#include "type_t.h"
#include "target_architecture.h"
#include "parser.h"
+#include "warning.h"
+#include "lang_features.h"
#include <assert.h>
#include <errno.h>
static const char *bufpos;
static strset_t stringset;
-static void error_prefix_at(const char *input_name, unsigned linenr)
-{
- fprintf(stderr, "%s:%u: Error: ", input_name, linenr);
-}
-
-static void error_prefix(void)
-{
- error_prefix_at(lexer_token.source_position.input_name,
- lexer_token.source_position.linenr);
-}
-
+/**
+ * Prints a parse error message at the current token.
+ *
+ * @param msg the error message
+ */
static void parse_error(const char *msg)
{
- error_prefix();
- fprintf(stderr, "%s\n", msg);
+ errorf(lexer_token.source_position, "%s", msg);
}
static inline void next_real_char(void)
{
- bufpos++;
- if(bufpos >= bufend) {
+ assert(bufpos <= bufend);
+ if (bufpos >= bufend) {
size_t s = fread(buf + MAX_PUTBACK, 1, sizeof(buf) - MAX_PUTBACK,
input);
if(s == 0) {
bufpos = buf + MAX_PUTBACK;
bufend = buf + MAX_PUTBACK + s;
}
- c = *(bufpos);
+ c = *bufpos++;
}
+/**
+ * Put a character back into the buffer.
+ *
+ * @param pc the character to put back
+ */
static inline void put_back(int pc)
{
- assert(bufpos >= buf);
- //assert(bufpos < buf+MAX_PUTBACK || *bufpos == pc);
-
- char *p = buf + (bufpos - buf);
- *p = (char) pc;
-
- /* going backwards in the buffer is legal as long as it's not more often
- * than MAX_PUTBACK */
- bufpos--;
+ assert(bufpos > buf);
+ *(--bufpos - buf + buf) = (char) pc;
#ifdef DEBUG_CHARS
printf("putback '%c'\n", pc);
next_char(); \
} \
lexer_token.source_position.linenr++; \
- code; \
+ code \
case '\n': \
next_char(); \
lexer_token.source_position.linenr++; \
- code;
+ code
#define eat(c_type) do { assert(c == c_type); next_char(); } while(0)
c = '\\';
}
+/**
+ * Set c to the next input character, ie.
+ * after expanding trigraphs.
+ */
static inline void next_char(void)
{
next_real_char();
case '8': \
case '9':
+/**
+ * Read a symbol from the input and build
+ * the lexer_token.
+ */
static void parse_symbol(void)
{
symbol_t *symbol;
static void parse_floating_suffix(void)
{
switch(c) {
- /* TODO: do something usefull with the suffixes... */
+ /* TODO: do something useful with the suffixes... */
case 'f':
case 'F':
next_char();
return v;
}
+/**
+ * Parses a hex number including hex floats and set the
+ * lexer_token.
+ */
static void parse_number_hex(void)
{
assert(c == 'x' || c == 'X');
parse_integer_suffix(true);
}
+/**
+ * Returns true if the given char is a octal digit.
+ *
+ * @param char the character to check
+ */
static inline bool is_octal_digit(int chr)
{
- return '0' <= chr && chr <= '7';
+ switch(chr) {
+ case '0':
+ case '1':
+ case '2':
+ case '3':
+ case '4':
+ case '5':
+ case '6':
+ case '7':
+ return true;
+ default:
+ return false;
+ }
}
+/**
+ * Parses a octal number and set the lexer_token.
+ */
static void parse_number_oct(void)
{
while(is_octal_digit(c)) {
parse_integer_suffix(true);
}
+/**
+ * Parses a decimal including float number and set the
+ * lexer_token.
+ */
static void parse_number_dec(void)
{
bool is_float = false;
obstack_free(&symbol_obstack, string);
}
+/**
+ * Parses a number and sets the lexer_token.
+ */
static void parse_number(void)
{
if (c == '0') {
}
}
+/**
+ * Returns the value of a digit.
+ * The only portable way to do it ...
