Saving a 32bit int

[Eugeneo]

I've always wondered what the best way to save a 32bit int is. What method if any would you use and why?

// Structure for 32 bit int
struct byteint
{
char b1;
char b2;
char b3;
char b4;
};

union I8
{
int32 I32;
struct byteint I8;
};

union I8 timer_value;



void Save1(void)
{
WRITE_EXT_EEPROM(ee_timer, timer_value.I8.b1);
WRITE_EXT_EEPROM(ee_timer+1, timer_value.I8.b2);
WRITE_EXT_EEPROM(ee_timer+2, timer_value.I8.b3);
WRITE_EXT_EEPROM(ee_timer+3, timer_value.I8.b4);
}

void Save2(void)
{
int temp=0;
int temp16=0;

do
{
if(temp==0)
temp16=timer_value.I8.b1;
if(temp==1)
temp16=timer_value.I8.b2;
if(temp==2)
temp16=timer_value.I8.b3;
if(temp==3)
temp16=timer_value.I8.b4;

WRITE_EXT_EEPROM(ee_timer+temp, temp16);
temp+=1;
}while(temp<4);
}

void Save3(void)
{
int temp=0;
do
{
WRITE_EXT_EEPROM(ee_timer+temp, timer_value.I32>>((temp+1)*8));
temp+=1;
}while(temp<4);
}

void Save4(void)
{
int temp=0;
do
{
WRITE_EXT_EEPROM(ee_timer+temp, *(&timer_value.I32+temp));
temp+=1;
}while(temp<4);
}

[Ttelmah]

Save1, or write the fourth version, using make8, and make32.

The union approach, or the CCS 'MAKEn' functions, do direct I/O to the individual bytes. The pointer approach, adds significant code, and time. There are probably another dozen versions you could actually generate, but generally these two are efficient. As an alternative approach:

[Eugeneo]