+/*
+ * This file is part of cparser.
+ * Copyright (C) 2007-2009 Matthias Braun <matze@braunis.de>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
+ * 02111-1307, USA.
+ */
#include <config.h>
+#include "diagnostic.h"
#include "lexer.h"
+#include "symbol_t.h"
#include "token_t.h"
#include "symbol_table_t.h"
#include "adt/error.h"
#include "adt/strset.h"
#include "adt/util.h"
+#include "types.h"
+#include "type_t.h"
+#include "target_architecture.h"
+#include "parser.h"
+#include "warning.h"
+#include "lang_features.h"
#include <assert.h>
#include <errno.h>
#include <string.h>
+#include <stdbool.h>
#include <ctype.h>
+#ifndef _WIN32
+#include <strings.h>
+#endif
+
//#define DEBUG_CHARS
#define MAX_PUTBACK 3
+#define BUF_SIZE 1024
-static int c;
-token_t lexer_token;
-symbol_t *symbol_L;
-static FILE *input;
-static char buf[1024 + MAX_PUTBACK];
-static const char *bufend;
-static const char *bufpos;
-static strset_t stringset;
+#if defined(_WIN32) || defined(__CYGWIN__)
+/* No strtold on windows and no replacement yet */
+#define strtold(s, e) strtod(s, e)
+#endif
-static void error_prefix_at(const char *input_name, unsigned linenr)
+static utf32 c;
+token_t lexer_token;
+symbol_t *symbol_L;
+static FILE *input;
+static utf32 buf[BUF_SIZE + MAX_PUTBACK];
+static const utf32 *bufend;
+static const utf32 *bufpos;
+static strset_t stringset;
+bool allow_dollar_in_symbol = true;
+
+/**
+ * Prints a parse error message at the current token.
+ *
+ * @param msg the error message
+ */
+static void parse_error(const char *msg)
{
- fprintf(stderr, "%s:%u: Error: ", input_name, linenr);
+ errorf(&lexer_token.source_position, "%s", msg);
}
-static void error_prefix(void)
+/**
+ * Prints an internal error message at the current token.
+ *
+ * @param msg the error message
+ */
+static NORETURN internal_error(const char *msg)
{
- error_prefix_at(lexer_token.source_position.input_name,
- lexer_token.source_position.linenr);
+ internal_errorf(&lexer_token.source_position, "%s", msg);
}
-static void parse_error(const char *msg)
+static size_t read_block(unsigned char *const read_buf, size_t const n)
{
- error_prefix();
- fprintf(stderr, "%s\n", msg);
+ size_t const s = fread(read_buf, 1, n, input);
+ if (s == 0) {
+ if (ferror(input))
+ parse_error("read from input failed");
+ buf[MAX_PUTBACK] = EOF;
+ bufpos = buf + MAX_PUTBACK;
+ bufend = buf + MAX_PUTBACK + 1;
+ }
+ return s;
}
-static inline void next_real_char(void)
+static void decode_iso_8859_1(void)
{
- bufpos++;
- if(bufpos >= bufend) {
- size_t s = fread(buf + MAX_PUTBACK, 1, sizeof(buf) - MAX_PUTBACK,
- input);
- if(s == 0) {
- c = EOF;
+ unsigned char read_buf[BUF_SIZE];
+ size_t const s = read_block(read_buf, sizeof(read_buf));
+ if (s == 0)
+ return;
+
+ unsigned char const *src = read_buf;
+ unsigned char const *end = read_buf + s;
+ utf32 *dst = buf + MAX_PUTBACK;
+ while (src != end)
+ *dst++ = *src++;
+
+ bufpos = buf + MAX_PUTBACK;
+ bufend = dst;
+}
+
+static void decode_iso_8859_15(void)
+{
+ unsigned char read_buf[BUF_SIZE];
+ size_t const s = read_block(read_buf, sizeof(read_buf));
+ if (s == 0)
+ return;
+
+ unsigned char const *src = read_buf;
+ unsigned char const *end = read_buf + s;
+ utf32 *dst = buf + MAX_PUTBACK;
+ while (src != end) {
+ utf32 tc = *src++;
+ switch (tc) {
+ case 0xA4: tc = 0x20AC; break; // €
+ case 0xA6: tc = 0x0160; break; // Š
+ case 0xA8: tc = 0x0161; break; // š
+ case 0xB4: tc = 0x017D; break; // Ž
+ case 0xB8: tc = 0x017E; break; // ž
+ case 0xBC: tc = 0x0152; break; // Œ
+ case 0xBD: tc = 0x0153; break; // œ
+ case 0xBE: tc = 0x0178; break; // Ÿ
+ }
+ *dst++ = tc;
+ }
+
+ bufpos = buf + MAX_PUTBACK;
+ bufend = dst;
+}
+
+static void decode_utf8(void)
+{
+ static utf32 part_decoded_min_code;
+ static utf32 part_decoded_char;