+ */
+static int digit_value(int digit) {
+ switch (digit) {
+ case '0': return 0;
+ case '1': return 1;
+ case '2': return 2;
+ case '3': return 3;
+ case '4': return 4;
+ case '5': return 5;
+ case '6': return 6;
+ case '7': return 7;
+ case '8': return 8;
+ case '9': return 9;
+ case 'a':
+ case 'A': return 10;
+ case 'b':
+ case 'B': return 11;
+ case 'c':
+ case 'C': return 12;
+ case 'd':
+ case 'D': return 13;
+ case 'e':
+ case 'E': return 14;
+ case 'f':
+ case 'F': return 15;
+ default:
+ panic("wrong character given");
+ }
+}
+
+/**
+ * Parses an octal character sequence.
+ *
+ * @param first_digit the already read first digit
+ */
static int parse_octal_sequence(const int first_digit)
{
assert(is_octal_digit(first_digit));
- int value = first_digit - '0';
+ int value = digit_value(first_digit);
if (!is_octal_digit(c)) return value;
- value = 8 * value + c - '0';
+ value = 8 * value + digit_value(c);
next_char();
if (!is_octal_digit(c)) return value;
- value = 8 * value + c - '0';
+ value = 8 * value + digit_value(c);
next_char();
return (char_type)value;
}
+/**
+ * Parses a hex character sequence.
+ */
static int parse_hex_sequence(void)
{
int value = 0;
- while(1) {
- if (c >= '0' && c <= '9') {
- value = 16 * value + c - '0';
- } else if ('A' <= c && c <= 'F') {
- value = 16 * value + c - 'A' + 10;
- } else if ('a' <= c && c <= 'f') {
- value = 16 * value + c - 'a' + 10;
- } else {
- break;
- }
+ while(isxdigit(c)) {
+ value = 16 * value + digit_value(c);
next_char();
}
return (char_type)value;
}
+/**
+ * Parse an escape sequence.
+ */
static int parse_escape_sequence(void)
{
eat('\\');
}
}
-const char *concat_strings(const char *s1, const char *s2)
+/**
+ * Concatenate two strings.
+ */
+string_t concat_strings(const string_t *const s1, const string_t *const s2)
{
- size_t len1 = strlen(s1);
- size_t len2 = strlen(s2);
+ const size_t len1 = s1->size - 1;
+ const size_t len2 = s2->size - 1;
- char *concat = obstack_alloc(&symbol_obstack, len1 + len2 + 1);
- memcpy(concat, s1, len1);
- memcpy(concat + len1, s2, len2 + 1);
+ char *const concat = obstack_alloc(&symbol_obstack, len1 + len2 + 1);
+ memcpy(concat, s1->begin, len1);
+ memcpy(concat + len1, s2->begin, len2 + 1);
+#if 0 /* TODO hash */
const char *result = strset_insert(&stringset, concat);
if(result != concat) {
obstack_free(&symbol_obstack, concat);
}
return result;
+#else
+ return (string_t){ concat, len1 + len2 + 1 };
+#endif
+}
+
+/**
+ * Concatenate a string and a wide string.
+ */
+wide_string_t concat_string_wide_string(const string_t *const s1, const wide_string_t *const s2)
+{
+ const size_t len1 = s1->size - 1;
+ const size_t len2 = s2->size - 1;
+
+ wchar_rep_t *const concat = obstack_alloc(&symbol_obstack, (len1 + len2 + 1) * sizeof(*concat));
+ const char *const src = s1->begin;
+ for (size_t i = 0; i != len1; ++i) {
+ concat[i] = src[i];
+ }
+ memcpy(concat + len1, s2->begin, (len2 + 1) * sizeof(*concat));
+
+ return (wide_string_t){ concat, len1 + len2 + 1 };
}
+/**
+ * Concatenate two wide strings.
