Finish up chapter 22, local variables; fix up indentation-based scoping

chunk.h

@@ -34,6 +34,10 @@ typedef enum {
 	OP_GET_GLOBAL_LONG,
 	OP_SET_GLOBAL,
 	OP_SET_GLOBAL_LONG,
+	OP_SET_LOCAL,
+	OP_SET_LOCAL_LONG,
+	OP_GET_LOCAL,
+	OP_GET_LOCAL_LONG,
 } KrkOpCode;
 
 /**

compiler.c

@@ -1,5 +1,6 @@
 #include <stdio.h>
 #include <stdlib.h>
+#include <string.h>
 
 #include "kuroko.h"
 #include "compiler.h"
@@ -35,16 +36,36 @@ typedef struct {
 	Precedence precedence;
 } ParseRule;
 
+typedef struct {
+	KrkToken name;
+	ssize_t depth;
+} Local;
+
+typedef struct {
+	Local  locals[256];
+	size_t localCount;
+	size_t scopeDepth;
+} Compiler;
+
 Parser parser;
+Compiler * current = NULL;
+
 KrkChunk * compilingChunk;
 static KrkChunk * currentChunk() {
 	return compilingChunk;
 }
 
+static void initCompiler(Compiler * compiler) {
+	compiler->localCount = 0;
+	compiler->scopeDepth = 0;
+	current = compiler;
+}
+
 static void parsePrecedence(Precedence precedence);
-static size_t parseVariable(const char * errorMessage);
+static ssize_t parseVariable(const char * errorMessage);
 static void defineVariable(size_t global);
-static size_t identifierConstant(KrkToken * name);
+static ssize_t identifierConstant(KrkToken * name);
+static ssize_t resolveLocal(Compiler * compiler, KrkToken * name);
 static ParseRule * getRule(KrkTokenType type);
 static void expression();
 static void statement();
@@ -249,9 +270,49 @@ static void expressionStatement() {
 	emitByte(OP_POP);
 }
 
+static void block() {
+	consume(TOKEN_COLON, "Blocks must start with a colon; where's your colon?");
+
+	if (match(TOKEN_EOL)) {
+		/* Begin actual blocks */
+		if (check(TOKEN_INDENTATION)) {
+			/* TODO: Check if this is correctly indented more than the current block */
+			size_t currentIndentation = parser.current.length;
+			do {
+				if (parser.current.length != currentIndentation) break;
+				advance();
+				declaration();
+			} while (check(TOKEN_INDENTATION));
+		} else {
+			errorAtCurrent("Expected indentation for block");
+		}
+	} else {
+		errorAtCurrent("Unsupported single-line block");
+	}
+}
+
+static void beginScope() {
+	current->scopeDepth++;
+}
+
+static void endScope() {
+	current->scopeDepth--;
+	while (current->localCount > 0 &&
+	       current->locals[current->localCount - 1].depth > current->scopeDepth) {
+		emitByte(OP_POP);
+		current->localCount--;
+	}
+}
+
 static void statement() {
+	if (match(TOKEN_EOL)) return; /* Meaningless blank line */
+
 	if (match(TOKEN_PRINT)) {
 		printStatement();
+	} else if (match(TOKEN_BLOCK)) {
+		beginScope();
+		block();
+		endScope();
 	} else {
 		expressionStatement();
 	}
@@ -288,13 +349,22 @@ static void string(int canAssign) {
 	else { emitBytes(opc ## _LONG, arg >> 16); emitBytes(arg >> 8, arg); } } while (0)
 
 static void namedVariable(KrkToken name, int canAssign) {
-	size_t arg = identifierConstant(&name);
-
-	if (canAssign && match(TOKEN_EQUAL)) {
-		expression();
-		EMIT_CONSTANT_OP(OP_SET_GLOBAL, arg);
+	ssize_t arg = resolveLocal(current, &name);
+	if (arg != -1) {
+		if (canAssign && match(TOKEN_EQUAL)) {
+			expression();
+			EMIT_CONSTANT_OP(OP_SET_LOCAL, arg);
+		} else {
+			EMIT_CONSTANT_OP(OP_GET_LOCAL, arg);
+		}
 	} else {
-		EMIT_CONSTANT_OP(OP_GET_GLOBAL, arg);
+		arg = identifierConstant(&name);
+		if (canAssign && match(TOKEN_EQUAL)) {
+			expression();
+			EMIT_CONSTANT_OP(OP_SET_GLOBAL, arg);
+		} else {
+			EMIT_CONSTANT_OP(OP_GET_GLOBAL, arg);
+		}
 	}
 }
 