+ static size_t part_decoded_rest_len;
+
+ do {
+ unsigned char read_buf[BUF_SIZE];
+ size_t const s = read_block(read_buf, sizeof(read_buf));
+ if (s == 0) {
+ if (part_decoded_rest_len > 0)
+ parse_error("incomplete input char at end of input");
return;
}
+
+ unsigned char const *src = read_buf;
+ unsigned char const *end = read_buf + s;
+ utf32 *dst = buf + MAX_PUTBACK;
+ utf32 decoded;
+ utf32 min_code;
+
+ if (part_decoded_rest_len != 0) {
+ min_code = part_decoded_min_code;
+ decoded = part_decoded_char;
+ size_t const rest_len = part_decoded_rest_len;
+ part_decoded_rest_len = 0;
+ switch (rest_len) {
+ case 4: goto realign;
+ case 3: goto three_more;
+ case 2: goto two_more;
+ default: goto one_more;
+ }
+ }
+
+ while (src != end) {
+ if ((*src & 0x80) == 0) {
+ decoded = *src++;
+ } else if ((*src & 0xE0) == 0xC0) {
+ min_code = 0x80;
+ decoded = *src++ & 0x1F;
+one_more:
+ if (src == end) {
+ part_decoded_min_code = min_code;
+ part_decoded_char = decoded;
+ part_decoded_rest_len = 1;
+ break;
+ }
+ if ((*src & 0xC0) == 0x80) {
+ decoded = (decoded << 6) | (*src++ & 0x3F);
+ } else {
+ goto invalid_char;
+ }
+ if (decoded < min_code ||
+ decoded > 0x10FFFF ||
+ (0xD800 <= decoded && decoded < 0xE000) || // high/low surrogates
+ (0xFDD0 <= decoded && decoded < 0xFDF0) || // noncharacters
+ (decoded & 0xFFFE) == 0xFFFE) { // noncharacters
+ parse_error("invalid byte sequence in input");
+ }
+ } else if ((*src & 0xF0) == 0xE0) {
+ min_code = 0x800;
+ decoded = *src++ & 0x0F;
+two_more:
+ if (src == end) {
+ part_decoded_min_code = min_code;
+ part_decoded_char = decoded;
+ part_decoded_rest_len = 2;
+ break;
+ }
+ if ((*src & 0xC0) == 0x80) {
+ decoded = (decoded << 6) | (*src++ & 0x3F);
+ } else {
+ goto invalid_char;
+ }
+ goto one_more;
+ } else if ((*src & 0xF8) == 0xF0) {
+ min_code = 0x10000;
+ decoded = *src++ & 0x07;
+three_more:
+ if (src == end) {
+ part_decoded_min_code = min_code;
+ part_decoded_char = decoded;
+ part_decoded_rest_len = 3;
+ break;
+ }
+ if ((*src & 0xC0) == 0x80) {
+ decoded = (decoded << 6) | (*src++ & 0x3F);
+ } else {
+ goto invalid_char;
+ }
+ goto two_more;
+ } else {
+invalid_char:
+ parse_error("invalid byte sequence in input");
+realign:
+ do {
+ ++src;
+ if (src == end) {
+ part_decoded_rest_len = 4;
+ break;
+ }
+ } while ((*src & 0xC0) == 0x80 || (*src & 0xF8) == 0xF8);
+ continue;
+ }
+ *dst++ = decoded;
+ }
+
bufpos = buf + MAX_PUTBACK;
- bufend = buf + MAX_PUTBACK + s;
+ bufend = dst;
+ } while (bufpos == bufend);
+}
+
+static void decode_windows_1252(void)
+{
+ unsigned char read_buf[BUF_SIZE];
+ size_t const s = read_block(read_buf, sizeof(read_buf));
+ if (s == 0)
+ return;
+
+ unsigned char const *src = read_buf;
+ unsigned char const *end = read_buf + s;
+ utf32 *dst = buf + MAX_PUTBACK;
+ while (src != end) {
+ utf32 tc = *src++;
+ switch (tc) {
+ case 0x80: tc = 0x20AC; break; // €
+ case 0x82: tc = 0x201A; break; // ‚
+ case 0x83: tc = 0x0192; break; // ƒ
+ case 0x84: tc = 0x201E; break; // „
+ case 0x85: tc = 0x2026; break; // …
+ case 0x86: tc = 0x2020; break; // †
+ case 0x87: tc = 0x2021; break; // ‡
+ case 0x88: tc = 0x02C6; break; // ˆ
+ case 0x89: tc = 0x2030; break; // ‰
+ case 0x8A: tc = 0x0160; break; // Š
+ case 0x8B: tc = 0x2039; break; // ‹
+ case 0x8C: tc = 0x0152; break; // Œ
+ case 0x8E: tc = 0x017D; break; // Ž
+ case 0x91: tc = 0x2018; break; // ‘
+ case 0x92: tc = 0x2019; break; // ’
+ case 0x93: tc = 0x201C; break; // “
+ case 0x94: tc = 0x201D; break; // ”
+ case 0x95: tc = 0x2022; break; // •
+ case 0x96: tc = 0x2013; break; // –
+ case 0x97: tc = 0x2014; break; // —
+ case 0x98: tc = 0x02DC; break; // ˜
+ case 0x99: tc = 0x2122; break; // ™
+ case 0x9A: tc = 0x0161; break; // š
+ case 0x9B: tc = 0x203A; break; // ›
