Add Kym parsing support for array literals
This commit is contained in:
parent
5e52fcc795
commit
27b809f2c5
|
|
@ -25,6 +25,7 @@ const Opcode = enum (u8) {
|
|||
push_integer,
|
||||
push_float,
|
||||
push_object,
|
||||
push_array,
|
||||
push_table,
|
||||
|
||||
not,
|
||||
|
|
@ -126,19 +127,32 @@ pub fn compile_expression(self: *Self, expression: ast.Expression) types.Runtime
|
|||
try self.emit_object(try self.intern(literal));
|
||||
},
|
||||
|
||||
.table_literal => |literal| {
|
||||
if (literal.values.len > coral.math.max_int(@typeInfo(types.Integer).Int)) {
|
||||
.array_literal => |elements| {
|
||||
if (elements.values.len > coral.math.max_int(@typeInfo(types.Integer).Int)) {
|
||||
return error.OutOfMemory;
|
||||
}
|
||||
|
||||
for (literal.values) |field| {
|
||||
try self.compile_expression(field.expression.*);
|
||||
for (elements.values) |element_expression| {
|
||||
try self.compile_expression(element_expression);
|
||||
}
|
||||
|
||||
try self.emit_opcode(.push_array);
|
||||
try self.emit_integer(@intCast(types.Integer, elements.values.len));
|
||||
},
|
||||
|
||||
.table_literal => |fields| {
|
||||
if (fields.values.len > coral.math.max_int(@typeInfo(types.Integer).Int)) {
|
||||
return error.OutOfMemory;
|
||||
}
|
||||
|
||||
for (fields.values) |field| {
|
||||
try self.compile_expression(field.expression);
|
||||
try self.emit_opcode(.push_object);
|
||||
try self.emit_object(try self.intern(field.identifier));
|
||||
}
|
||||
|
||||
try self.emit_opcode(.push_table);
|
||||
try self.emit_integer(@intCast(types.Integer, literal.values.len));
|
||||
try self.emit_integer(@intCast(types.Integer, fields.values.len));
|
||||
},
|
||||
|
||||
.binary_operation => |operation| {
|
||||
|
|
|
|||
|
|
@ -4,6 +4,8 @@ const tokens = @import("./tokens.zig");
|
|||
|
||||
const types = @import("./types.zig");
|
||||
|
||||
const ArrayElements = coral.list.Stack(Expression);
|
||||
|
||||
pub const BinaryOperation = enum {
|
||||
addition,
|
||||
subtraction,
|
||||
|
|
@ -33,7 +35,7 @@ pub const ParsedExpression = union (enum) {
|
|||
};
|
||||
|
||||
if (tokenizer.current_token == .symbol_plus) {
|
||||
if (!tokenizer.step()) {
|
||||
if (!tokenizer.step(.ignore_newlines)) {
|
||||
return ParsedExpression{.invalid = "expected right-hand side of expression after `+`"};
|
||||
}
|
||||
|
||||
|
|
@ -55,7 +57,7 @@ pub const ParsedExpression = union (enum) {
|
|||
}
|
||||
|
||||
if (tokenizer.current_token == .symbol_minus) {
|
||||
if (!tokenizer.step()) {
|
||||
if (!tokenizer.step(.ignore_newlines)) {
|
||||
return ParsedExpression{.invalid = "expected right-hand side of expression after `-`"};
|
||||
}
|
||||
|
||||
|
|
@ -101,7 +103,7 @@ pub const ParsedExpression = union (enum) {
|
|||
};
|
||||
|
||||
if (tokenizer.current_token == .symbol_greater_than) {
|
||||
if (!tokenizer.step()) {
|
||||
if (!tokenizer.step(.ignore_newlines)) {
|
||||
return ParsedExpression{.invalid = "expected right-hand side of expression after `>`"};
|
||||
}
|
||||
|
||||
|
|
@ -123,7 +125,7 @@ pub const ParsedExpression = union (enum) {
|
|||
}
|
||||
|
||||
if (tokenizer.current_token == .symbol_greater_equals) {
|
||||
if (!tokenizer.step()) {
|
||||
if (!tokenizer.step(.ignore_newlines)) {
|
||||
