case T_MINUSMINUS: \
case T_PLUSPLUS: \
case T_STRING_LITERAL: \
- case T_WIDE_CHARACTER_CONSTANT: \
case T___FUNCDNAME__: \
case T___FUNCSIG__: \
- case T___FUNCTION__: \
case T___PRETTY_FUNCTION__: \
case T___alignof__: \
case T___builtin_classify_type: \
[EXPR_LITERAL_BOOLEAN] = sizeof(literal_expression_t),
[EXPR_LITERAL_INTEGER] = sizeof(literal_expression_t),
[EXPR_LITERAL_FLOATINGPOINT] = sizeof(literal_expression_t),
- [EXPR_LITERAL_CHARACTER] = sizeof(literal_expression_t),
- [EXPR_LITERAL_WIDE_CHARACTER] = sizeof(literal_expression_t),
+ [EXPR_LITERAL_CHARACTER] = sizeof(string_literal_expression_t),
[EXPR_STRING_LITERAL] = sizeof(string_literal_expression_t),
- [EXPR_WIDE_STRING_LITERAL] = sizeof(string_literal_expression_t),
[EXPR_COMPOUND_LITERAL] = sizeof(compound_literal_expression_t),
[EXPR_CALL] = sizeof(call_expression_t),
[EXPR_UNARY_FIRST] = sizeof(unary_expression_t),
{
static const size_t sizes[] = {
[INITIALIZER_VALUE] = sizeof(initializer_value_t),
- [INITIALIZER_STRING] = sizeof(initializer_string_t),
- [INITIALIZER_WIDE_STRING] = sizeof(initializer_wide_string_t),
+ [INITIALIZER_STRING] = sizeof(initializer_value_t),
[INITIALIZER_LIST] = sizeof(initializer_list_t),
[INITIALIZER_DESIGNATOR] = sizeof(initializer_designator_t)
};
msg, type1, type2);
}
-/**
- * Expect the current token is the expected token.
- * If not, generate an error and skip until the next anchor.
- */
-static void expect(token_kind_t const expected)
+static bool skip_till(token_kind_t const expected, char const *const context)
{
if (UNLIKELY(token.kind != expected)) {
- parse_error_expected(NULL, expected, NULL);
+ parse_error_expected(context, expected, NULL);
add_anchor_token(expected);
eat_until_anchor();
rem_anchor_token(expected);
if (token.kind != expected)
- return;
+ return false;
}
- eat(expected);
+ return true;
+}
+
+/**
+ * Expect the current token is the expected token.
+ * If not, generate an error and skip until the next anchor.
+ */
+static void expect(token_kind_t const expected)
+{
+ if (skip_till(expected, NULL))
+ eat(expected);
}
static symbol_t *expect_identifier(char const *const context, source_position_t *const pos)
{
- if (token.kind != T_IDENTIFIER) {
- parse_error_expected(context, T_IDENTIFIER, NULL);
- add_anchor_token(T_IDENTIFIER);
- eat_until_anchor();
- rem_anchor_token(T_IDENTIFIER);
- if (token.kind != T_IDENTIFIER)
- return NULL;
- }
+ if (!skip_till(T_IDENTIFIER, context))
+ return NULL;
symbol_t *const sym = token.base.symbol;
if (pos)
*pos = *HERE;
{
/* FIXME Using the ast_obstack is a hack. Using the symbol_obstack is not
* possible, because other tokens are grown there alongside. */
- obstack_grow(&ast_obstack, s->begin, s->size - 1);
+ obstack_grow(&ast_obstack, s->begin, s->size);
}
static string_t finish_string(void)
{
obstack_1grow(&ast_obstack, '\0');
- size_t const size = obstack_object_size(&ast_obstack);
+ size_t const size = obstack_object_size(&ast_obstack) - 1;
char const *const string = obstack_finish(&ast_obstack);
return (string_t){ string, size };
}
return result;
}
-static string_t parse_string_literals(void)
+static string_t parse_string_literals(char const *const context)
{
+ if (!skip_till(T_STRING_LITERAL, context))
+ return (string_t){ "", 0 };
+
string_encoding_t enc;
source_position_t const pos = *HERE;
string_t const res = concat_string_literals(&enc);
return;
case EXPR_LITERAL_CASES:
+ case EXPR_LITERAL_CHARACTER:
case EXPR_ERROR:
case EXPR_STRING_LITERAL:
- case EXPR_WIDE_STRING_LITERAL:
case EXPR_COMPOUND_LITERAL: // TODO init?