+ case 0x9C: tc = 0x0153; break; // œ
+ case 0x9E: tc = 0x017E; break; // ž
+ case 0x9F: tc = 0x0178; break; // Ÿ
+ }
+ *dst++ = tc;
}
- c = *(bufpos);
+
+ bufpos = buf + MAX_PUTBACK;
+ bufend = dst;
}
-static inline void put_back(int pc)
+typedef void (*decoder_t)(void);
+
+static decoder_t decoder = decode_utf8;
+
+typedef struct named_decoder_t {
+ char const *name;
+ decoder_t decoder;
+} named_decoder_t;
+
+static named_decoder_t const decoders[] = {
+ { "CP819", decode_iso_8859_1 }, // offical alias
+ { "IBM819", decode_iso_8859_1 }, // offical alias
+ { "ISO-8859-1", decode_iso_8859_1 }, // offical alias
+ { "ISO-8859-15", decode_iso_8859_15 }, // offical name
+ { "ISO8859-1", decode_iso_8859_1 },
+ { "ISO8859-15", decode_iso_8859_15 },
+ { "ISO_8859-1", decode_iso_8859_1 }, // offical alias
+ { "ISO_8859-15", decode_iso_8859_15 }, // offical alias
+ { "ISO_8859-1:1987", decode_iso_8859_1 }, // offical name
+ { "Latin-9", decode_iso_8859_15 }, // offical alias
+ { "UTF-8", decode_utf8 }, // offical name
+ { "csISOLatin1", decode_iso_8859_1 }, // offical alias
+ { "cp1252", decode_windows_1252 },
+ { "iso-ir-100", decode_iso_8859_1 }, // offical alias
+ { "l1", decode_iso_8859_1 }, // offical alias
+ { "latin1", decode_iso_8859_1 }, // offical alias
+ { "windows-1252", decode_windows_1252 }, // official name
+
+ { NULL, NULL }
+};
+
+void select_input_encoding(char const* const encoding)
{
- assert(bufpos >= buf);
- assert(bufpos < buf+MAX_PUTBACK || *bufpos == pc);
+ for (named_decoder_t const *i = decoders; i->name != NULL; ++i) {
+ if (strcasecmp(encoding, i->name) != 0)
+ continue;
+ decoder = i->decoder;
+ return;
+ }
+ fprintf(stderr, "error: input encoding \"%s\" not supported\n", encoding);
+}
- char *p = buf + (bufpos - buf);
- *p = pc;
+static inline void next_real_char(void)
+{
+ assert(bufpos <= bufend);
+ if (bufpos >= bufend) {
+ if (input == NULL) {
+ c = EOF;
+ return;
+ }
+ decoder();
+ }
+ c = *bufpos++;
+}
- /* going backwards in the buffer is legal as long as it's not more often
- * than MAX_PUTBACK */
- bufpos--;
+/**
+ * Put a character back into the buffer.
+ *
+ * @param pc the character to put back
+ */
+static inline void put_back(utf32 const pc)
+{
+ assert(bufpos > buf);
+ *(--bufpos - buf + buf) = pc;
#ifdef DEBUG_CHARS
- printf("putback '%c'\n", pc);
+ printf("putback '%lc'\n", pc);
#endif
}
#define MATCH_NEWLINE(code) \
case '\r': \
next_char(); \
- if(c == '\n') { \
+ if (c == '\n') { \
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)
+#define eat(c_type) do { assert(c == c_type); next_char(); } while (0)
static void maybe_concat_lines(void)
{
eat('\\');
- switch(c) {
+ switch (c) {
MATCH_NEWLINE(return;)
default:
c = '\\';
}
+/**
+ * Set c to the next input character, ie.
+ * after expanding trigraphs.
+ */
static inline void next_char(void)
{
next_real_char();
-#if 0
/* filter trigraphs */
- if(UNLIKELY(c == '\\')) {
+ if (UNLIKELY(c == '\\')) {
maybe_concat_lines();
goto end_of_next_char;
}
- if(LIKELY(c != '?'))
+ if (LIKELY(c != '?'))
goto end_of_next_char;
next_real_char();
- if(LIKELY(c != '?')) {
+ if (LIKELY(c != '?')) {
put_back(c);
c = '?';
goto end_of_next_char;
}
next_real_char();
- switch(c) {
+ switch (c) {
case '=': c = '#'; break;
case '(': c = '['; break;
case '/': c = '\\'; maybe_concat_lines(); break;
case '>': c = '}'; break;
case '-': c = '~'; break;
default:
- put_back('?');
put_back(c);
+ put_back('?');
c = '?';
break;
}
-end_of_next_char:
-#endif
- (void) maybe_concat_lines;
+end_of_next_char:;
#ifdef DEBUG_CHARS
printf("nchar '%c'\n", c);
-#else
- ;
#endif
}
#define SYMBOL_CHARS \
+ case '$': if (!allow_dollar_in_symbol) goto dollar_sign; \
case 'a': \
case 'b': \
case 'c': \
case '8': \
case '9':
+/**
+ * Read a symbol from the input and build
+ * the lexer_token.
