			break;
+		}
+
+		return Format.read(tag,encoding,offset);
+		// throw new BadFormatException(ErrorMessages.badTagMessage(tag));
+	}
+
+
+	protected ANumeric readNumeric(byte tag, Blob blob, int offset) throws BadFormatException {
+		if (tag<0x19) return CVMLong.read(tag,blob,offset);
+		if (tag == 0x19) return CVMBigInteger.read(blob,offset);
+		if (tag == Tag.DOUBLE) return CVMDouble.read(tag,blob,offset);
+
+		throw new BadFormatException(ErrorMessages.badTagMessage(tag));
+	}
+
+	protected ACell readBasicObject(byte tag, Blob blob, int offset)  throws BadFormatException{
+		switch (tag) {
+			case Tag.SYMBOL: return Symbol.read(blob,offset);
+			case Tag.KEYWORD: return Keyword.read(blob,offset);
+			case Tag.BLOB: return Blobs.read(blob,offset);
+			case Tag.STRING: return Strings.read(blob,offset);
+		} 
+
+		if ((tag&Tag.CHAR_MASK)==Tag.CHAR_BASE) {
+			int len=CVMChar.byteCountFromTag(tag);
+			if (len>4) throw new BadFormatException("Can't read char type with length: " + len);
+			return CVMChar.read(len, blob,offset); // skip tag byte
+		}
+
+		throw new BadFormatException(ErrorMessages.badTagMessage(tag));
+	}
+
+	protected ACell readDataStructure(byte tag, Blob b, int pos) throws BadFormatException {
+		if (tag == Tag.VECTOR) return Vectors.read(b,pos);
+
+		if (tag == Tag.MAP) return Maps.read(b,pos);
+
+		if (tag == Tag.SYNTAX) return Syntax.read(b,pos);
+
+		if (tag == Tag.SET) return Sets.read(b,pos);
+
+		if (tag == Tag.LIST) return List.read(b,pos);
+
+		if (tag == Tag.INDEX) return Index.read(b,pos);
+
+		throw new BadFormatException("Can't read data structure with tag byte: " + tag);
+	}
+
+	protected SignedData<?> readSignedData(byte tag,Blob blob, int offset) throws BadFormatException {
+		if (tag==Tag.SIGNED_DATA) return SignedData.read(blob,offset,true);	
+		if (tag==Tag.SIGNED_DATA_SHORT) return SignedData.read(blob,offset,false);	
+		throw new BadFormatException(ErrorMessages.badTagMessage(tag));
+	}
+
+	protected ARecord readSparseRecord(byte tag, Blob encoding, int offset) throws BadFormatException {
+		// TODO spare records
+		throw new BadFormatException(ErrorMessages.TODO);
 	}
 
+	protected ACell readCodedData(byte tag, Blob encoding, int offset) throws BadFormatException {
+		// TODO Change delegation to proper read
+		return Format.read(tag,encoding,offset);
+	}
+
+	protected ACell readDenseRecord(byte tag, Blob encoding, int offset) throws BadFormatException {
+		// TODO Change delegation to proper read
+		return Format.read(tag,encoding,offset);
+	}
+
+	protected ACell readExtension(byte tag, Blob blob, int offset) throws BadFormatException {
+		// We expect a VLQ Count following the tag
+		long code=Format.readVLQCount(blob,offset+1);
+		return ExtensionValue.create(tag, code);
+	}
+
+
+
+
 	/**
 	 * Reads a cell value from a Blob of data, allowing for non-embedded branches following the first cell
 	 * @param data Data to decode

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

@@ -1,9 +1,25 @@
 package convex.core.data;
 
+import convex.core.exceptions.BadFormatException;
+import convex.core.lang.Core;
+
+/**
+ * Encoder for CVM values and data structures
+ */
 public class CVMEncoder extends CAD3Encoder {
 
 	public static final CVMEncoder INSTANCE = new CVMEncoder();
 
+	@Override
+	public ACell read(Blob encoding,int offset) throws BadFormatException {
+		return super.read(encoding,offset);
+	}
 
-
+	protected ACell readExtension(byte tag, Blob blob, int offset) throws BadFormatException {
+		// We expect a VLQ Count following the tag
+		long code=Format.readVLQCount(blob,offset+1);
+		if (tag == Tag.CORE_DEF) return Core.fromCode(code);
+
+		return ExtensionValue.create(tag, code);
+	}
 }

