X-Git-Url: https://git.xonotic.org/?p=xonotic%2Fgmqcc.git;a=blobdiff_plain;f=lexer.c;h=f1ef767f09d36a4e02cb041dc4bfb7a0a96095b7;hp=3660190946f4b4bfd5b273a0e7a71395e78a17b2;hb=f9b1d057b5247f2ecbaa7259ed1b7bcb56e45922;hpb=36cdffe1b3c8770083803214ec77c06fd773bc8e diff --git a/lexer.c b/lexer.c index 3660190..f1ef767 100644 --- a/lexer.c +++ b/lexer.c @@ -47,7 +47,10 @@ static const char *keywords_fg[] = { "switch", "case", "default", "struct", "union", "break", "continue", - "typedef" + "typedef", + "goto", + + "__builtin_debug_printtype" }; static size_t num_keywords_fg = sizeof(keywords_fg) / sizeof(keywords_fg[0]); @@ -77,14 +80,14 @@ bool lexwarn(lex_file *lex, int warntype, const char *fmt, ...) if (!OPTS_WARN(warntype)) return false; - if (opts_werror) + if (opts.werror) lvl = LVL_ERROR; va_start(ap, fmt); - con_vprintmsg(lvl, lex->name, lex->sline, "warning", fmt, ap); + con_vprintmsg(lvl, lex->name, lex->sline, (opts.werror ? "error" : "warning"), fmt, ap); va_end(ap); - return opts_werror; + return opts.werror; } @@ -438,7 +441,7 @@ static bool lex_try_pragma(lex_file *lex) goto unroll; } - for (ch = lex_getch(lex); vec_size(param) < 32 && ch != ')' && ch != '\n'; ch = lex_getch(lex)) + for (ch = lex_getch(lex); vec_size(param) < 1024 && ch != ')' && ch != '\n'; ch = lex_getch(lex)) vec_push(param, ch); vec_push(param, 0); @@ -489,6 +492,7 @@ unroll: vec_pop(command); } vec_free(command); + lex_ungetch(lex, ' '); } if (command) { vec_pop(command); @@ -497,6 +501,7 @@ unroll: vec_pop(command); } vec_free(command); + lex_ungetch(lex, ' '); } if (pragma) { vec_pop(pragma); @@ -544,10 +549,10 @@ 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) +static int lex_skipwhite(lex_file *lex, bool hadwhite) { int ch = 0; - bool haswhite = false; + bool haswhite = hadwhite; do { @@ -739,6 +744,7 @@ static bool lex_finish_frames(lex_file *lex) static int GMQCC_WARN lex_finish_string(lex_file *lex, int quote) { int ch = 0; + int nextch; while (ch != EOF) { @@ -775,7 +781,61 @@ static int GMQCC_WARN lex_finish_string(lex_file *lex, int quote) case 't': ch = '\t'; break; case 'f': ch = '\f'; break; case 'v': ch = '\v'; break; + case 'x': + case 'X': + /* same procedure as in fteqcc */ + ch = 0; + nextch = lex_getch(lex); + if (nextch >= '0' && nextch <= '9') + ch += nextch - '0'; + else if (nextch >= 'a' && nextch <= 'f') + ch += nextch - 'a' + 10; + else if (nextch >= 'A' && nextch <= 'F') + ch += nextch - 'A' + 10; + else { + lexerror(lex, "bad character code"); + lex_ungetch(lex, nextch); + return (lex->tok.ttype = TOKEN_ERROR); + } + + ch *= 0x10; + nextch = lex_getch(lex); + if (nextch >= '0' && nextch <= '9') + ch += nextch - '0'; + else if (nextch >= 'a' && nextch <= 'f') + ch += nextch - 'a' + 10; + else if (nextch >= 'A' && nextch <= 'F') + ch += nextch - 'A' + 10; + else { + lexerror(lex, "bad character code"); + lex_ungetch(lex, nextch); + return (lex->tok.ttype = TOKEN_ERROR); + } + break; + + /* fteqcc support */ + case '0': case '1': case '2': case '3': + case '4': case '5': case '6': case '7': + case '8': case '9': + ch = 18 + ch - '0'; + break; + case '<': ch = 29; break; + case '-': ch = 30; break; + case '>': ch = 31; break; + case '[': ch = 16; break; + case ']': ch = 17; break; + case '{': + ch = 0; + for (nextch = lex_getch(lex); nextch != '}'; nextch = lex_getch(lex)) { + ch = ch * 10 + nextch - '0'; + if (nextch < '0' || nextch > '9' || ch > 255) { + lexerror(lex, "bad character code"); + return (lex->tok.ttype = TOKEN_ERROR); + } + } + break; case '\n': ch = '\n'; break; + default: lexwarn(lex, WARN_UNKNOWN_CONTROL_SEQUENCE, "unrecognized control sequence: \\%c", ch); /* so we just add the character plus backslash no matter what it actually is */ @@ -799,7 +859,10 @@ static int GMQCC_WARN lex_finish_digit(lex_file *lex, int lastch) int ch = lastch; /* parse a number... */ - lex->tok.ttype = TOKEN_INTCONST; + if (ch == '.') + lex->tok.ttype = TOKEN_FLOATCONST; + else + lex->tok.ttype = TOKEN_INTCONST; lex_tokench(lex, ch); @@ -832,7 +895,7 @@ static int GMQCC_WARN lex_finish_digit(lex_file *lex, int lastch) } } /* NOT else, '.' can come from above as well */ - if (ch == '.' && !ishex) + if (lex->tok.ttype != TOKEN_FLOATCONST && ch == '.' && !ishex) { /* Allow floating comma in non-hex mode */ lex->tok.ttype = TOKEN_FLOATCONST; @@ -869,6 +932,7 @@ static int GMQCC_WARN lex_finish_digit(lex_file *lex, int lastch) int lex_do(lex_file *lex) { int ch, nextch, thirdch; + bool hadwhite = false; lex_token_new(lex); #if 0 @@ -877,7 +941,8 @@ int lex_do(lex_file *lex) #endif while (true) { - ch = lex_skipwhite(lex); + ch = lex_skipwhite(lex, hadwhite); + hadwhite = true; if (!lex->flags.mergelines || ch != '\\') break; ch = lex_getch(lex); @@ -908,7 +973,7 @@ int lex_do(lex_file *lex) } /* modelgen / spiritgen commands */ - if (ch == '$') { + if (ch == '$' && !lex->flags.preprocessing) { const char *v; size_t frame; @@ -1077,6 +1142,18 @@ int lex_do(lex_file *lex) break; } + if (ch == '.') { + nextch = lex_getch(lex); + /* digits starting with a dot */ + if (isdigit(nextch)) { + lex_ungetch(lex, nextch); + lex->tok.ttype = lex_finish_digit(lex, ch); + lex_endtoken(lex); + return lex->tok.ttype; + } + lex_ungetch(lex, nextch); + } + if (lex->flags.noops) { /* Detect characters early which are normally @@ -1146,7 +1223,7 @@ int lex_do(lex_file *lex) lex_tokench(lex, ch); nextch = lex_getch(lex); - if (nextch == ch || nextch == '=') { + if (nextch == '=' || (nextch == ch && ch != '!')) { lex_tokench(lex, nextch); } else if (ch == '-' && nextch == '>') { lex_tokench(lex, nextch); @@ -1227,7 +1304,7 @@ int lex_do(lex_file *lex) if (!strcmp(v, keywords_qc[kw])) return (lex->tok.ttype = TOKEN_KEYWORD); } - if (opts_standard != COMPILER_QCC) { + if (opts.standard != COMPILER_QCC) { for (kw = 0; kw < num_keywords_fg; ++kw) { if (!strcmp(v, keywords_fg[kw])) return (lex->tok.ttype = TOKEN_KEYWORD); @@ -1249,7 +1326,7 @@ int lex_do(lex_file *lex) while (!lex->flags.preprocessing && lex->tok.ttype == TOKEN_STRINGCONST) { /* Allow c style "string" "continuation" */ - ch = lex_skipwhite(lex); + ch = lex_skipwhite(lex, false); if (ch != '"') { lex_ungetch(lex, ch); break; @@ -1276,6 +1353,7 @@ int lex_do(lex_file *lex) lex_tokench(lex, ch); lex_endtoken(lex); + lex->tok.ttype = TOKEN_CHARCONST; /* It's a vector if we can successfully scan 3 floats */ #ifdef WIN32 if (sscanf_s(lex->tok.value, " %f %f %f ", @@ -1288,6 +1366,14 @@ int lex_do(lex_file *lex) { lex->tok.ttype = TOKEN_VECTORCONST; } + else + { + if (!lex->flags.preprocessing && strlen(lex->tok.value) > 1) { + if (lexwarn(lex, WARN_MULTIBYTE_CHARACTER, "multibyte character: `%s`", lex->tok.value)) + return (lex->tok.ttype = TOKEN_ERROR); + } + lex->tok.constval.i = lex->tok.value[0]; + } return lex->tok.ttype; } @@ -1299,6 +1385,12 @@ int lex_do(lex_file *lex) return lex->tok.ttype; } + if (lex->flags.preprocessing) { + lex_tokench(lex, ch); + lex_endtoken(lex); + return (lex->tok.ttype = ch); + } + lexerror(lex, "unknown token"); return (lex->tok.ttype = TOKEN_ERROR); }