Initial platform / compiler specific code refactoring.

[xonotic/gmqcc.git] / lexer.c

diff --git a/lexer.c b/lexer.c
index a00814dc093e8fef32e3df97f4e01db419fe7822..2d36afbc8f639d83cb52202f16f9dd9320701841 100644 (file)
--- a/lexer.c
+++ b/lexer.c
@@ -37,8 +37,6 @@ static const char *keywords_qc[] = {
     "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",
@@ -49,12 +47,10 @@ static const char *keywords_fg[] = {
 
     "__builtin_debug_printtype"
 };
-static size_t num_keywords_fg = sizeof(keywords_fg) / sizeof(keywords_fg[0]);
 
 /*
  * Lexer code
  */
-
 static char* *lex_filenames;
 
 static void lexerror(lex_file *lex, const char *fmt, ...)
@@ -63,7 +59,7 @@ static void lexerror(lex_file *lex, const char *fmt, ...)
 
     va_start(ap, fmt);
     if (lex)
-        con_vprintmsg(LVL_ERROR, lex->ctx.file, lex->sline, lex->ctx.column, "parse error", fmt, ap);
+        con_vprintmsg(LVL_ERROR, lex->name, lex->sline, lex->column, "parse error", fmt, ap);
     else
         con_vprintmsg(LVL_ERROR, "", 0, 0, "parse error", fmt, ap);
     va_end(ap);
@@ -71,11 +67,16 @@ static void lexerror(lex_file *lex, const char *fmt, ...)
 
 static bool lexwarn(lex_file *lex, int warntype, const char *fmt, ...)
 {
-    bool    r;
-    va_list ap;
+    bool      r;
+    lex_ctx_t ctx;
+    va_list   ap;
+
+    ctx.file   = lex->name;
+    ctx.line   = lex->sline;
+    ctx.column = lex->column;
 
     va_start(ap, fmt);
-    r = vcompile_warning(lex->ctx, warntype, fmt, ap);
+    r = vcompile_warning(ctx, warntype, fmt, ap);
     va_end(ap);
     return r;
 }
@@ -169,7 +170,9 @@ static void lex_token_new(lex_file *lex)
         vec_shrinkto(lex->tok.value, 0);
 
     lex->tok.constval.t  = 0;
-    lex->tok.ctx         = &lex->ctx; /* this is 'shallow' */
+    lex->tok.ctx.line    = lex->sline;
+    lex->tok.ctx.file    = lex->name;
+    lex->tok.ctx.column  = lex->column;
 #endif
 }
 #endif
@@ -193,14 +196,14 @@ lex_file* lex_open(const char *file)
 
     memset(lex, 0, sizeof(*lex));
 
-    lex->file        = in;
-    lex->ctx.file    = util_strdup(file);
-    lex->ctx.line    = 1; /* we start counting at 1 */
-    lex->ctx.column  = 0;
-    lex->peekpos     = 0;
-    lex->eof         = false;
+    lex->file    = in;
+    lex->name    = util_strdup(file);
+    lex->line    = 1; /* we start counting at 1 */
+    lex->column  = 0;
+    lex->peekpos = 0;
+    lex->eof     = false;
 
-    vec_push(lex_filenames, (char *)lex->ctx.file);
+    vec_push(lex_filenames, lex->name);
     return lex;
 }
 
@@ -221,13 +224,13 @@ lex_file* lex_open_string(const char *str, size_t len, const char *name)
     lex->open_string_length = len;
     lex->open_string_pos    = 0;
 
-    lex->ctx.file    = util_strdup(name ? name : "<string-source>");
-    lex->ctx.line    = 1; /* we start counting at 1 */
-    lex->ctx.column  = 0;
-    lex->peekpos     = 0;
-    lex->eof         = false;
+    lex->name    = util_strdup(name ? name : "<string-source>");
+    lex->line    = 1; /* we start counting at 1 */
+    lex->peekpos = 0;
+    lex->eof     = false;
+    lex->column  = 0;
 
