X-Git-Url: https://git.xonotic.org/?p=xonotic%2Fgmqcc.git;a=blobdiff_plain;f=lexer.c;h=42b21a1a6cd7de46bd45aa21b227cc412a4fbdec;hp=513499a708bbe35a23c4e68f20cc8f09952b022c;hb=e11a17b40805b93e1e2ae80129c9d235b6d6119b;hpb=990dcb106a86ae880f55471412ff51356d56465c

diff --git a/lexer.c b/lexer.c
index 513499a..42b21a1 100644
--- a/lexer.c
+++ b/lexer.c
@@ -1,3 +1,25 @@
+/*
+ * Copyright (C) 2012, 2013
+ *     Wolfgang Bumiller
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy of
+ * this software and associated documentation files (the "Software"), to deal in
+ * the Software without restriction, including without limitation the rights to
+ * use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies
+ * of the Software, and to permit persons to whom the Software is furnished to do
+ * so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in all
+ * copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
@@ -6,36 +28,63 @@
 #include "gmqcc.h"
 #include "lexer.h"
 
+/*
+ * List of Keywords
+ */
+
+/* original */
+static const char *keywords_qc[] = {
+    "for", "do", "while",
+    "if", "else",
+    "local",
+    "return",
+    "const"
+};
+static size_t num_keywords_qc = sizeof(keywords_qc) / sizeof(keywords_qc[0]);
+
+/* For fte/gmgqcc */
+static const char *keywords_fg[] = {
+    "switch", "case", "default",
+    "struct", "union",
+    "break", "continue",
+    "typedef",
+    "goto",
+
+    "__builtin_debug_printtype"
+};
+static size_t num_keywords_fg = sizeof(keywords_fg) / sizeof(keywords_fg[0]);
+
+/*
+ * Lexer code
+ */
+
 char* *lex_filenames;
 
 void lexerror(lex_file *lex, const char *fmt, ...)
 {
-	va_list ap;
+    va_list ap;
 
-	va_start(ap, fmt);
-	if (lex)
+    va_start(ap, fmt);
+    if (lex)
         con_vprintmsg(LVL_ERROR, lex->name, lex->sline, "parse error", fmt, ap);
     else
         con_vprintmsg(LVL_ERROR, "", 0, "parse error", fmt, ap);
-	va_end(ap);
+    va_end(ap);
 }
 
 bool lexwarn(lex_file *lex, int warntype, const char *fmt, ...)
 {
-	va_list ap;
-	int lvl = LVL_WARNING;
-
-    if (!OPTS_WARN(warntype))
-        return false;
-
-    if (opts_werror)
-	    lvl = LVL_ERROR;
+    bool    r;
+    lex_ctx ctx;
+    va_list ap;
 
-	va_start(ap, fmt);
-    con_vprintmsg(lvl, lex->name, lex->sline, "warning", fmt, ap);
-	va_end(ap);
+    ctx.file = lex->name;
+    ctx.line = lex->sline;
 
-	return opts_werror;
+    va_start(ap, fmt);
+    r = vcompile_warning(ctx, warntype, fmt, ap);
+    va_end(ap);
+    return r;
 }
 
 
@@ -135,7 +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);
@@ -144,7 +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;
     }
@@ -209,7 +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);
@@ -242,12 +291,16 @@ static int lex_try_trigraph(lex_file *lex, int old)
 {
     int c2, c3;
     c2 = lex_fgetc(lex);
+    if (!lex->push_line && c2 == '\n')
+        lex->line++;
     if (c2 != '?') {
         lex_ungetch(lex, c2);
         return old;
     }
 
     c3 = lex_fgetc(lex);
+    if (!lex->push_line && c3 == '\n')
+        lex->line++;
     switch (c3) {
         case '=': return '#';
         case '/': return '\\';
@@ -269,6 +322,11 @@ 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 == '>')
@@ -289,13 +347,13 @@ static int lex_getch(lex_file *lex)
 
