How can you make a C compiler assume decimal literals (e.g. 1.23) are float and not double?

Question

In my source code, if I write 1.23 as a literal, e.g. doThis(1.23), gcc assumes it's a double.

Rather than type doThis((float) 1.23), is there a way to use floats for decimal literals/constants unless otherwise specified in an individual source file?

Mega-bonus points, is there a way that works across (nearly) every C compiler?

There's no way to do it for all "plain" constants in a source file, as you seem to want. You have to use either a cast or a suffix. — AnT stands with Russia, Oct 11 '10 at 06:34
The suffix works great... I'd completely forgotten about it! — HoboBen, Oct 11 '10 at 06:38
[Make C floating point literals float (rather than double)](https://stackoverflow.com/q/32266864/995714). But note that this is not recommended for most cases — phuclv, Sep 10 '18 at 17:20

score 7 · Accepted Answer · answered Oct 11 '10 at 06:26

7

Yes, the standard way is to write 1.23f. It works with every C compiler, since it is defined in ISO C99 section 6.4.4.2 Floating constants. ISO C90 and K&R have similar definitions.

answered Oct 11 '10 at 06:26

Roland Illig

40,703
10
88
121

score 5 · Answer 2 · answered Oct 11 '10 at 06:26

5

try:

float fred = 0.37f;

answered Oct 11 '10 at 06:26

Bas Bossink

9,388
4
41
53

3

Yoda says: 'Do, or do not. There is no try'. – Jonathan Leffler Oct 11 '10 at 06:32

score 2 · Answer 3 · answered Oct 11 '10 at 06:27

2

try 123.4F for a float constant

answered Oct 11 '10 at 06:27

Peter Miehle

5,984
2
38
55

score 2 · Answer 4 · answered Dec 30 '10 at 22:06

2

Also gcc has the option -fsingle-precision-constant that tells the compiler to treat constants as single precision. See http://gcc.gnu.org/wiki/FloatingPointMath

answered Dec 30 '10 at 22:06

Paul Beusterien

27,542
6
83
139

How can you make a C compiler assume decimal literals (e.g. 1.23) are float and not double?

4 Answers4