X-Git-Url: https://git.xonotic.org/?p=xonotic%2Fgmqcc.git;a=blobdiff_plain;f=lexer.c;h=e506aec31ef8c105f7c3ef5de7a74eefd560d809;hp=17af7b6f4b92ad7ca30c2da8bd4f5a2161afbb7b;hb=6bc29a1601b6326dad807ac4f1d6db9e37adb3b6;hpb=f65dbda6595f5fe33f263767c37e78871770e06c diff --git a/lexer.c b/lexer.c index 17af7b6..e506aec 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]); @@ -71,20 +74,17 @@ void lexerror(lex_file *lex, const char *fmt, ...) bool lexwarn(lex_file *lex, int warntype, const char *fmt, ...) { + bool r; + lex_ctx ctx; va_list ap; - int lvl = LVL_WARNING; - - if (!OPTS_WARN(warntype)) - return false; - if (opts_werror) - lvl = LVL_ERROR; + ctx.file = lex->name; + ctx.line = lex->sline; va_start(ap, fmt); - con_vprintmsg(lvl, lex->name, lex->sline, "warning", fmt, ap); + r = vcompile_warning(ctx, warntype, fmt, ap); va_end(ap); - - return opts_werror; + return r; } @@ -184,7 +184,7 @@ static void lex_token_new(lex_file *lex) lex_file* lex_open(const char *file) { lex_file *lex; - FILE *in = util_fopen(file, "rb"); + FILE *in = file_open(file, "rb"); if (!in) { lexerror(NULL, "open failed: '%s'\n", file); @@ -193,7 +193,7 @@ lex_file* lex_open(const char *file) lex = (lex_file*)mem_a(sizeof(*lex)); if (!lex) { - fclose(in); + file_close(in); lexerror(NULL, "out of memory\n"); return NULL; } @@ -258,7 +258,7 @@ void lex_close(lex_file *lex) vec_free(lex->modelname); if (lex->file) - fclose(lex->file); + file_close(lex->file); #if 0 if (lex->tok) token_delete(lex->tok); @@ -318,6 +318,11 @@ static int lex_try_digraph(lex_file *lex, int ch) { int c2; 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') + lex->line++; if (ch == '<' && c2 == ':') return '['; else if (ch == ':' && c2 == '>') @@ -433,7 +438,7 @@ 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); @@ -484,6 +489,7 @@ unroll: vec_pop(command); } vec_free(command); + lex_ungetch(lex, ' '); } if (command) { vec_pop(command); @@ -492,6 +498,7 @@ unroll: vec_pop(command); } vec_free(command); + lex_ungetch(lex, ' '); } if (pragma) { vec_pop(pragma); @@ -539,10 +546,10 @@ 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 { @@ -734,6 +741,7 @@ 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) { @@ -770,6 +778,61 @@ 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 */ @@ -793,7 +856,10 @@ 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); @@ -826,7 +892,7 @@ 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; @@ -863,6 +929,7 @@ 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 @@ -871,7 +938,8 @@ 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); @@ -902,7 +970,7 @@ int lex_do(lex_file *lex) } /* modelgen / spiritgen commands */ - if (ch == '$') { + if (ch == '$' && !lex->flags.preprocessing) { const char *v; size_t frame; @@ -993,7 +1061,6 @@ int lex_do(lex_file *lex) if (rc < 0) return (lex->tok.ttype = TOKEN_FATAL); - v = lex->tok.value; if (lex->modelname) { frame_macro m; m.value = lex->framevalue; @@ -1048,6 +1115,15 @@ int lex_do(lex_file *lex) switch (ch) { case '[': + nextch = lex_getch(lex); + if (nextch == '[') { + lex_tokench(lex, ch); + lex_tokench(lex, nextch); + lex_endtoken(lex); + return (lex->tok.ttype = TOKEN_ATTRIBUTE_OPEN); + } + lex_ungetch(lex, nextch); + /* FALL THROUGH */ case '(': case ':': case '?': @@ -1057,11 +1133,23 @@ int lex_do(lex_file *lex) return (lex->tok.ttype = ch); else return (lex->tok.ttype = TOKEN_OPERATOR); + + case ']': + if (lex->flags.noops) { + nextch = lex_getch(lex); + if (nextch == ']') { + lex_tokench(lex, ch); + lex_tokench(lex, nextch); + lex_endtoken(lex); + return (lex->tok.ttype = TOKEN_ATTRIBUTE_CLOSE); + } + lex_ungetch(lex, nextch); + } + /* FALL THROUGH */ case ')': case ';': case '{': case '}': - case ']': case '#': lex_tokench(lex, ch); @@ -1071,6 +1159,18 @@ 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 @@ -1140,7 +1240,7 @@ 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); @@ -1221,7 +1321,7 @@ 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); @@ -1243,7 +1343,7 @@ 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; @@ -1270,8 +1370,9 @@ 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 +#ifdef _MSC_VER if (sscanf_s(lex->tok.value, " %f %f %f ", &lex->tok.constval.v.x, &lex->tok.constval.v.y, &lex->tok.constval.v.z) == 3) #else @@ -1282,6 +1383,14 @@ 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; } @@ -1293,6 +1402,12 @@ 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); }