From: TimePath Date: Sat, 2 Sep 2017 12:38:52 +0000 (+1000) Subject: Classy Tokens X-Git-Url: https://git.xonotic.org/?a=commitdiff_plain;h=efbf047ade4eec013ed1a8f1823189cc3d3c9565;p=xonotic%2Fgmqcc.git Classy Tokens --- diff --git a/ftepp.cpp b/ftepp.cpp index 190978f..a85d857 100644 --- a/ftepp.cpp +++ b/ftepp.cpp @@ -14,7 +14,7 @@ struct ppcondition { }; struct pptoken { - int token; + Token token; char *value; /* a copy from the lexer */ union { @@ -37,7 +37,7 @@ struct ppmacro { struct ftepp_t { lex_file *lex; - int token; + Token token; unsigned int errors; bool output_on; ppcondition *conditions; @@ -322,7 +322,7 @@ static GMQCC_INLINE void ftepp_macro_delete(ftepp_t *ftepp, const char *name) util_htrm(ftepp->macros, name, (void (*)(void*))&ppmacro_delete); } -static GMQCC_INLINE int ftepp_next(ftepp_t *ftepp) +static GMQCC_INLINE Token ftepp_next(ftepp_t *ftepp) { return (ftepp->token = lex_do(ftepp->lex)); } @@ -330,10 +330,10 @@ static GMQCC_INLINE int ftepp_next(ftepp_t *ftepp) /* Important: this does not skip newlines! */ static bool ftepp_skipspace(ftepp_t *ftepp) { - if (ftepp->token != TOKEN_WHITE) + if (ftepp->token != Token::WHITE) return true; - while (ftepp_next(ftepp) == TOKEN_WHITE) {} - if (ftepp->token >= TOKEN_EOF) { + while (ftepp_next(ftepp) == Token::WHITE) {} + if (ftepp->token >= Token::END) { ftepp_error(ftepp, "unexpected end of preprocessor directive"); return false; } @@ -343,12 +343,12 @@ static bool ftepp_skipspace(ftepp_t *ftepp) /* this one skips EOLs as well */ static bool ftepp_skipallwhite(ftepp_t *ftepp) { - if (ftepp->token != TOKEN_WHITE && ftepp->token != TOKEN_EOL) + if (ftepp->token != Token::WHITE && ftepp->token != Token::EOL) return true; do { ftepp_next(ftepp); - } while (ftepp->token == TOKEN_WHITE || ftepp->token == TOKEN_EOL); - if (ftepp->token >= TOKEN_EOF) { + } while (ftepp->token == Token::WHITE || ftepp->token == Token::EOL); + if (ftepp->token >= Token::END) { ftepp_error(ftepp, "unexpected end of preprocessor directive"); return false; } @@ -364,15 +364,15 @@ static bool ftepp_define_params(ftepp_t *ftepp, ppmacro *macro) ftepp_next(ftepp); if (!ftepp_skipspace(ftepp)) return false; - if (ftepp->token == ')') + if (ftepp->token == Token::PAREN_CLOSE) break; switch (ftepp->token) { - case TOKEN_IDENT: - case TOKEN_TYPENAME: - case TOKEN_KEYWORD: + case Token::IDENT: + case Token::TYPENAME: + case Token::KEYWORD: vec_push(macro->params, util_strdup(ftepp_tokval(ftepp))); break; - case TOKEN_DOTS: + case Token::DOTS: macro->variadic = true; break; default: @@ -382,13 +382,13 @@ static bool ftepp_define_params(ftepp_t *ftepp, ppmacro *macro) ftepp_next(ftepp); if (!ftepp_skipspace(ftepp)) return false; - if (macro->variadic && ftepp->token != ')') { + if (macro->variadic && ftepp->token != Token::PAREN_CLOSE) { ftepp_error(ftepp, "cannot have parameters after the variadic parameters"); return false; } - } while (ftepp->token == ','); + } while (ftepp->token == Token::COMMA); - if (ftepp->token != ')') { + if (ftepp->token != Token::PAREN_CLOSE) { ftepp_error(ftepp, "expected closing paren after macro parameter list"); return false; } @@ -400,25 +400,25 @@ static bool ftepp_define_params(ftepp_t *ftepp, ppmacro *macro) static bool ftepp_define_body(ftepp_t *ftepp, ppmacro *macro) { pptoken *ptok; - while (ftepp->token != TOKEN_EOL && ftepp->token < TOKEN_EOF) { + while (ftepp->token != Token::EOL && ftepp->token < Token::END) { bool subscript = false; size_t index = 0; if (macro->variadic && !strcmp(ftepp_tokval(ftepp), "__VA_ARGS__")) { - subscript = !!(ftepp_next(ftepp) == '#'); + subscript = !!(ftepp_next(ftepp) == Token::HASH); - if (subscript && ftepp_next(ftepp) != '#') { + if (subscript && ftepp_next(ftepp) != Token::HASH) { ftepp_error(ftepp, "expected `##` in __VA_ARGS__ for subscripting"); return false; } else if (subscript) { - if (ftepp_next(ftepp) == '[') { - if (ftepp_next(ftepp) != TOKEN_INTCONST) { + if (ftepp_next(ftepp) == Token::BRACKET_OPEN) { + if (ftepp_next(ftepp) != Token::INTCONST) { ftepp_error(ftepp, "expected index for __VA_ARGS__ subscript"); return false; } index = (int)strtol(ftepp_tokval(ftepp), nullptr, 10); - if (ftepp_next(ftepp) != ']') { + if (ftepp_next(ftepp) != Token::BRACKET_CLOSE) { ftepp_error(ftepp, "expected `]` in __VA_ARGS__ subscript"); return false; } @@ -427,7 +427,7 @@ static bool ftepp_define_body(ftepp_t *ftepp, ppmacro *macro) * mark it as an array to be handled later as such and not * as traditional __VA_ARGS__ */ - ftepp->token = TOKEN_VA_ARGS_ARRAY; + ftepp->token = Token::VA_ARGS_ARRAY; ptok = pptoken_make(ftepp); ptok->constval.i = index; vec_push(macro->output, ptok); @@ -437,15 +437,15 @@ static bool ftepp_define_body(ftepp_t *ftepp, ppmacro *macro) return false; } } else { - int old = ftepp->token; - ftepp->token = TOKEN_VA_ARGS; + auto old = ftepp->token; + ftepp->token = Token::VA_ARGS; ptok = pptoken_make(ftepp); vec_push(macro->output, ptok); ftepp->token = old; } } else if (macro->variadic && !strcmp(ftepp_tokval(ftepp), "__VA_COUNT__")) { - ftepp->token = TOKEN_VA_COUNT; + ftepp->token = Token::VA_COUNT; ptok = pptoken_make(ftepp); vec_push(macro->output, ptok); ftepp_next(ftepp); @@ -456,7 +456,7 @@ static bool ftepp_define_body(ftepp_t *ftepp, ppmacro *macro) } } /* recursive expansion can cause EOFs here */ - if (ftepp->token != TOKEN_EOL && ftepp->token != TOKEN_EOF) { + if (ftepp->token != Token::EOL && ftepp->token != Token::END) { ftepp_error(ftepp, "unexpected junk after macro or unexpected end of file"); return false; } @@ -492,9 +492,9 @@ static bool ftepp_define(ftepp_t *ftepp) return false; switch (ftepp->token) { - case TOKEN_IDENT: - case TOKEN_TYPENAME: - case TOKEN_KEYWORD: + case Token::IDENT: + case Token::TYPENAME: + case Token::KEYWORD: if (OPTS_FLAG(FTEPP_MATHDEFS)) { for (i = 0; i < GMQCC_ARRAY_COUNT(ftepp_math_constants); i++) { if (!strcmp(ftepp_math_constants[i][0], ftepp_tokval(ftepp))) { @@ -528,7 +528,7 @@ static bool ftepp_define(ftepp_t *ftepp) (void)ftepp_next(ftepp); - if (ftepp->token == '(') { + if (ftepp->token == Token::PAREN_OPEN) { macro->has_params = true; if (!ftepp_define_params(ftepp, macro)) { ppmacro_delete(macro); @@ -588,34 +588,34 @@ static bool ftepp_macro_call_params(ftepp_t *ftepp, macroparam **out_params) if (!ftepp_skipallwhite(ftepp)) return false; - while (ftepp->token != ')') { + while (ftepp->token != Token::PAREN_CLOSE) { mp.tokens = nullptr; if (!ftepp_skipallwhite(ftepp)) return false; - while (parens || ftepp->token != ',') { - if (ftepp->token == '(') + while (parens || ftepp->token != Token::COMMA) { + if (ftepp->token == Token::PAREN_OPEN) ++parens; - else if (ftepp->token == ')') { + else if (ftepp->token == Token::PAREN_CLOSE) { if (!parens) break; --parens; } ptok = pptoken_make(ftepp); vec_push(mp.tokens, ptok); - if (ftepp_next(ftepp) >= TOKEN_EOF) { + if (ftepp_next(ftepp) >= Token::END) { ftepp_error(ftepp, "unexpected end of file in macro call"); goto on_error; } } vec_push(params, mp); mp.tokens = nullptr; - if (ftepp->token == ')') + if (ftepp->token == Token::PAREN_CLOSE) break; - if (ftepp->token != ',') { + if (ftepp->token != Token::COMMA) { ftepp_error(ftepp, "expected closing paren or comma in macro call"); goto on_error; } - if (ftepp_next(ftepp) >= TOKEN_EOF) { + if (ftepp_next(ftepp) >= Token::END) { ftepp_error(ftepp, "unexpected end of file in macro call"); goto on_error; } @@ -650,7 +650,7 @@ static void ftepp_stringify_token(ftepp_t *ftepp, pptoken *token) const char *ch; chs[1] = 0; switch (token->token) { - case TOKEN_STRINGCONST: + case Token::STRINGCONST: ch = token->value; while (*ch) { /* in preprocessor mode strings already are string, @@ -668,10 +668,10 @@ static void ftepp_stringify_token(ftepp_t *ftepp, pptoken *token) ++ch; } break; - /*case TOKEN_WHITE: + /*case Token::WHITE: ftepp_out(ftepp, " ", false); break;*/ - case TOKEN_EOL: + case Token::EOL: ftepp_out(ftepp, "\\n", false); break; default: @@ -706,7 +706,7 @@ static void ftepp_param_out(ftepp_t *ftepp, macroparam *param) pptoken *out; for (i = 0; i < vec_size(param->tokens); ++i) { out = param->tokens[i]; - if (out->token == TOKEN_EOL) + if (out->token == Token::EOL) ftepp_out(ftepp, "\n", false); else { ppmacro *find = ftepp_macro_find(ftepp, out->value); @@ -736,7 +736,7 @@ static bool ftepp_macro_expand(ftepp_t *ftepp, ppmacro *macro, macroparam *param bool old_inmacro; bool strip = false; - int nextok; + Token nextok; if (vararg_start < vec_size(params)) varargs = vec_size(params) - vararg_start; @@ -751,9 +751,9 @@ static bool ftepp_macro_expand(ftepp_t *ftepp, ppmacro *macro, macroparam *param for (o = 0; o < vec_size(macro->output); ++o) { pptoken *out = macro->output[o]; switch (out->token) { - case TOKEN_VA_ARGS: + case Token::VA_ARGS: if (!macro->variadic) { - ftepp_error(ftepp, "internal preprocessor error: TOKEN_VA_ARGS in non-variadic macro"); + ftepp_error(ftepp, "internal preprocessor error: Token::VA_ARGS in non-variadic macro"); vec_free(old_string); return false; } @@ -768,7 +768,7 @@ static bool ftepp_macro_expand(ftepp_t *ftepp, ppmacro *macro, macroparam *param } break; - case TOKEN_VA_ARGS_ARRAY: + case Token::VA_ARGS_ARRAY: if ((size_t)out->constval.i >= varargs) { ftepp_error(ftepp, "subscript of `[%u]` is out of bounds for `__VA_ARGS__`", out->constval.i); vec_free(old_string); @@ -778,33 +778,33 @@ static bool ftepp_macro_expand(ftepp_t *ftepp, ppmacro *macro, macroparam *param ftepp_param_out(ftepp, ¶ms[out->constval.i + vararg_start]); break; - case TOKEN_VA_COUNT: + case Token::VA_COUNT: util_asprintf(&buffer, "%d", varargs); ftepp_out(ftepp, buffer, false); mem_d(buffer); break; - case TOKEN_IDENT: - case TOKEN_TYPENAME: - case TOKEN_KEYWORD: + case Token::IDENT: + case Token::TYPENAME: + case Token::KEYWORD: if (!macro_params_find(macro, out->value, &pi)) { ftepp_out(ftepp, out->value, false); break; } else ftepp_param_out(ftepp, ¶ms[pi]); break; - case '#': + case Token::HASH: if (o + 1 < vec_size(macro->output)) { nextok = macro->output[o+1]->token; - if (nextok == '#') { + if (nextok == Token::HASH) { /* raw concatenation */ ++o; strip = true; break; } - if ( (nextok == TOKEN_IDENT || - nextok == TOKEN_KEYWORD || - nextok == TOKEN_TYPENAME) && + if ( (nextok == Token::IDENT || + nextok == Token::KEYWORD || + nextok == Token::TYPENAME) && macro_params_find(macro, macro->output[o+1]->value, &pi)) { ++o; @@ -815,13 +815,13 @@ static bool ftepp_macro_expand(ftepp_t *ftepp, ppmacro *macro, macroparam *param } ftepp_out(ftepp, "#", false); break; - case TOKEN_EOL: + case Token::EOL: ftepp_out(ftepp, "\n", false); break; default: buffer = out->value; #define buffer_stripable(X) ((X) == ' ' || (X) == '\t') - if (vec_size(macro->output) > o + 1 && macro->output[o+1]->token == '#' && buffer_stripable(*buffer)) + if (vec_size(macro->output) > o + 1 && macro->output[o+1]->token == Token::HASH && buffer_stripable(*buffer)) buffer++; if (strip) { while (buffer_stripable(*buffer)) buffer++; @@ -907,7 +907,7 @@ static bool ftepp_macro_call(ftepp_t *ftepp, ppmacro *macro) if (!ftepp_skipallwhite(ftepp)) return false; - if (ftepp->token != '(') { + if (ftepp->token != Token::PAREN_OPEN) { ftepp_error(ftepp, "expected macro parameters in parenthesis"); return false; } @@ -974,14 +974,14 @@ static bool ftepp_if_value(ftepp_t *ftepp, bool *out, double *value_out) if (!ftepp_skipspace(ftepp)) return false; - while (ftepp->token == '!') { + while (ftepp->token == Token::NOT) { wasnot = true; ftepp_next(ftepp); if (!ftepp_skipspace(ftepp)) return false; } - if (ftepp->token == TOKEN_OPERATOR && !strcmp(ftepp_tokval(ftepp), "-")) + if (ftepp->token == Token::OPERATOR && !strcmp(ftepp_tokval(ftepp), "-")) { wasneg = true; ftepp_next(ftepp); @@ -990,23 +990,23 @@ static bool ftepp_if_value(ftepp_t *ftepp, bool *out, double *value_out) } switch (ftepp->token) { - case TOKEN_IDENT: - case TOKEN_TYPENAME: - case TOKEN_KEYWORD: + case Token::IDENT: + case Token::TYPENAME: + case Token::KEYWORD: if (!strcmp(ftepp_tokval(ftepp), "defined")) { ftepp_next(ftepp); if (!ftepp_skipspace(ftepp)) return false; - if (ftepp->token != '(') { + if (ftepp->token != Token::PAREN_OPEN) { ftepp_error(ftepp, "`defined` keyword in #if requires a macro name in parenthesis"); return false; } ftepp_next(ftepp); if (!ftepp_skipspace(ftepp)) return false; - if (ftepp->token != TOKEN_IDENT && - ftepp->token != TOKEN_TYPENAME && - ftepp->token != TOKEN_KEYWORD) + if (ftepp->token != Token::IDENT && + ftepp->token != Token::TYPENAME && + ftepp->token != Token::KEYWORD) { ftepp_error(ftepp, "defined() used on an unexpected token type"); return false; @@ -1016,7 +1016,7 @@ static bool ftepp_if_value(ftepp_t *ftepp, bool *out, double *value_out) ftepp_next(ftepp); if (!ftepp_skipspace(ftepp)) return false; - if (ftepp->token != ')') { + if (ftepp->token != Token::PAREN_CLOSE) { ftepp_error(ftepp, "expected closing paren"); return false; } @@ -1030,11 +1030,11 @@ static bool ftepp_if_value(ftepp_t *ftepp, bool *out, double *value_out) } else { /* This does not expand recursively! */ switch (macro->output[0]->token) { - case TOKEN_INTCONST: + case Token::INTCONST: *value_out = macro->output[0]->constval.i; *out = !!(macro->output[0]->constval.i); break; - case TOKEN_FLOATCONST: + case Token::FLOATCONST: *value_out = macro->output[0]->constval.f; *out = !!(macro->output[0]->constval.f); break; @@ -1044,24 +1044,24 @@ static bool ftepp_if_value(ftepp_t *ftepp, bool *out, double *value_out) } } break; - case TOKEN_STRINGCONST: + case Token::STRINGCONST: *value_out = 0; *out = false; break; - case TOKEN_INTCONST: + case Token::INTCONST: *value_out = ftepp->lex->tok.constval.i; *out = !!(ftepp->lex->tok.constval.i); break; - case TOKEN_FLOATCONST: + case Token::FLOATCONST: *value_out = ftepp->lex->tok.constval.f; *out = !!(ftepp->lex->tok.constval.f); break; - case '(': + case Token::PAREN_OPEN: ftepp_next(ftepp); if (!ftepp_if_expr(ftepp, out, value_out)) return false; - if (ftepp->token != ')') { + if (ftepp->token != Token::PAREN_CLOSE) { ftepp_error(ftepp, "expected closing paren in #if expression"); return false; } @@ -1099,7 +1099,7 @@ static bool ftepp_if_expr(ftepp_t *ftepp, bool *out, double *value_out) if (!ftepp_if_op(ftepp)) return false; - if (ftepp->token == ')' || ftepp->token != TOKEN_OPERATOR) + if (ftepp->token == Token::PAREN_CLOSE || ftepp->token != Token::OPERATOR) return true; /* FTEQCC is all right-associative and no precedence here */ @@ -1111,7 +1111,7 @@ static bool ftepp_if_expr(ftepp_t *ftepp, bool *out, double *value_out) double nextvalue; (void)nextvalue; - if (!ftepp_next(ftepp)) + if (ftepp_next(ftepp) == Token::NONE) return false; if (!ftepp_if_expr(ftepp, &next, &nextvalue)) return false; @@ -1136,7 +1136,7 @@ static bool ftepp_if_expr(ftepp_t *ftepp, bool *out, double *value_out) const char opc1 = ftepp_tokval(ftepp)[1]; double other; - if (!ftepp_next(ftepp)) + if (ftepp_next(ftepp) == Token::NONE) return false; if (!ftepp_if_expr(ftepp, &next, &other)) return false; @@ -1173,7 +1173,7 @@ static bool ftepp_if(ftepp_t *ftepp, ppcondition *cond) if (!ftepp_skipspace(ftepp)) return false; - if (ftepp->token == TOKEN_EOL) { + if (ftepp->token == Token::EOL) { ftepp_error(ftepp, "expected expression for #if-directive"); return false; } @@ -1197,9 +1197,9 @@ static bool ftepp_ifdef(ftepp_t *ftepp, ppcondition *cond) return false; switch (ftepp->token) { - case TOKEN_IDENT: - case TOKEN_TYPENAME: - case TOKEN_KEYWORD: + case Token::IDENT: + case Token::TYPENAME: + case Token::KEYWORD: macro = ftepp_macro_find(ftepp, ftepp_tokval(ftepp)); break; default: @@ -1211,7 +1211,7 @@ static bool ftepp_ifdef(ftepp_t *ftepp, ppcondition *cond) if (!ftepp_skipspace(ftepp)) return false; /* relaxing this condition - if (ftepp->token != TOKEN_EOL && ftepp->token != TOKEN_EOF) { + if (ftepp->token != Token::EOL && ftepp->token != Token::END) { ftepp_error(ftepp, "stray tokens after #ifdef"); return false; } @@ -1231,9 +1231,9 @@ static bool ftepp_undef(ftepp_t *ftepp) if (ftepp->output_on) { switch (ftepp->token) { - case TOKEN_IDENT: - case TOKEN_TYPENAME: - case TOKEN_KEYWORD: + case Token::IDENT: + case Token::TYPENAME: + case Token::KEYWORD: ftepp_macro_delete(ftepp, ftepp_tokval(ftepp)); break; default: @@ -1246,7 +1246,7 @@ static bool ftepp_undef(ftepp_t *ftepp) if (!ftepp_skipspace(ftepp)) return false; /* relaxing this condition - if (ftepp->token != TOKEN_EOL && ftepp->token != TOKEN_EOF) { + if (ftepp->token != Token::EOL && ftepp->token != Token::END) { ftepp_error(ftepp, "stray tokens after #ifdef"); return false; } @@ -1334,11 +1334,11 @@ static bool ftepp_directive_warning(ftepp_t *ftepp) { return false; /* handle the odd non string constant case so it works like C */ - if (ftepp->token != TOKEN_STRINGCONST) { + if (ftepp->token != Token::STRINGCONST) { bool store = false; vec_append(message, 8, "#warning"); ftepp_next(ftepp); - while (ftepp->token != TOKEN_EOL) { + while (ftepp->token != Token::EOL) { vec_append(message, strlen(ftepp_tokval(ftepp)), ftepp_tokval(ftepp)); ftepp_next(ftepp); } @@ -1365,10 +1365,10 @@ static void ftepp_directive_error(ftepp_t *ftepp) { return; /* handle the odd non string constant case so it works like C */ - if (ftepp->token != TOKEN_STRINGCONST) { + if (ftepp->token != Token::STRINGCONST) { vec_append(message, 6, "#error"); ftepp_next(ftepp); - while (ftepp->token != TOKEN_EOL) { + while (ftepp->token != Token::EOL) { vec_append(message, strlen(ftepp_tokval(ftepp)), ftepp_tokval(ftepp)); ftepp_next(ftepp); } @@ -1393,10 +1393,10 @@ static void ftepp_directive_message(ftepp_t *ftepp) { return; /* handle the odd non string constant case so it works like C */ - if (ftepp->token != TOKEN_STRINGCONST) { + if (ftepp->token != Token::STRINGCONST) { vec_append(message, 8, "#message"); ftepp_next(ftepp); - while (ftepp->token != TOKEN_EOL) { + while (ftepp->token != Token::EOL) { vec_append(message, strlen(ftepp_tokval(ftepp)), ftepp_tokval(ftepp)); ftepp_next(ftepp); } @@ -1433,7 +1433,7 @@ static bool ftepp_include(ftepp_t *ftepp) if (!ftepp_skipspace(ftepp)) return false; - if (ftepp->token != TOKEN_STRINGCONST) { + if (ftepp->token != Token::STRINGCONST) { ppmacro *macro = ftepp_macro_find(ftepp, ftepp_tokval(ftepp)); if (macro) { char *backup = ftepp->output_string; @@ -1514,7 +1514,7 @@ static bool ftepp_include(ftepp_t *ftepp) (void)ftepp_next(ftepp); if (!ftepp_skipspace(ftepp)) return false; - if (ftepp->token != TOKEN_EOL) { + if (ftepp->token != Token::EOL) { ftepp_error(ftepp, "stray tokens after #include"); return false; } @@ -1553,9 +1553,9 @@ static bool ftepp_hash(ftepp_t *ftepp) return false; switch (ftepp->token) { - case TOKEN_KEYWORD: - case TOKEN_IDENT: - case TOKEN_TYPENAME: + case Token::KEYWORD: + case Token::IDENT: + case Token::TYPENAME: if (!strcmp(ftepp_tokval(ftepp), "define")) { ftepp_inmacro(ftepp, "define"); return ftepp_define(ftepp); @@ -1684,16 +1684,16 @@ static bool ftepp_hash(ftepp_t *ftepp) default: ftepp_error(ftepp, "unexpected preprocessor token: `%s`", ftepp_tokval(ftepp)); return false; - case TOKEN_EOL: + case Token::EOL: ftepp_errorat(ftepp, ctx, "empty preprocessor directive"); return false; - case TOKEN_EOF: + case Token::END: ftepp_error(ftepp, "missing newline at end of file", ftepp_tokval(ftepp)); return false; /* Builtins! Don't forget the builtins! */ - case TOKEN_INTCONST: - case TOKEN_FLOATCONST: + case Token::INTCONST: + case Token::FLOATCONST: ftepp_out(ftepp, "#", false); return true; } @@ -1717,12 +1717,12 @@ static bool ftepp_preprocess(ftepp_t *ftepp) ftepp_next(ftepp); do { - if (ftepp->token >= TOKEN_EOF) + if (ftepp->token >= Token::END) break; switch (ftepp->token) { - case TOKEN_KEYWORD: - case TOKEN_IDENT: - case TOKEN_TYPENAME: + case Token::KEYWORD: + case Token::IDENT: + case Token::TYPENAME: /* is it a predef? */ if (OPTS_FLAG(FTEPP_PREDEFS)) { char *(*predef)(ftepp_t*) = ftepp_predef(ftepp_tokval(ftepp)); @@ -1747,30 +1747,30 @@ static bool ftepp_preprocess(ftepp_t *ftepp) break; } if (!ftepp_macro_call(ftepp, macro)) - ftepp->token = TOKEN_ERROR; + ftepp->token = Token::ERROR; break; - case '#': + case Token::HASH: if (!newline) { ftepp_out(ftepp, ftepp_tokval(ftepp), false); ftepp_next(ftepp); break; } ftepp->lex->flags.mergelines = true; - if (ftepp_next(ftepp) >= TOKEN_EOF) { + if (ftepp_next(ftepp) >= Token::END) { ftepp_error(ftepp, "error in preprocessor directive"); - ftepp->token = TOKEN_ERROR; + ftepp->token = Token::ERROR; break; } if (!ftepp_hash(ftepp)) - ftepp->token = TOKEN_ERROR; + ftepp->token = Token::ERROR; ftepp->lex->flags.mergelines = false; break; - case TOKEN_EOL: + case Token::EOL: newline = true; ftepp_out(ftepp, "\n", true); ftepp_next(ftepp); break; - case TOKEN_WHITE: + case Token::WHITE: /* same as default but don't set newline=false */ ftepp_out(ftepp, ftepp_tokval(ftepp), true); ftepp_next(ftepp); @@ -1781,13 +1781,13 @@ static bool ftepp_preprocess(ftepp_t *ftepp) ftepp_next(ftepp); break; } - } while (!ftepp->errors && ftepp->token < TOKEN_EOF); + } while (!ftepp->errors && ftepp->token < Token::END); /* force a 0 at the end but don't count it as added to the output */ vec_push(ftepp->output_string, 0); vec_shrinkby(ftepp->output_string, 1); - return (ftepp->token == TOKEN_EOF); + return (ftepp->token == Token::END); } /* Like in parser.c - files keep the previous state so we have one global diff --git a/gmqcc.h b/gmqcc.h index 520a1cf..07f8c95 100644 --- a/gmqcc.h +++ b/gmqcc.h @@ -176,7 +176,7 @@ char *stat_mem_strdup(const char *, bool); #define util_islower(a) (((unsigned)(a)-'a') < 26) #define util_isupper(a) (((unsigned)(a)-'A') < 26) #define util_isprint(a) (((unsigned)(a)-0x20) < 0x5F) -#define util_isspace(a) (((a) >= 9 && (a) <= 13) || (a) == ' ') +#define util_isspace(a) (((a) >= 9 && (a) <= 13) || (a) == Token::WS) bool util_strupper(const char *); bool util_strdigit(const char *); @@ -734,14 +734,6 @@ prog_section_def_t* prog_getdef (qc_program_t *prog, qcint_t off); qcany_t* prog_getedict (qc_program_t *prog, qcint_t e); qcint_t prog_tempstring(qc_program_t *prog, const char *_str); - -/* parser.c */ -struct parser_t; -parser_t *parser_create(void); -bool parser_compile_file(parser_t &parser, const char *); -bool parser_compile_string(parser_t &parser, const char *, const char *, size_t); -bool parser_finish(parser_t &parser, const char *); - /* ftepp.c */ struct ftepp_t; ftepp_t *ftepp_create (void); diff --git a/lexer.cpp b/lexer.cpp index 34fc71b..0bc28cb 100644 --- a/lexer.cpp +++ b/lexer.cpp @@ -71,8 +71,8 @@ static void lex_token_new(lex_file *lex) lex->tok.ctx.column = lex->column; } -static void lex_ungetch(lex_file *lex, int ch); -static int lex_getch(lex_file *lex); +static void lex_ungetch(lex_file *lex, Token ch); +static Token lex_getch(lex_file *lex); lex_file* lex_open(const char *file) { @@ -103,9 +103,9 @@ lex_file* lex_open(const char *file) /* handle BOM */ if ((read = (lex_getch(lex) << 16) | (lex_getch(lex) << 8) | lex_getch(lex)) != 0xEFBBBF) { - lex_ungetch(lex, (read & 0x0000FF)); - lex_ungetch(lex, (read & 0x00FF00) >> 8); - lex_ungetch(lex, (read & 0xFF0000) >> 16); + lex_ungetch(lex, static_cast((read & 0x0000FF))); + lex_ungetch(lex, static_cast((read & 0x00FF00) >> 8)); + lex_ungetch(lex, static_cast((read & 0xFF0000) >> 16)); } else { /* * otherwise the lexer has advanced 3 bytes for the BOM, we need @@ -175,19 +175,21 @@ void lex_close(lex_file *lex) -static int lex_fgetc(lex_file *lex) +static Token lex_fgetc(lex_file *lex) { if (lex->file) { lex->column++; - return fgetc(lex->file); + auto c = fgetc(lex->file); + return c == EOF ? Token::END : static_cast(c); } if (lex->open_string) { if (lex->open_string_pos >= lex->open_string_length) - return EOF; + return Token::END; lex->column++; - return lex->open_string[lex->open_string_pos++]; + auto c = lex->open_string[lex->open_string_pos++]; + return static_cast(c); } - return EOF; + return Token::END; } /* Get or put-back data @@ -195,36 +197,35 @@ static int lex_fgetc(lex_file *lex) * are working on. * The are merely wrapping get/put in order to count line numbers. */ -static int lex_try_trigraph(lex_file *lex, int old) +static Token lex_try_trigraph(lex_file *lex, Token old) { - int c2, c3; - c2 = lex_fgetc(lex); - if (!lex->push_line && c2 == '\n') { + auto c2 = lex_fgetc(lex); + if (!lex->push_line && c2 == Token::LF) { lex->line++; lex->column = 0; } - if (c2 != '?') { + if (c2 != Token::QUESTION) { lex_ungetch(lex, c2); return old; } - c3 = lex_fgetc(lex); - if (!lex->push_line && c3 == '\n') { + auto c3 = lex_fgetc(lex); + if (!lex->push_line && c3 == Token::LF) { lex->line++; lex->column = 0; } switch (c3) { - case '=': return '#'; - case '/': return '\\'; - case '\'': return '^'; - case '(': return '['; - case ')': return ']'; - case '!': return '|'; - case '<': return '{'; - case '>': return '}'; - case '-': return '~'; + case Token::EQ: return Token::HASH; + case Token::DIV: return Token::BACKSLASH; + case Token::QUOT_SINGLE: return Token::XOR; + case Token::PAREN_OPEN: return Token::BRACKET_OPEN; + case Token::PAREN_CLOSE: return Token::BRACKET_CLOSE; + case Token::NOT: return Token::OR; + case Token::LT: return Token::BRACE_OPEN; + case Token::GT: return Token::BRACE_CLOSE; + case Token::SUB: return Token::BITNOT; default: lex_ungetch(lex, c3); lex_ungetch(lex, c2); @@ -232,59 +233,56 @@ static int lex_try_trigraph(lex_file *lex, int old) } } -static int lex_try_digraph(lex_file *lex, int ch) +static Token lex_try_digraph(lex_file *lex, Token ch) { - int c2; - c2 = lex_fgetc(lex); + auto c2 = lex_fgetc(lex); /* we just used fgetc() so count lines * need to offset a \n the ungetch would recognize */ - if (!lex->push_line && c2 == '\n') + if (!lex->push_line && c2 == Token::LF) lex->line++; - if (ch == '<' && c2 == ':') - return '['; - else if (ch == ':' && c2 == '>') - return ']'; - else if (ch == '<' && c2 == '%') - return '{'; - else if (ch == '%' && c2 == '>') - return '}'; - else if (ch == '%' && c2 == ':') - return '#'; + if (ch == Token::LT && c2 == Token::COLON) + return Token::BRACKET_OPEN; + else if (ch == Token::COLON && c2 == Token::GT) + return Token::BRACKET_CLOSE; + else if (ch == Token::LT && c2 == Token::MOD) + return Token::BRACE_OPEN; + else if (ch == Token::MOD && c2 == Token::GT) + return Token::BRACE_CLOSE; + else if (ch == Token::MOD && c2 == Token::COLON) + return Token::HASH; lex_ungetch(lex, c2); return ch; } -static int lex_getch(lex_file *lex) +static Token lex_getch(lex_file *lex) { - int ch; - if (lex->peekpos) { lex->peekpos--; - if (!lex->push_line && lex->peek[lex->peekpos] == '\n') { + if (!lex->push_line && lex->peek[lex->peekpos] == Token::LF) { lex->line++; lex->column = 0; } return lex->peek[lex->peekpos]; } - ch = lex_fgetc(lex); - if (!lex->push_line && ch == '\n') { + auto ch = lex_fgetc(lex); + if (!lex->push_line && ch == Token::LF) { lex->line++; lex->column = 0; } - else if (ch == '?') + else if (ch == Token::QUESTION) return lex_try_trigraph(lex, ch); - else if (!lex->flags.nodigraphs && (ch == '<' || ch == ':' || ch == '%')) + else if (!lex->flags.nodigraphs && (ch == Token::LT || ch == Token::COLON || ch == Token::MOD)) return lex_try_digraph(lex, ch); return ch; } -static void lex_ungetch(lex_file *lex, int ch) +static void lex_ungetch(lex_file *lex, Token ch) { lex->peek[lex->peekpos++] = ch; lex->column--; - if (!lex->push_line && ch == '\n') { + if (!lex->push_line && ch == Token::LF) { lex->line--; lex->column = 0; } @@ -328,7 +326,6 @@ static void lex_endtoken(lex_file *lex) static bool lex_try_pragma(lex_file *lex) { - int ch; char *pragma = nullptr; char *command = nullptr; char *param = nullptr; @@ -339,8 +336,8 @@ static bool lex_try_pragma(lex_file *lex) line = lex->line; - ch = lex_getch(lex); - if (ch != '#') { + auto ch = lex_getch(lex); + if (ch != Token::HASH) { lex_ungetch(lex, ch); return false; } @@ -349,7 +346,7 @@ static bool lex_try_pragma(lex_file *lex) vec_push(pragma, ch); vec_push(pragma, 0); - if (ch != ' ' || strcmp(pragma, "pragma")) { + if (ch != Token::WS|| strcmp(pragma, "pragma")) { lex_ungetch(lex, ch); goto unroll; } @@ -358,16 +355,16 @@ static bool lex_try_pragma(lex_file *lex) vec_push(command, ch); vec_push(command, 0); - if (ch != '(') { + if (ch != Token::PAREN_OPEN) { lex_ungetch(lex, ch); goto unroll; } - for (ch = lex_getch(lex); vec_size(param) < 1024 && ch != ')' && ch != '\n'; ch = lex_getch(lex)) + for (ch = lex_getch(lex); vec_size(param) < 1024 && ch != Token::PAREN_CLOSE && ch != Token::LF; ch = lex_getch(lex)) vec_push(param, ch); vec_push(param, 0); - if (ch != ')') { + if (ch != Token::PAREN_CLOSE) { lex_ungetch(lex, ch); goto unroll; } @@ -402,7 +399,7 @@ static bool lex_try_pragma(lex_file *lex) goto unroll; lex->line = line; - while (ch != '\n' && ch != EOF) + while (ch != Token::LF && ch != Token::END) ch = lex_getch(lex); vec_free(command); vec_free(param); @@ -413,30 +410,30 @@ unroll: if (command) { vec_pop(command); while (vec_size(command)) { - lex_ungetch(lex, (unsigned char)vec_last(command)); + lex_ungetch(lex, static_cast(vec_last(command))); vec_pop(command); } vec_free(command); - lex_ungetch(lex, ' '); + lex_ungetch(lex, Token::WS); } if (param) { vec_pop(param); while (vec_size(param)) { - lex_ungetch(lex, (unsigned char)vec_last(param)); + lex_ungetch(lex, static_cast(vec_last(param))); vec_pop(param); } vec_free(param); - lex_ungetch(lex, ' '); + lex_ungetch(lex, Token::WS); } if (pragma) { vec_pop(pragma); while (vec_size(pragma)) { - lex_ungetch(lex, (unsigned char)vec_last(pragma)); + lex_ungetch(lex, static_cast(vec_last(pragma))); vec_pop(pragma); } vec_free(pragma); } - lex_ungetch(lex, '#'); + lex_ungetch(lex, Token::HASH); lex->line = line; return false; @@ -474,30 +471,30 @@ printf( "line one\n" * here is to store the line of the first character after skipping * the initial whitespace in lex->sline, this happens in lex_do. */ -static int lex_skipwhite(lex_file *lex, bool hadwhite) +static Token lex_skipwhite(lex_file *lex, bool hadwhite) { - int ch = 0; + Token ch; bool haswhite = hadwhite; do { ch = lex_getch(lex); - while (ch != EOF && util_isspace(ch)) { - if (ch == '\n') { + while (ch != Token::END && util_isspace(ch)) { + if (ch == Token::LF) { if (lex_try_pragma(lex)) continue; } if (lex->flags.preprocessing) { - if (ch == '\n') { + if (ch == Token::LF) { /* end-of-line */ /* see if there was whitespace first */ if (haswhite) { /* (vec_size(lex->tok.value)) { */ lex_ungetch(lex, ch); lex_endtoken(lex); - return TOKEN_WHITE; + return Token::WHITE; } /* otherwise return EOL */ - return TOKEN_EOL; + return Token::EOL; } haswhite = true; lex_tokench(lex, ch); @@ -505,75 +502,75 @@ static int lex_skipwhite(lex_file *lex, bool hadwhite) ch = lex_getch(lex); } - if (ch == '/') { + if (ch == Token::DIV) { ch = lex_getch(lex); - if (ch == '/') + if (ch == Token::DIV) { /* one line comment */ ch = lex_getch(lex); if (lex->flags.preprocessing) { haswhite = true; - lex_tokench(lex, ' '); - lex_tokench(lex, ' '); + lex_tokench(lex, Token::WS); + lex_tokench(lex, Token::WS); } - while (ch != EOF && ch != '\n') { + while (ch != Token::END && ch != Token::LF) { if (lex->flags.preprocessing) - lex_tokench(lex, ' '); /* ch); */ + lex_tokench(lex, Token::WS); /* ch); */ ch = lex_getch(lex); } if (lex->flags.preprocessing) { - lex_ungetch(lex, '\n'); + lex_ungetch(lex, Token::LF); lex_endtoken(lex); - return TOKEN_WHITE; + return Token::WHITE; } continue; } - if (ch == '*') + if (ch == Token::MUL) { /* multiline comment */ if (lex->flags.preprocessing) { haswhite = true; - lex_tokench(lex, ' '); - lex_tokench(lex, ' '); + lex_tokench(lex, Token::WS); + lex_tokench(lex, Token::WS); } - while (ch != EOF) + while (ch != Token::END) { ch = lex_getch(lex); - if (ch == '*') { + if (ch == Token::MUL) { ch = lex_getch(lex); - if (ch == '/') { + if (ch == Token::DIV) { if (lex->flags.preprocessing) { - lex_tokench(lex, ' '); - lex_tokench(lex, ' '); + lex_tokench(lex, Token::WS); + lex_tokench(lex, Token::WS); } break; } lex_ungetch(lex, ch); } if (lex->flags.preprocessing) { - if (ch == '\n') - lex_tokench(lex, '\n'); + if (ch == Token::LF) + lex_tokench(lex, Token::LF); else - lex_tokench(lex, ' '); + lex_tokench(lex, Token::WS); } } - ch = ' '; /* cause TRUE in the isspace check */ + ch = Token::WS; /* cause TRUE in the isspace check */ continue; } /* Otherwise roll back to the slash and break out of the loop */ lex_ungetch(lex, ch); - ch = '/'; + ch = Token::DIV; break; } - } while (ch != EOF && util_isspace(ch)); + } while (ch != Token::END && util_isspace(ch)); if (haswhite) { lex_endtoken(lex); lex_ungetch(lex, ch); - return TOKEN_WHITE; + return Token::WHITE; } return ch; } @@ -581,10 +578,8 @@ static int lex_skipwhite(lex_file *lex, bool hadwhite) /* Get a token */ static bool GMQCC_WARN lex_finish_ident(lex_file *lex) { - int ch; - - ch = lex_getch(lex); - while (ch != EOF && isident(ch)) + auto ch = lex_getch(lex); + while (ch != Token::END && isident(ch)) { lex_tokench(lex, ch); ch = lex_getch(lex); @@ -599,15 +594,13 @@ static bool GMQCC_WARN lex_finish_ident(lex_file *lex) /* read one ident for the frame list */ static int lex_parse_frame(lex_file *lex) { - int ch; - lex_token_new(lex); - ch = lex_getch(lex); - while (ch != EOF && ch != '\n' && util_isspace(ch)) + auto ch = lex_getch(lex); + while (ch != Token::END && ch != Token::LF && util_isspace(ch)) ch = lex_getch(lex); - if (ch == '\n') + if (ch == Token::LF) return 1; if (!isident_start(ch)) { @@ -656,38 +649,38 @@ static bool lex_finish_frames(lex_file *lex) return false; } -static int GMQCC_WARN lex_finish_string(lex_file *lex, int quote) +static Token GMQCC_WARN lex_finish_string(lex_file *lex, int quote) { utf8ch_t chr = 0; int ch = 0, texttype = 0; - int nextch; + Token nextch = Token::NONE; bool hex; bool oct; char u8buf[8]; /* way more than enough */ int u8len, uc; - while (ch != EOF) + while (ch != Token::END) { ch = lex_getch(lex); if (ch == quote) - return TOKEN_STRINGCONST; + return Token::STRINGCONST; if (lex->flags.preprocessing && ch == '\\') { lex_tokench(lex, ch); ch = lex_getch(lex); - if (ch == EOF) { + if (ch == Token::END) { lexerror(lex, "unexpected end of file"); - lex_ungetch(lex, EOF); /* next token to be TOKEN_EOF */ - return (lex->tok.ttype = TOKEN_ERROR); + lex_ungetch(lex, Token::END); /* next token to be Token::END */ + return (lex->tok.ttype = Token::ERROR); } lex_tokench(lex, ch); } else if (ch == '\\') { ch = lex_getch(lex); - if (ch == EOF) { + if (ch == Token::END) { lexerror(lex, "unexpected end of file"); - lex_ungetch(lex, EOF); /* next token to be TOKEN_EOF */ - return (lex->tok.ttype = TOKEN_ERROR); + lex_ungetch(lex, Token::END); /* next token to be Token::END */ + return (lex->tok.ttype = Token::ERROR); } switch (ch) { @@ -714,7 +707,7 @@ static int GMQCC_WARN lex_finish_string(lex_file *lex, int quote) else { lexerror(lex, "bad character code"); lex_ungetch(lex, nextch); - return (lex->tok.ttype = TOKEN_ERROR); + return (lex->tok.ttype = Token::ERROR); } ch *= 0x10; @@ -728,7 +721,7 @@ static int GMQCC_WARN lex_finish_string(lex_file *lex, int quote) else { lexerror(lex, "bad character code"); lex_ungetch(lex, nextch); - return (lex->tok.ttype = TOKEN_ERROR); + return (lex->tok.ttype = Token::ERROR); } break; @@ -750,13 +743,13 @@ static int GMQCC_WARN lex_finish_string(lex_file *lex, int quote) oct = (nextch == '0'); if (!hex && !oct) lex_ungetch(lex, nextch); - for (nextch = lex_getch(lex); nextch != '}'; nextch = lex_getch(lex)) { + for (nextch = lex_getch(lex); nextch != Token::BRACE_CLOSE; nextch = lex_getch(lex)) { if (!hex && !oct) { if (nextch >= '0' && nextch <= '9') chr = chr * 10 + nextch - '0'; else { lexerror(lex, "bad character code"); - return (lex->tok.ttype = TOKEN_ERROR); + return (lex->tok.ttype = Token::ERROR); } } else if (!oct) { if (nextch >= '0' && nextch <= '9') @@ -767,20 +760,20 @@ static int GMQCC_WARN lex_finish_string(lex_file *lex, int quote) chr = chr * 0x10 + nextch - 'A' + 10; else { lexerror(lex, "bad character code"); - return (lex->tok.ttype = TOKEN_ERROR); + return (lex->tok.ttype = Token::ERROR); } } else { if (nextch >= '0' && nextch <= '9') chr = chr * 8 + chr - '0'; else { lexerror(lex, "bad character code"); - return (lex->tok.ttype = TOKEN_ERROR); + return (lex->tok.ttype = Token::ERROR); } } if (chr > 0x10FFFF || (!OPTS_FLAG(UTF8) && chr > 255)) { lexerror(lex, "character code out of range"); - return (lex->tok.ttype = TOKEN_ERROR); + return (lex->tok.ttype = Token::ERROR); } } if (OPTS_FLAG(UTF8) && chr >= 128) { @@ -824,26 +817,26 @@ static int GMQCC_WARN lex_finish_string(lex_file *lex, int quote) lex_tokench(lex, ch); } lexerror(lex, "unexpected end of file within string constant"); - lex_ungetch(lex, EOF); /* next token to be TOKEN_EOF */ - return (lex->tok.ttype = TOKEN_ERROR); + lex_ungetch(lex, Token::END); /* next token to be Token::END */ + return (lex->tok.ttype = Token::ERROR); } -static int GMQCC_WARN lex_finish_digit(lex_file *lex, int lastch) +static Token GMQCC_WARN lex_finish_digit(lex_file *lex, Token lastch) { bool ishex = false; - int ch = lastch; + Token ch = lastch; /* parse a number... */ - if (ch == '.') - lex->tok.ttype = TOKEN_FLOATCONST; + if (ch == Token::DOT) + lex->tok.ttype = Token::FLOATCONST; else - lex->tok.ttype = TOKEN_INTCONST; + lex->tok.ttype = Token::INTCONST; lex_tokench(lex, ch); ch = lex_getch(lex); - if (ch != '.' && !util_isdigit(ch)) + if (ch != Token::DOT && !util_isdigit(ch)) { if (lastch != '0' || ch != 'x') { @@ -860,7 +853,7 @@ static int GMQCC_WARN lex_finish_digit(lex_file *lex, int lastch) /* EOF would have been caught above */ - if (ch != '.') + if (ch != Token::DOT) { lex_tokench(lex, ch); ch = lex_getch(lex); @@ -871,10 +864,10 @@ static int GMQCC_WARN lex_finish_digit(lex_file *lex, int lastch) } } /* NOT else, '.' can come from above as well */ - if (lex->tok.ttype != TOKEN_FLOATCONST && ch == '.' && !ishex) + if (lex->tok.ttype != Token::FLOATCONST && ch == Token::DOT && !ishex) { /* Allow floating comma in non-hex mode */ - lex->tok.ttype = TOKEN_FLOATCONST; + lex->tok.ttype = Token::FLOATCONST; lex_tokench(lex, ch); /* continue digits-only */ @@ -887,27 +880,27 @@ static int GMQCC_WARN lex_finish_digit(lex_file *lex, int lastch) } /* put back the last character */ /* but do not put back the trailing 'f' or a float */ - if (lex->tok.ttype == TOKEN_FLOATCONST && ch == 'f') + if (lex->tok.ttype == Token::FLOATCONST && ch == 'f') ch = lex_getch(lex); /* generally we don't want words to follow numbers: */ if (isident(ch)) { lexerror(lex, "unexpected trailing characters after number"); - return (lex->tok.ttype = TOKEN_ERROR); + return (lex->tok.ttype = Token::ERROR); } lex_ungetch(lex, ch); lex_endtoken(lex); - if (lex->tok.ttype == TOKEN_FLOATCONST) + if (lex->tok.ttype == Token::FLOATCONST) lex->tok.constval.f = strtod(lex->tok.value, nullptr); else lex->tok.constval.i = strtol(lex->tok.value, nullptr, 0); return lex->tok.ttype; } -int lex_do(lex_file *lex) +Token lex_do(lex_file *lex) { - int ch, nextch, thirdch; + Token ch, nextch, thirdch; bool hadwhite = false; lex_token_new(lex); @@ -915,22 +908,21 @@ int lex_do(lex_file *lex) while (true) { ch = lex_skipwhite(lex, hadwhite); hadwhite = true; - if (!lex->flags.mergelines || ch != '\\') + if (!lex->flags.mergelines || ch != Token::BACKSLASH) break; ch = lex_getch(lex); - if (ch == '\r') + if (ch == Token::CR) ch = lex_getch(lex); - if (ch != '\n') { + if (ch != Token::LF) { lex_ungetch(lex, ch); - ch = '\\'; + ch = Token::BACKSLASH; break; } /* we reached a linemerge */ lex_tokench(lex, '\n'); - continue; } - if (lex->flags.preprocessing && (ch == TOKEN_WHITE || ch == TOKEN_EOL || ch == TOKEN_FATAL)) { + if (lex->flags.preprocessing && (ch == Token::WHITE || ch == Token::EOL || ch == Token::FATAL)) { return (lex->tok.ttype = ch); } @@ -939,15 +931,15 @@ int lex_do(lex_file *lex) lex->tok.ctx.file = lex->name; if (lex->eof) - return (lex->tok.ttype = TOKEN_FATAL); + return (lex->tok.ttype = Token::FATAL); - if (ch == EOF) { + if (ch == Token::END) { lex->eof = true; - return (lex->tok.ttype = TOKEN_EOF); + return (lex->tok.ttype = Token::END); } /* modelgen / spiritgen commands */ - if (ch == '$' && !lex->flags.preprocessing) { + if (ch == Token::DOLLAR && !lex->flags.preprocessing) { const char *v; size_t frame; @@ -958,7 +950,7 @@ int lex_do(lex_file *lex) } lex_tokench(lex, ch); if (!lex_finish_ident(lex)) - return (lex->tok.ttype = TOKEN_ERROR); + return (lex->tok.ttype = Token::ERROR); lex_endtoken(lex); /* skip the known commands */ v = lex->tok.value; @@ -971,14 +963,14 @@ int lex_do(lex_file *lex) * which the parser is unaware of */ if (!lex_finish_frames(lex)) - return (lex->tok.ttype = TOKEN_ERROR); + return (lex->tok.ttype = Token::ERROR); return lex_do(lex); } if (!strcmp(v, "framevalue")) { ch = lex_getch(lex); - while (ch != EOF && util_isspace(ch) && ch != '\n') + while (ch != Token::END && util_isspace(ch) && ch != Token::LF) ch = lex_getch(lex); if (!util_isdigit(ch)) { @@ -989,7 +981,7 @@ int lex_do(lex_file *lex) lex_token_new(lex); lex->tok.ttype = lex_finish_digit(lex, ch); lex_endtoken(lex); - if (lex->tok.ttype != TOKEN_INTCONST) { + if (lex->tok.ttype != Token::INTCONST) { lexerror(lex, "$framevalue requires an integer parameter"); return lex_do(lex); } @@ -1010,7 +1002,7 @@ int lex_do(lex_file *lex) return lex_do(lex); } if (rc < 0) - return (lex->tok.ttype = TOKEN_FATAL); + return (lex->tok.ttype = Token::FATAL); v = lex->tok.value; for (frame = 0; frame < vec_size(lex->frames); ++frame) { @@ -1036,7 +1028,7 @@ int lex_do(lex_file *lex) return lex_do(lex); } if (rc < 0) - return (lex->tok.ttype = TOKEN_FATAL); + return (lex->tok.ttype = Token::FATAL); if (lex->modelname) { frame_macro m; @@ -1058,7 +1050,7 @@ int lex_do(lex_file *lex) vec_free(lex->frames); /* skip line (fteqcc does it too) */ ch = lex_getch(lex); - while (ch != EOF && ch != '\n') + while (ch != Token::END && ch != Token::LF) ch = lex_getch(lex); return lex_do(lex); } @@ -1072,7 +1064,7 @@ int lex_do(lex_file *lex) { /* skip line */ ch = lex_getch(lex); - while (ch != EOF && ch != '\n') + while (ch != Token::END && ch != Token::LF) ch = lex_getch(lex); return lex_do(lex); } @@ -1080,7 +1072,7 @@ int lex_do(lex_file *lex) for (frame = 0; frame < vec_size(lex->frames); ++frame) { if (!strcmp(v, lex->frames[frame].name)) { lex->tok.constval.i = lex->frames[frame].value; - return (lex->tok.ttype = TOKEN_INTCONST); + return (lex->tok.ttype = Token::INTCONST); } } @@ -1091,44 +1083,44 @@ int lex_do(lex_file *lex) /* single-character tokens */ switch (ch) { - case '[': + case Token::BRACKET_OPEN: nextch = lex_getch(lex); - if (nextch == '[') { + if (nextch == Token::BRACKET_OPEN) { lex_tokench(lex, ch); lex_tokench(lex, nextch); lex_endtoken(lex); - return (lex->tok.ttype = TOKEN_ATTRIBUTE_OPEN); + return (lex->tok.ttype = Token::ATTRIBUTE_OPEN); } lex_ungetch(lex, nextch); /* FALL THROUGH */ - case '(': - case ':': - case '?': + case Token::PAREN_OPEN: + case Token::COLON: + case Token::QUESTION: lex_tokench(lex, ch); lex_endtoken(lex); if (lex->flags.noops) return (lex->tok.ttype = ch); else - return (lex->tok.ttype = TOKEN_OPERATOR); + return (lex->tok.ttype = Token::OPERATOR); - case ']': + case Token::BRACKET_CLOSE: if (lex->flags.noops) { nextch = lex_getch(lex); - if (nextch == ']') { + if (nextch == Token::BRACKET_CLOSE) { lex_tokench(lex, ch); lex_tokench(lex, nextch); lex_endtoken(lex); - return (lex->tok.ttype = TOKEN_ATTRIBUTE_CLOSE); + return (lex->tok.ttype = Token::ATTRIBUTE_CLOSE); } lex_ungetch(lex, nextch); } /* FALL THROUGH */ - case ')': - case ';': - case '{': - case '}': + case Token::PAREN_CLOSE: + case Token::SEMICOLON: + case Token::BRACE_OPEN: + case Token::BRACE_CLOSE: - case '#': + case Token::HASH: lex_tokench(lex, ch); lex_endtoken(lex); return (lex->tok.ttype = ch); @@ -1136,7 +1128,7 @@ int lex_do(lex_file *lex) break; } - if (ch == '.') { + if (ch == Token::DOT) { nextch = lex_getch(lex); /* digits starting with a dot */ if (util_isdigit(nextch)) { @@ -1155,17 +1147,17 @@ int lex_do(lex_file *lex) */ switch (ch) { - case '*': - case '/': - case '<': - case '>': - case '=': - case '&': - case '|': - case '^': - case '~': - case ',': - case '!': + case Token::MUL: + case Token::DIV: + case Token::LT: + case Token::GT: + case Token::EQ: + case Token::AND: + case Token::OR: + case Token::XOR: + case Token::BITNOT: + case Token::COMMA: + case Token::NOT: lex_tokench(lex, ch); lex_endtoken(lex); return (lex->tok.ttype = ch); @@ -1174,72 +1166,72 @@ int lex_do(lex_file *lex) } } - if (ch == '.') + if (ch == Token::DOT) { lex_tokench(lex, ch); /* peak ahead once */ nextch = lex_getch(lex); - if (nextch != '.') { + if (nextch != Token::DOT) { lex_ungetch(lex, nextch); lex_endtoken(lex); if (lex->flags.noops) return (lex->tok.ttype = ch); else - return (lex->tok.ttype = TOKEN_OPERATOR); + return (lex->tok.ttype = Token::OPERATOR); } /* peak ahead again */ nextch = lex_getch(lex); - if (nextch != '.') { + if (nextch != Token::DOT) { lex_ungetch(lex, nextch); - lex_ungetch(lex, '.'); + lex_ungetch(lex, Token::DOT); lex_endtoken(lex); if (lex->flags.noops) return (lex->tok.ttype = ch); else - return (lex->tok.ttype = TOKEN_OPERATOR); + return (lex->tok.ttype = Token::OPERATOR); } /* fill the token to be "..." */ lex_tokench(lex, ch); lex_tokench(lex, ch); lex_endtoken(lex); - return (lex->tok.ttype = TOKEN_DOTS); + return (lex->tok.ttype = Token::DOTS); } - if (ch == ',' || ch == '.') { + if (ch == Token::COMMA || ch == Token::DOT) { lex_tokench(lex, ch); lex_endtoken(lex); - return (lex->tok.ttype = TOKEN_OPERATOR); + return (lex->tok.ttype = Token::OPERATOR); } - if (ch == '+' || ch == '-' || /* ++, --, +=, -= and -> as well! */ - ch == '>' || ch == '<' || /* <<, >>, <=, >= and >< as well! */ - ch == '=' || ch == '!' || /* <=>, ==, != */ - ch == '&' || ch == '|' || /* &&, ||, &=, |= */ - ch == '~' || ch == '^' /* ~=, ~, ^ */ + if (ch == Token::ADD || ch == Token::SUB || /* ++, --, +=, -= and -> as well! */ + ch == Token::GT || ch == Token::LT|| /* <<, >>, <=, >= and >< as well! */ + ch == Token::EQ || ch == Token::NOT || /* <=>, ==, != */ + ch == Token::AND || ch == Token::OR || /* &&, ||, &=, |= */ + ch == Token::BITNOT || ch == Token::XOR /* ~=, ~, ^ */ ) { lex_tokench(lex, ch); nextch = lex_getch(lex); - if ((nextch == '=' && ch != '<') || (nextch == '<' && ch == '>')) + if ((nextch == Token::EQ && ch != Token::LT) || (nextch == Token::LT && ch == Token::GT)) lex_tokench(lex, nextch); - else if (nextch == ch && ch != '!') { + else if (nextch == ch && ch != Token::NOT) { lex_tokench(lex, nextch); - if ((thirdch = lex_getch(lex)) == '=') + if ((thirdch = lex_getch(lex)) == Token::EQ) lex_tokench(lex, thirdch); else lex_ungetch(lex, thirdch); - } else if (ch == '<' && nextch == '=') { + } else if (ch == Token::LT && nextch == Token::EQ) { lex_tokench(lex, nextch); - if ((thirdch = lex_getch(lex)) == '>') + if ((thirdch = lex_getch(lex)) == Token::GT) lex_tokench(lex, thirdch); else lex_ungetch(lex, thirdch); - } else if (ch == '-' && nextch == '>') { + } else if (ch == Token::SUB && nextch == Token::GT) { lex_tokench(lex, nextch); - } else if (ch == '&' && nextch == '~') { + } else if (ch == Token::AND && nextch == Token::BITNOT) { thirdch = lex_getch(lex); - if (thirdch != '=') { + if (thirdch != Token::EQ) { lex_ungetch(lex, thirdch); lex_ungetch(lex, nextch); } @@ -1249,10 +1241,10 @@ int lex_do(lex_file *lex) } } else if (lex->flags.preprocessing && - ch == '-' && util_isdigit(nextch)) + ch == Token::SUB && util_isdigit(nextch)) { lex->tok.ttype = lex_finish_digit(lex, nextch); - if (lex->tok.ttype == TOKEN_INTCONST) + if (lex->tok.ttype == Token::INTCONST) lex->tok.constval.i = -lex->tok.constval.i; else lex->tok.constval.f = -lex->tok.constval.f; @@ -1263,27 +1255,27 @@ int lex_do(lex_file *lex) } lex_endtoken(lex); - return (lex->tok.ttype = TOKEN_OPERATOR); + return (lex->tok.ttype = Token::OPERATOR); } - if (ch == '*' || ch == '/') /* *=, /= */ + if (ch == Token::MUL || ch == Token::DIV) /* *=, /= */ { lex_tokench(lex, ch); nextch = lex_getch(lex); - if (nextch == '=' || nextch == '*') { + if (nextch == Token::EQ || nextch == Token::MUL) { lex_tokench(lex, nextch); } else lex_ungetch(lex, nextch); lex_endtoken(lex); - return (lex->tok.ttype = TOKEN_OPERATOR); + return (lex->tok.ttype = Token::OPERATOR); } - if (ch == '%') { + if (ch == Token::MOD) { lex_tokench(lex, ch); lex_endtoken(lex); - return (lex->tok.ttype = TOKEN_OPERATOR); + return (lex->tok.ttype = Token::OPERATOR); } if (isident_start(ch)) @@ -1293,42 +1285,42 @@ int lex_do(lex_file *lex) lex_tokench(lex, ch); if (!lex_finish_ident(lex)) { /* error? */ - return (lex->tok.ttype = TOKEN_ERROR); + return (lex->tok.ttype = Token::ERROR); } lex_endtoken(lex); - lex->tok.ttype = TOKEN_IDENT; + lex->tok.ttype = Token::IDENT; v = lex->tok.value; if (!strcmp(v, "void")) { - lex->tok.ttype = TOKEN_TYPENAME; + lex->tok.ttype = Token::TYPENAME; lex->tok.constval.t = TYPE_VOID; } else if (!strcmp(v, "int")) { - lex->tok.ttype = TOKEN_TYPENAME; + lex->tok.ttype = Token::TYPENAME; lex->tok.constval.t = TYPE_INTEGER; } else if (!strcmp(v, "float")) { - lex->tok.ttype = TOKEN_TYPENAME; + lex->tok.ttype = Token::TYPENAME; lex->tok.constval.t = TYPE_FLOAT; } else if (!strcmp(v, "string")) { - lex->tok.ttype = TOKEN_TYPENAME; + lex->tok.ttype = Token::TYPENAME; lex->tok.constval.t = TYPE_STRING; } else if (!strcmp(v, "entity")) { - lex->tok.ttype = TOKEN_TYPENAME; + lex->tok.ttype = Token::TYPENAME; lex->tok.constval.t = TYPE_ENTITY; } else if (!strcmp(v, "vector")) { - lex->tok.ttype = TOKEN_TYPENAME; + lex->tok.ttype = Token::TYPENAME; lex->tok.constval.t = TYPE_VECTOR; } else if (!strcmp(v, "_length")) { - lex->tok.ttype = TOKEN_OPERATOR; + lex->tok.ttype = Token::OPERATOR; } else { size_t kw; for (kw = 0; kw < GMQCC_ARRAY_COUNT(keywords_qc); ++kw) { if (!strcmp(v, keywords_qc[kw])) - return (lex->tok.ttype = TOKEN_KEYWORD); + return (lex->tok.ttype = Token::KEYWORD); } if (OPTS_OPTION_U32(OPTION_STANDARD) != COMPILER_QCC) { for (kw = 0; kw < GMQCC_ARRAY_COUNT(keywords_fg); ++kw) { if (!strcmp(v, keywords_fg[kw])) - return (lex->tok.ttype = TOKEN_KEYWORD); + return (lex->tok.ttype = Token::KEYWORD); } } } @@ -1336,52 +1328,52 @@ int lex_do(lex_file *lex) return lex->tok.ttype; } - if (ch == '"') + if (ch == Token::QUOT_DOUBLE) { lex->flags.nodigraphs = true; if (lex->flags.preprocessing) lex_tokench(lex, ch); - lex->tok.ttype = lex_finish_string(lex, '"'); + lex->tok.ttype = lex_finish_string(lex, Token::QUOT_DOUBLE); if (lex->flags.preprocessing) lex_tokench(lex, ch); - while (!lex->flags.preprocessing && lex->tok.ttype == TOKEN_STRINGCONST) + while (!lex->flags.preprocessing && lex->tok.ttype == Token::STRINGCONST) { /* Allow c style "string" "continuation" */ ch = lex_skipwhite(lex, false); - if (ch != '"') { + if (ch != Token::QUOT_DOUBLE) { lex_ungetch(lex, ch); break; } - lex->tok.ttype = lex_finish_string(lex, '"'); + lex->tok.ttype = lex_finish_string(lex, Token::QUOT_DOUBLE); } lex->flags.nodigraphs = false; lex_endtoken(lex); return lex->tok.ttype; } - if (ch == '\'') + if (ch == Token::QUOT_SINGLE) { /* we parse character constants like string, - * but return TOKEN_CHARCONST, or a vector type if it fits... + * but return Token::CHARCONST, or a vector type if it fits... * Likewise actual unescaping has to be done by the parser. * The difference is we don't allow 'char' 'continuation'. */ if (lex->flags.preprocessing) lex_tokench(lex, ch); - lex->tok.ttype = lex_finish_string(lex, '\''); + lex->tok.ttype = lex_finish_string(lex, Token::QUOT_SINGLE); if (lex->flags.preprocessing) lex_tokench(lex, ch); lex_endtoken(lex); - lex->tok.ttype = TOKEN_CHARCONST; + lex->tok.ttype = Token::CHARCONST; /* It's a vector if we can successfully scan 3 floats */ if (util_sscanf(lex->tok.value, " %f %f %f ", &lex->tok.constval.v.x, &lex->tok.constval.v.y, &lex->tok.constval.v.z) == 3) { - lex->tok.ttype = TOKEN_VECTORCONST; + lex->tok.ttype = Token::VECTORCONST; } else { @@ -1393,7 +1385,7 @@ int lex_do(lex_file *lex) ( OPTS_FLAG(UTF8) ? "invalid multibyte character sequence `%s`" : "multibyte character: `%s`" ), lex->tok.value)) - return (lex->tok.ttype = TOKEN_ERROR); + return (lex->tok.ttype = Token::ERROR); } else lex->tok.constval.i = u8char; @@ -1413,11 +1405,11 @@ int lex_do(lex_file *lex) } if (lex->flags.preprocessing) { - lex_tokench(lex, ch); + lex_tokench(lex, static_cast(ch)); lex_endtoken(lex); return (lex->tok.ttype = ch); } lexerror(lex, "unknown token: `%c`", ch); - return (lex->tok.ttype = TOKEN_ERROR); + return (lex->tok.ttype = Token::ERROR); } diff --git a/lexer.h b/lexer.h index 8ed85a9..94db3d2 100644 --- a/lexer.h +++ b/lexer.h @@ -2,63 +2,103 @@ #define GMQCC_LEXER_HDR #include "gmqcc.h" -struct token { - int ttype; - char *value; - union { - vec3_t v; - int i; - qcfloat_t f; - qc_type t; /* type */ - } constval; - lex_ctx_t ctx; -}; - /* Lexer * */ -enum { +enum Token : int { // todo: enum class /* Other tokens which we can return: */ - TOKEN_NONE = 0, - TOKEN_START = 128, + NONE = 0, + + CR = '\r', + LF = '\n', + WS = ' ', + BACKSLASH = '\\', + + HASH = '#', + DOLLAR = '$', + + DOT = '.', + COMMA = ',', + COLON = ':', + SEMICOLON = ';', + + AND = '&', + OR = '|', + XOR = '^', + BITNOT = '~', + NOT = '!', + + LT = '<', + GT = '>', + EQ = '=', - TOKEN_IDENT, + MUL = '*', + DIV = '/', + MOD = '%', - TOKEN_TYPENAME, + ADD = '+', + SUB = '-', - TOKEN_OPERATOR, + QUOT_SINGLE = '\'', + QUOT_DOUBLE = '"', - TOKEN_KEYWORD, /* loop */ + QUESTION = '?', - TOKEN_DOTS, /* 3 dots, ... */ + BRACE_OPEN = '{', BRACE_CLOSE = '}', + BRACKET_OPEN = '[', BRACKET_CLOSE = ']', + PAREN_OPEN = '(', PAREN_CLOSE = ')', - TOKEN_ATTRIBUTE_OPEN, /* [[ */ - TOKEN_ATTRIBUTE_CLOSE, /* ]] */ + START = 128, - TOKEN_VA_ARGS, /* for the ftepp only */ - TOKEN_VA_ARGS_ARRAY, /* for the ftepp only */ - TOKEN_VA_COUNT, /* to get the count of vaargs */ + IDENT, - TOKEN_STRINGCONST, /* not the typename but an actual "string" */ - TOKEN_CHARCONST, - TOKEN_VECTORCONST, - TOKEN_INTCONST, - TOKEN_FLOATCONST, + TYPENAME, - TOKEN_WHITE, - TOKEN_EOL, + OPERATOR, + + KEYWORD, /* loop */ + + DOTS, /* 3 dots, ... */ + + ATTRIBUTE_OPEN, /* [[ */ + ATTRIBUTE_CLOSE, /* ]] */ + + VA_ARGS, /* for the ftepp only */ + VA_ARGS_ARRAY, /* for the ftepp only */ + VA_COUNT, /* to get the count of vaargs */ + + STRINGCONST, /* not the typename but an actual "string" */ + CHARCONST, + VECTORCONST, + INTCONST, + FLOATCONST, + + WHITE, + EOL, /* if we add additional tokens before this, the exposed API * should not be broken anyway, but EOF/ERROR/... should * still be at the bottom */ - TOKEN_EOF = 1024, + END = 1024, - /* We use '< TOKEN_ERROR', so TOKEN_FATAL must come after it and any + /* We use '< ERROR', so FATAL must come after it and any * other error related tokens as well */ - TOKEN_ERROR, - TOKEN_FATAL /* internal error, eg out of memory */ + ERROR, + FATAL /* internal error, eg out of memory */ +}; + +struct token { + Token ttype; + char *value; + union { + vec3_t v; + int i; + qcfloat_t f; + qc_type t; /* type */ + } constval; + lex_ctx_t ctx; }; struct frame_macro { @@ -77,7 +117,7 @@ struct lex_file { size_t sline; /* line at the start of a token */ size_t column; - int peek[256]; + Token peek[256]; size_t peekpos; bool eof; @@ -101,7 +141,7 @@ struct lex_file { lex_file* lex_open (const char *file); lex_file* lex_open_string(const char *str, size_t len, const char *name); void lex_close(lex_file *lex); -int lex_do (lex_file *lex); +Token lex_do (lex_file *lex); void lex_cleanup(void); /* Parser diff --git a/opts.cpp b/opts.cpp index c35fd73..c5014a2 100644 --- a/opts.cpp +++ b/opts.cpp @@ -2,6 +2,7 @@ #include #include "gmqcc.h" +#include "lexer.h" const unsigned int opts_opt_oflag[COUNT_OPTIMIZATIONS+1] = { # define GMQCC_TYPE_OPTIMIZATIONS @@ -210,7 +211,7 @@ static size_t opts_ini_parse ( char *read_name; char *read_value; - while (util_getline(&line, &linesize, filehandle) != EOF) { + while (util_getline(&line, &linesize, filehandle) != Token::END) { parse_beg = line; /* handle BOM */ diff --git a/parser.cpp b/parser.cpp index 3eefe10..b2957d4 100644 --- a/parser.cpp +++ b/parser.cpp @@ -65,9 +65,9 @@ static bool parser_next(parser_t &parser) { /* lex_do kills the previous token */ parser.tok = lex_do(parser.lex); - if (parser.tok == TOKEN_EOF) + if (parser.tok == Token::END) return true; - if (parser.tok >= TOKEN_ERROR) { + if (parser.tok >= Token::ERROR) { parseerror(parser, "lex error"); return false; } @@ -77,17 +77,6 @@ static bool parser_next(parser_t &parser) #define parser_tokval(p) ((p).lex->tok.value) #define parser_token(p) (&((p).lex->tok)) -char *parser_strdup(const char *str) -{ - if (str && !*str) { - /* actually dup empty strings */ - char *out = (char*)mem_a(1); - *out = 0; - return out; - } - return util_strdup(str); -} - static ast_expression* parser_find_field(parser_t &parser, const char *name) { return (ast_expression*)util_htget(parser.htfields, name); } @@ -1427,11 +1416,11 @@ static bool parser_close_paren(parser_t &parser, shunt *sy) static void parser_reclassify_token(parser_t &parser) { size_t i; - if (parser.tok >= TOKEN_START) + if (parser.tok >= Token::START) return; for (i = 0; i < operator_count; ++i) { if (!strcmp(parser_tokval(parser), operators[i].op)) { - parser.tok = TOKEN_OPERATOR; + parser.tok = Token::OPERATOR; return; } } @@ -1473,7 +1462,7 @@ static ast_expression* parse_vararg_do(parser_t &parser) return out; } - if (!parser_next(parser) || (parser.tok != TOKEN_IDENT && parser.tok != TOKEN_TYPENAME)) { + if (!parser_next(parser) || (parser.tok != Token::IDENT && parser.tok != Token::TYPENAME)) { ast_unref(idx); parseerror(parser, "expected typename for vararg"); return nullptr; @@ -1528,7 +1517,7 @@ bool ftepp_predef_exists(const char *name); static bool parse_sya_operand(parser_t &parser, shunt *sy, bool with_labels) { if (OPTS_FLAG(TRANSLATABLE_STRINGS) && - parser.tok == TOKEN_IDENT && + parser.tok == Token::IDENT && !strcmp(parser_tokval(parser), "_")) { /* a translatable string */ @@ -1540,7 +1529,7 @@ static bool parse_sya_operand(parser_t &parser, shunt *sy, bool with_labels) return false; } parser.lex->flags.noops = false; - if (!parser_next(parser) || parser.tok != TOKEN_STRINGCONST) { + if (!parser_next(parser) || parser.tok != Token::STRINGCONST) { parseerror(parser, "expected a constant string in translatable-string extension"); return false; } @@ -1555,7 +1544,7 @@ static bool parse_sya_operand(parser_t &parser, shunt *sy, bool with_labels) } return true; } - else if (parser.tok == TOKEN_DOTS) + else if (parser.tok == Token::DOTS) { ast_expression *va; if (!OPTS_FLAG(VARIADIC_ARGS)) { @@ -1568,35 +1557,35 @@ static bool parse_sya_operand(parser_t &parser, shunt *sy, bool with_labels) sy->out.push_back(syexp(parser_ctx(parser), va)); return true; } - else if (parser.tok == TOKEN_FLOATCONST) { + else if (parser.tok == Token::FLOATCONST) { ast_expression *val = parser.m_fold.constgen_float((parser_token(parser)->constval.f), false); if (!val) return false; sy->out.push_back(syexp(parser_ctx(parser), val)); return true; } - else if (parser.tok == TOKEN_INTCONST || parser.tok == TOKEN_CHARCONST) { + else if (parser.tok == Token::INTCONST || parser.tok == Token::CHARCONST) { ast_expression *val = parser.m_fold.constgen_float((qcfloat_t)(parser_token(parser)->constval.i), false); if (!val) return false; sy->out.push_back(syexp(parser_ctx(parser), val)); return true; } - else if (parser.tok == TOKEN_STRINGCONST) { + else if (parser.tok == Token::STRINGCONST) { ast_expression *val = parser.m_fold.constgen_string(parser_tokval(parser), false); if (!val) return false; sy->out.push_back(syexp(parser_ctx(parser), val)); return true; } - else if (parser.tok == TOKEN_VECTORCONST) { + else if (parser.tok == Token::VECTORCONST) { ast_expression *val = parser.m_fold.constgen_vector(parser_token(parser)->constval.v); if (!val) return