X-Git-Url: https://git.xonotic.org/?a=blobdiff_plain;f=lexer.c;h=b6d5ceb6d3bd94aed4e4cc0bb0d2926d932b9b6a;hb=7011a2ef2ceaf9652cea0b0e61319b338667888b;hp=db0a8f4d9bf541bdd25597c5047e040d7dc36fc5;hpb=685fa54daf0b3fdd92d9424e7fcab1ded9fb3c29;p=xonotic%2Fgmqcc.git

diff --git a/lexer.c b/lexer.c
index db0a8f4..b6d5ceb 100644
--- a/lexer.c
+++ b/lexer.c
@@ -20,10 +20,8 @@
  * 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>
-#include <stdarg.h>
+#include <stdlib.h>
 
 #include "gmqcc.h"
 #include "lexer.h"
@@ -40,8 +38,6 @@ 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",
@@ -52,34 +48,33 @@ 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;
 
-char* *lex_filenames;
-
-void lexerror(lex_file *lex, const char *fmt, ...)
+static void lexerror(lex_file *lex, const char *fmt, ...)
 {
     va_list ap;
 
     va_start(ap, fmt);
     if (lex)
-        con_vprintmsg(LVL_ERROR, lex->name, lex->sline, "parse error", fmt, ap);
+        con_vprintmsg(LVL_ERROR, lex->name, lex->sline, lex->column, "parse error", fmt, ap);
     else
-        con_vprintmsg(LVL_ERROR, "", 0, "parse error", fmt, ap);
+        con_vprintmsg(LVL_ERROR, "", 0, 0, "parse error", fmt, ap);
     va_end(ap);
 }
 
-bool lexwarn(lex_file *lex, int warntype, const char *fmt, ...)
+static bool lexwarn(lex_file *lex, int warntype, const char *fmt, ...)
 {
-    bool    r;
-    lex_ctx ctx;
-    va_list ap;
+    bool      r;
+    lex_ctx_t ctx;
+    va_list   ap;
 
-    ctx.file = lex->name;
-    ctx.line = lex->sline;
+    ctx.file   = lex->name;
+    ctx.line   = lex->sline;
+    ctx.column = lex->column;
 
     va_start(ap, fmt);
     r = vcompile_warning(ctx, warntype, fmt, ap);
@@ -174,17 +169,19 @@ static void lex_token_new(lex_file *lex)
 #else
     if (lex->tok.value)
         vec_shrinkto(lex->tok.value, 0);
+
     lex->tok.constval.t  = 0;
-    lex->tok.ctx.line = lex->sline;
-    lex->tok.ctx.file = lex->name;
+    lex->tok.ctx.line    = lex->sline;
+    lex->tok.ctx.file    = lex->name;
+    lex->tok.ctx.column  = lex->column;
 #endif
 }
 #endif
 
 lex_file* lex_open(const char *file)
 {
-    lex_file *lex;
-    FILE *in = fs_file_open(file, "rb");
+    lex_file  *lex;
+    fs_file_t *in = fs_file_open(file, "rb");
 
     if (!in) {
         lexerror(NULL, "open failed: '%s'\n", file);
@@ -200,12 +197,12 @@ lex_file* lex_open(const char *file)
 
     memset(lex, 0, sizeof(*lex));
 
-    lex->file = in;
-    lex->name = util_strdup(file);
-    lex->line = 1; /* we start counting at 1 */
-
+    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;
+    lex->eof     = false;
 
     vec_push(lex_filenames, lex->name);
     return lex;
@@ -228,11 +225,11 @@ 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->name = util_strdup(name ? name : "<string-source>");
-    lex->line = 1; /* we start counting at 1 */
-
+    lex->name    = util_strdup(name ? name : "<string-source>");
+    lex->line    = 1; /* we start counting at 1 */
     lex->peekpos = 0;
-    lex->eof = false;
+    lex->eof     = false;
+    lex->column  = 0;
 
     vec_push(lex_filenames, lex->name);
 