return ParsedExpression{.invalid = "expected right-hand side of expression after `>=`"};
|
||||
}
|
||||
|
||||
|
|
@ -145,7 +147,7 @@ pub const ParsedExpression = union (enum) {
|
|||
}
|
||||
|
||||
if (tokenizer.current_token == .symbol_less_than) {
|
||||
if (!tokenizer.step()) {
|
||||
if (!tokenizer.step(.ignore_newlines)) {
|
||||
return ParsedExpression{.invalid = "expected right-hand side of expression after `<`"};
|
||||
}
|
||||
|
||||
|
|
@ -167,7 +169,7 @@ pub const ParsedExpression = union (enum) {
|
|||
}
|
||||
|
||||
if (tokenizer.current_token == .symbol_less_equals) {
|
||||
if (!tokenizer.step()) {
|
||||
if (!tokenizer.step(.ignore_newlines)) {
|
||||
return ParsedExpression{.invalid = "expected right-hand side of expression after `<=`"};
|
||||
}
|
||||
|
||||
|
|
@ -213,7 +215,7 @@ pub const ParsedExpression = union (enum) {
|
|||
};
|
||||
|
||||
if (tokenizer.current_token == .symbol_double_equals) {
|
||||
if (!tokenizer.step()) {
|
||||
if (!tokenizer.step(.ignore_newlines)) {
|
||||
return ParsedExpression{.invalid = "expected right-hand side of expression after `==`"};
|
||||
}
|
||||
|
||||
|
|
@ -249,7 +251,7 @@ pub const ParsedExpression = union (enum) {
|
|||
|
||||
switch (tokenizer.current_token) {
|
||||
.symbol_paren_left => {
|
||||
if (!tokenizer.step()) {
|
||||
if (!tokenizer.step(.ignore_newlines)) {
|
||||
return ParsedExpression{.invalid = "expected an expression after `(`"};
|
||||
}
|
||||
|
||||
|
|
@ -263,7 +265,7 @@ pub const ParsedExpression = union (enum) {
|
|||
expression.deinit(allocator);
|
||||
};
|
||||
|
||||
if ((!tokenizer.step()) or (tokenizer.current_token != .symbol_paren_right)) {
|
||||
if ((!tokenizer.step(.ignore_newlines)) or (tokenizer.current_token != .symbol_paren_right)) {
|
||||
return ParsedExpression{.invalid = "expected a closing `)` after expression"};
|
||||
}
|
||||
|
||||
|
|
@ -278,10 +280,7 @@ pub const ParsedExpression = union (enum) {
|
|||
}
|
||||
},
|
||||
|
||||
.integer => |value| {
|
||||
defer _ = tokenizer.step();
|
||||
|
||||
return ParsedExpression{
|
||||
.integer => |value| return ParsedExpression{
|
||||
.valid = .{
|
||||
.integer_literal = coral.utf8.parse_int(
|
||||
@typeInfo(types.Integer).Int,
|
||||
|
|
@ -295,13 +294,9 @@ pub const ParsedExpression = union (enum) {
|
|||
};
|
||||
},
|
||||
},
|
||||
};
|
||||
},
|
||||
|
||||
.real => |value| {
|
||||
defer _ = tokenizer.step();
|
||||
|
||||
return ParsedExpression{
|
||||
.real => |value| return ParsedExpression{
|
||||
.valid = .{
|
||||
.float_literal = coral.utf8.parse_float(
|
||||
@typeInfo(types.Float).Float,
|
||||
|
|
@ -314,19 +309,58 @@ pub const ParsedExpression = union (enum) {
|
|||
};
|
||||
},
|
||||
},
|
||||
},
|
||||
|
||||
.string => |value| return ParsedExpression{
|
||||
.valid = .{.string_literal = value},
|
||||
},
|
||||
|
||||
.symbol_bracket_left => {
|
||||
if (tokenizer.step(.ignore_newlines)) {
|
||||
return ParsedExpression{.invalid = "unexpected end of array literal"};
|
||||
}
|
||||
|
||||
var is_invalid = true;
|
||||
var array_elements = try ArrayElements.init(allocator, 0);
|
||||
|
||||
defer if (is_invalid) {
|
||||
array_elements.deinit(allocator);
|
||||
};
|
||||
|
||||
while (true) {
|
||||
switch (tokenizer.current_token) {
|
||||