case EXPR_SIZEOF:
case EXPR_CLASSIFY_TYPE:
}
}
-static initializer_t *initializer_from_string(array_type_t *const type,
- const string_t *const string)
-{
- /* TODO: check len vs. size of array type */
- (void) type;
-
- initializer_t *initializer = allocate_initializer_zero(INITIALIZER_STRING);
- initializer->string.string = *string;
-
- return initializer;
-}
-
-static initializer_t *initializer_from_wide_string(array_type_t *const type,
- const string_t *const string)
-{
- /* TODO: check len vs. size of array type */
- (void) type;
-
- initializer_t *const initializer =
- allocate_initializer_zero(INITIALIZER_WIDE_STRING);
- initializer->wide_string.string = *string;
-
- return initializer;
-}
-
/**
* Build an initializer from a given expression.
*/
{
/* TODO check that expression is a constant expression */
- /* §6.7.8.14/15 char array may be initialized by string literals */
- type_t *type = skip_typeref(orig_type);
- type_t *expr_type_orig = expression->base.type;
- type_t *expr_type = skip_typeref(expr_type_orig);
+ type_t *const type = skip_typeref(orig_type);
- if (is_type_array(type) && expr_type->kind == TYPE_POINTER) {
+ /* §6.7.8.14/15 char array may be initialized by string literals */
+ if (expression->kind == EXPR_STRING_LITERAL && is_type_array(type)) {
array_type_t *const array_type = &type->array;
type_t *const element_type = skip_typeref(array_type->element_type);
-
- if (element_type->kind == TYPE_ATOMIC) {
- atomic_type_kind_t akind = element_type->atomic.akind;
- switch (expression->kind) {
- case EXPR_STRING_LITERAL:
- if (akind == ATOMIC_TYPE_CHAR
- || akind == ATOMIC_TYPE_SCHAR
- || akind == ATOMIC_TYPE_UCHAR) {
- return initializer_from_string(array_type,
- &expression->string_literal.value);
- }
- break;
-
- case EXPR_WIDE_STRING_LITERAL: {
- type_t *bare_wchar_type = skip_typeref(type_wchar_t);
- if (get_unqualified_type(element_type) == bare_wchar_type) {
- return initializer_from_wide_string(array_type,
- &expression->string_literal.value);
- }
- break;
+ switch (expression->string_literal.encoding) {
+ case STRING_ENCODING_CHAR: {
+ if (is_type_atomic(element_type, ATOMIC_TYPE_CHAR) ||
+ is_type_atomic(element_type, ATOMIC_TYPE_SCHAR) ||
+ is_type_atomic(element_type, ATOMIC_TYPE_UCHAR)) {
+ goto make_string_init;
}
+ break;
+ }
- default:
- break;
+ case STRING_ENCODING_WIDE: {
+ type_t *bare_wchar_type = skip_typeref(type_wchar_t);
+ if (get_unqualified_type(element_type) == bare_wchar_type) {
+make_string_init:;
+ initializer_t *const init = allocate_initializer_zero(INITIALIZER_STRING);
+ init->value.value = expression;
+ return init;
}
+ break;
+ }
}
}
}
/* handle { "string" } special case */
- if ((expression->kind == EXPR_STRING_LITERAL
- || expression->kind == EXPR_WIDE_STRING_LITERAL)
- && outer_type != NULL) {
+ if (expression->kind == EXPR_STRING_LITERAL && outer_type != NULL) {
sub = initializer_from_expression(outer_type, expression);
if (sub != NULL) {
next_if(',');
size = max_index + 1;
break;
- case INITIALIZER_STRING:
- size = result->string.string.size;
- break;
-
- case INITIALIZER_WIDE_STRING:
- size = result->wide_string.string.size;
+ case INITIALIZER_STRING: {
+ string_literal_expression_t const *const str = get_init_string(result);
+ size = get_string_len(str->encoding, &str->value) + 1;
break;
+ }
case INITIALIZER_DESIGNATOR:
case INITIALIZER_VALUE:
} else {
errorf(pos, "multiple datatypes in declaration");
}
- goto end_error;
+ specifiers->type = type_error_type;
+ return;
}
}
if (specifiers->attributes != NULL)
type = handle_type_attributes(specifiers->attributes, type);
specifiers->type = type;
- return;
