typedef token_kind_t as int
[cparser] / preprocessor.c
index 26bc3ef..af95d90 100644 (file)
@@ -99,7 +99,7 @@ static const char       *printed_input_name = NULL;
 static source_position_t expansion_pos;
 static pp_definition_t  *current_expansion  = NULL;
 static strset_t          stringset;
-static preprocessor_token_kind_t last_token;
+static token_kind_t      last_token;
 
 static searchpath_entry_t *searchpath;
 
@@ -218,17 +218,18 @@ static inline void put_back(utf32 const pc)
        --input.position.colno;
 }
 
-#define MATCH_NEWLINE(code)                   \
-       case '\r':                                \
-               next_char();                          \
-               if (input.c == '\n') {                \
-       case '\n':                                \
-                       next_char();                      \
-               }                                     \
-               info.whitespace = 0;                  \
-               ++input.position.lineno;              \
-               input.position.colno = 1;             \
-               code
+#define NEWLINE \
+       '\r': \
+               next_char(); \
+               if (input.c == '\n') { \
+       case '\n': \
+                       next_char(); \
+               } \
+               info.whitespace = 0; \
+               ++input.position.lineno; \
+               input.position.colno = 1; \
+               goto newline; \
+               newline // Let it look like an ordinary case label.
 
 #define eat(c_type) (assert(input.c == c_type), next_char())
 
@@ -237,9 +238,8 @@ static void maybe_concat_lines(void)
        eat('\\');
 
        switch (input.c) {
-       MATCH_NEWLINE(
+       case NEWLINE:
                return;
-       )
 
        default:
                break;
@@ -451,13 +451,18 @@ static const char *identify_string(char *string)
        return result;
 }
 
-static string_t make_string(char *string, size_t len)
+static string_t sym_make_string(string_encoding_t const enc)
 {
-       const char *result = identify_string(string);
-       return (string_t) {result, len};
+       obstack_1grow(&symbol_obstack, '\0');
+       size_t      const len    = obstack_object_size(&symbol_obstack) - 1;
+       char       *const string = obstack_finish(&symbol_obstack);
+       char const *const result = identify_string(string);
+       return (string_t){ result, len, enc };
 }
 
-static void parse_string(utf32 const delimiter, preprocessor_token_kind_t const kind, string_encoding_t const enc, char const *const context)
+static void parse_string(utf32 const delimiter, token_kind_t const kind,
+                         string_encoding_t const enc,
+                         char const *const context)
 {
        const unsigned start_linenr = input.position.lineno;
 
@@ -485,10 +490,9 @@ static void parse_string(utf32 const delimiter, preprocessor_token_kind_t const
                        break;
                }
 
-               MATCH_NEWLINE(
+               case NEWLINE:
                        errorf(&pp_token.base.source_position, "newline while parsing %s", context);
                        break;
-               )
 
                case EOF: {
                        source_position_t source_position;
@@ -510,14 +514,9 @@ static void parse_string(utf32 const delimiter, preprocessor_token_kind_t const
                }
        }
 
-end_of_string:;
-       obstack_1grow(&symbol_obstack, '\0');
-       size_t const size   = obstack_object_size(&symbol_obstack) - 1;
-       char  *const string = obstack_finish(&symbol_obstack);
-
-       pp_token.kind            = kind;
-       pp_token.string.encoding = enc;
-       pp_token.string.string   = make_string(string, size);
+end_of_string:
+       pp_token.kind           = kind;
+       pp_token.literal.string = sym_make_string(enc);
 }
 
 static void parse_string_literal(string_encoding_t const enc)
@@ -528,13 +527,13 @@ static void parse_string_literal(string_encoding_t const enc)
 static void parse_character_constant(string_encoding_t const enc)
 {
        parse_string('\'', TP_CHARACTER_CONSTANT, enc, "character constant");
-       if (pp_token.string.string.size == 0) {
+       if (pp_token.literal.string.size == 0) {
                parse_error("empty character constant");
        }
 }
 
-#define SYMBOL_CHARS_WITHOUT_E_P \
-       case 'a': \
+#define SYMBOL_CASES_WITHOUT_E_P \
+            'a': \
        case 'b': \
        case 'c': \
        case 'd': \
@@ -582,17 +581,17 @@ static void parse_character_constant(string_encoding_t const enc)
        case 'X': \
        case 'Y': \
        case 'Z': \
-       case '_':
+       case '_'
 