@@ -271,14 +268,17 @@ void lex_close(lex_file *lex)
 
 static int lex_fgetc(lex_file *lex)
 {
-    if (lex->file)
+    if (lex->file) {
+        lex->column++;
         return fs_file_getc(lex->file);
+    }
     if (lex->open_string) {
         if (lex->open_string_pos >= lex->open_string_length)
-            return EOF;
+            return FS_FILE_EOF;
+        lex->column++;
         return lex->open_string[lex->open_string_pos++];
     }
-    return EOF;
+    return FS_FILE_EOF;
 }
 
 /* Get or put-back data
@@ -291,16 +291,22 @@ static int lex_try_trigraph(lex_file *lex, int old)
 {
     int c2, c3;
     c2 = lex_fgetc(lex);
-    if (!lex->push_line && c2 == '\n')
+    if (!lex->push_line && c2 == '\n') {
         lex->line++;
+        lex->column = 0;
+    }
+
     if (c2 != '?') {
         lex_ungetch(lex, c2);
         return old;
     }
 
     c3 = lex_fgetc(lex);
-    if (!lex->push_line && c3 == '\n')
+    if (!lex->push_line && c3 == '\n') {
         lex->line++;
+        lex->column = 0;
+    }
+
     switch (c3) {
         case '=': return '#';
         case '/': return '\\';
@@ -347,14 +353,18 @@ static int lex_getch(lex_file *lex)
 
     if (lex->peekpos) {
         lex->peekpos--;
-        if (!lex->push_line && lex->peek[lex->peekpos] == '\n')
+        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')
+    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 == '%'))
@@ -365,8 +375,11 @@ static int lex_getch(lex_file *lex)
 static void lex_ungetch(lex_file *lex, int ch)
 {
     lex->peek[lex->peekpos++] = ch;
-    if (!lex->push_line && ch == '\n')
+    lex->column--;
+    if (!lex->push_line && ch == '\n') {
         lex->line--;
+        lex->column = 0;
+    }
 }
 
 /* classify characters
@@ -376,12 +389,12 @@ static void lex_ungetch(lex_file *lex, int ch)
 /* Idents are alphanumberic, but they start with alpha or _ */
 static bool isident_start(int ch)
 {
-    return isalpha(ch) || ch == '_';
+    return util_isalpha(ch) || ch == '_';
 }
 
 static bool isident(int ch)
 {
-    return isident_start(ch) || isdigit(ch);
+    return isident_start(ch) || util_isdigit(ch);
 }
 
 /* isxdigit_only is used when we already know it's not a digit
@@ -481,7 +494,7 @@ static bool lex_try_pragma(lex_file *lex)
         goto unroll;
 
     lex->line = line;
-    while (ch != '\n' && ch != EOF)
+    while (ch != '\n' && ch != FS_FILE_EOF)
         ch = lex_getch(lex);
     vec_free(command);
     vec_free(param);
@@ -561,7 +574,7 @@ static int lex_skipwhite(lex_file *lex, bool hadwhite)
     do
     {
         ch = lex_getch(lex);
-        while (ch != EOF && isspace(ch)) {
+        while (ch != FS_FILE_EOF && util_isspace(ch)) {
             if (ch == '\n') {
                 if (lex_try_pragma(lex))
                     continue;
@@ -601,7 +614,7 @@ static int lex_skipwhite(lex_file *lex, bool hadwhite)
                     lex_tokench(lex, ' ');
                 }
 
-                while (ch != EOF && ch != '\n') {
+                while (ch != FS_FILE_EOF && ch != '\n') {
                     if (lex->flags.preprocessing)
                         lex_tokench(lex, ' '); /* ch); */
                     ch = lex_getch(lex);
@@ -626,7 +639,7 @@ static int lex_skipwhite(lex_file *lex, bool hadwhite)
                     lex_tokench(lex, ' ');
                 }
 