-
-end_error:
- specifiers->type = type_error_type;
}
static type_qualifiers_t parse_type_qualifiers(void)
case EXPR_REFERENCE:
case EXPR_ENUM_CONSTANT:
case EXPR_LITERAL_CASES:
+ case EXPR_LITERAL_CHARACTER:
case EXPR_STRING_LITERAL:
- case EXPR_WIDE_STRING_LITERAL:
case EXPR_COMPOUND_LITERAL: // TODO descend into initialisers
case EXPR_LABEL_ADDRESS:
case EXPR_CLASSIFY_TYPE:
}
case INITIALIZER_STRING:
- case INITIALIZER_WIDE_STRING:
case INITIALIZER_DESIGNATOR: // designators have no payload
return true;
}
symbol_t *iter_symbol = iter->base.symbol;
if (iter_symbol == NULL) {
type_t *type = iter->declaration.type;
- if (type->kind != TYPE_COMPOUND_STRUCT
- && type->kind != TYPE_COMPOUND_UNION)
+ if (!is_type_compound(type))
continue;
compound_t *sub_compound = type->compound.compound;
static expression_parser_function_t expression_parsers[T_LAST_TOKEN];
-static type_t *get_string_type(void)
+static type_t *get_string_type(string_encoding_t const enc)
{
- return is_warn_on(WARN_WRITE_STRINGS) ? type_const_char_ptr : type_char_ptr;
-}
-
-static type_t *get_wide_string_type(void)
-{
- return is_warn_on(WARN_WRITE_STRINGS) ? type_const_wchar_t_ptr : type_wchar_t_ptr;
+ bool const warn = is_warn_on(WARN_WRITE_STRINGS);
+ switch (enc) {
+ case STRING_ENCODING_CHAR: return warn ? type_const_char_ptr : type_char_ptr;
+ case STRING_ENCODING_WIDE: return warn ? type_const_wchar_t_ptr : type_wchar_t_ptr;
+ }
+ panic("invalid string encoding");
}
/**
*/
static expression_t *parse_string_literal(void)
{
- string_encoding_t enc;
- source_position_t const pos = *HERE;
- string_t const res = concat_string_literals(&enc);
-
- expression_t *literal;
- if (enc != STRING_ENCODING_CHAR) {
- literal = allocate_expression_zero(EXPR_WIDE_STRING_LITERAL);
- literal->base.type = get_wide_string_type();
- } else {
- literal = allocate_expression_zero(EXPR_STRING_LITERAL);
- literal->base.type = get_string_type();
- }
- literal->base.source_position = pos;
- literal->literal.value = res;
-
- return literal;
+ expression_t *const expr = allocate_expression_zero(EXPR_STRING_LITERAL);
+ expr->string_literal.value = concat_string_literals(&expr->string_literal.encoding);
+ expr->base.type = get_string_type(expr->string_literal.encoding);
+ return expr;
}
/**
*/
static expression_t *parse_character_constant(void)
{
- expression_t *literal = allocate_expression_zero(EXPR_LITERAL_CHARACTER);
- literal->base.type = c_mode & _CXX ? type_char : type_int;
- literal->literal.value = token.string.string;
+ expression_t *const literal = allocate_expression_zero(EXPR_LITERAL_CHARACTER);
+ literal->string_literal.encoding = token.string.encoding;
+ literal->string_literal.value = token.string.string;
+
+ size_t const size = get_string_len(token.string.encoding, &token.string.string);
+ switch (token.string.encoding) {
+ case STRING_ENCODING_CHAR:
+ literal->base.type = c_mode & _CXX ? type_char : type_int;
+ if (size > 1) {
+ if (!GNU_MODE && !(c_mode & _C99)) {
+ errorf(HERE, "more than 1 character in character constant");
+ } else {
+ literal->base.type = type_int;
+ warningf(WARN_MULTICHAR, HERE, "multi-character character constant");
+ }
+ }
+ break;
- size_t len = literal->literal.value.size;
- if (len > 1) {
- if (!GNU_MODE && !(c_mode & _C99)) {
- errorf(HERE, "more than 1 character in character constant");
- } else {
- literal->base.type = type_int;
+ case STRING_ENCODING_WIDE:
+ literal->base.type = type_int;
+ if (size > 1) {
warningf(WARN_MULTICHAR, HERE, "multi-character character constant");
}
+ break;
}
eat(T_CHARACTER_CONSTANT);
return literal;
}
-/**
- * Parse a wide character constant.
