False positive on MISRA 18.8

[Marco Ferragatta]

Hello,
before all, thank you for this great software!

Running the software I got some false positives, and one of them is related to MISRA 18.8 rule.
When the length of an array is determined using a #define I get the rule violation. It seems related to some sort of preprocessing problem, in fact if the #define is in the same file in which the array is declared there is no error. Otherwise, if the #define is located in a header which will be included, the rule violation is raised.
In my embedded project I have to deal with a large amount of generated code, that I can't modify, which make a wide use of #defines that recall to other #defines in another include file.

I have attached a small project that reproduces the error.

I've tried to launch the tests from command line also, this is the result:

        C:\Program Files\Cppcheck>cppcheck -rp --addon="C:\Program Files\Cppcheck\addons\misra.json" C:\Temp\Test_CppCheck\Sources -I C:\Temp\Test_CppCheck\Sources
        Checking C:\Temp\Test_CppCheck\Sources\main.c ...
        Checking C:\Temp\Test_CppCheck\Sources\main.c: DEFINESL1_H...
        Checking C:\Temp\Test_CppCheck\Sources\main.c: DEFINESL2_H...
        main.c:8:17: style: Required: Variable-length array types shall not be used. [misra-c2012-18.8]

                        ^
        main.c:7:17: style: Required: Variable-length array types shall not be used. [misra-c2012-18.8]

                        ^

Finally, if check the configuration from command line, I got no errors:

        C:\Program Files\Cppcheck>cppcheck -rp --check-config --addon="C:\Program Files\Cppcheck\addons\misra.json" C:\Temp\Test_CppCheck\Sources -I C:\Temp\Test_CppCheck\Sources
        Checking C:\Temp\Test_CppCheck\Sources\main.c ...

        C:\Program Files\Cppcheck>

[Georgiy Komarov]

Hello, Marco,

Thanks for reporting this. I can confirm this false positive.

I created the following ticket: https://trac.cppcheck.net/ticket/10197

[Marco Ferragatta]

Hello Georgiy,
thanks to you for taking into account my report.
Checking the embedded code of my application I've found other cases: for some of them I'm pretty sure they are false positives, for some others I'm suspicious. Now I'm a little busy at work, but as soon as I've some time I will provide a separate report for each of them with attached a simple project that can reproduce the false positive (or the suspicious one).

Have a nice day!