@@ -374,16 +444,68 @@ static void parsePrecedence(Precedence precedence) {
 	}
 }
 
-static size_t identifierConstant(KrkToken * name) {
+static ssize_t identifierConstant(KrkToken * name) {
 	return krk_addConstant(currentChunk(), OBJECT_VAL(copyString(name->start, name->length)));
 }
 
-static size_t parseVariable(const char * errorMessage) {
+static int identifiersEqual(KrkToken * a, KrkToken * b) {
+	return (a->length == b->length && memcmp(a->start, b->start, a->length) == 0);
+}
+
+static ssize_t resolveLocal(Compiler * compiler, KrkToken * name) {
+	for (ssize_t i = compiler->localCount - 1; i >= 0; i--) {
+		Local * local = &compiler->locals[i];
+		if (identifiersEqual(name, &local->name)) {
+			if (local->depth == -1) {
+				error("can not initialize value recursively (are you shadowing something?)");
+			}
+			return i;
+		}
+	}
+	return -1;
+}
+
+static void addLocal(KrkToken name) {
+	if (current->localCount == 256) {
+		error("too many locals");
+		return;
+	}
+	Local * local = &current->locals[current->localCount++];
+	local->name = name;
+	local->depth = -1;
+}
+
+static void declareVariable() {
+	if (current->scopeDepth == 0) return;
+	KrkToken * name = &parser.previous;
+	/* Detect duplicate definition */
+	for (ssize_t i = current->localCount - 1; i >= 0; i--) {
+		Local * local = &current->locals[i];
+		if (local->depth != -1 && local->depth < current->scopeDepth) break;
+		if (identifiersEqual(name, &local->name)) error("Duplicate definition");
+	}
+	addLocal(*name);
+}
+
+static ssize_t parseVariable(const char * errorMessage) {
 	consume(TOKEN_IDENTIFIER, errorMessage);
+
+	declareVariable();
+	if (current->scopeDepth > 0) return 0;
+
 	return identifierConstant(&parser.previous);
 }
 
+static void markInitialized() {
+	current->locals[current->localCount - 1].depth = current->scopeDepth;
+}
+
 static void defineVariable(size_t global) {
+	if (current->scopeDepth > 0) {
+		markInitialized();
+		return;
+	}
+
 	EMIT_CONSTANT_OP(OP_DEFINE_GLOBAL, global);
 }
 
@@ -393,6 +515,8 @@ static ParseRule * getRule(KrkTokenType type) {
 
 int krk_compile(const char * src, KrkChunk * chunk) {
 	krk_initScanner(src);
+	Compiler compiler;
+	initCompiler(&compiler);
 	compilingChunk = chunk;
 
 	parser.hadError = 0;

debug.c

@@ -22,6 +22,13 @@ void krk_disassembleChunk(KrkChunk * chunk, const char * name) {
 	krk_printValue(stderr, chunk->constants.values[constant]); \
 	fprintf(stderr,"' (type=%s)\n", typeName(chunk->constants.values[constant])); \
 	return offset + 4; }
+#define OPERAND(opc) case opc: { uint32_t operand = chunk->code[offset + 1]; \
+	fprintf(stderr, "%-16s %4d\n", #opc, (int)operand); \
+	return offset + 2; } \
+	case opc ## _LONG: { uint32_t operand = (chunk->code[offset + 1] << 16) | \
+	(chunk->code[offset + 2] << 8) | (chunk->code[offset + 3]); \
+	fprintf(stderr, "%-16s %4d\n", #opc "_LONG", (int)operand); \
+	return offset + 4; }
 
 size_t krk_disassembleInstruction(KrkChunk * chunk, size_t offset) {
 	fprintf(stderr, "%04u ", (unsigned int)offset);
@@ -52,6 +59,8 @@ size_t krk_disassembleInstruction(KrkChunk * chunk, size_t offset) {
 		CONSTANT(OP_CONSTANT)
 		CONSTANT(OP_GET_GLOBAL)
 		CONSTANT(OP_SET_GLOBAL)
+		OPERAND(OP_SET_LOCAL)
+		OPERAND(OP_GET_LOCAL)
 		default:
 			fprintf(stderr, "Unknown opcode: %02x\n", opcode);
 			return offset + 1;

kuroko.c

@@ -8,8 +8,21 @@
 int main(int argc, char * argv[]) {
 	krk_initVM();
 