+ */
static void parse_symbol(void)
{
- symbol_t *symbol;
- char *string;
-
- obstack_1grow(&symbol_obstack, c);
+ obstack_1grow(&symbol_obstack, (char) c);
next_char();
- while(1) {
- switch(c) {
+ while (true) {
+ switch (c) {
DIGITS
SYMBOL_CHARS
- obstack_1grow(&symbol_obstack, c);
+ obstack_1grow(&symbol_obstack, (char) c);
next_char();
break;
default:
+dollar_sign:
goto end_symbol;
}
}
end_symbol:
obstack_1grow(&symbol_obstack, '\0');
- string = obstack_finish(&symbol_obstack);
- symbol = symbol_table_insert(string);
+ char *string = obstack_finish(&symbol_obstack);
+ symbol_t *symbol = symbol_table_insert(string);
- lexer_token.type = symbol->ID;
- lexer_token.v.symbol = symbol;
+ lexer_token.type = symbol->ID;
+ lexer_token.symbol = symbol;
- if(symbol->string != string) {
+ if (symbol->string != string) {
obstack_free(&symbol_obstack, string);
}
}
-static void parse_integer_suffix(void)
+/**
+ * parse suffixes like 'LU' or 'f' after numbers
+ */
+static void parse_number_suffix(void)
{
- if(c == 'U' || c == 'U') {
- /* TODO do something with the suffixes... */
- next_char();
- if(c == 'L' || c == 'l') {
- next_char();
- if(c == 'L' || c == 'l') {
- next_char();
- }
- }
- } else if(c == 'l' || c == 'L') {
- next_char();
- if(c == 'l' || c == 'L') {
- next_char();
- if(c == 'u' || c == 'U') {
- next_char();
- }
- } else if(c == 'u' || c == 'U') {
+ assert(obstack_object_size(&symbol_obstack) == 0);
+ while (true) {
+ switch (c) {
+ SYMBOL_CHARS
+ obstack_1grow(&symbol_obstack, (char) c);
next_char();
+ break;
+ default:
+ dollar_sign:
+ goto finish_suffix;
}
}
+finish_suffix:
+ if (obstack_object_size(&symbol_obstack) == 0) {
+ lexer_token.symbol = NULL;
+ return;
+ }
+
+ obstack_1grow(&symbol_obstack, '\0');
+ char *string = obstack_finish(&symbol_obstack);
+ symbol_t *symbol = symbol_table_insert(string);
+
+ if (symbol->string != string) {
+ obstack_free(&symbol_obstack, string);
+ }
+ lexer_token.symbol = symbol;
}
-static void parse_floating_suffix(void)
+static string_t identify_string(char *string, size_t len)
{
- switch(c) {
- /* TODO: do something usefull with the suffixes... */
- case 'f':
- case 'F':
- case 'l':
- case 'L':
- next_char();
- break;
- default:
- break;
+ /* TODO hash */
+#if 0
+ const char *result = strset_insert(&stringset, concat);
+ if (result != concat) {
+ obstack_free(&symbol_obstack, concat);
}
+#else
+ const char *result = string;
+#endif
+ return (string_t) {result, len};
}
+/**
+ * Parses a hex number including hex floats and set the
+ * lexer_token.
+ */
static void parse_number_hex(void)
{
- assert(c == 'x' || c == 'X');
- next_char();
+ bool is_float = false;
+ bool has_digits = false;
- if (!isdigit(c) &&
- !('A' <= c && c <= 'F') &&
- !('a' <= c && c <= 'f')) {
- parse_error("premature end of hex number literal");
- lexer_token.type = T_ERROR;
- return;
+ assert(obstack_object_size(&symbol_obstack) == 0);
+ while (isxdigit(c)) {
+ has_digits = true;
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
}
- int value = 0;
- while(1) {
- if (isdigit(c)) {
- 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 {
- parse_integer_suffix();
+ if (c == '.') {
+ is_float = true;
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
- lexer_token.type = T_INTEGER;
- lexer_token.v.intvalue = value;
- return;
+ while (isxdigit(c)) {
+ has_digits = true;
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
}
- next_char();
}
-
- if(c == '.' || c == 'p' || c == 'P') {
+ if (c == 'p' || c == 'P') {
+ is_float = true;
+ obstack_1grow(&symbol_obstack, (char) c);
next_char();
- panic("Hex floating point numbers not implemented yet");
- }
-}
-static void parse_number_oct(void)
-{
- int value = 0;
- while(c >= '0' && c <= '7') {
- value = 8 * value + c - '0';
- next_char();
- }
- if (c == '8' || c == '9') {
- parse_error("invalid octal number");
- lexer_token.type = T_ERROR;
- return;
- }
+ if (c == '-' || c == '+') {
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
+ }
- lexer_token.type = T_INTEGER;
- lexer_token.v.intvalue = value;
+ while (isxdigit(c)) {
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
+ }
+ } else if (is_float) {
+ errorf(&lexer_token.source_position,
+ "hexadecimal floatingpoint constant requires an exponent");
+ }
- parse_integer_suffix();
-}
+ size_t size = obstack_object_size(&symbol_obstack);
+ char *string = obstack_finish(&symbol_obstack);
+ lexer_token.literal = identify_string(string, size);
-static void parse_floatingpoint_exponent(long double value)
-{
- unsigned int expo = 0;
- long double factor = 10.;
+ lexer_token.type =
+ is_float ? T_FLOATINGPOINT_HEXADECIMAL : T_INTEGER_HEXADECIMAL;
- if(c == '-') {
- next_char();
- factor = 0.1;
- } else if(c == '+') {
- next_char();
+ if (!has_digits) {
+ errorf(&lexer_token.source_position, "invalid number literal '0x%S'",
+ &lexer_token.literal);
+ lexer_token.literal.begin = "0";
+ lexer_token.literal.size = 1;
}
- while(c >= '0' && c <= '9') {
- expo = 10 * expo + (c - '0');
- next_char();
- }
+ parse_number_suffix();
+}
- while(1) {
- if(expo & 1)
- value *= factor;
- expo >>= 1;
- if(expo == 0)
- break;
- factor *= factor;
+/**
+ * Returns true if the given char is a octal digit.