false; sy->out.push_back(syexp(parser_ctx(parser), val)); return true; } - else if (parser.tok == TOKEN_IDENT) + else if (parser.tok == Token::IDENT) { const char *ctoken = parser_tokval(parser); ast_expression *prev = sy->out.size() ? sy->out.back().out : nullptr; @@ -1713,12 +1702,12 @@ static ast_expression* parse_expression_leave(parser_t &parser, bool stopatcomma while (true) { - if (parser.tok == TOKEN_TYPENAME) { + if (parser.tok == Token::TYPENAME) { parseerror(parser, "unexpected typename `%s`", parser_tokval(parser)); goto onerr; } - if (parser.tok == TOKEN_OPERATOR) + if (parser.tok == Token::OPERATOR) { /* classify the operator */ const oper_info *op; @@ -1742,13 +1731,13 @@ static ast_expression* parse_expression_leave(parser_t &parser, bool stopatcomma /* when declaring variables, a comma starts a new variable */ if (op->id == opid1(',') && sy.paren.empty() && stopatcomma) { /* fixup the token */ - parser.tok = ','; + parser.tok = Token::COMMA; break; } /* a colon without a pervious question mark cannot be a ternary */ if (!ternaries && op->id == opid2(':','?')) { - parser.tok = ':'; + parser.tok = Token::COLON; break; } @@ -1924,7 +1913,7 @@ static ast_expression* parse_expression_leave(parser_t &parser, bool stopatcomma else { /* in this case we might want to allow constant string concatenation */ bool concatenated = false; - if (parser.tok == TOKEN_STRINGCONST && sy.out.size()) { + if (parser.tok == Token::STRINGCONST && sy.out.size()) { ast_expression *lexpr = sy.out.back().out; if (ast_istype(lexpr, ast_value)) { ast_value *last = (ast_value*)lexpr; @@ -2130,7 +2119,7 @@ static bool parse_if(parser_t &parser, ast_block *block, ast_expression **out) parseerror(parser, "expected condition or 'not'"); return false; } - if (parser.tok == TOKEN_IDENT && !strcmp(parser_tokval(parser), "not")) { + if (parser.tok == Token::IDENT && !strcmp(parser_tokval(parser), "not")) { ifnot = true; if (!parser_next(parser)) { parseerror(parser, "expected condition in parenthesis"); @@ -2217,7 +2206,7 @@ static bool parse_while(parser_t &parser, ast_block *block, ast_expression **out if (parser.tok == ':') { if (!OPTS_FLAG(LOOP_LABELS)) parseerror(parser, "labeled loops not activated, try using -floop-labels"); - if (!parser_next(parser) || parser.tok != TOKEN_IDENT) { + if (!parser_next(parser) || parser.tok != Token::IDENT) { parseerror(parser, "expected loop label"); return false; } @@ -2318,7 +2307,7 @@ static bool parse_dowhile(parser_t &parser, ast_block *block, ast_expression **o if (parser.tok == ':') { if (!OPTS_FLAG(LOOP_LABELS)) parseerror(parser, "labeled loops not activated, try using -floop-labels"); - if (!parser_next(parser) || parser.tok != TOKEN_IDENT) { + if (!parser_next(parser) || parser.tok != Token::IDENT) { parseerror(parser, "expected loop label"); return false; } @@ -2364,7 +2353,7 @@ static bool parse_dowhile_go(parser_t &parser, ast_block *block, ast_expression return false; /* expect the "while" */ - if (parser.tok != TOKEN_KEYWORD || + if (parser.tok != Token::KEYWORD || strcmp(parser_tokval(parser), "while")) { parseerror(parser, "expected 'while' and condition"); @@ -2438,7 +2427,7 @@ static bool parse_for(parser_t &parser, ast_block *block, ast_expression **out) if (parser.tok == ':') { if (!OPTS_FLAG(LOOP_LABELS)) parseerror(parser, "labeled loops not activated, try using -floop-labels"); - if (!parser_next(parser) || parser.tok != TOKEN_IDENT) { + if (!parser_next(parser) || parser.tok != Token::IDENT) { parseerror(parser, "expected loop label"); return false; } @@ -2497,10 +2486,10 @@ static bool parse_for_go(parser_t &parser, ast_block *block, ast_expression **ou } typevar = nullptr; - if (parser.tok == TOKEN_IDENT) + if (parser.tok == Token::IDENT) typevar = parser_find_typedef(parser, parser_tokval(parser), 0); - if (typevar || parser.tok == TOKEN_TYPENAME) { + if (typevar || parser.tok == Token::TYPENAME) { if (!parse_variable(parser, block, true, CV_VAR, typevar, false, false, 0, nullptr)) goto onerr; } @@ -2709,7 +2698,7 @@ static bool parse_break_continue(parser_t &parser, ast_block *block, ast_express parseerror(parser, "`break` can only be used inside loops or switches"); } - if (parser.tok == TOKEN_IDENT) { + if (parser.tok == Token::IDENT) { if (!OPTS_FLAG(LOOP_LABELS)) parseerror(parser, "labeled loops not activated, try using -floop-labels"); i = loops.size(); @@ -2771,7 +2760,7 @@ static bool parse_qualifiers(parser_t &parser, bool with_local, int *cvq, bool * for (;;) { size_t i; - if (parser.tok == TOKEN_ATTRIBUTE_OPEN) { + if (parser.tok == Token::ATTRIBUTE_OPEN) { had_attrib = true; /* parse an attribute */ if (!parser_next(parser)) { @@ -2783,7 +2772,7 @@ static bool parse_qualifiers(parser_t &parser, bool with_local, int *cvq, bool * for (i = 0; i < GMQCC_ARRAY_COUNT(attributes); i++) { if (!strcmp(parser_tokval(parser), attributes[i].name)) { flags |= attributes[i].flag; - if (!parser_next(parser) || parser.tok != TOKEN_ATTRIBUTE_CLOSE) { + if (!parser_next(parser) || parser.tok != Token::ATTRIBUTE_CLOSE) { parseerror(parser, "`%s` attribute has no parameters, expected `]]`", attributes[i].name); *cvq = CV_WRONG; @@ -2799,7 +2788,7 @@ static bool parse_qualifiers(parser_t &parser, bool with_local, int *cvq, bool * if (!strcmp(parser_tokval(parser), "noref")) { had_noref = true; - if (!parser_next(parser) || parser.tok != TOKEN_ATTRIBUTE_CLOSE) { + if (!parser_next(parser) || parser.tok != Token::ATTRIBUTE_CLOSE) { parseerror(parser, "`noref` attribute has no parameters, expected `]]`"); *cvq = CV_WRONG; return false; @@ -2815,7 +2804,7 @@ static bool parse_qualifiers(parser_t &parser, bool with_local, int *cvq, bool * } if (parser.tok == '(') { - if (!parser_next(parser) || parser.tok != TOKEN_STRINGCONST) { + if (!parser_next(parser) || parser.tok != Token::STRINGCONST) { parseerror(parser, "`alias` attribute missing parameter"); goto argerr; } @@ -2838,7 +2827,7 @@ static bool parse_qualifiers(parser_t &parser, bool with_local, int *cvq, bool * } } - if (parser.tok != TOKEN_ATTRIBUTE_CLOSE) { + if (parser.tok != Token::ATTRIBUTE_CLOSE) { parseerror(parser, "`alias` attribute expected `]]`"); goto argerr; } @@ -2853,7 +2842,7 @@ static bool parse_qualifiers(parser_t &parser, bool with_local, int *cvq, bool * } if (parser.tok == '(') { - if (!parser_next(parser) || parser.tok != TOKEN_STRINGCONST) { + if (!parser_next(parser) || parser.tok != Token::STRINGCONST) { parseerror(parser, "`deprecated` attribute missing parameter"); goto argerr; } @@ -2876,7 +2865,7 @@ static bool parse_qualifiers(parser_t &parser, bool with_local, int *cvq, bool * } } /* no message */ - if (parser.tok != TOKEN_ATTRIBUTE_CLOSE) { + if (parser.tok != Token::ATTRIBUTE_CLOSE) { parseerror(parser, "`deprecated` attribute expected `]]`"); argerr: /* ugly */ @@ -2904,7 +2893,7 @@ static bool parse_qualifiers(parser_t &parser, bool with_local, int *cvq, bool * } if (parser.tok != ')') { do { - if (parser.tok != TOKEN_IDENT) + if (parser.tok != Token::IDENT) goto bad_coverage_arg; if (!strcmp(parser_tokval(parser), "block")) flags |= AST_FLAG_BLOCK_COVERAGE; @@ -2931,7 +2920,7 @@ static bool parse_qualifiers(parser_t &parser, bool with_local, int *cvq, bool * { /* Skip tokens until we hit a ]] */ (void)!parsewarning(parser, WARN_UNKNOWN_ATTRIBUTE, "unknown attribute starting with `%s`", parser_tokval(parser)); - while (parser.tok != TOKEN_ATTRIBUTE_CLOSE) { + while (parser.tok != Token::ATTRIBUTE_CLOSE) { if (!parser_next(parser)) { parseerror(parser, "error inside attribute"); *cvq = CV_WRONG; @@ -2994,7 +2983,7 @@ static bool parse_switch(parser_t &parser, ast_block *block, ast_expression **ou if (parser.tok == ':') { if (!OPTS_FLAG(LOOP_LABELS)) parseerror(parser, "labeled loops not activated, try using -floop-labels"); - if (!parser_next(parser) || parser.tok != TOKEN_IDENT) { + if (!parser_next(parser) || parser.tok != Token::IDENT) { parseerror(parser, "expected loop label"); return false; } @@ -3081,9 +3070,9 @@ static bool parse_switch_go(parser_t &parser, ast_block *block, ast_expression * parser_enterblock(parser); while (true) { typevar = nullptr; - if (parser.tok == TOKEN_IDENT) + if (parser.tok == Token::IDENT) typevar = parser_find_typedef(parser, parser_tokval(parser), 0); - if (typevar || parser.tok == TOKEN_TYPENAME) { + if (typevar || parser.tok == Token::TYPENAME) { if (!parse_variable(parser, block, true, CV_NONE, typevar, false, false, 0, nullptr)) { delete switchnode; return false; @@ -3215,7 +3204,7 @@ static bool parse_switch_go(parser_t &parser, ast_block *block, ast_expression * ast_expression *expr; if (parser.tok == '}') break; - if (parser.tok == TOKEN_KEYWORD) { + if (parser.tok == Token::KEYWORD) { if (!strcmp(parser_tokval(parser), "case") || !strcmp(parser_tokval(parser), "default")) { @@ -3295,7 +3284,7 @@ static bool parse_goto(parser_t &parser, ast_expression **out) if (!parser_next(parser)) return false; - if (parser.tok != TOKEN_IDENT) { + if (parser.tok != Token::IDENT) { ast_expression *expression; /* could be an expression i.e computed goto :-) */ @@ -3348,33 +3337,33 @@ static bool parse_skipwhite(parser_t &parser) do { if (!parser_next(parser)) return false; - } while (parser.tok == TOKEN_WHITE && parser.tok < TOKEN_ERROR); - return parser.tok < TOKEN_ERROR; + } while (parser.tok == Token::WHITE && parser.tok < Token::ERROR); + return parser.tok < Token::ERROR; } static bool parse_eol(parser_t &parser) { if (!parse_skipwhite(parser)) return false; - return parser.tok == TOKEN_EOL; + return parser.tok == Token::EOL; } static bool parse_pragma_do(parser_t &parser) { if (!parser_next(parser) || - parser.tok != TOKEN_IDENT || + parser.tok != Token::IDENT || strcmp(parser_tokval(parser), "pragma")) { parseerror(parser, "expected `pragma` keyword after `#`, got `%s`", parser_tokval(parser)); return false; } - if (!parse_skipwhite(parser) || parser.tok != TOKEN_IDENT) { + if (!parse_skipwhite(parser) || parser.tok != Token::IDENT) { parseerror(parser, "expected pragma, got `%s`", parser_tokval(parser)); return false; } if (!strcmp(parser_tokval(parser), "noref")) { - if (!parse_skipwhite(parser) || parser.tok != TOKEN_INTCONST) { + if (!parse_skipwhite(parser) || parser.tok != Token::INTCONST) { parseerror(parser, "`noref` pragma requires an argument: 0 or 1"); return false; } @@ -3404,7 +3393,7 @@ static bool parse_pragma(parser_t &parser) parser.lex->flags.preprocessing = true; parser.lex->flags.mergelines = true; auto ret = parse_pragma_do(parser); - if (parser.tok != TOKEN_EOL) { + if (parser.tok != Token::EOL) { parseerror(parser, "junk after pragma"); ret = false; } @@ -3427,10 +3416,10 @@ static bool parse_statement(parser_t &parser, ast_block *block, ast_expression * *out = nullptr; - if (parser.tok == TOKEN_IDENT) + if (parser.tok == Token::IDENT) typevar = parser_find_typedef(parser, parser_tokval(parser), 0); - if (typevar || parser.tok == TOKEN_TYPENAME || parser.tok == '.' || parser.tok == TOKEN_DOTS) + if (typevar || parser.tok == Token::TYPENAME || parser.tok == '.' || parser.tok == Token::DOTS) { /* local variable */ if (!block) { @@ -3451,7 +3440,7 @@ static bool parse_statement(parser_t &parser, ast_block *block, ast_expression * return false; return parse_variable(parser, block, false, cvq, nullptr, noref, is_static, qflags, vstring); } - else if (parser.tok == TOKEN_KEYWORD) + else if (parser.tok == Token::KEYWORD) { if (!strcmp(parser_tokval(parser), "__builtin_debug_printtype")) { @@ -3463,7 +3452,7 @@ static bool parse_statement(parser_t &parser, ast_block *block, ast_expression * return false; } - if (parser.tok == TOKEN_IDENT && (tdef = parser_find_typedef(parser, parser_tokval(parser), 0))) + if (parser.tok == Token::IDENT && (tdef = parser_find_typedef(parser, parser_tokval(parser), 0))) { ast_type_to_string(tdef, ty, sizeof(ty)); con_out("__builtin_debug_printtype: `%s`=`%s`\n", tdef->m_name.c_str(), ty); @@ -3563,7 +3552,7 @@ static bool parse_statement(parser_t &parser, ast_block *block, ast_expression * parseerror(parser, "expected label name"); return false; } - if (parser.tok != TOKEN_IDENT) { + if (parser.tok != Token::IDENT) { parseerror(parser, "label must be an identifier"); return false; } @@ -3634,7 +3623,7 @@ static bool parse_enum(parser_t &parser) /* enumeration attributes (can add more later) */ if (parser.tok == ':') { - if (!parser_next(parser) || parser.tok != TOKEN_IDENT){ + if (!parser_next(parser) || parser.tok != Token::IDENT){ parseerror(parser, "expected `flag` or `reverse` for enumeration attribute"); return false; } @@ -3659,7 +3648,7 @@ static bool parse_enum(parser_t &parser) } while (true) { - if (!parser_next(parser) || parser.tok != TOKEN_IDENT) { + if (!parser_next(parser) || parser.tok != Token::IDENT) { if (parser.tok == '}') { /* allow an empty enum */ break; @@ -3758,7 +3747,7 @@ static bool parse_block_into(parser_t &parser, ast_block *block) goto cleanup; } - while (parser.tok != TOKEN_EOF && parser.tok < TOKEN_ERROR) + while (parser.tok != Token::END && parser.tok < Token::ERROR) { ast_expression *expr = nullptr; if (parser.tok == '}') @@ -3923,7 +3912,7 @@ static bool parse_function_body(parser_t &parser, ast_value *var) return false; } - if (parser.tok == TOKEN_IDENT && !parser_find_var(parser, parser_tokval(parser))) + if (parser.tok == Token::IDENT && !parser_find_var(parser, parser_tokval(parser))) { /* qc allows the use of not-yet-declared functions here * - this automatically creates a prototype */ @@ -4634,11 +4623,11 @@ static ast_value *parse_parameter_list(parser_t &parser, ast_value *var) if (is_varargs) { /* '...' indicates a varargs function */ variadic = true; - if (parser.tok != ')' && parser.tok != TOKEN_IDENT) { + if (parser.tok != ')' && parser.tok != Token::IDENT) { parseerror(parser, "`...` must be the last parameter of a variadic function declaration"); goto on_error; } - if (parser.tok == TOKEN_IDENT) { + if (parser.tok == Token::IDENT) { argcounter = util_strdup(parser_tokval(parser)); if (!parser_next(parser) || parser.tok != ')') { parseerror(parser, "`...` must be the last parameter of a variadic function declaration"); @@ -4654,15 +4643,15 @@ static ast_value *parse_parameter_list(parser_t &parser, ast_value *var) goto on_error; } /* type-restricted varargs */ - if (parser.tok == TOKEN_DOTS) { + if (parser.tok == Token::DOTS) { variadic = true; varparam = params.back().release(); params.pop_back(); - if (!parser_next(parser) || (parser.tok != ')' && parser.tok != TOKEN_IDENT)) { + if (!parser_next(parser) || (parser.tok != ')' && parser.tok != Token::IDENT)) { parseerror(parser, "`...` must be the last parameter of a variadic function declaration"); goto on_error; } - if (parser.tok == TOKEN_IDENT) { + if (parser.tok == Token::IDENT) { argcounter = util_strdup(parser_tokval(parser)); param->m_name = argcounter; if (!parser_next(parser) || parser.tok != ')') { @@ -4808,14 +4797,14 @@ static ast_value *parse_typename(parser_t &parser, ast_value **storebase, ast_va bool wasarray = false; size_t morefields = 0; - bool vararg = (parser.tok == TOKEN_DOTS); + bool vararg = (parser.tok == Token::DOTS); ctx = parser_ctx(parser); /* types may start with a dot */ - if (parser.tok == '.' || parser.tok == TOKEN_DOTS) { + if (parser.tok == '.' || parser.tok == Token::DOTS) { isfield = true; - if (parser.tok == TOKEN_DOTS) + if (parser.tok == Token::DOTS) morefields += 2; /* if we parsed a dot we need a typename now */ if (!parser_next(parser)) { @@ -4829,7 +4818,7 @@ static ast_value *parse_typename(parser_t &parser, ast_value **storebase, ast_va while (true) { if (parser.tok == '.') ++morefields; - else if (parser.tok == TOKEN_DOTS) + else if (parser.tok == Token::DOTS) morefields += 3; else break; @@ -4840,9 +4829,9 @@ static ast_value *parse_typename(parser_t &parser, ast_value **storebase, ast_va } } } - if (parser.tok == TOKEN_IDENT) + if (parser.tok == Token::IDENT) cached_typedef = parser_find_typedef(parser, parser_tokval(parser), 0); - if (!cached_typedef && parser.tok != TOKEN_TYPENAME) { + if (!cached_typedef && parser.tok != Token::TYPENAME) { if (vararg && is_vararg) { *is_vararg = true; return nullptr; @@ -4898,10 +4887,10 @@ static ast_value *parse_typename(parser_t &parser, ast_value **storebase, ast_va } /* there may be a name now */ - if (parser.tok == TOKEN_IDENT || parser.tok == TOKEN_KEYWORD) { + if (parser.tok == Token::IDENT || parser.tok == Token::KEYWORD) { if (!strcmp(parser_tokval(parser), "break")) (void)!parsewarning(parser, WARN_BREAKDEF, "break definition ignored (suggest removing it)"); - else if (parser.tok == TOKEN_KEYWORD) + else if (parser.tok == Token::KEYWORD) goto leave; name = util_strdup(parser_tokval(parser)); @@ -5664,7 +5653,7 @@ skipvar: /* we only want the integral part anyways */ builtin_num = integral; - } else if (parser.tok == TOKEN_INTCONST) { + } else if (parser.tok == Token::INTCONST) { builtin_num = parser_token(parser)->constval.i; } else { parseerror(parser, "builtin number must be a compile time constant"); @@ -5832,7 +5821,7 @@ another: break; } - if (parser.tok != TOKEN_IDENT) { + if (parser.tok != Token::IDENT) { parseerror(parser, "expected another variable"); break; } @@ -5878,9 +5867,9 @@ cleanup: static bool parser_global_statement(parser_t &parser) { ast_value *istype = nullptr; - if ((parser.tok == TOKEN_IDENT && (istype = parser_find_typedef(parser, parser_tokval(parser), 0)) != nullptr) - || parser.tok == TOKEN_TYPENAME - || parser.tok == '.' || parser.tok == TOKEN_DOTS) { + if ((parser.tok == Token::IDENT && (istype = parser_find_typedef(parser, parser_tokval(parser), 0)) != nullptr) + || parser.tok == Token::TYPENAME + || parser.tok == '.' || parser.tok == Token::DOTS) { return parse_variable(parser, nullptr, false, CV_NONE, istype, false, false, 0, nullptr); } @@ -5895,11 +5884,11 @@ static bool parser_global_statement(parser_t &parser) return parse_variable(parser, nullptr, false, cvq, nullptr, noref, is_static, qflags, vstring); } - if (parser.tok == TOKEN_IDENT && strcmp(parser_tokval(parser), "enum") == 0) { + if (parser.tok == Token::IDENT && strcmp(parser_tokval(parser), "enum") == 0) { return parse_enum(parser); } - if (parser.tok == TOKEN_KEYWORD) { + if (parser.tok == Token::KEYWORD) { if (strcmp(parser_tokval(parser), "typedef") == 0) { if (!parser_next(parser)) { parseerror(parser, "expected type definition after 'typedef'"); @@ -6004,7 +5993,7 @@ static void generate_checksum(parser_t &parser, ir_builder &ir) parser_t::parser_t() : lex(nullptr) - , tok(0) + , tok(Token::NONE) , ast_cleaned(false) , translated(0) , crc_globals(0) @@ -6083,9 +6072,9 @@ static bool parser_compile(parser_t &parser) compile_errors = true; goto cleanup; } - while (parser.tok != TOKEN_EOF && parser.tok < TOKEN_ERROR) { + while (parser.tok != Token::END && parser.tok < Token::ERROR) { if (parser_global_statement(parser)) continue; - if (parser.tok == TOKEN_EOF) { + if (parser.tok == Token::END) { parseerror(parser, "unexpected end of file"); } else if (compile_errors) { parseerror(parser, "there have been errors, bailing out"); diff --git a/parser.h b/parser.h index 1e45de3..4f227b6 100644 --- a/parser.h +++ b/parser.h @@ -7,8 +7,6 @@ #include "intrin.h" #include "fold.h" -struct parser_t; - #define parser_ctx(p) ((p).lex->tok.ctx) struct parser_t { @@ -18,7 +16,7 @@ struct parser_t { void remove_ast(); lex_file *lex; - int tok; + Token tok; bool ast_cleaned; @@ -76,9 +74,21 @@ struct parser_t { intrin m_intrin; }; - /* parser.c */ -char *parser_strdup (const char *str); +inline char *parser_strdup(const char *str) +{ + if (str && !*str) { + /* actually dup empty strings */ + auto *out = reinterpret_cast(mem_a(1)); + *out = 0; + return out; + } + return util_strdup(str); +} ast_expression *parser_find_global(parser_t &parser, const char *name); +parser_t *parser_create(); +bool parser_compile_file(parser_t &parser, const char *); +bool parser_compile_string(parser_t &parser, const char *, const char *, size_t); +bool parser_finish(parser_t &parser, const char *); #endif