     if (lex->peekpos) {
         lex->peekpos--;
-        if (lex->peek[lex->peekpos] == '\n')
+        if (!lex->push_line && lex->peek[lex->peekpos] == '\n')
             lex->line++;
         return lex->peek[lex->peekpos];
     }
 
     ch = lex_fgetc(lex);
-    if (ch == '\n')
+    if (!lex->push_line && ch == '\n')
         lex->line++;
     else if (ch == '?')
         return lex_try_trigraph(lex, ch);
@@ -307,7 +365,7 @@ static int lex_getch(lex_file *lex)
 static void lex_ungetch(lex_file *lex, int ch)
 {
     lex->peek[lex->peekpos++] = ch;
-    if (ch == '\n')
+    if (!lex->push_line && ch == '\n')
         lex->line--;
 }
 
@@ -347,6 +405,119 @@ static void lex_endtoken(lex_file *lex)
     vec_shrinkby(lex->tok.value, 1);
 }
 
+static bool lex_try_pragma(lex_file *lex)
+{
+    int ch;
+    char *pragma  = NULL;
+    char *command = NULL;
+    char *param   = NULL;
+    size_t line;
+
+    if (lex->flags.preprocessing)
+        return false;
+
+    line = lex->line;
+
+    ch = lex_getch(lex);
+    if (ch != '#') {
+        lex_ungetch(lex, ch);
+        return false;
+    }
+
+    for (ch = lex_getch(lex); vec_size(pragma) < 8 && ch >= 'a' && ch <= 'z'; ch = lex_getch(lex))
+        vec_push(pragma, ch);
+    vec_push(pragma, 0);
+
+    if (ch != ' ' || strcmp(pragma, "pragma")) {
+        lex_ungetch(lex, ch);
+        goto unroll;
+    }
+
+    for (ch = lex_getch(lex); vec_size(command) < 32 && ch >= 'a' && ch <= 'z'; ch = lex_getch(lex))
+        vec_push(command, ch);
+    vec_push(command, 0);
+
+    if (ch != '(') {
+        lex_ungetch(lex, ch);
+        goto unroll;
+    }
+
+    for (ch = lex_getch(lex); vec_size(param) < 1024 && ch != ')' && ch != '\n'; ch = lex_getch(lex))
+        vec_push(param, ch);
+    vec_push(param, 0);
+
+    if (ch != ')') {
+        lex_ungetch(lex, ch);
+        goto unroll;
+    }
+
+    if (!strcmp(command, "push")) {
+        if (!strcmp(param, "line")) {
+            lex->push_line++;
+            if (lex->push_line == 1)
+                --line;
+        }
+        else
+            goto unroll;
+    }
+    else if (!strcmp(command, "pop")) {
+        if (!strcmp(param, "line")) {
+            if (lex->push_line)
+                lex->push_line--;
+            if (lex->push_line == 0)
+                --line;
+        }
+        else
+            goto unroll;
+    }
+    else if (!strcmp(command, "file")) {
+        lex->name = util_strdup(param);
+        vec_push(lex_filenames, lex->name);
+    }
+    else if (!strcmp(command, "line")) {
+        line = strtol(param, NULL, 0)-1;
+    }
+    else
+        goto unroll;
+
+    lex->line = line;
+    while (ch != '\n' && ch != EOF)
+        ch = lex_getch(lex);
+    return true;
+
+unroll:
+    if (command) {
+        vec_pop(command);
+        while (vec_size(command)) {
+            lex_ungetch(lex, (unsigned char)vec_last(command));
+            vec_pop(command);
+        }
+        vec_free(command);
+        lex_ungetch(lex, ' ');
+    }
+    if (command) {
+        vec_pop(command);
+        while (vec_size(command)) {
+            lex_ungetch(lex, (unsigned char)vec_last(command));
+            vec_pop(command);
+        }
+        vec_free(command);
+        lex_ungetch(lex, ' ');
+    }
+    if (pragma) {
+        vec_pop(pragma);
+        while (vec_size(pragma)) {
+            lex_ungetch(lex, (unsigned char)vec_last(pragma));
+            vec_pop(pragma);
+        }
+        vec_free(pragma);
+    }
+    lex_ungetch(lex, '#');
+
+    lex->line = line;
+    return false;
+}
+
 /* Skip whitespace and comments and return the first
  * non-white character.
  * As this makes use of the above getch() ungetch() functions,
@@ -379,15 +550,19 @@ 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
     {
         ch = lex_getch(lex);
         while (ch != EOF && isspace(ch)) {
+            if (ch == '\n') {
+                if (lex_try_pragma(lex))
+                    continue;
+            }
             if (lex->flags.preprocessing) {
                 if (ch == '\n') {
                     /* end-of-line */
@@ -415,13 +590,17 @@ static int lex_skipwhite(lex_file *lex)
 
