Re: [PATCH] conf, tap: Silence two false positive invalidFunctionArg from cppcheck

On Fri, 17 Feb 2023 09:29:51 -0500
Laine Stump <laine(a)redhat.com> wrote:

...
  On 2/16/23 5:53 PM, David Gibson wrote:
  On Thu, Feb 16, 2023 at 07:22:10PM +0100, Stefano
Brivio wrote:  
  The newly introduced die() calls exit(), but
cppcheck doesn't see it
 and warns about possibly invalid arguments used after the check which
 triggers die(). Add return statements to silence the warnings.

 Signed-off-by: Stefano Brivio <sbrivio(a)redhat.com>   
 
 Oof, that's super ugly.  Any chance that cppcheck will recognize the
 ((noreturn)) attribute if we added it to die()?   
 
 Why is this only a problem in these two files? 
Because that's where we would use a variable as a system call argument,
if the check possibly leading to die() didn't exist.

...
  (and is there a "make check" target that I
should have been running and 
 haven't?) 
make cppcheck, make clang-tidy, and possibly the tests (see
test/README.md) -- but it's not mandatory, especially if you're an
occasional contributor, I'm running them anyway.

...
  Requiring an extra "return" after die() kind
of removes the advantage of 
 using it over err(). :-/ If we have to do that, it would be more 
 straightforward to just use err() followed by exit() directly. 
We don't have to (see the rest of this thread): we could simply define
die() as err() with an additional return statement *outside* err()
itself, instead of passing 'doexit' as parameter. Probably something
like:

do {
	err(...)
	exit(EXIT_FAILURE);
	return;
} while (0)

-- 
Stefano