.symbol_bracket_right => {
|
||||
_ = tokenizer.step(.ignore_newlines);
|
||||
is_invalid = false;
|
||||
|
||||
return ParsedExpression{
|
||||
.valid = .{.array_literal = array_elements},
|
||||
};
|
||||
},
|
||||
|
||||
.string => |value| {
|
||||
defer _ = tokenizer.step();
|
||||
else => {
|
||||
if (!tokenizer.step(.ignore_newlines)) {
|
||||
return ParsedExpression{.invalid = "expected `]` or expression after `[`"};
|
||||
}
|
||||
|
||||
return ParsedExpression{
|
||||
.valid = .{.string_literal = value},
|
||||
};
|
||||
var parsed_expression = try ParsedExpression.init(allocator, tokenizer);
|
||||
|
||||
switch (parsed_expression) {
|
||||
.valid => |*expression| {
|
||||
errdefer expression.deinit(allocator);
|
||||
|
||||
try array_elements.push_one(allocator, expression.*);
|
||||
},
|
||||
|
||||
.invalid => |detail| return ParsedExpression{.invalid = detail},
|
||||
}
|
||||
},
|
||||
}
|
||||
}
|
||||
},
|
||||
|
||||
.symbol_brace_left => {
|
||||
if (tokenizer.step()) {
|
||||
if (!tokenizer.step(.ignore_newlines)) {
|
||||
return ParsedExpression{.invalid = "unexpected end of table literal"};
|
||||
}
|
||||
|
||||
|
|
@ -340,7 +374,7 @@ pub const ParsedExpression = union (enum) {
|
|||
while (true) {
|
||||
switch (tokenizer.current_token) {
|
||||
.symbol_brace_right => {
|
||||
_ = tokenizer.step();
|
||||
_ = tokenizer.step(.ignore_newlines);
|
||||
is_invalid = false;
|
||||
|
||||
return ParsedExpression{
|
||||
|
|
@ -351,11 +385,11 @@ pub const ParsedExpression = union (enum) {
|
|||
.local => |identifier| {
|
||||
const key = identifier;
|
||||
|
||||
if (!tokenizer.step() or tokenizer.current_token != .symbol_equals) {
|
||||
if (!tokenizer.step(.ignore_newlines) or tokenizer.current_token != .symbol_equals) {
|
||||
return ParsedExpression{.invalid = "expected `=` after identifier"};
|
||||
}
|
||||
|
||||
if (!tokenizer.step()) {
|
||||
if (!tokenizer.step(.ignore_newlines)) {
|
||||
return ParsedExpression{.invalid = "unexpected end after `=`"};
|
||||
}
|
||||
|
||||
|
|
@ -367,48 +401,40 @@ pub const ParsedExpression = union (enum) {
|
|||
|
||||
try table_fields.push_one(allocator, .{
|
||||
.identifier = key,
|
||||
.expression = expression,
|
||||
.expression = expression.*,
|
||||
});
|
||||
},
|
||||
|
||||
.invalid => |details| return ParsedExpression{.invalid = details},
|
||||
}
|
||||
},
|
||||
|
||||
.string => |identifier| {
|
||||
const key = identifier;
|
||||
|
||||
if (!tokenizer.step() or tokenizer.current_token != .symbol_equals) {
|
||||
return ParsedExpression{.invalid = "expected `=` after identifier"};
|
||||
if (!tokenizer.step(.ignore_newlines)) {
|
||||
return ParsedExpression{.invalid = "unexpected end after expression"};
|
||||
}
|
||||
|
||||
if (!tokenizer.step()) {
|
||||
return ParsedExpression{.invalid = "unexpected end after `=`"};
|
||||
}
|
||||
switch (tokenizer.current_token) {
|
||||
.symbol_comma => _ = tokenizer.step(.ignore_newlines),
|
||||
|
||||
var parsed_expression = try init(allocator, tokenizer);
|
||||
.symbol_brace_right => {
|
||||
_ = tokenizer.step(.ignore_newlines);
|
||||
is_invalid = false;
|
||||
|
||||
switch (parsed_expression) {
|
||||
.valid => |*expression| {
|
||||
errdefer expression.deinit(allocator);
|
||||
|
||||