                 if (lex->flags.preprocessing) {
                     haswhite = true;
+                    /*
                     lex_tokench(lex, '/');
                     lex_tokench(lex, '/');
+                    */
+                    lex_tokench(lex, ' ');
+                    lex_tokench(lex, ' ');
                 }
 
                 while (ch != EOF && ch != '\n') {
                     if (lex->flags.preprocessing)
-                        lex_tokench(lex, ch);
+                        lex_tokench(lex, ' '); /* ch); */
                     ch = lex_getch(lex);
                 }
                 if (lex->flags.preprocessing) {
@@ -436,8 +615,12 @@ static int lex_skipwhite(lex_file *lex)
                 /* multiline comment */
                 if (lex->flags.preprocessing) {
                     haswhite = true;
+                    /*
                     lex_tokench(lex, '/');
                     lex_tokench(lex, '*');
+                    */
+                    lex_tokench(lex, ' ');
+                    lex_tokench(lex, ' ');
                 }
 
                 while (ch != EOF)
@@ -447,14 +630,22 @@ static int lex_skipwhite(lex_file *lex)
                         ch = lex_getch(lex);
                         if (ch == '/') {
                             if (lex->flags.preprocessing) {
+                                /*
                                 lex_tokench(lex, '*');
                                 lex_tokench(lex, '/');
+                                */
+                                lex_tokench(lex, ' ');
+                                lex_tokench(lex, ' ');
                             }
                             break;
                         }
+                        lex_ungetch(lex, ch);
                     }
                     if (lex->flags.preprocessing) {
-                        lex_tokench(lex, ch);
+                        if (ch == '\n')
+                            lex_tokench(lex, '\n');
+                        else
+                            lex_tokench(lex, ' '); /* ch); */
                     }
                 }
                 ch = ' '; /* cause TRUE in the isspace check */
@@ -549,11 +740,18 @@ static bool lex_finish_frames(lex_file *lex)
         vec_shrinkto(lex->tok.value, 0);
         vec_push(lex->frames, m);
     } while (true);
+
+    return false;
 }
 
 static int GMQCC_WARN lex_finish_string(lex_file *lex, int quote)
 {
+    uchar_t chr;
     int ch = 0;
+    int nextch;
+    bool hex;
+    char u8buf[8]; /* way more than enough */
+    int  u8len, uc;
 
     while (ch != EOF)
     {
@@ -561,7 +759,17 @@ static int GMQCC_WARN lex_finish_string(lex_file *lex, int quote)
         if (ch == quote)
             return TOKEN_STRINGCONST;
 
