case T_MINUSMINUS: \
case T_PLUSPLUS: \
case T_STRING_LITERAL: \
- case T_WIDE_CHARACTER_CONSTANT: \
- case T_WIDE_STRING_LITERAL: \
case T___FUNCDNAME__: \
case T___FUNCSIG__: \
case T___FUNCTION__: \
[EXPR_LITERAL_CHARACTER] = sizeof(literal_expression_t),
[EXPR_LITERAL_WIDE_CHARACTER] = sizeof(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),
return parse_subexpression(PREC_ASSIGNMENT);
}
-static void warn_string_concat(const source_position_t *pos)
+static void append_string(string_t const *const s)
{
- warningf(WARN_TRADITIONAL, pos, "traditional C rejects string constant concatenation");
+ /* 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);
}
-static string_t parse_string_literals(void)
+static string_t finish_string(void)
{
- assert(token.kind == T_STRING_LITERAL);
- string_t result = token.string.string;
+ obstack_1grow(&ast_obstack, '\0');
+ size_t const size = obstack_object_size(&ast_obstack) - 1;
+ char const *const string = obstack_finish(&ast_obstack);
+ return (string_t){ string, size };
+}
- eat(T_STRING_LITERAL);
+static string_t concat_string_literals(string_encoding_t *const out_enc)
+{
+ assert(token.kind == T_STRING_LITERAL);
- while (token.kind == T_STRING_LITERAL) {
- warn_string_concat(HERE);
- result = concat_strings(&result, &token.string.string);
+ string_t result;
+ string_encoding_t enc = token.string.encoding;
+ if (look_ahead(1)->kind == T_STRING_LITERAL) {
+ append_string(&token.string.string);
+ eat(T_STRING_LITERAL);
+ warningf(WARN_TRADITIONAL, HERE, "traditional C rejects string constant concatenation");
+ do {
+ if (token.string.encoding != STRING_ENCODING_CHAR) {
+ enc = token.string.encoding;
+ }
+ append_string(&token.string.string);
+ eat(T_STRING_LITERAL);
+ } while (token.kind == T_STRING_LITERAL);
+ result = finish_string();
+ } else {
+ result = token.string.string;
eat(T_STRING_LITERAL);
}
+ *out_enc = enc;
return result;
}
+static string_t parse_string_literals(void)
+{
+ string_encoding_t enc;
+ source_position_t const pos = *HERE;
+ string_t const res = concat_string_literals(&enc);
+
+ if (enc != STRING_ENCODING_CHAR) {
+ errorf(&pos, "expected plain string literal, got wide string literal");
+ }
+
+ return res;
+}
+
static attribute_t *allocate_attribute_zero(attribute_kind_t kind)
{
attribute_t *attribute = allocate_ast_zero(sizeof(*attribute));
case EXPR_LITERAL_CASES:
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:
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 (element_type->kind == TYPE_ATOMIC && expression->kind == EXPR_STRING_LITERAL) {
+ switch (expression->string_literal.encoding) {
+ case STRING_ENCODING_CHAR: {
+ atomic_type_kind_t const akind = element_type->atomic.akind;
if (akind == ATOMIC_TYPE_CHAR
|| akind == ATOMIC_TYPE_SCHAR
|| akind == ATOMIC_TYPE_UCHAR) {
&expression->string_literal.value);
}
break;
+ }
- case EXPR_WIDE_STRING_LITERAL: {
+ case STRING_ENCODING_WIDE: {
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,
}
break;
}
-
- default:
- break;
}
}
}
}
} else {
expression_t *array_index = designator->array_index;
- assert(designator->array_index != NULL);
+ if (is_constant_expression(array_index) != EXPR_CLASS_CONSTANT)
+ return true;
if (!is_type_array(type)) {
if (is_type_valid(type)) {
}
/* 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(',');
break;
case INITIALIZER_STRING:
- size = result->string.string.size;
+ size = result->string.string.size + 1;
break;
case INITIALIZER_WIDE_STRING:
case EXPR_ENUM_CONSTANT:
case EXPR_LITERAL_CASES:
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:
static expression_parser_function_t expression_parsers[T_LAST_TOKEN];
-static type_t *get_string_type(void)
-{
- return is_warn_on(WARN_WRITE_STRINGS) ? type_const_char_ptr : type_char_ptr;
-}
-
-static type_t *get_wide_string_type(void)
+static type_t *get_string_type(string_encoding_t const enc)
{
- 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)
{
- source_position_t begin = *HERE;
- string_t res = token.string.string;
- bool is_wide = (token.kind == T_WIDE_STRING_LITERAL);
-
- next_token();
- while (token.kind == T_STRING_LITERAL
- || token.kind == T_WIDE_STRING_LITERAL) {
- warn_string_concat(HERE);
- res = concat_strings(&res, &token.string.string);
- next_token();
- is_wide |= token.kind == T_WIDE_STRING_LITERAL;
- }
-
- expression_t *literal;
- if (is_wide) {
- 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 = begin;
- 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;
-
- 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;
- warningf(WARN_MULTICHAR, HERE, "multi-character character constant");
+ expression_t *literal;
+ switch (token.string.encoding) {
+ case STRING_ENCODING_CHAR: {
+ literal = allocate_expression_zero(EXPR_LITERAL_CHARACTER);
+ literal->base.type = c_mode & _CXX ? type_char : type_int;
+ literal->literal.value = token.string.string;
+
+ 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;
+ 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;
+ case STRING_ENCODING_WIDE: {
+ 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");
+ size_t len = wstrlen(&literal->literal.value);
+ if (len > 1) {
+ warningf(WARN_MULTICHAR, HERE, "multi-character character constant");
+ }
+ break;
+ }
}
- eat(T_WIDE_CHARACTER_CONSTANT);
+ eat(T_CHARACTER_CONSTANT);
return literal;
}
}
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 = expression->string_literal.value.size + 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:
- case T_WIDE_STRING_LITERAL: return parse_string_literal();
+ 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);
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_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;
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_WIDE_STRING_LITERAL);
add_anchor_token(T__Bool);
add_anchor_token(T__Complex);
add_anchor_token(T__Imaginary);
rem_anchor_token(T__Imaginary);
rem_anchor_token(T__Complex);
rem_anchor_token(T__Bool);
- rem_anchor_token(T_WIDE_STRING_LITERAL);
- rem_anchor_token(T_WIDE_CHARACTER_CONSTANT);
rem_anchor_token(T_STRING_LITERAL);
rem_anchor_token(T_PLUSPLUS);
rem_anchor_token(T_MINUSMINUS);
projects / cparser / blobdiff