+ */
+wide_string_t concat_wide_strings(const wide_string_t *const s1, const wide_string_t *const s2)
+{
+ const size_t len1 = s1->size - 1;
+ const size_t len2 = s2->size - 1;
+
+ wchar_rep_t *const concat = obstack_alloc(&symbol_obstack, (len1 + len2 + 1) * sizeof(*concat));
+ memcpy(concat, s1->begin, len1 * sizeof(*concat));
+ memcpy(concat + len1, s2->begin, (len2 + 1) * sizeof(*concat));
+
+ return (wide_string_t){ concat, len1 + len2 + 1 };
+}
+
+/**
+ * Concatenate a wide string and a string.
+ */
+wide_string_t concat_wide_string_string(const wide_string_t *const s1, const string_t *const s2)
+{
+ const size_t len1 = s1->size - 1;
+ const size_t len2 = s2->size - 1;
+
+ wchar_rep_t *const concat = obstack_alloc(&symbol_obstack, (len1 + len2 + 1) * sizeof(*concat));
+ memcpy(concat, s1->begin, len1 * sizeof(*concat));
+ const char *const src = s2->begin;
+ for (size_t i = 0; i != len2 + 1; ++i) {
+ concat[i] = src[i];
+ }
+
+ return (wide_string_t){ concat, len1 + len2 + 1 };
+}
+
+/**
+ * Parse a string literal and set lexer_token.
+ */
static void parse_string_literal(void)
{
- unsigned start_linenr = lexer_token.source_position.linenr;
- char *string;
- const char *result;
+ const unsigned start_linenr = lexer_token.source_position.linenr;
- assert(c == '"');
- next_char();
+ eat('"');
int tc;
while(1) {
obstack_1grow(&symbol_obstack, (char) tc);
break;
- case EOF:
- error_prefix_at(lexer_token.source_position.input_name,
- start_linenr);
- fprintf(stderr, "string has no end\n");
+ case EOF: {
+ source_position_t source_position;
+ source_position.input_name = lexer_token.source_position.input_name;
+ source_position.linenr = start_linenr;
+ errorf(source_position, "string has no end");
lexer_token.type = T_ERROR;
return;
+ }
case '"':
next_char();
/* add finishing 0 to the string */
obstack_1grow(&symbol_obstack, '\0');
- string = obstack_finish(&symbol_obstack);
+ const size_t size = (size_t)obstack_object_size(&symbol_obstack);
+ const char *const string = obstack_finish(&symbol_obstack);
+#if 0 /* TODO hash */
/* check if there is already a copy of the string */
result = strset_insert(&stringset, string);
if(result != string) {
obstack_free(&symbol_obstack, string);
}
+#else
+ const char *const result = string;
+#endif
- lexer_token.type = T_STRING_LITERAL;
- lexer_token.v.string = result;
+ lexer_token.type = T_STRING_LITERAL;
+ lexer_token.v.string.begin = result;
+ lexer_token.v.string.size = size;
}
+/**
+ * Parse a wide character constant and set lexer_token.
+ */
static void parse_wide_character_constant(void)
{
eat('\'');
lexer_token.datatype = type_wchar_t;
}
+/**
+ * Parse a wide string literal and set lexer_token.
+ */
static void parse_wide_string_literal(void)
{
const unsigned start_linenr = lexer_token.source_position.linenr;
while(1) {
switch(c) {
- case '\\': {
- wchar_rep_t tc = parse_escape_sequence();
- obstack_grow(&symbol_obstack, &tc, sizeof(tc));
- break;
- }
+ case '\\': {
+ wchar_rep_t tc = parse_escape_sequence();
+ obstack_grow(&symbol_obstack, &tc, sizeof(tc));
+ break;
+ }
- case EOF:
- error_prefix_at(lexer_token.source_position.input_name,
- start_linenr);
- fprintf(stderr, "string has no end\n");
- lexer_token.type = T_ERROR;
- return;
+ case EOF: {
+ source_position_t source_position;
+ source_position.input_name = lexer_token.source_position.input_name;
+ source_position.linenr = start_linenr;
+ errorf(source_position, "string has no end");
+ lexer_token.type = T_ERROR;
+ return;
+ }
- case '"':
- next_char();
- goto end_of_string;
+ case '"':
+ next_char();
+ goto end_of_string;
- default: {
- wchar_rep_t tc = c;
- obstack_grow(&symbol_obstack, &tc, sizeof(tc));
- next_char();
- break;
- }
+ default: {
+ wchar_rep_t tc = c;
+ obstack_grow(&symbol_obstack, &tc, sizeof(tc));
+ next_char();
+ break;
+ }
}
}
lexer_token.v.wide_string.size = size;
}
+/**
+ * Parse a character constant and set lexer_token.