+ *
+ * @param char the character to check
+ */
+static bool is_octal_digit(utf32 chr)
+{
+ switch (chr) {
+ case '0':
+ case '1':
+ case '2':
+ case '3':
+ case '4':
+ case '5':
+ case '6':
+ case '7':
+ return true;
+ default:
+ return false;
}
-
- lexer_token.type = T_FLOATINGPOINT;
- lexer_token.v.floatvalue = value;
-
- parse_floating_suffix();
}
-static void parse_floatingpoint_fract(int integer_part)
+/**
+ * Parses a number and sets the lexer_token.
+ */
+static void parse_number(void)
{
- long double value = integer_part;
- long double factor = 1.;
+ bool is_float = false;
+ bool has_digits = false;
- while(c >= '0' && c <= '9') {
- factor *= 0.1;
- value += (c - '0') * factor;
+ assert(obstack_object_size(&symbol_obstack) == 0);
+ if (c == '0') {
next_char();
+ if (c == 'x' || c == 'X') {
+ next_char();
+ parse_number_hex();
+ return;
+ } else {
+ has_digits = true;
+ }
+ obstack_1grow(&symbol_obstack, '0');
}
- if(c == 'e' || c == 'E') {
+ while (isdigit(c)) {
+ has_digits = true;
+ obstack_1grow(&symbol_obstack, (char) c);
next_char();
- parse_floatingpoint_exponent(value);
- return;
}
- lexer_token.type = T_FLOATINGPOINT;
- lexer_token.v.floatvalue = value;
+ if (c == '.') {
+ is_float = true;
+ obstack_1grow(&symbol_obstack, '.');
+ next_char();
- parse_floating_suffix();
-}
+ while (isdigit(c)) {
+ has_digits = true;
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
+ }
+ }
+ if (c == 'e' || c == 'E') {
+ is_float = true;
+ obstack_1grow(&symbol_obstack, 'e');
+ next_char();
-static void parse_number_dec(void)
-{
- int value = 0;
+ if (c == '-' || c == '+') {
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
+ }
- while(isdigit(c)) {
- value = 10 * value + c - '0';
- next_char();
+ while (isdigit(c)) {
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
+ }
}
- if(c == '.') {
- next_char();
- parse_floatingpoint_fract(value);
- return;
+ size_t size = obstack_object_size(&symbol_obstack);
+ char *string = obstack_finish(&symbol_obstack);
+ lexer_token.literal = identify_string(string, size);
+
+ /* is it an octal number? */
+ if (is_float) {
+ lexer_token.type = T_FLOATINGPOINT;
+ } else if (string[0] == '0') {
+ lexer_token.type = T_INTEGER_OCTAL;
+
+ /* check for invalid octal digits */
+ for (size_t i= 0; i < size; ++i) {
+ char t = string[i];
+ if (t == '8' || t == '9')
+ errorf(&lexer_token.source_position,
+ "invalid digit '%c' in octal number", t);
+ }
+ } else {
+ lexer_token.type = T_INTEGER;
}
- if(c == 'e' || c == 'E') {
- next_char();
- parse_floatingpoint_exponent(value);
- return;
+
+ if (!has_digits) {
+ errorf(&lexer_token.source_position, "invalid number literal '%S'",
+ &lexer_token.literal);
}
- parse_integer_suffix();
- lexer_token.type = T_INTEGER;
- lexer_token.v.intvalue = value;
+ parse_number_suffix();
}
-static void parse_number(void)
+/**
+ * Returns the value of a digit.
+ * The only portable way to do it ...
+ */
+static int digit_value(utf32 const digit)
{
- if (c == '0') {
- next_char();
- switch (c) {
- case 'X':
- case 'x':
- parse_number_hex();
- break;
- case '0':
- case '1':
- case '2':
- case '3':
- case '4':
- case '5':
- case '6':
- case '7':
- parse_number_oct();
- break;
- case '.':
- next_char();
- parse_floatingpoint_fract(0);
- break;
- case 'e':
- case 'E':
- parse_floatingpoint_exponent(0);
- break;
- case '8':
- case '9':
- next_char();
- parse_error("invalid octal number");
- lexer_token.type = T_ERROR;
- return;
- default:
- put_back(c);
- c = '0';
- parse_number_dec();
- return;
- }
- } else {
- parse_number_dec();
+ 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:
+ internal_error("wrong character given");
}
}
-static int parse_octal_sequence(void)
+/**
+ * Parses an octal character sequence.
