Atmel CAVR-4 Manual De Usuario

CAVR-4

Part 2. Compiler reference

The 

char

 type is by default unsigned in the compiler, but the 

--char_is_signed

 

compiler option allows you to make it signed. Note, however, that the library is compiled 
with the 

char

 type as unsigned.

The 

wchar_t

 data type is an integer type whose range of values can represent distinct 

codes for all members of the largest extended character set specified among the 
supported locals.

The 

wchar_t

 data type is supported by default in the C++ language. To use the 

wchar_t

 type also in C source code, you must include the file 

stddef.h

 from the 

runtime library.

Note: The IAR CLIB Library has only rudimentary support for 

wchar_t

.

In ISO/ANSI C, 

int

 and 

unsigned

 

int

 can be used as the base type for integer 

bitfields. In the AVR IAR C/C++ Compiler, any integer type can be used as the base type 
when language extensions are enabled.

Bitfields in expressions will have the same data type as the integer base type.

By default, the compiler places bitfield members from the least significant to the most 
significant bit in the container type. 

By using the directive 

#pragma

 

bitfields=reversed

, the bitfield members are 

placed from the most significant to the least significant bit.

FLOATING-POINT TYPES 

In the AVR IAR C/C++ Compiler, floating-point values are represented in standard 
IEEE format.

The ranges and sizes for the different floating-point types are:

float 

32 bits

±1.18E-38 to ±3.39E+38

8 bits

23 bits

double 

*

32 bits (default)

±1.18E-38 to ±3.39E+38

8 bits

23 bits

double 

*

64 bits

±2.23E-308 to ±1.79E+308

11 bits

52 bits

long double 

*

32 bits

±1.18E-38 to ±3.39E+38

8 bits

23 bits

long double 

*

64 bits

±2.23E-308 to ±1.79E+308

11 bits

52 bits