+ */
static void parse_character_constant(void)
{
+ const unsigned start_linenr = lexer_token.source_position.linenr;
+
eat('\'');
- int found_char = 0;
+ int tc;
while(1) {
switch(c) {
case '\\':
- found_char = parse_escape_sequence();
+ tc = parse_escape_sequence();
+ obstack_1grow(&symbol_obstack, (char) tc);
break;
MATCH_NEWLINE(
break;
)
+ case EOF: {
+ source_position_t source_position;
+ source_position.input_name = lexer_token.source_position.input_name;
+ source_position.linenr = start_linenr;
+ errorf(source_position, "EOF while parsing character constant");
+ lexer_token.type = T_ERROR;
+ return;
+ }
+
case '\'':
next_char();
goto end_of_char_constant;
- case EOF:
- parse_error("EOF while parsing character constant");
- lexer_token.type = T_ERROR;
- return;
-
default:
- if(found_char != 0) {
- parse_error("more than 1 characters in character "
- "constant");
- goto end_of_char_constant;
- } else {
- found_char = c;
- next_char();
- }
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
break;
+
}
}
-end_of_char_constant:
- lexer_token.type = T_INTEGER;
- lexer_token.v.intvalue = found_char;
- lexer_token.datatype = type_int;
+end_of_char_constant:;
+ const size_t size = (size_t)obstack_object_size(&symbol_obstack);
+ const char *const string = obstack_finish(&symbol_obstack);
+
+ lexer_token.type = T_CHARS;
+ lexer_token.v.string.begin = string;
+ lexer_token.v.string.size = size;
+ lexer_token.datatype = type_int;
}
+/**
+ * Skip a multiline comment.
+ */
static void skip_multiline_comment(void)
{
unsigned start_linenr = lexer_token.source_position.linenr;
while(1) {
switch(c) {
+ case '/':
+ next_char();
+ if (c == '*') {
+ /* TODO: nested comment, warn here */
+ }
+ break;
case '*':
next_char();
if(c == '/') {
MATCH_NEWLINE(break;)
- case EOF:
- error_prefix_at(lexer_token.source_position.input_name,
- start_linenr);
- fprintf(stderr, "at end of file while looking for comment end\n");
+ case EOF: {
+ source_position_t source_position;
+ source_position.input_name = lexer_token.source_position.input_name;
+ source_position.linenr = start_linenr;
+ errorf(source_position, "at end of file while looking for comment end");
return;
+ }
default:
next_char();
}
}
+/**
+ * Skip a single line comment.
+ */
static void skip_line_comment(void)
{
while(1) {
}
}
+/** The current preprocessor token. */
static token_t pp_token;
+/**
+ * Read the next preprocessor token.
+ */
static inline void next_pp_token(void)
{
lexer_next_preprocessing_token();
pp_token = lexer_token;
}
+/**
+ * Eat all preprocessor tokens until newline.
+ */
static void eat_until_newline(void)
{
while(pp_token.type != '\n' && pp_token.type != T_EOF) {
}
}
-static void error_directive(void)
-{
- error_prefix();
- fprintf(stderr, "#error directive: \n");
-
- /* parse pp-tokens until new-line */
-}
-
+/**
+ * Handle the define directive.
+ */
static void define_directive(void)
{
lexer_next_preprocessing_token();
}
}
+/**
+ * Handle the ifdef directive.
+ */
static void ifdef_directive(int is_ifndef)
{
(void) is_ifndef;
//extect_newline();
}
+/**
+ * Handle the endif directive.
+ */
static void endif_directive(void)
{
//expect_newline();
}
+/**
+ * Parse the line directive.