-	krk_interpret("let breakfast = \"beignets\"\nlet beverage = \"café au lait\"\nbreakfast = \"beignets with \" + beverage\nprint breakfast");
+	FILE * f = fopen("test.krk","r");
+	if (!f) return 1;
+	fseek(f, 0, SEEK_END);
+	size_t size = ftell(f);
+	fseek(f, 0, SEEK_SET);
+
+	char * buf = malloc(size+1);
+	fread(buf, 1, size, f);
+	fclose(f);
+	buf[size] = '\0';
+
+	krk_interpret(buf);
 
 	krk_freeVM();
+
+	free(buf);
 	return 0;
 }

scanner.c

@@ -81,7 +81,13 @@ static void skipWhitespace() {
 static KrkToken makeIndentation() {
 	while (!isAtEnd() && peek() == ' ') advance();
 	if (peek() == '\n') {
-		return errorToken("Empty indentation line is invalid.");
+		/* Pretend we didn't see this line */
+		return makeToken(TOKEN_INDENTATION);
+	}
+	if (peek() == '#') {
+		KrkToken out = makeToken(TOKEN_INDENTATION);
+		while (!isAtEnd() && peek() != '\n') advance();
+		return out;
 	}
 	return makeToken(TOKEN_INDENTATION);
 }
@@ -175,6 +181,7 @@ static KrkTokenType identifierType() {
 	switch (*scanner.start) {
 		case 'a': return checkKeyword(1, "nd", TOKEN_AND);
 		case 'c': return checkKeyword(1, "lass", TOKEN_CLASS);
+		case 'b': return checkKeyword(1, "lock", TOKEN_BLOCK);
 		case 'd': return checkKeyword(1, "ef", TOKEN_DEF);
 		case 'e': return checkKeyword(1, "lse", TOKEN_ELSE);
 		case 'f': return checkKeyword(1, "or", TOKEN_FOR);
@@ -210,9 +217,8 @@ KrkToken krk_scanToken() {
 	/* If at start of line, do thing */
 	if (scanner.startOfLine && peek() == ' ') {
 		scanner.start = scanner.cur;
-		return makeIndentation();
-	} else {
 		scanner.startOfLine = 0;
+		return makeIndentation();
 	}
 
 	/* Eat whitespace */
@@ -228,10 +234,18 @@ KrkToken krk_scanToken() {
 
 	if (c == '\n') {
 		scanner.line++;
-		scanner.startOfLine = 1;
-		return makeToken(TOKEN_EOL);
+		if (scanner.startOfLine) {
+			/* Ignore completely blank lines */
+			return makeToken(TOKEN_RETRY);
+		} else {
+			scanner.startOfLine = 1;
+			return makeToken(TOKEN_EOL);
+		}
 	}
 
+	/* Not indentation, not a linefeed on an empty line, must be not be start of line any more */
+	scanner.startOfLine = 0;
+
 	if (isAlpha(c)) return identifier();
 	if (isDigit(c)) return number(c);
 

scanner.h

@@ -36,6 +36,7 @@ typedef enum {
 
 	TOKEN_INDENTATION,
 
+	TOKEN_BLOCK, /* temporary XXX remove me */
 	TOKEN_RETRY,
 	TOKEN_EOL,
 

test.krk

@@ -0,0 +1,16 @@
+print "Hello, world"
+let a = "Hello,"
+let b = "world"
+print a + " " + b
+block:
+    let a = "This is a new variable"
+    print "I'm in a block"
+    # This should work
+    print a
+    
+    print b
+
+    print "I am still in the block"
+
+print a
+print b

value.c

@@ -10,7 +10,7 @@ void krk_initValueArray(KrkValueArray * array) {
 }
 
 void krk_writeValueArray(KrkValueArray * array, KrkValue value) {
-	if (array->capacity < array->count - 1) {
+	if (array->capacity < array->count + 1) {
 		int old = array->capacity;
 		array->capacity = GROW_CAPACITY(old);
 		array->values = GROW_ARRAY(KrkValue, array->values, old, array->capacity);

vm.c

@@ -267,6 +267,18 @@ static KrkValue run() {
 				}
 				break;
 			}
+			case OP_GET_LOCAL_LONG:
+			case OP_GET_LOCAL: {
+				uint32_t slot = readBytes((opcode == OP_GET_LOCAL ? 1 : 3));
+				krk_push(vm.stack[slot]);
+				break;
+			}
+			case OP_SET_LOCAL_LONG:
+			case OP_SET_LOCAL: {
+				uint32_t slot = readBytes((opcode == OP_SET_LOCAL ? 1 : 3));
+				vm.stack[slot] = krk_peep(0);
+				break;
+			}
 		}
 	}