+ *
+ * @param first_digit the already read first digit
+ */
+static utf32 parse_octal_sequence(utf32 const first_digit)
{
- int value = 0;
- while(1) {
- if(c < '0' || c > '7')
- break;
- value = 8 * value + c - '0';
- next_char();
- }
-
+ assert(is_octal_digit(first_digit));
+ utf32 value = digit_value(first_digit);
+ if (!is_octal_digit(c)) return value;
+ value = 8 * value + digit_value(c);
+ next_char();
+ if (!is_octal_digit(c)) return value;
+ value = 8 * value + digit_value(c);
+ next_char();
return value;
}
-static int parse_hex_sequence(void)
+/**
+ * Parses a hex character sequence.
+ */
+static utf32 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;
- }
+ utf32 value = 0;
+ while (isxdigit(c)) {
+ value = 16 * value + digit_value(c);
next_char();
}
-
return value;
}
-static int parse_escape_sequence(void)
+/**
+ * Parse an escape sequence.
+ */
+static utf32 parse_escape_sequence(void)
{
eat('\\');
- int ec = c;
+ utf32 const ec = c;
next_char();
- switch(ec) {
+ switch (ec) {
case '"': return '"';
- case '\'': return'\'';
+ case '\'': return '\'';
case '\\': return '\\';
case '?': return '\?';
case 'a': return '\a';
case '5':
case '6':
case '7':
- return parse_octal_sequence();
+ return parse_octal_sequence(ec);
case EOF:
parse_error("reached end of file while parsing escape sequence");
return EOF;
- default:
- parse_error("unknown escape sequence");
+ /* \E is not documented, but handled, by GCC. It is acceptable according
+ * to §6.11.4, whereas \e is not. */
+ case 'E':
+ case 'e':
+ if (c_mode & _GNUC)
+ return 27; /* hopefully 27 is ALWAYS the code for ESCAPE */
+ break;
+ case 'u':
+ case 'U':
+ parse_error("universal character parsing not implemented yet");
return EOF;
+ default:
+ break;
}
+ /* §6.4.4.4:8 footnote 64 */
+ parse_error("unknown escape sequence");
+ return EOF;
}
-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);
- const char *result = strset_insert(&stringset, concat);
- if(result != concat) {
- obstack_free(&symbol_obstack, concat);
- }
+ return identify_string(concat, len1 + len2 + 1);
+}
+
+string_t make_string(const char *string)
+{
+ size_t len = strlen(string) + 1;
+ char *const space = obstack_alloc(&symbol_obstack, len);
+ memcpy(space, string, len);
- return result;
+ return identify_string(space, len);
}
+static void grow_symbol(utf32 const tc)
+{
+ struct obstack *const o = &symbol_obstack;
+ if (tc < 0x80U) {
+ obstack_1grow(o, tc);
+ } else if (tc < 0x800) {
+ obstack_1grow(o, 0xC0 | (tc >> 6));
+ obstack_1grow(o, 0x80 | (tc & 0x3F));
+ } else if (tc < 0x10000) {
+ obstack_1grow(o, 0xE0 | ( tc >> 12));
+ obstack_1grow(o, 0x80 | ((tc >> 6) & 0x3F));
+ obstack_1grow(o, 0x80 | ( tc & 0x3F));
+ } else {
+ obstack_1grow(o, 0xF0 | ( tc >> 18));
+ obstack_1grow(o, 0x80 | ((tc >> 12) & 0x3F));
+ obstack_1grow(o, 0x80 | ((tc >> 6) & 0x3F));
+ obstack_1grow(o, 0x80 | ( tc & 0x3F));
+ }
+}
+
+/**
+ * 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) {
- switch(c) {
- case '\\':
- tc = parse_escape_sequence();
+ while (true) {
+ switch (c) {
+ case '\\': {
+ utf32 const tc = parse_escape_sequence();
+ if (tc >= 0x100) {
+ warningf(&lexer_token.source_position,
+ "escape sequence out of range");
+ }
obstack_1grow(&symbol_obstack, 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();
goto end_of_string;
default:
- obstack_1grow(&symbol_obstack, c);
+ grow_symbol(c);
next_char();
break;
}
/* 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);
+ char *string = obstack_finish(&symbol_obstack);
- /* check if there is already a copy of the string */
- result = strset_insert(&stringset, string);
- if(result != string) {
- obstack_free(&symbol_obstack, string);
+ lexer_token.type = T_STRING_LITERAL;
+ lexer_token.literal = identify_string(string, size);
+}
+
+/**
+ * Parse a wide character constant and set lexer_token.