-#define SYMBOL_CHARS \
-       SYMBOL_CHARS_WITHOUT_E_P \
+#define SYMBOL_CASES \
+            SYMBOL_CASES_WITHOUT_E_P: \
        case 'e': \
        case 'p': \
        case 'E': \
-       case 'P':
+       case 'P'
 
-#define DIGITS \
-       case '0':  \
+#define DIGIT_CASES \
+            '0':  \
        case '1':  \
        case '2':  \
        case '3':  \
@@ -601,7 +600,7 @@ static void parse_character_constant(string_encoding_t const enc)
        case '6':  \
        case '7':  \
        case '8':  \
-       case '9':
+       case '9'
 
 /**
  * returns next final token from a preprocessor macro expansion
@@ -689,9 +688,8 @@ static void skip_multiline_comment(void)
                        }
                        break;
 
-               MATCH_NEWLINE(
+               case NEWLINE:
                        break;
-               )
 
                case EOF: {
                        source_position_t source_position;
@@ -717,10 +715,9 @@ static void skip_whitespace(void)
                        next_char();
                        continue;
 
-               MATCH_NEWLINE(
+               case NEWLINE:
                        info.at_line_begin = true;
                        return;
-               )
 
                case '/':
                        next_char();
@@ -743,10 +740,10 @@ static void skip_whitespace(void)
        }
 }
 
-static void eat_pp(preprocessor_token_kind_t const type)
+static inline void eat_pp(token_kind_t const kind)
 {
-       (void) type;
-       assert(pp_token.kind == type);
+       assert(pp_token.kind == kind);
+       (void) kind;
        next_preprocessing_token();
 }
 
@@ -757,8 +754,8 @@ static void parse_symbol(void)
 
        while (true) {
                switch (input.c) {
-               DIGITS
-               SYMBOL_CHARS
+               case DIGIT_CASES:
+               case SYMBOL_CASES:
                        obstack_1grow(&symbol_obstack, (char) input.c);
                        next_char();
                        break;
@@ -803,8 +800,8 @@ static void parse_number(void)
        while (true) {
                switch (input.c) {
                case '.':
-               DIGITS
-               SYMBOL_CHARS_WITHOUT_E_P
+               case DIGIT_CASES:
+               case SYMBOL_CASES_WITHOUT_E_P:
                        obstack_1grow(&symbol_obstack, (char) input.c);
                        next_char();
                        break;
@@ -827,12 +824,8 @@ static void parse_number(void)
        }
 
 end_number:
-       obstack_1grow(&symbol_obstack, '\0');
-       size_t  size   = obstack_object_size(&symbol_obstack);
-       char   *string = obstack_finish(&symbol_obstack);
-
-       pp_token.kind          = TP_NUMBER;
-       pp_token.number.number = make_string(string, size);
+       pp_token.kind           = TP_NUMBER;
+       pp_token.literal.string = sym_make_string(STRING_ENCODING_CHAR);
 }
 
 
@@ -880,17 +873,16 @@ restart:
                next_char();
                goto restart;
 
-       MATCH_NEWLINE(
+       case NEWLINE:
                info.at_line_begin = true;
                info.had_whitespace = true;
                goto restart;
-       )
 
-       SYMBOL_CHARS
+       case SYMBOL_CASES:
                parse_symbol();
                return;
 
-       DIGITS
+       case DIGIT_CASES:
                parse_number();
                return;
 
@@ -1034,7 +1026,6 @@ restart:
        case '~':
        case ';':
        case ',':
-       case '\\':
                pp_token.kind = input.c;
                next_char();
                return;
@@ -1143,20 +1134,20 @@ static void emit_pp_token(void)
                fputs(pp_token.base.symbol->string, out);
                break;
        case TP_NUMBER:
-               fputs(pp_token.number.number.begin, out);
+               fputs(pp_token.literal.string.begin, out);
                break;
 
        case TP_STRING_LITERAL:
-               fputs(get_string_encoding_prefix(pp_token.string.encoding), out);
+               fputs(get_string_encoding_prefix(pp_token.literal.string.encoding), out);
                fputc('"', out);
-               fputs(pp_token.string.string.begin, out);
+               fputs(pp_token.literal.string.begin, out);
                fputc('"', out);
                break;
 