-        if (!lex->flags.preprocessing && ch == '\\') {
+        if (lex->flags.preprocessing && ch == '\\') {
+            lex_tokench(lex, ch);
+            ch = lex_getch(lex);
+            if (ch == EOF) {
+                lexerror(lex, "unexpected end of file");
+                lex_ungetch(lex, EOF); /* next token to be TOKEN_EOF */
+                return (lex->tok.ttype = TOKEN_ERROR);
+            }
+            lex_tokench(lex, ch);
+        }
+        else if (ch == '\\') {
             ch = lex_getch(lex);
             if (ch == EOF) {
                 lexerror(lex, "unexpected end of file");
@@ -571,6 +779,8 @@ static int GMQCC_WARN lex_finish_string(lex_file *lex, int quote)
 
             switch (ch) {
             case '\\': break;
+            case '\'': break;
+            case '"':  break;
             case 'a':  ch = '\a'; break;
             case 'b':  ch = '\b'; break;
             case 'r':  ch = '\r'; break;
@@ -578,6 +788,100 @@ 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 '{':
+                chr = 0;
+                nextch = lex_getch(lex);
+                hex = (nextch == 'x');
+                if (!hex)
+                    lex_ungetch(lex, nextch);
+                for (nextch = lex_getch(lex); nextch != '}'; nextch = lex_getch(lex)) {
+                    if (!hex) {
+                        if (nextch >= '0' && nextch <= '9')
+                            chr = chr * 10 + nextch - '0';
+                        else {
+                            lexerror(lex, "bad character code");
+                            return (lex->tok.ttype = TOKEN_ERROR);
+                        }
+                    } else {
+                        if (nextch >= '0' && nextch <= '9')
+                            chr = chr * 0x10 + nextch - '0';
+                        else if (nextch >= 'a' && nextch <= 'f')
+                            chr = chr * 0x10 + nextch - 'a' + 10;
+                        else if (nextch >= 'A' && nextch <= 'F')
+                            chr = chr * 0x10 + nextch - 'A' + 10;
+                        else {
+                            lexerror(lex, "bad character code");
+                            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);
+                    }
+                }
+                if (OPTS_FLAG(UTF8) && chr >= 128) {
+                    u8len = u8_fromchar(chr, u8buf, sizeof(u8buf));
+                    if (!u8len)
+                        ch = 0;
+                    else {
+                        --u8len;
+                        for (uc = 0; uc < u8len; ++uc)
+                            lex_tokench(lex, u8buf[uc]);
+                        /* the last character will be inserted with the tokench() call
+                         * below the switch
+                         */
+                        ch = u8buf[uc];
+                    }
+                }
+                else
+                    ch = chr;
+                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 */
@@ -601,7 +905,10 @@ 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);
 
@@ -634,7 +941,7 @@ 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;
@@ -670,7 +977,8 @@ static int GMQCC_WARN lex_finish_digit(lex_file *lex, int lastch)
 
 int lex_do(lex_file *lex)
 {
-    int ch, nextch;
+    int ch, nextch, thirdch;
+    bool hadwhite = false;
 
     lex_token_new(lex);
 #if 0
@@ -678,15 +986,30 @@ int lex_do(lex_file *lex)
         return TOKEN_FATAL;
 #endif
 
-    ch = lex_skipwhite(lex);
-    lex->sline = lex->line;
-    lex->tok.ctx.line = lex->sline;
-    lex->tok.ctx.file = lex->name;
+    while (true) {
+        ch = lex_skipwhite(lex, hadwhite);
+        hadwhite = true;
+        if (!lex->flags.mergelines || ch != '\\')
+            break;
+        ch = lex_getch(lex);
+        if (ch != '\n') {
+            lex_ungetch(lex, ch);
+            ch = '\\';
+            break;
+        }
+        /* we reached a linemerge */
+        lex_tokench(lex, '\n');
+        continue;
+    }
 
     if (lex->flags.preprocessing && (ch == TOKEN_WHITE || ch == TOKEN_EOL || ch == TOKEN_FATAL)) {
         return (lex->tok.ttype = ch);
     }
 
+    lex->sline = lex->line;
+    lex->tok.ctx.line = lex->sline;
+    lex->tok.ctx.file = lex->name;
+
     if (lex->eof)
         return (lex->tok.ttype = TOKEN_FATAL);
 