try table_fields.push_one(allocator, .{
|
||||
.identifier = key,
|
||||
.expression = expression,
|
||||
});
|
||||
return ParsedExpression{
|
||||
.valid = .{.table_literal = table_fields},
|
||||
};
|
||||
},
|
||||
|
||||
.invalid => |details| return ParsedExpression{.invalid = details},
|
||||
else => return ParsedExpression{.invalid = "expected `,` or `}` after expression"},
|
||||
}
|
||||
},
|
||||
|
||||
else => return ParsedExpression{.invalid = "expected `}` or fields in table expression"}
|
||||
else => return ParsedExpression{.invalid = "expected `}` or fields in table literal"},
|
||||
}
|
||||
}
|
||||
},
|
||||
|
||||
.symbol_minus => {
|
||||
if (!tokenizer.step()) {
|
||||
if (!tokenizer.step(.ignore_newlines)) {
|
||||
return ParsedExpression{.invalid = "expected expression after numeric negation (`-`)"};
|
||||
}
|
||||
|
||||
|
|
@ -431,7 +457,7 @@ pub const ParsedExpression = union (enum) {
|
|||
},
|
||||
|
||||
.symbol_bang => {
|
||||
if (!tokenizer.step()) {
|
||||
if (!tokenizer.step(.ignore_newlines)) {
|
||||
return ParsedExpression{.invalid = "expected expression after boolean negation (`!`)"};
|
||||
}
|
||||
|
||||
|
|
@ -470,7 +496,7 @@ pub const ParsedExpression = union (enum) {
|
|||
};
|
||||
|
||||
if (tokenizer.current_token == .symbol_asterisk) {
|
||||
if (!tokenizer.step()) {
|
||||
if (!tokenizer.step(.ignore_newlines)) {
|
||||
return ParsedExpression{.invalid = "expected right-hand side of expression after `*`"};
|
||||
}
|
||||
|
||||
|
|
@ -492,7 +518,7 @@ pub const ParsedExpression = union (enum) {
|
|||
}
|
||||
|
||||
if (tokenizer.current_token == .symbol_forward_slash) {
|
||||
if (!tokenizer.step()) {
|
||||
if (!tokenizer.step(.ignore_newlines)) {
|
||||
return ParsedExpression{.invalid = "expected right-hand side of expression after `/`"};
|
||||
}
|
||||
|
||||
|
|
@ -540,16 +566,14 @@ pub const ParsedStatements = union (enum) {
|
|||
statements_list.deinit(allocator);
|
||||
};
|
||||
|
||||
while (tokenizer.step()) {
|
||||
while (tokenizer.step(.ignore_newlines)) {
|
||||
switch (tokenizer.current_token) {
|
||||
.newline => {},
|
||||
|
||||
.keyword_return => {
|
||||
if (has_returned) {
|
||||
return ParsedStatements{.invalid = "cannot return more than once per function scope"};
|
||||
}
|
||||
|
||||
if (tokenizer.step() and (tokenizer.current_token != .newline)) {
|
||||
if (tokenizer.step(.include_newlines) and (tokenizer.current_token != .newline)) {
|
||||
var parsed_expression = try ParsedExpression.init(allocator, tokenizer);
|
||||
|
||||
switch (parsed_expression) {
|
||||
|
|
@ -569,16 +593,14 @@ pub const ParsedStatements = union (enum) {
|
|||
try statements_list.push_one(allocator, .return_nothing);
|
||||
}
|
||||
|
||||
if (tokenizer.step() and tokenizer.current_token != .newline) {
|
||||
return ParsedStatements{.invalid = "expected newline after expression"};
|
||||
if (tokenizer.step(.ignore_newlines) and tokenizer.current_token != .newline) {
|
||||
return ParsedStatements{.invalid = "expected newline"};
|
||||
}
|
||||
|
||||
has_returned = true;
|
||||
},
|
||||
|
||||
else => {
|
||||
return ParsedStatements{.invalid = "invalid statement"};
|
||||
},
|
||||
else => return ParsedStatements{.invalid = "invalid statement"},
|
||||
}
|