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

@@ -39,6 +39,7 @@
 import convex.core.store.AStore;
 import convex.core.store.Stores;
 import convex.core.util.Bits;
+import convex.core.util.ErrorMessages;
 import convex.core.util.Trees;
 import convex.core.util.Utils;
 
@@ -580,7 +581,7 @@ public static <T extends ACell> T read(String hexString) throws BadFormatExcepti
 	 * @throws BadFormatException If encoding is invalid for the given tag
 	 */
 	@SuppressWarnings("unchecked")
-	private static <T extends ACell> T read(byte tag, Blob blob, int offset) throws BadFormatException {
+	static <T extends ACell> T read(byte tag, Blob blob, int offset) throws BadFormatException {
 
 		// Fast paths for common one-byte instances. TODO: might switch have better performance if compiled correctly into a table?
 		if (tag==Tag.NULL) return null;
@@ -616,27 +617,22 @@ private static <T extends ACell> T read(byte tag, Blob blob, int offset) throws
 		} catch (Exception e) {
 			throw new BadFormatException("Unexpected Exception when decoding ("+tag+"): "+e.getMessage(), e);
 		}
-		throw new BadFormatException(badTagMessage(tag));
+		throw new BadFormatException(ErrorMessages.badTagMessage(tag));
 	}
 
 
 	private static <T extends ACell> SignedData<T> readSignedData(byte tag,Blob blob, int offset) throws BadFormatException {
 		if (tag==Tag.SIGNED_DATA) return SignedData.read(blob,offset,true);	
 		if (tag==Tag.SIGNED_DATA_SHORT) return SignedData.read(blob,offset,false);	
-		throw new BadFormatException(badTagMessage(tag));
-	}
-
-	private static String badTagMessage(byte tag) {
-		return "Unrecognised tag byte 0x"+Utils.toHexString(tag);
+		throw new BadFormatException(ErrorMessages.badTagMessage(tag));
 	}
 
 	private static ANumeric readNumeric(byte tag, Blob blob, int offset) throws BadFormatException {
-		// TODO Auto-generated method stub
 		if (tag<0x19) return CVMLong.read(tag,blob,offset);
 		if (tag == 0x19) return CVMBigInteger.read(blob,offset);
 		if (tag == Tag.DOUBLE) return CVMDouble.read(tag,blob,offset);
 
-		throw new BadFormatException(badTagMessage(tag));
+		throw new BadFormatException(ErrorMessages.badTagMessage(tag));
 	}
 
 	private static ACell readBasicObject(byte tag, Blob blob, int offset)  throws BadFormatException{
@@ -653,7 +649,7 @@ private static ACell readBasicObject(byte tag, Blob blob, int offset)  throws Ba
 			return CVMChar.read(len, blob,offset); // skip tag byte
 		}
 
-		throw new BadFormatException(badTagMessage(tag));
+		throw new BadFormatException(ErrorMessages.badTagMessage(tag));
 	}
 
 
@@ -691,7 +687,7 @@ private static <T extends ARecord> T readRecord(byte tag, Blob b, int pos) throw
 		if (tag == Tag.PEER_STATUS) return (T) PeerStatus.read(b,pos);
 		if (tag == Tag.ACCOUNT_STATUS) return (T) AccountStatus.read(b,pos); 
 
-		throw new BadFormatException(badTagMessage(tag));
+		throw new BadFormatException(ErrorMessages.badTagMessage(tag));
 	}
 
 	@SuppressWarnings("unchecked")
@@ -712,7 +708,7 @@ private static <T extends ACell> T readTransaction(byte tag, Blob b, int pos) th
 
 		// Might be a generic Dense Record
 		DenseRecord dr=DenseRecord.read(tag,b,pos);
-		if (dr==null) throw new BadFormatException(badTagMessage(tag));
+		if (dr==null) throw new BadFormatException(ErrorMessages.badTagMessage(tag));
 		return (T) dr;
 	}
 

convex-core/src/main/java/convex/core/util/ErrorMessages.java

@@ -19,6 +19,8 @@ public class ErrorMessages {
 
 
 
+	public static final String TODO = "Not yet implemented.";
+
 	public static String immutable(Object a) {
 		return "Object is immutable: "+a.getClass();
 	}
@@ -61,4 +63,8 @@ public static ErrorValue nobody(Address address) {
 
 	public static ErrorValue INVALID_NUMERIC = ErrorValue.create(ErrorCodes.ARGUMENT,"Invalid numeric result");
 
+	public static String badTagMessage(byte tag) {
+		return "Unrecognised tag byte 0x"+Utils.toHexString(tag);
+	}
+
 }