- */
-static expression_t *parse_wide_character_constant(void)
-{
- expression_t *literal = allocate_expression_zero(EXPR_LITERAL_WIDE_CHARACTER);
- literal->base.type = type_int;
- literal->literal.value = token.string.string;
-
- size_t len = wstrlen(&literal->literal.value);
- if (len > 1) {
- warningf(WARN_MULTICHAR, HERE, "multi-character character constant");
- }
-
- eat(T_WIDE_CHARACTER_CONSTANT);
- return literal;
-}
-
static entity_t *create_implicit_function(symbol_t *symbol, source_position_t const *const pos)
{
type_t *ntype = allocate_type_zero(TYPE_FUNCTION);
}
case EXPR_STRING_LITERAL: {
- size_t size = expression->string_literal.value.size;
- return make_array_type(type_char, size, TYPE_QUALIFIER_NONE);
- }
-
- case EXPR_WIDE_STRING_LITERAL: {
- size_t size = wstrlen(&expression->string_literal.value);
- return make_array_type(type_wchar_t, size, TYPE_QUALIFIER_NONE);
+ size_t const size = get_string_len(expression->string_literal.encoding, &expression->string_literal.value) + 1;
+ type_t *const elem = get_unqualified_type(expression->base.type->pointer.points_to);
+ return make_array_type(elem, size, TYPE_QUALIFIER_NONE);
}
case EXPR_COMPOUND_LITERAL:
case T_INTEGER:
case T_FLOATINGPOINT: return parse_number_literal();
case T_CHARACTER_CONSTANT: return parse_character_constant();
- case T_WIDE_CHARACTER_CONSTANT: return parse_wide_character_constant();
case T_STRING_LITERAL: return parse_string_literal();
- case T___FUNCTION__:
case T___func__: return parse_function_keyword(FUNCNAME_FUNCTION);
case T___PRETTY_FUNCTION__: return parse_function_keyword(FUNCNAME_PRETTY_FUNCTION);
case T___FUNCSIG__: return parse_function_keyword(FUNCNAME_FUNCSIG);
type_left = type;
}
- if (type_left->kind != TYPE_COMPOUND_STRUCT &&
- type_left->kind != TYPE_COMPOUND_UNION) {
-
+ if (!is_type_compound(type_left)) {
if (is_type_valid(type_left) && !saw_error) {
errorf(&pos,
"request for member '%Y' in something not a struct or union, but '%T'",
expr = expr->unary.value;
}
- if (expr->kind == EXPR_STRING_LITERAL
- || expr->kind == EXPR_WIDE_STRING_LITERAL) {
+ if (expr->kind == EXPR_STRING_LITERAL) {
source_position_t const *const pos = &expr->base.source_position;
warningf(WARN_ADDRESS, pos, "comparison with string literal results in unspecified behaviour");
}
case EXPR_LITERAL_MS_NOOP: return true;
case EXPR_LITERAL_BOOLEAN:
case EXPR_LITERAL_CHARACTER:
- case EXPR_LITERAL_WIDE_CHARACTER:
case EXPR_LITERAL_INTEGER:
case EXPR_LITERAL_FLOATINGPOINT:
case EXPR_STRING_LITERAL: return false;
- case EXPR_WIDE_STRING_LITERAL: return false;
case EXPR_CALL: {
const call_expression_t *const call = &expr->call;
{
expression_parser_function_t *entry = &expression_parsers[token_kind];
- if (entry->parser != NULL) {
- diagnosticf("for token '%k'\n", (token_kind_t)token_kind);
- panic("trying to register multiple expression parsers for a token");
- }
+ assert(!entry->parser);
entry->parser = parser;
}
{
expression_parser_function_t *entry = &expression_parsers[token_kind];
- if (entry->infix_parser != NULL) {
- diagnosticf("for token '%k'\n", (token_kind_t)token_kind);
- panic("trying to register multiple infix expression parsers for a "
- "token");
- }
+ assert(!entry->infix_parser);
entry->infix_parser = parser;
entry->infix_precedence = precedence;
}
return NULL;
}
- argument->constraints = parse_string_literals();
+ argument->constraints = parse_string_literals("asm argument");
add_anchor_token(')');
expect('(');