||||
}
|
||||
|
||||
|
|
@ -599,6 +621,7 @@ pub const Expression = union (enum) {
|
|||
integer_literal: types.Integer,
|
||||
float_literal: types.Float,
|
||||
string_literal: []const u8,
|
||||
array_literal: ArrayElements,
|
||||
table_literal: TableFields,
|
||||
grouped_expression: *Expression,
|
||||
|
||||
|
|
@ -617,13 +640,20 @@ pub const Expression = union (enum) {
|
|||
switch (self.*) {
|
||||
.nil_literal, .true_literal, .false_literal, .integer_literal, .float_literal, .string_literal => {},
|
||||
|
||||
.table_literal => |*literal| {
|
||||
for (literal.values) |field| {
|
||||
field.expression.deinit(allocator);
|
||||
coral.io.deallocate(allocator, field.expression);
|
||||
.array_literal => |*elements| {
|
||||
for (elements.values) |*element_expression| {
|
||||
element_expression.deinit(allocator);
|
||||
}
|
||||
|
||||
literal.deinit(allocator);
|
||||
elements.deinit(allocator);
|
||||
},
|
||||
|
||||
.table_literal => |*fields| {
|
||||
for (fields.values) |*field| {
|
||||
field.expression.deinit(allocator);
|
||||
}
|
||||
|
||||
fields.deinit(allocator);
|
||||
},
|
||||
|
||||
.grouped_expression => |expression| {
|
||||
|
|
@ -716,7 +746,7 @@ pub const Statements = struct {
|
|||
|
||||
const TableFields = coral.list.Stack(struct {
|
||||
identifier: []const u8,
|
||||
expression: *Expression,
|
||||
expression: Expression,
|
||||
});
|
||||
|
||||
pub const UnaryOperation = enum {
|
||||
|
|
|
|||
|
|
@ -43,7 +43,6 @@ pub const Token = union(enum) {
|
|||
pub fn text(self: Token) []const u8 {
|
||||
return switch (self) {
|
||||
.unknown => |unknown| @ptrCast([*]const u8, &unknown)[0 .. 1],
|
||||
.newline => "newline",
|
||||
.identifier_global => |identifier| identifier,
|
||||
.identifier_local => |identifier| identifier,
|
||||
|
||||
|
|
@ -83,22 +82,18 @@ pub const Tokenizer = struct {
|
|||
previous_token: Token = .newline,
|
||||
current_token: Token = .newline,
|
||||
|
||||
pub fn has_next(self: Tokenizer) bool {
|
||||
return self.source.len != 0;
|
||||
}
|
||||
|
||||
pub fn step(self: *Tokenizer) bool {
|
||||
pub fn step(self: *Tokenizer, newline_rules: enum { ignore_newlines, include_newlines }) bool {
|
||||
self.previous_token = self.current_token;
|
||||
|
||||
var cursor = @as(usize, 0);
|
||||
|
||||
defer self.source = self.source[cursor ..];
|
||||
|
||||
while (self.has_next()) switch (self.source[cursor]) {
|
||||
while (cursor < self.source.len) switch (self.source[cursor]) {
|
||||
'#' => {
|
||||
cursor += 1;
|
||||
|
||||
while (self.has_next() and (self.source[cursor] == '\n')) {
|
||||
while (cursor < self.source.len and self.source[cursor] == '\n') {
|
||||
cursor += 1;
|
||||
}
|
||||
},
|
||||
|
|
@ -107,24 +102,31 @@ pub const Tokenizer = struct {
|
|||
|
||||
'\n' => {
|
||||
cursor += 1;
|
||||
|
||||
switch (newline_rules) {
|
||||
.include_newlines => {
|
||||
self.lines_stepped += 1;
|
||||
self.current_token = .newline;
|
||||
|
||||
return true;
|
||||
},
|
||||
|
||||
else => {},
|
||||
}
|
||||
},
|
||||
|
||||
'0' ... '9' => {
|
||||
const begin = cursor;
|
||||
|
||||
cursor += 1;
|
||||
|
||||
while (self.has_next()) switch (self.source[cursor]) {
|
||||
while (cursor < self.source.len) switch (self.source[cursor]) {
|
||||
'0' ... '9' => cursor += 1,
