使用Shell脚本中的-D名称编译选项

Question

I'm attempting to write a shell script that will run several tests of my C++ program, redefining a macro each time it runs. I'm trying to use the -D name preprocessor option (see https://gcc.gnu.org/onlinedocs/gcc/Preprocessor-Options.html ), but I'm consistently getting the warning that the macro is being redefined, and then the program executes without redefining it.

My script is as follows:

#!/bin/bash

#NUMS is number of subdivisions:
for subdiv in 10 100 500 1000
do
    echo NUMS = $subdiv
    g++ -D NUMS=$subdiv project01.cpp -o project01 -lm -fopenmp
    ./project01 >> bezier_results.txt
done

In my C++ file, project01.cpp, I state:

#define NUMS 1

I've tried leaving out the '1', but that produces errors as well. It's clear that the script isn't actually redefining the macro. Any thoughts? Thanks!

Answer 1

By defining the macro on the command-line with -DNUMS=100 you've provided a default value.

You're code is then overriding that default value when you do:

#define NUMS 1

The compiler warning is telling you exactly what has happened. Instead consider coding something like this:

#if !defined NUMS
#define NUMS 1
#endif

Now the compiler will only redefine NUMS as 1 when it hasn't already been defined (ie elsewhere in your source, or in this case the command-line.)