-                while (ch != EOF)
+                while (ch != FS_FILE_EOF)
                 {
                     ch = lex_getch(lex);
                     if (ch == '*') {
@@ -659,7 +672,7 @@ static int lex_skipwhite(lex_file *lex, bool hadwhite)
             ch = '/';
             break;
         }
-    } while (ch != EOF && isspace(ch));
+    } while (ch != FS_FILE_EOF && util_isspace(ch));
 
     if (haswhite) {
         lex_endtoken(lex);
@@ -675,7 +688,7 @@ static bool GMQCC_WARN lex_finish_ident(lex_file *lex)
     int ch;
 
     ch = lex_getch(lex);
-    while (ch != EOF && isident(ch))
+    while (ch != FS_FILE_EOF && isident(ch))
     {
         lex_tokench(lex, ch);
         ch = lex_getch(lex);
@@ -695,7 +708,7 @@ static int lex_parse_frame(lex_file *lex)
     lex_token_new(lex);
 
     ch = lex_getch(lex);
-    while (ch != EOF && ch != '\n' && isspace(ch))
+    while (ch != FS_FILE_EOF && ch != '\n' && util_isspace(ch))
         ch = lex_getch(lex);
 
     if (ch == '\n')
@@ -749,14 +762,15 @@ 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;
+    bool oct;
     char u8buf[8]; /* way more than enough */
     int  u8len, uc;
 
-    while (ch != EOF)
+    while (ch != FS_FILE_EOF)
     {
         ch = lex_getch(lex);
         if (ch == quote)
@@ -765,18 +779,18 @@ static int GMQCC_WARN lex_finish_string(lex_file *lex, int quote)
         if (lex->flags.preprocessing && ch == '\\') {
             lex_tokench(lex, ch);
             ch = lex_getch(lex);
-            if (ch == EOF) {
+            if (ch == FS_FILE_EOF) {
                 lexerror(lex, "unexpected end of file");
-                lex_ungetch(lex, EOF); /* next token to be TOKEN_EOF */
+                lex_ungetch(lex, FS_FILE_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) {
+            if (ch == FS_FILE_EOF) {
                 lexerror(lex, "unexpected end of file");
-                lex_ungetch(lex, EOF); /* next token to be TOKEN_EOF */
+                lex_ungetch(lex, FS_FILE_EOF); /* next token to be TOKEN_EOF */
                 return (lex->tok.ttype = TOKEN_ERROR);
             }
 
@@ -838,17 +852,18 @@ static int GMQCC_WARN lex_finish_string(lex_file *lex, int quote)
                 chr = 0;
                 nextch = lex_getch(lex);
                 hex = (nextch == 'x');
-                if (!hex)
+                oct = (nextch == '0');
+                if (!hex && !oct)
                     lex_ungetch(lex, nextch);
                 for (nextch = lex_getch(lex); nextch != '}'; nextch = lex_getch(lex)) {
-                    if (!hex) {
+                    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);
                         }
-                    } else {
+                    } else if (!oct) {
                         if (nextch >= '0' && nextch <= '9')
                             chr = chr * 0x10 + nextch - '0';
                         else if (nextch >= 'a' && nextch <= 'f')
@@ -859,6 +874,13 @@ static int GMQCC_WARN lex_finish_string(lex_file *lex, int quote)
                             lexerror(lex, "bad character code");
                             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);
+                        }
                     }
                     if (chr > 0x10FFFF || (!OPTS_FLAG(UTF8) && chr > 255))
                     {
@@ -867,14 +889,16 @@ 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->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];
@@ -897,7 +921,7 @@ 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 */
+    lex_ungetch(lex, FS_FILE_EOF); /* next token to be TOKEN_EOF */
     return (lex->tok.ttype = TOKEN_ERROR);
 }
 
