[clang-tidy] Variable is non-const and globally accessible, consider making it const

[MaJerle]

I have a module based static variable, that clang reports as globally accessible

//...
static osMutexId_t app_sys_mutex;
//

[{
	"message": "variable 'app_sys_mutex' is non-const and globally accessible, consider making it const",
	"source": "C/C++",
}]

https://clang.llvm.org/extra/clang-tidy/checks/cppcoreguidelines/avoid-non-const-global-variables.html
In the link above, there is no example if static keyword.

It is a module specific variable, that must be module global.
Why does it report it as globally accessible, if it is static?

Version: 17.0.5

[PiotrZSL]

"Enforcement (Simple) Report all non-const variables declared at namespace scope and global pointers/references to non-const data."
If you do not agree with this rule, simply disable check or report issue here: https://github.com/isocpp/CppCoreGuidelines/issues

[nolange]

"Enforcement (Simple) Report all non-const variables declared at namespace scope and global pointers/references to non-const data."

static variables and variables defined in an anonymous namespace have file scope, so are not subject to this rule which covers global and namespace scope.
This is how I read the rule, and many sentences just don't apply if the intent was to include file scope ("The initialization of global objects is not totally ordered")

[PiotrZSL]

"Reason Non-const global variables hide dependencies and make the dependencies subject to unpredictable changes."
"Note The rule against global variables applies to namespace scope variables as well."

This rule is mainly to avoid passing data between functions/classes via global objects.
In such case it doesn't make a difference if this is shared between files or not.

[nolange]

yes, it makes all the difference, a static global can only be accessed in a single file and thus is a small managable scope. a more global "global" would end up in a header and thats when the dangerous stuff happens
of course bugs can always happen, but with that mindset you need to prohibit variables in function scope as well.

are you involved in the guidelines themselves? to me this rule is pretty clearly not implemented as intended.

[PiotrZSL]

No, i'm not involved.
Static variables in functions are not too big problem, as they are only used in that function.
Here problem is if developer instead of using class members or function parameter would decide to pass data via global variables, for that's what this check/guideline is checking. You can still do some global static variable in some class, and use it as global, but thats more explicit than just moving some local variable outside function.

[PiotrZSL]

Still some options could always be added to check to make it more ... configurable. If you see some options that would align check to your needs, just describe them.

[dzid26]

How can this be avoided in C ? Only passing variables via argument?