|
||||
|
||||
'.' => {
|
||||
cursor += 1;
|
||||
|
||||
while (self.has_next()) switch (self.source[cursor]) {
|
||||
while (cursor < self.source.len) switch (self.source[cursor]) {
|
||||
'0' ... '9' => cursor += 1,
|
||||
else => break,
|
||||
};
|
||||
|
|
@ -159,40 +161,52 @@ pub const Tokenizer = struct {
|
|||
switch (identifier[0]) {
|
||||
'n' => if (coral.io.ends_with(identifier, "il")) {
|
||||
self.current_token = .keyword_nil;
|
||||
|
||||
return true;
|
||||
},
|
||||
|
||||
'f' => if (coral.io.ends_with(identifier, "alse")) {
|
||||
self.current_token = .keyword_false;
|
||||
|
||||
return true;
|
||||
},
|
||||
|
||||
't' => if (coral.io.ends_with(identifier, "rue")) {
|
||||
self.current_token = .keyword_true;
|
||||
|
||||
return true;
|
||||
},
|
||||
|
||||
'r' => if (coral.io.ends_with(identifier, "eturn")) {
|
||||
self.current_token = .keyword_return;
|
||||
|
||||
return true;
|
||||
},
|
||||
|
||||
's' => if (coral.io.ends_with(identifier, "elf")) {
|
||||
self.current_token = .keyword_self;
|
||||
|
||||
return true;
|
||||
},
|
||||
|
||||
else => self.current_token = .{.local = identifier},
|
||||
else => {},
|
||||
}
|
||||
|
||||
self.current_token = .{.local = identifier};
|
||||
|
||||
return true;
|
||||
},
|
||||
|
||||
'@' => {
|
||||
cursor += 1;
|
||||
|
||||
if (self.has_next()) switch (self.source[cursor]) {
|
||||
if (cursor < self.source.len) switch (self.source[cursor]) {
|
||||
'A'...'Z', 'a'...'z', '_' => {
|
||||
const begin = cursor;
|
||||
|
||||
cursor += 1;
|
||||
|
||||
while (self.has_next()) switch (self.source[cursor]) {
|
||||
while (cursor < self.source.len) switch (self.source[cursor]) {
|
||||
'0'...'9', 'A'...'Z', 'a'...'z', '_' => cursor += 1,
|
||||
else => break,
|
||||
};
|
||||
|
|
@ -209,7 +223,7 @@ pub const Tokenizer = struct {
|
|||
|
||||
cursor += 1;
|
||||
|
||||
while (self.has_next()) switch (self.source[cursor]) {
|
||||
while (cursor < self.source.len) switch (self.source[cursor]) {
|
||||
'"' => break,
|
||||
else => cursor += 1,
|
||||
};
|
||||
|
|
@ -235,7 +249,7 @@ pub const Tokenizer = struct {
|
|||
|
||||
cursor += 1;
|
||||
|
||||
while (self.has_next()) switch (self.source[cursor]) {
|
||||
while (cursor < self.source.len) switch (self.source[cursor]) {
|
||||
'"' => break,
|
||||
else => cursor += 1,
|
||||
};
|
||||
|
|
@ -319,7 +333,7 @@ pub const Tokenizer = struct {
|
|||
'=' => {
|
||||
cursor += 1;
|
||||
|
||||
if (self.has_next()) {
|
||||
if (cursor < self.source.len) {
|
||||
switch (self.source[cursor]) {
|
||||
'=' => {
|
||||
cursor += 1;
|
||||
|
|
@ -347,7 +361,7 @@ pub const Tokenizer = struct {
|
|||
'<' => {
|
||||
cursor += 1;
|
||||
|
||||
if (self.has_next() and (self.source[cursor] == '=')) {
|
||||
if (cursor < self.source.len and (self.source[cursor] == '=')) {
|
||||
cursor += 1;
|
||||
self.current_token = .symbol_less_equals;
|
||||
|
||||
|
|
@ -362,7 +376,7 @@ pub const Tokenizer = struct {
|
|||
'>' => {
|
||||
cursor += 1;
|
||||
|
||||
if (self.has_next() and (self.source[cursor] == '=')) {
|
||||
if (cursor < self.source.len and (self.source[cursor] == '=')) {
|
||||
cursor += 1;
|
||||
self.current_token = .symbol_greater_equals;
|
||||
|
||||
|
|
|
|||
Loading…
Reference in New Issue