@@ -696,7 +1019,7 @@ int lex_do(lex_file *lex)
     }
 
     /* modelgen / spiritgen commands */
-    if (ch == '$') {
+    if (ch == '$' && !lex->flags.preprocessing) {
         const char *v;
         size_t frame;
 
@@ -787,7 +1110,6 @@ 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;
@@ -802,9 +1124,9 @@ int lex_do(lex_file *lex)
 
         if (!strcmp(v, "flush"))
         {
-            size_t frame;
-            for (frame = 0; frame < vec_size(lex->frames); ++frame)
-                mem_d(lex->frames[frame].name);
+            size_t fi;
+            for (fi = 0; fi < vec_size(lex->frames); ++fi)
+                mem_d(lex->frames[fi].name);
             vec_free(lex->frames);
             /* skip line (fteqcc does it too) */
             ch = lex_getch(lex);
@@ -842,18 +1164,41 @@ 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 '?':
             lex_tokench(lex, ch);
             lex_endtoken(lex);
             if (lex->flags.noops)
                 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);
@@ -863,6 +1208,18 @@ 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
@@ -870,8 +1227,10 @@ int lex_do(lex_file *lex)
          */
         switch (ch)
         {
+            /*
             case '+':
             case '-':
+            */
             case '*':
             case '/':
             case '<':
@@ -889,31 +1248,37 @@ int lex_do(lex_file *lex)
             default:
                 break;
         }
+    }
 
-        if (ch == '.')
-        {
-            lex_tokench(lex, ch);
-            /* peak ahead once */
-            nextch = lex_getch(lex);
-            if (nextch != '.') {
-                lex_ungetch(lex, nextch);
-                lex_endtoken(lex);
-                return (lex->tok.ttype = ch);
-            }
-            /* peak ahead again */
-            nextch = lex_getch(lex);
-            if (nextch != '.') {
-                lex_ungetch(lex, nextch);
-                lex_ungetch(lex, nextch);
-                lex_endtoken(lex);
+    if (ch == '.')
+    {
+        lex_tokench(lex, ch);
+        /* peak ahead once */
+        nextch = lex_getch(lex);
+        if (nextch != '.') {
+            lex_ungetch(lex, nextch);
+            lex_endtoken(lex);
+            if (lex->flags.noops)
                 return (lex->tok.ttype = ch);
-            }
-            /* fill the token to be "..." */
-            lex_tokench(lex, ch);
-            lex_tokench(lex, ch);
+            else
+                return (lex->tok.ttype = TOKEN_OPERATOR);
+        }
+        /* peak ahead again */
+        nextch = lex_getch(lex);
+        if (nextch != '.') {
+            lex_ungetch(lex, nextch);
+            lex_ungetch(lex, '.');
             lex_endtoken(lex);
-            return (lex->tok.ttype = TOKEN_DOTS);
+            if (lex->flags.noops)
+                return (lex->tok.ttype = ch);
+            else
+                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);
     }
 
     if (ch == ',' || ch == '.') {
@@ -923,17 +1288,39 @@ int lex_do(lex_file *lex)
     }
 
     if (ch == '+' || ch == '-' || /* ++, --, +=, -=  and -> as well! */
-        ch == '>' || ch == '<' || /* <<, >>, <=, >= */
-        ch == '=' || ch == '!' || /* ==, != */
-        ch == '&' || ch == '|')   /* &&, ||, &=, |= */
-    {
+        ch == '>' || ch == '<' || /* <<, >>, <=, >=                  */
+        ch == '=' || ch == '!' || /* ==, !=                          */
+        ch == '&' || ch == '|' || /* &&, ||, &=, |=                  */
+        ch == '~'                 /* ~=, ~                           */
+    )  {
         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);
+        } else if (ch == '&' && nextch == '~') {
+            thirdch = lex_getch(lex);
+            if (thirdch != '=') {
+                lex_ungetch(lex, thirdch);
+                lex_ungetch(lex, nextch);
+            }
+            else {
+                lex_tokench(lex, nextch);
+                lex_tokench(lex, thirdch);
+            }
+        }
+        else if (lex->flags.preprocessing &&
+                 ch == '-' && isdigit(nextch))
+        {
+            lex->tok.ttype = lex_finish_digit(lex, nextch);
+            if (lex->tok.ttype == TOKEN_INTCONST)
+                lex->tok.constval.i = -lex->tok.constval.i;
+            else
+                lex->tok.constval.f = -lex->tok.constval.f;
+            lex_endtoken(lex);
+            return lex->tok.ttype;
         } else
             lex_ungetch(lex, nextch);
 