expression_t *expression = parse_expression();
while (token.kind == T_STRING_LITERAL) {
asm_clobber_t *clobber = allocate_ast_zero(sizeof(clobber[0]));
- clobber->clobber = parse_string_literals();
+ clobber->clobber = parse_string_literals(NULL);
*anchor = clobber;
anchor = &clobber->next;
asm_statement_t *asm_statement = &statement->asms;
eat(T_asm);
+ add_anchor_token(')');
+ add_anchor_token(':');
+ add_anchor_token(T_STRING_LITERAL);
if (next_if(T_volatile))
asm_statement->is_volatile = true;
expect('(');
- add_anchor_token(')');
- if (token.kind != T_STRING_LITERAL) {
- parse_error_expected("after asm(", T_STRING_LITERAL, NULL);
- goto end_of_asm;
- }
- asm_statement->asm_text = parse_string_literals();
+ rem_anchor_token(T_STRING_LITERAL);
+ asm_statement->asm_text = parse_string_literals("asm statement");
- add_anchor_token(':');
- if (!next_if(':')) {
- rem_anchor_token(':');
- goto end_of_asm;
- }
+ if (next_if(':'))
+ asm_statement->outputs = parse_asm_arguments(true);
- asm_statement->outputs = parse_asm_arguments(true);
- if (!next_if(':')) {
- rem_anchor_token(':');
- goto end_of_asm;
- }
+ if (next_if(':'))
+ asm_statement->inputs = parse_asm_arguments(false);
- asm_statement->inputs = parse_asm_arguments(false);
- if (!next_if(':')) {
- rem_anchor_token(':');
- goto end_of_asm;
- }
rem_anchor_token(':');
+ if (next_if(':'))
+ asm_statement->clobbers = parse_asm_clobbers();
- asm_statement->clobbers = parse_asm_clobbers();
-
-end_of_asm:
rem_anchor_token(')');
expect(')');
expect(';');
return is_local_variable(entity);
}
-/**
- * Check if a given expression represents a local variable and
- * return its declaration then, else return NULL.
- */
-entity_t *expression_is_variable(const expression_t *expression)
-{
- if (expression->base.kind != EXPR_REFERENCE) {
- return NULL;
- }
- entity_t *entity = expression->reference.entity;
- if (entity->kind != ENTITY_VARIABLE)
- return NULL;
-
- return entity;
-}
-
static void err_or_warn(source_position_t const *const pos, char const *const msg)
{
if (c_mode & _CXX || strict_mode) {
add_anchor_token(T_MINUSMINUS);
add_anchor_token(T_PLUSPLUS);
add_anchor_token(T_STRING_LITERAL);
- add_anchor_token(T_WIDE_CHARACTER_CONSTANT);
add_anchor_token(T__Bool);
add_anchor_token(T__Complex);
add_anchor_token(T__Imaginary);
- add_anchor_token(T___FUNCTION__);
add_anchor_token(T___PRETTY_FUNCTION__);
add_anchor_token(T___alignof__);
add_anchor_token(T___attribute__);
rem_anchor_token(T___attribute__);
rem_anchor_token(T___alignof__);
rem_anchor_token(T___PRETTY_FUNCTION__);
- rem_anchor_token(T___FUNCTION__);
rem_anchor_token(T__Imaginary);
rem_anchor_token(T__Complex);
rem_anchor_token(T__Bool);
- rem_anchor_token(T_WIDE_CHARACTER_CONSTANT);
rem_anchor_token(T_STRING_LITERAL);
rem_anchor_token(T_PLUSPLUS);
rem_anchor_token(T_MINUSMINUS);
expect('(');
rem_anchor_token(T_STRING_LITERAL);
- statement->asms.asm_text = parse_string_literals();
+ statement->asms.asm_text = parse_string_literals("global asm");
statement->base.next = unit->global_asm;
unit->global_asm = statement;
eat(T_extern);
source_position_t const pos = *HERE;
- char const *const linkage = parse_string_literals().begin;
+ char const *const linkage = parse_string_literals(NULL).begin;
linkage_kind_t old_linkage = current_linkage;
linkage_kind_t new_linkage;
{
environment_stack = NEW_ARR_F(stack_entry_t, 0);
label_stack = NEW_ARR_F(stack_entry_t, 0);
- diagnostic_count = 0;
error_count = 0;
warning_count = 0;
projects / cparser / blobdiff