        case TP_CHARACTER_CONSTANT:
-               fputs(get_string_encoding_prefix(pp_token.string.encoding), out);
+               fputs(get_string_encoding_prefix(pp_token.literal.string.encoding), out);
                fputc('\'', out);
-               fputs(pp_token.string.string.begin, out);
+               fputs(pp_token.literal.string.begin, out);
                fputc('\'', out);
                break;
        default:
@@ -1200,7 +1191,7 @@ static bool pp_tokens_equal(const token_t *token1, const token_t *token2)
        case TP_NUMBER:
        case TP_CHARACTER_CONSTANT:
        case TP_STRING_LITERAL:
-               return strings_equal(&token1->string.string, &token2->string.string);
+               return strings_equal(&token1->literal.string, &token2->literal.string);
 
        default:
                return true;
@@ -1350,16 +1341,16 @@ static void parse_undef_directive(void)
        eat_pp_directive();
 }
 
+/** behind an #include we can have the special headername lexems.
+ * They're only allowed behind an #include so they're not recognized
+ * by the normal next_preprocessing_token. We handle them as a special
+ * exception here */
 static void parse_headername(void)
 {
        const source_position_t start_position = input.position;
-       string_t                string         = {NULL, 0};
+       string_t                string         = { NULL, 0, STRING_ENCODING_CHAR };
        assert(obstack_object_size(&symbol_obstack) == 0);
 
-       /* behind an #include we can have the special headername lexems.
-        * They're only allowed behind an #include so they're not recognized
-        * by the normal next_preprocessing_token. We handle them as a special
-        * exception here */
        if (info.at_line_begin) {
                parse_error("expected headername after #include");
                goto finish_error;
@@ -1375,12 +1366,10 @@ parse_name:
                next_char();
                while (true) {
                        switch (input.c) {
+                       case NEWLINE:
                        case EOF:
-                               /* fallthrough */
-                       MATCH_NEWLINE(
                                errorf(&pp_token.base.source_position, "header name without closing '%c'", (char)delimiter);
                                goto finish_error;
-                       )
 
                        default:
                                if (input.c == delimiter) {
@@ -1402,15 +1391,12 @@ parse_name:
        }
 
 finished_headername:
-       obstack_1grow(&symbol_obstack, '\0');
-       const size_t size       = (size_t)obstack_object_size(&symbol_obstack);
-       char *const  headername = obstack_finish(&symbol_obstack);
-       string                  = make_string(headername, size);
+       string = sym_make_string(STRING_ENCODING_CHAR);
 
 finish_error:
        pp_token.base.source_position = start_position;
        pp_token.kind                 = TP_HEADERNAME;
-       pp_token.string.string        = string;
+       pp_token.literal.string       = string;
 }
 
 static bool do_include(bool system_include, const char *headername)
@@ -1467,9 +1453,7 @@ static void skip_till_newline(void)
        /* skip till newline */
        while (true) {
                switch (input.c) {
-               MATCH_NEWLINE(
-                       return;
-               )
+               case NEWLINE:
                case EOF:
                        return;
                }
@@ -1484,7 +1468,7 @@ static bool parse_include_directive(void)
        skip_whitespace();
        bool system_include = input.c == '<';
        parse_headername();
-       string_t headername = pp_token.string.string;
+       string_t headername = pp_token.literal.string;
        if (headername.begin == NULL) {
                eat_pp_directive();
                return false;
@@ -1507,10 +1491,9 @@ static bool parse_include_directive(void)
        /* switch inputs */
        emit_newlines();
        push_input();
-       bool res = do_include(system_include, pp_token.string.string.begin);
+       bool res = do_include(system_include, pp_token.literal.string.begin);
        if (!res) {
-               errorf(&pp_token.base.source_position,
-                      "failed including '%S': %s", pp_token.string, strerror(errno));
+               errorf(&pp_token.base.source_position, "failed including '%S': %s", &pp_token.literal, strerror(errno));
                pop_restore_input();
                return false;
        }
@@ -1574,7 +1557,7 @@ static void parse_ifdef_ifndef_directive(void)
                condition = true;
        } else {
                /* evaluate wether we are in true or false case */
-               condition = !pp_token.base.symbol->pp_definition == is_ifndef;
+               condition = (bool)!pp_token.base.symbol->pp_definition == is_ifndef;
 
                next_preprocessing_token();
 
@@ -1615,7 +1598,7 @@ static void parse_else_directive(void)
        if (conditional->in_else) {
                errorf(&pp_token.base.source_position,
                       "#else after #else (condition started %P)",
-                      conditional->source_position);
+                      &conditional->source_position);
                skip_mode = true;
                return;
        }