@@ -916,7 +940,7 @@ static int GMQCC_WARN lex_finish_digit(lex_file *lex, int lastch)
     lex_tokench(lex, ch);
 
     ch = lex_getch(lex);
-    if (ch != '.' && !isdigit(ch))
+    if (ch != '.' && !util_isdigit(ch))
     {
         if (lastch != '0' || ch != 'x')
         {
@@ -937,7 +961,7 @@ static int GMQCC_WARN lex_finish_digit(lex_file *lex, int lastch)
     {
         lex_tokench(lex, ch);
         ch = lex_getch(lex);
-        while (isdigit(ch) || (ishex && isxdigit_only(ch)))
+        while (util_isdigit(ch) || (ishex && isxdigit_only(ch)))
         {
             lex_tokench(lex, ch);
             ch = lex_getch(lex);
@@ -952,7 +976,7 @@ static int GMQCC_WARN lex_finish_digit(lex_file *lex, int lastch)
 
         /* continue digits-only */
         ch = lex_getch(lex);
-        while (isdigit(ch))
+        while (util_isdigit(ch))
         {
             lex_tokench(lex, ch);
             ch = lex_getch(lex);
@@ -1018,7 +1042,7 @@ int lex_do(lex_file *lex)
     if (lex->eof)
         return (lex->tok.ttype = TOKEN_FATAL);
 
-    if (ch == EOF) {
+    if (ch == FS_FILE_EOF) {
         lex->eof = true;
         return (lex->tok.ttype = TOKEN_EOF);
     }
@@ -1055,10 +1079,10 @@ int lex_do(lex_file *lex)
         if (!strcmp(v, "framevalue"))
         {
             ch = lex_getch(lex);
-            while (ch != EOF && isspace(ch) && ch != '\n')
+            while (ch != FS_FILE_EOF && util_isspace(ch) && ch != '\n')
                 ch = lex_getch(lex);
 
-            if (!isdigit(ch)) {
+            if (!util_isdigit(ch)) {
                 lexerror(lex, "$framevalue requires an integer parameter");
                 return lex_do(lex);
             }
@@ -1135,7 +1159,7 @@ 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 != FS_FILE_EOF && ch != '\n')
                 ch = lex_getch(lex);
             return lex_do(lex);
         }
@@ -1149,7 +1173,7 @@ int lex_do(lex_file *lex)
         {
             /* skip line */
             ch = lex_getch(lex);
-            while (ch != EOF && ch != '\n')
+            while (ch != FS_FILE_EOF && ch != '\n')
                 ch = lex_getch(lex);
             return lex_do(lex);
         }
@@ -1216,7 +1240,7 @@ int lex_do(lex_file *lex)
     if (ch == '.') {
         nextch = lex_getch(lex);
         /* digits starting with a dot */
-        if (isdigit(nextch)) {
+        if (util_isdigit(nextch)) {
             lex_ungetch(lex, nextch);
             lex->tok.ttype = lex_finish_digit(lex, ch);
             lex_endtoken(lex);
@@ -1293,15 +1317,17 @@ 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 == '~' || ch == '^'    /* ~=, ~, ^                        */
     )  {
         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);
@@ -1324,7 +1350,7 @@ int lex_do(lex_file *lex)
             }
         }
         else if (lex->flags.preprocessing &&
-                 ch == '-' && isdigit(nextch))
+                 ch == '-' && util_isdigit(nextch))
         {
             lex->tok.ttype = lex_finish_digit(lex, nextch);
             if (lex->tok.ttype == TOKEN_INTCONST)
@@ -1403,12 +1429,12 @@ 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);
                 }
@@ -1457,14 +1483,10 @@ 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 (util_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;
@@ -1472,9 +1494,9 @@ 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`" ),
@@ -1491,7 +1513,7 @@ int lex_do(lex_file *lex)
         return lex->tok.ttype;
     }
 
-    if (isdigit(ch))
+    if (util_isdigit(ch))
     {
         lex->tok.ttype = lex_finish_digit(lex, ch);
         lex_endtoken(lex);
@@ -1504,6 +1526,6 @@ 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);
 }