glsl: Disallow float literals with the 'f' suffix but no point or exponent
authorNeil Roberts <neil@linux.intel.com>
Wed, 26 Nov 2014 17:15:01 +0000 (17:15 +0000)
committerNeil Roberts <neil@linux.intel.com>
Thu, 27 Nov 2014 11:43:17 +0000 (11:43 +0000)
According to the GLSL spec float literals like â€˜1f’ shouldn't be allowed
without adding a decimal point or an exponent. Apparently the AMD driver also
disallows this so it seems unlikely that anything would be relying on it.

Reviewed-by: Matt Turner <mattst88@gmail.com>
src/glsl/glsl_lexer.ll

index ad6d32f9bc0c89b9b83174af62d4ca48df41c173..419a07b1daf8c92fc256b8084df5bc67974aec80 100644 (file)
@@ -466,10 +466,6 @@ layout             {
                            yylval->real = _mesa_strtof(yytext, NULL);
                            return FLOATCONSTANT;
                        }
-[0-9]+[fF]             {
-                           yylval->real = _mesa_strtof(yytext, NULL);
-                           return FLOATCONSTANT;
-                       }
 
 true                   {
                            yylval->n = 1;