-    vec_push(lex_filenames, (char*)lex->ctx.file);
+    vec_push(lex_filenames, lex->name);
 
     return lex;
 }
@@ -265,13 +268,13 @@ void lex_close(lex_file *lex)
 static int lex_fgetc(lex_file *lex)
 {
     if (lex->file) {
-        lex->ctx.column++;
+        lex->column++;
         return fs_file_getc(lex->file);
     }
     if (lex->open_string) {
         if (lex->open_string_pos >= lex->open_string_length)
             return EOF;
-        lex->ctx.column++;
+        lex->column++;
         return lex->open_string[lex->open_string_pos++];
     }
     return EOF;
@@ -288,8 +291,8 @@ static int lex_try_trigraph(lex_file *lex, int old)
     int c2, c3;
     c2 = lex_fgetc(lex);
     if (!lex->push_line && c2 == '\n') {
-        lex->ctx.line++;
-        lex->ctx.column = 0;
+        lex->line++;
+        lex->column = 0;
     }
 
     if (c2 != '?') {
@@ -299,8 +302,8 @@ static int lex_try_trigraph(lex_file *lex, int old)
 
     c3 = lex_fgetc(lex);
     if (!lex->push_line && c3 == '\n') {
-        lex->ctx.line++;
-        lex->ctx.column = 0;
+        lex->line++;
+        lex->column = 0;
     }
 
     switch (c3) {
@@ -328,7 +331,7 @@ static int lex_try_digraph(lex_file *lex, int ch)
      * need to offset a \n the ungetch would recognize
      */
     if (!lex->push_line && c2 == '\n')
-        lex->ctx.line++;
+        lex->line++;
     if      (ch == '<' && c2 == ':')
         return '[';
     else if (ch == ':' && c2 == '>')
@@ -349,14 +352,18 @@ static int lex_getch(lex_file *lex)
 
     if (lex->peekpos) {
         lex->peekpos--;
-        if (!lex->push_line && lex->peek[lex->peekpos] == '\n')
-            lex->ctx.line++;
+        if (!lex->push_line && lex->peek[lex->peekpos] == '\n') {
+            lex->line++;
+            lex->column = 0;
+        }
         return lex->peek[lex->peekpos];
     }
 
     ch = lex_fgetc(lex);
-    if (!lex->push_line && ch == '\n')
-        lex->ctx.line++;
+    if (!lex->push_line && ch == '\n') {
+        lex->line++;
+        lex->column = 0;
+    }
     else if (ch == '?')
         return lex_try_trigraph(lex, ch);
     else if (!lex->flags.nodigraphs && (ch == '<' || ch == ':' || ch == '%'))
@@ -367,10 +374,10 @@ static int lex_getch(lex_file *lex)
 static void lex_ungetch(lex_file *lex, int ch)
 {
     lex->peek[lex->peekpos++] = ch;
-    lex->ctx.column--;
+    lex->column--;
     if (!lex->push_line && ch == '\n') {
-        lex->ctx.line--;
-        lex->ctx.column = 0;
+        lex->line--;
+        lex->column = 0;
     }
 }
 
@@ -421,7 +428,7 @@ static bool lex_try_pragma(lex_file *lex)
     if (lex->flags.preprocessing)
         return false;
 
-    line = lex->ctx.line;
+    line = lex->line;
 
     ch = lex_getch(lex);
     if (ch != '#') {
@@ -476,8 +483,8 @@ static bool lex_try_pragma(lex_file *lex)
             goto unroll;
     }
     else if (!strcmp(command, "file")) {
-        lex->ctx.file = util_strdup(param);
-        vec_push(lex_filenames, (char*)lex->ctx.file);
+        lex->name = util_strdup(param);
+        vec_push(lex_filenames, lex->name);
     }
     else if (!strcmp(command, "line")) {
         line = strtol(param, NULL, 0)-1;
@@ -485,7 +492,7 @@ static bool lex_try_pragma(lex_file *lex)
     else
         goto unroll;
 
-    lex->ctx.line = line;
+    lex->line = line;
     while (ch != '\n' && ch != EOF)
         ch = lex_getch(lex);
     vec_free(command);
@@ -522,7 +529,7 @@ unroll:
     }
     lex_ungetch(lex, '#');
 
