-1

If I want my program to be able to run in debug mode, is it a good idea to make it accept a flag such as -D=DEBUG when I run the program? I currently have a DEBUG variable in a .h file but that I can't change after I compile.

#define DEBUG false

I'd like to be able to run my program both in debug mode and "stable" mode. Is getops a good option for doing this since I already use getops for accepting the args ./a.out --help and ./a.out --version ?

  • 2
    I'm sure what you're suggesting would work, but I think there's a reason people don't do it. When you use preprocessor directives like #define and #ifdef you actually change which parts of your code are compiled and make it to the final binary. People typically use this approach so that debug code doesn't make it to a production release (even if they are 99% sure it won't be invoked). Leaving debug code in a production build is just a liability. Also, I'm pretty sure you miss out on tons of compiler optimisations when you release in debug mode. – MetaFight Apr 30 '16 at 10:37
5

You are better off to make a clear distinction between "compile time debug" mode (which is what you control by #define DEBUG), and specific "debug features" which should be available even when you compile with #define DEBUG false, and could be enabled or disabled at run time. Better call the latter differently, name the features (like "logging mode", "validating mode" etc.), and provide different command line flags for them.

That gives you the opportunity to enable these features even when you deployed the program into production environment, where you might not be able to use a debugger.

4

The whole point of compiling in "debug mode" is that it includes debug symbols in the executable to enable the use of a debugger, at the cost of foregoing some performance optimizations that aren't possible with debug symbols. If you don't plan on running that executable in a debugger, then "being in debug mode" doesn't gain you anything.

It's likely that you're conflating debug mode (which is something you do at compile time) with other features that actually are useful to have in a release binary, such as a -v flag (passed by the user at runtime) for increased verbosity in command line tools, or a diagnostic mode in video games (triggered by the user during execution) to help players identify lag sources on their server. Whether these should be command line arguments or something else depends on the application.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.