+ */
+static void parse_wide_character_constant(void)
+{
+ const unsigned start_linenr = lexer_token.source_position.linenr;
+
+ eat('\'');
+
+ while (true) {
+ switch (c) {
+ case '\\': {
+ const utf32 tc = parse_escape_sequence();
+ grow_symbol(tc);
+ break;
+ }
+
+ MATCH_NEWLINE(
+ parse_error("newline while parsing character constant");
+ break;
+ )
+
+ case '\'':
+ next_char();
+ goto end_of_wide_char_constant;
+
+ case EOF: {
+ source_position_t source_position = lexer_token.source_position;
+ source_position.linenr = start_linenr;
+ errorf(&source_position, "EOF while parsing character constant");
+ lexer_token.type = T_ERROR;
+ return;
+ }
+
+ default:
+ grow_symbol(c);
+ next_char();
+ break;
+ }
}
- lexer_token.type = T_STRING_LITERAL;
- lexer_token.v.string = result;
+end_of_wide_char_constant:;
+ size_t size = (size_t) obstack_object_size(&symbol_obstack);
+ char *string = obstack_finish(&symbol_obstack);
+
+ lexer_token.type = T_WIDE_CHARACTER_CONSTANT;
+ lexer_token.literal = identify_string(string, size);
}
+/**
+ * Parse a wide string literal and set lexer_token.
+ */
+static void parse_wide_string_literal(void)
+{
+ parse_string_literal();
+ if (lexer_token.type == T_STRING_LITERAL)
+ lexer_token.type = T_WIDE_STRING_LITERAL;
+}
+
+/**
+ * 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;
- while(1) {
- switch(c) {
- case '\\':
- found_char = parse_escape_sequence();
+ while (true) {
+ switch (c) {
+ case '\\': {
+ utf32 const tc = parse_escape_sequence();
+ if (tc >= 0x100) {
+ warningf(&lexer_token.source_position,
+ "escape sequence out of range");
+ }
+ obstack_1grow(&symbol_obstack, tc);
break;
+ }
MATCH_NEWLINE(
parse_error("newline while parsing character constant");
next_char();
goto end_of_char_constant;
- case EOF:
- parse_error("EOF while parsing character constant");
+ 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;
+ }
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();
- }
+ grow_symbol(c);
+ next_char();
break;
+
}
}
-end_of_char_constant:
- lexer_token.type = T_INTEGER;
- lexer_token.v.intvalue = found_char;
+end_of_char_constant:;
+ const size_t size = (size_t)obstack_object_size(&symbol_obstack);
+ char *const string = obstack_finish(&symbol_obstack);
+
+ lexer_token.type = T_CHARACTER_CONSTANT;
+ lexer_token.literal = identify_string(string, size);
}
+/**
+ * Skip a multiline comment.
+ */
static void skip_multiline_comment(void)
{
unsigned start_linenr = lexer_token.source_position.linenr;
- while(1) {
- switch(c) {
+ while (true) {
+ switch (c) {
+ case '/':
+ next_char();
+ if (c == '*') {
+ /* nested comment, warn here */
+ if (warning.comment) {
+ warningf(&lexer_token.source_position, "'/*' within comment");
+ }
+ }
+ break;
case '*':
next_char();
- if(c == '/') {
+ if (c == '/') {
next_char();
return;
}
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) {
- switch(c) {
+ while (true) {
+ switch (c) {
case EOF:
return;
case '\r':
return;
+ case '\\':
+ next_char();
+ if (c == '\n' || c == '\r') {
+ if (warning.comment)
+ warningf(&lexer_token.source_position, "multi-line comment");
+ return;
+ }
+ break;
+
default:
next_char();
break;
}
}
+/** 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) {
+ while (pp_token.type != '\n' && pp_token.type != T_EOF) {
next_pp_token();
}
}
-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();
- if(lexer_token.type != T_IDENTIFIER) {
+ if (lexer_token.type != T_IDENTIFIER) {
parse_error("expected identifier after #define\n");
eat_until_newline();
}
}
+/**
+ * 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) {
+ if (pp_token.type != T_INTEGER) {
parse_error("expected integer");
} else {
- lexer_token.source_position.linenr = pp_token.v.intvalue - 1;
+ lexer_token.source_position.linenr = atoi(pp_token.literal.begin);
next_pp_token();
}
- if(pp_token.type == T_STRING_LITERAL) {
- lexer_token.source_position.input_name = pp_token.v.string;
+ if (pp_token.type == T_STRING_LITERAL) {
+ lexer_token.source_position.input_name = pp_token.literal.begin;
next_pp_token();
}
eat_until_newline();
}
+/**
+ * STDC pragmas.
+ */
+typedef enum stdc_pragma_kind_t {
+ STDC_UNKNOWN,
+ STDC_FP_CONTRACT,
+ STDC_FENV_ACCESS,
+ STDC_CX_LIMITED_RANGE
+} stdc_pragma_kind_t;
+
+/**
+ * STDC pragma values.
+ */
+typedef enum stdc_pragma_value_kind_t {
+ 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.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.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.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);
- symbol_t *symbol = pp_token.v.symbol;
+ symbol_t *symbol = pp_token.symbol;
- switch(symbol->pp_ID) {
+ switch (symbol->pp_ID) {
case TP_include:
printf("include - enable header name parsing!\n");
break;
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:
+ parse_pragma();
break;
}
}
+/**
+ * Parse a preprocessor directive.