-    lex->ctx.line = line;
+    lex->line = line;
     return false;
 }
 
@@ -754,7 +761,7 @@ static bool lex_finish_frames(lex_file *lex)
 
 static int GMQCC_WARN lex_finish_string(lex_file *lex, int quote)
 {
-    uchar_t chr;
+    utf8ch_t chr = 0;
     int ch = 0;
     int nextch;
     bool hex;
@@ -872,15 +879,16 @@ static int GMQCC_WARN lex_finish_string(lex_file *lex, int quote)
                     }
                 }
                 if (OPTS_FLAG(UTF8) && chr >= 128) {
-                    u8len = u8_fromchar(chr, u8buf, sizeof(u8buf));
+                    u8len = utf8_from(u8buf, chr);
                     if (!u8len)
                         ch = 0;
                     else {
                         --u8len;
-                        lex->ctx.column += u8len;
+                        lex->column += u8len;
                         for (uc = 0; uc < u8len; ++uc)
                             lex_tokench(lex, u8buf[uc]);
-                        /* the last character will be inserted with the tokench() call
+                        /*
+                         * the last character will be inserted with the tokench() call
                          * below the switch
                          */
                         ch = u8buf[uc];
@@ -1017,9 +1025,9 @@ int lex_do(lex_file *lex)
         return (lex->tok.ttype = ch);
     }
 
-    lex->sline         = lex->ctx.line;
-    lex->tok.ctx->line = lex->sline;
-    lex->tok.ctx->file = lex->ctx.file;
+    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);
@@ -1299,7 +1307,7 @@ int lex_do(lex_file *lex)
     }
 
     if (ch == '+' || ch == '-' || /* ++, --, +=, -=  and -> as well! */
-        ch == '>' || ch == '<' || /* <<, >>, <=, >=                  */
+        ch == '>' || ch == '<' || /* <<, >>, <=, >=  and >< as well! */
         ch == '=' || ch == '!' || /* <=>, ==, !=                     */
         ch == '&' || ch == '|' || /* &&, ||, &=, |=                  */
         ch == '~' || ch == '^'    /* ~=, ~, ^                        */
@@ -1307,7 +1315,9 @@ int lex_do(lex_file *lex)
         lex_tokench(lex, ch);
 
         nextch = lex_getch(lex);
-        if ((nextch == '=' && ch != '<') || (nextch == ch && ch != '!')) {
+        if ((nextch == '=' && ch != '<') ||
+            (nextch == ch  && ch != '!') ||
+            (nextch == '<' && ch == '>')) {
             lex_tokench(lex, nextch);
         } else if (ch == '<' && nextch == '=') {
             lex_tokench(lex, nextch);
@@ -1409,12 +1419,12 @@ int lex_do(lex_file *lex)
             lex->tok.constval.t = TYPE_VECTOR;
         } else {
             size_t kw;
-            for (kw = 0; kw < num_keywords_qc; ++kw) {
+            for (kw = 0; kw < GMQCC_ARRAY_COUNT(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) {
+                for (kw = 0; kw < GMQCC_ARRAY_COUNT(keywords_fg); ++kw) {
                     if (!strcmp(v, keywords_fg[kw]))
                         return (lex->tok.ttype = TOKEN_KEYWORD);
                 }
@@ -1463,14 +1473,10 @@ int lex_do(lex_file *lex)
         lex_endtoken(lex);
 
         lex->tok.ttype = TOKEN_CHARCONST;
-         /* It's a vector if we can successfully scan 3 floats */
-#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
-        if (sscanf(lex->tok.value, " %f %f %f ",
+
+        /* It's a vector if we can successfully scan 3 floats */
+        if (platform_sscanf(lex->tok.value, " %f %f %f ",
                    &lex->tok.constval.v.x, &lex->tok.constval.v.y, &lex->tok.constval.v.z) == 3)
-#endif
 
         {
              lex->tok.ttype = TOKEN_VECTORCONST;
@@ -1478,9 +1484,9 @@ int lex_do(lex_file *lex)
         else
         {
             if (!lex->flags.preprocessing && strlen(lex->tok.value) > 1) {
-                uchar_t u8char;
+                utf8ch_t u8char;
                 /* check for a valid utf8 character */
-                if (!OPTS_FLAG(UTF8) || !u8_analyze(lex->tok.value, NULL, NULL, &u8char, 8)) {
+                if (!OPTS_FLAG(UTF8) || !utf8_to(&u8char, (const unsigned char *)lex->tok.value, 8)) {
                     if (lexwarn(lex, WARN_MULTIBYTE_CHARACTER,
                                 ( OPTS_FLAG(UTF8) ? "invalid multibyte character sequence `%s`"
                                                   : "multibyte character: `%s`" ),
@@ -1510,6 +1516,6 @@ int lex_do(lex_file *lex)
         return (lex->tok.ttype = ch);
     }
 
-    lexerror(lex, "unknown token: `%s`", lex->tok.value);
+    lexerror(lex, "unknown token: `%c`", ch);
     return (lex->tok.ttype = TOKEN_ERROR);
 }