+ */
static void parse_line_directive(void)
{
if(pp_token.type != T_INTEGER) {
next_pp_token();
}
if(pp_token.type == T_STRING_LITERAL) {
- lexer_token.source_position.input_name = pp_token.v.string;
+ lexer_token.source_position.input_name = pp_token.v.string.begin;
next_pp_token();
}
eat_until_newline();
}
+/**
+ * STDC pragmas.
+ */
+typedef enum {
+ STDC_UNKNOWN,
+ STDC_FP_CONTRACT,
+ STDC_FENV_ACCESS,
+ STDC_CX_LIMITED_RANGE
+} stdc_pragma_kind_t;
+
+/**
+ * STDC pragma values.
+ */
+typedef enum {
+ STDC_VALUE_UNKNOWN,
+ STDC_VALUE_ON,
+ STDC_VALUE_OFF,
+ STDC_VALUE_DEFAULT
+} stdc_pragma_value_kind_t;
+
+/**
+ * Parse a pragma directive.
+ */
+static void parse_pragma(void) {
+ bool unknown_pragma = true;
+
+ next_pp_token();
+ if (pp_token.v.symbol->pp_ID == TP_STDC) {
+ stdc_pragma_kind_t kind = STDC_UNKNOWN;
+ /* a STDC pragma */
+ if (c_mode & _C99) {
+ next_pp_token();
+
+ switch (pp_token.v.symbol->pp_ID) {
+ case TP_FP_CONTRACT:
+ kind = STDC_FP_CONTRACT;
+ break;
+ case TP_FENV_ACCESS:
+ kind = STDC_FENV_ACCESS;
+ break;
+ case TP_CX_LIMITED_RANGE:
+ kind = STDC_CX_LIMITED_RANGE;
+ break;
+ default:
+ break;
+ }
+ if (kind != STDC_UNKNOWN) {
+ stdc_pragma_value_kind_t value = STDC_VALUE_UNKNOWN;
+ next_pp_token();
+ switch (pp_token.v.symbol->pp_ID) {
+ case TP_ON:
+ value = STDC_VALUE_ON;
+ break;
+ case TP_OFF:
+ value = STDC_VALUE_OFF;
+ break;
+ case TP_DEFAULT:
+ value = STDC_VALUE_DEFAULT;
+ break;
+ default:
+ break;
+ }
+ if (value != STDC_VALUE_UNKNOWN) {
+ unknown_pragma = false;
+ } else {
+ errorf(pp_token.source_position, "bad STDC pragma argument");
+ }
+ }
+ }
+ } else {
+ unknown_pragma = true;
+ }
+ eat_until_newline();
+ if (unknown_pragma && warning.unknown_pragmas) {
+ warningf(pp_token.source_position, "encountered unknown #pragma");
+ }
+}
+
+/**
+ * Parse a preprocessor non-null directive.
+ */
static void parse_preprocessor_identifier(void)
{
assert(pp_token.type == T_IDENTIFIER);
case TP_elif:
case TP_undef:
case TP_error:
- error_directive();
+ /* TODO; output the rest of the line */
+ parse_error("#error directive: ");
break;
case TP_pragma:
- warningf(lexer_token.source_position, "encountered unknown #pragma");
- eat_until_newline();
+ parse_pragma();
break;
}
}
+/**
+ * Parse a preprocessor directive.
+ */
static void parse_preprocessor_directive(void)
{
next_pp_token();
case T_INTEGER:
parse_line_directive();
break;
+ case '\n':
+ /* NULL directive, see ยง 6.10.7 */
+ break;
default:
parse_error("invalid preprocessor directive");
eat_until_newline();
default:
next_char();
- error_prefix();
- fprintf(stderr, "unknown character '%c' found\n", c);
+ errorf(lexer_token.source_position, "unknown character '%c' found\n", c);
lexer_token.type = T_ERROR;
return;
}
lexer_token.source_position.input_name = input_name;
symbol_L = symbol_table_insert("L");
+ bufpos = NULL;
+ bufend = NULL;
/* place a virtual \n at the beginning so the lexer knows that we're
* at the beginning of a line */
projects / cparser / blobdiff