+ */
static void parse_preprocessor_directive(void)
{
next_pp_token();
- switch(pp_token.type) {
+ switch (pp_token.type) {
case T_IDENTIFIER:
parse_preprocessor_identifier();
break;
case T_INTEGER:
parse_line_directive();
break;
+ case '\n':
+ /* NULL directive, see §6.10.7 */
+ break;
default:
parse_error("invalid preprocessor directive");
eat_until_newline();
#define MAYBE_PROLOG \
next_char(); \
- while(1) { \
- switch(c) {
+ while (true) { \
+ switch (c) {
#define MAYBE(ch, set_type) \
case ch: \
lexer_token.type = set_type; \
return;
+/* must use this as last thing */
+#define MAYBE_MODE(ch, set_type, mode) \
+ case ch: \
+ if (c_mode & mode) { \
+ next_char(); \
+ lexer_token.type = set_type; \
+ return; \
+ } \
+ /* fallthrough */
+
#define ELSE_CODE(code) \
default: \
- code; \
+ code \
} \
- } /* end of while(1) */ \
+ } /* end of while (true) */ \
break;
#define ELSE(set_type) \
void lexer_next_preprocessing_token(void)
{
- while(1) {
- switch(c) {
+ while (true) {
+ switch (c) {
case ' ':
case '\t':
next_char();
SYMBOL_CHARS
parse_symbol();
/* might be a wide string ( L"string" ) */
- if(c == '"' && (lexer_token.type == T_IDENTIFIER &&
- lexer_token.v.symbol == symbol_L)) {
- parse_string_literal();
- return;
+ if (lexer_token.symbol == symbol_L) {
+ switch (c) {
+ case '"': parse_wide_string_literal(); break;
+ case '\'': parse_wide_character_constant(); break;
+ }
}
return;
case '.':
MAYBE_PROLOG
+ DIGITS
+ put_back(c);
+ c = '.';
+ parse_number();
+ return;
+
case '.':
MAYBE_PROLOG
MAYBE('.', T_DOTDOTDOT)
ELSE('/')
case '%':
MAYBE_PROLOG
- MAYBE('>', T_PERCENTGREATER)
+ MAYBE('>', '}')
MAYBE('=', T_PERCENTEQUAL)
case ':':
MAYBE_PROLOG
case '%':
MAYBE_PROLOG
- MAYBE(':', T_PERCENTCOLONPERCENTCOLON)
+ MAYBE(':', T_HASHHASH)
ELSE_CODE(
put_back(c);
c = '%';
- lexer_token.type = T_PERCENTCOLON;
+ lexer_token.type = '#';
return;
)
- ELSE(T_PERCENTCOLON)
+ ELSE('#')
ELSE('%')
case '<':
MAYBE_PROLOG
- MAYBE(':', T_LESSCOLON)
- MAYBE('%', T_LESSPERCENT)
+ MAYBE(':', '[')
+ MAYBE('%', '{')
MAYBE('=', T_LESSEQUAL)
case '<':
MAYBE_PROLOG
ELSE('|')
case ':':
MAYBE_PROLOG
- MAYBE('>', T_COLONGREATER)
+ MAYBE('>', ']')
+ MAYBE_MODE(':', T_COLONCOLON, _CXX)
ELSE(':')
case '=':
MAYBE_PROLOG
return;
default:
+dollar_sign:
+ errorf(&lexer_token.source_position, "unknown character '%c' found", c);
next_char();
- error_prefix();
- fprintf(stderr, "unknown character '%c' found\n", c);
lexer_token.type = T_ERROR;
return;
}
void lexer_next_token(void)
{
lexer_next_preprocessing_token();
- if(lexer_token.type != '\n')
- return;
+ while (lexer_token.type == '\n') {
newline_found:
- do {
lexer_next_preprocessing_token();
- } while(lexer_token.type == '\n');
+ }
- if(lexer_token.type == '#') {
+ if (lexer_token.type == '#') {
parse_preprocessor_directive();
goto newline_found;
}
void init_lexer(void)
{
strset_init(&stringset);
+ symbol_L = symbol_table_insert("L");
}
void lexer_open_stream(FILE *stream, const char *input_name)
lexer_token.source_position.linenr = 0;
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 */
+ c = '\n';
+}
+
+void lexer_open_buffer(const char *buffer, size_t len, const char *input_name)
+{
+ input = NULL;
+ lexer_token.source_position.linenr = 0;
+ lexer_token.source_position.input_name = input_name;
+
+#if 0 // TODO
+ bufpos = buffer;
+ bufend = buffer + len;
+#else
+ (void)buffer;
+ (void)len;
+ panic("builtin lexing not done yet");
+#endif
/* place a virtual \n at the beginning so the lexer knows that we're
* at the beginning of a line */
static __attribute__((unused))
void dbg_pos(const source_position_t source_position)
{
- fprintf(stdout, "%s:%d\n", source_position.input_name,
+ fprintf(stdout, "%s:%u\n", source_position.input_name,
source_position.linenr);
fflush(stdout);
}