@@ -995,29 +1382,17 @@ int lex_do(lex_file *lex)
         } else if (!strcmp(v, "vector")) {
             lex->tok.ttype = TOKEN_TYPENAME;
             lex->tok.constval.t = TYPE_VECTOR;
-        } else if (!strcmp(v, "for")  ||
-                 !strcmp(v, "while")  ||
-                 !strcmp(v, "do")     ||
-                 !strcmp(v, "if")     ||
-                 !strcmp(v, "else")   ||
-                 !strcmp(v, "local")  ||
-                 !strcmp(v, "return") ||
-                 !strcmp(v, "not")    ||
-                 !strcmp(v, "const"))
-        {
-            lex->tok.ttype = TOKEN_KEYWORD;
-        }
-        else if (opts_standard != COMPILER_QCC)
-        {
-            /* other standards reserve these keywords */
-            if (!strcmp(v, "switch") ||
-                !strcmp(v, "struct") ||
-                !strcmp(v, "union")  ||
-                !strcmp(v, "break")  ||
-                !strcmp(v, "continue") ||
-                !strcmp(v, "var"))
-            {
-                lex->tok.ttype = TOKEN_KEYWORD;
+        } else {
+            size_t kw;
+            for (kw = 0; kw < num_keywords_qc; ++kw) {
+                if (!strcmp(v, keywords_qc[kw]))
+                    return (lex->tok.ttype = TOKEN_KEYWORD);
+            }
+            if (OPTS_OPTION_U32(OPTION_STANDARD) != COMPILER_QCC) {
+                for (kw = 0; kw < num_keywords_fg; ++kw) {
+                    if (!strcmp(v, keywords_fg[kw]))
+                        return (lex->tok.ttype = TOKEN_KEYWORD);
+                }
             }
         }
 
@@ -1035,7 +1410,7 @@ 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;
@@ -1062,8 +1437,9 @@ 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
@@ -1074,6 +1450,24 @@ int lex_do(lex_file *lex)
         {
              lex->tok.ttype = TOKEN_VECTORCONST;
         }
+        else
+        {
+            if (!lex->flags.preprocessing && strlen(lex->tok.value) > 1) {
+                uchar_t u8char;
+                /* check for a valid utf8 character */
+                if (!OPTS_FLAG(UTF8) || !u8_analyze(lex->tok.value, NULL, NULL, &u8char, 8)) {
+                    if (lexwarn(lex, WARN_MULTIBYTE_CHARACTER,
+                                ( OPTS_FLAG(UTF8) ? "invalid multibyte character sequence `%s`"
+                                                  : "multibyte character: `%s`" ),
+                                lex->tok.value))
+                        return (lex->tok.ttype = TOKEN_ERROR);
+                }
+                else
+                    lex->tok.constval.i = u8char;
+            }
+            else
+                lex->tok.constval.i = lex->tok.value[0];
+        }
 
         return lex->tok.ttype;
     }
@@ -1085,6 +1479,12 @@ int lex_do(lex_file *lex)
         return lex->tok.ttype;
     }
 
-    lexerror(lex, "unknown token");
+    if (lex->flags.preprocessing) {
+        lex_tokench(lex, ch);
+        lex_endtoken(lex);
+        return (lex->tok.ttype = ch);
+    }
+
+    lexerror(lex, "unknown token: `%s`", lex->tok.value);
     return (lex->tok.ttype = TOKEN_ERROR);
 }