https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118246

--- Comment #3 from newbie-02 <newbie-02 at gmx dot de> ---
hello @ Andrew Pinski,  

thanks for fast reaction,  

a snippet that fails, compile with '-lquadmath' option,  
change commenting between FLT128_MIN and FLT128_TRUE_MIN and see that all is  
ok, except FLT128_TRUE_MIN:  
```  

#include <stdio.h>                                                             
        // reg. e.g. printf, 
#include <float.h>                                                             
        // reg. e.g. INT_MIN, 
#include <limits.h>                                                            
        // reg. e.g. xx_MIN, 
#include <quadmath.h>                                                          
        // reg. e.g. print of binary128, 

int main( int argc, char *argv[] ) 
{ 
        volatile long double x1l = 0.0l; 
        volatile __float128 x1q = 0.0q; 
        char str[ 45 ]; 

        x1l = LDBL_TRUE_MIN; 
        x1q = 6E-4966q; 
//      x1q = FLT128_MIN; 
        x1q = FLT128_TRUE_MIN; 
        quadmath_snprintf( str, sizeof(str), "%.36QE", x1q );
        printf( " x1q: %s \n", str ); 
        return 0; 
} 
```  
my gcc:  
cc (Debian 13.3.0-6) 13.3.0
my glibc:  
ldd (Debian GLIBC 2.40-2) 2.40

Reply via email to