Helping Your Compiler Handle the Size of Your Constants
By RickRamsey-Oracle on Jun 25, 2014
by Darryl Gove
When I use a constant in the following code, I get a warning:
On the other hand if I wrote:
Then then compiler will quite happily handle the constant.
The problem with the first bit of code is that it treats the value as a signed integer, and a signed integer can only hold 31 bits of precision plus a sign bit.
So how does the compiler decide how to represent a constant? The answer is interesting.
The compiler will attempt to fit a constant into the smallest value that it can. So it will try to fit the value into these types, in order: into an
long int, and then a
long long int.
In the above code sample, the compiler will find that 1 and 31 both fit very nicely into signed
ints. There's a shift left operation (<<) in the expression that produces a result of the same type as the
left operand. So the whole expression (1<<31) has type
signed int, which leads to the the warning.
To avoid the warning we can tell the compiler that this is an unsigned value. Either by typecasting the 1 to be unsigned in this manner:
or by declaring it as an unsigned value, like this:
More About Oracle Solaris StudioOracle Solaris Studio is a C, C++ and Fortran development tool suite, with compiler optimizations, multithread performance, and analysis tools for application development on Oracle Solaris, Oracle Linux, and Red Hat Enterprise Linux operating systems. Find out more about the Oracle Solaris Studio 12.4 Beta program here.
About the Photograph
Photograph of Zion National Park, Utah taken by Rick Ramsey in May 2014 on The Ride to the Sun Reunion.
|Follow Darryl on: Oracle Blog | Twitter||Follow OTN Garage on: Web | Facebook | Twitter | YouTube|