Possible check miss related to suppressed rules

[Enrique Malras]

Hi,

I am working with cppcheck version 2.4.1 + misra.py addon combined with --inline-suppr option for cppcheck. I am checking a boilerplate.c source code that haves 5 known misra violations.

I am able to generate the .dump file with the suppressions section, and next output is generated when misra.py parses the .dump file:

>python misra/misra.py src/boilerplate.c.dump
Checking src/boilerplate.c.dump...
Checking src/boilerplate.c.dump, config ...

MISRA rules violations found:
        Undefined: 5

MISRA rules violated:
        misra-c2012-15.5 (-): 1
        misra-c2012-20.1 (-): 4

As can be seen, the rules violation dissapear (as expected) but the total count of violations remain up to 5 (expected to be 0).

I could review the misra.py code, and it seems that it is missing some check on the reportError() function. I could solve this (at least working in my environment) by adding an extra check (call to cppcheckdata.is_suppressed()) in that function:

def reportError(self, location, num1, num2):
        ruleNum = num1 * 100 + num2

        if self.settings.verify:
            self.verify_actual.append(str(location.linenr) + ':' + str(num1) + '.' + str(num2))
        #EMalras: First statement checks if the rule has been included as 'global suppressed'
        #EMalras: Second statement checks if the rule has been suppressed because has been justified with some 'cppcheck-suppress' tag within the source code
        elif (self.isRuleSuppressed(location.file, location.linenr, ruleNum) or cppcheckdata.is_suppressed(location, '',  'misra-c2012-'+str(num1)+'.'+str(num2))):
            # Error is suppressed. Ignore
            self.suppressionStats.setdefault(ruleNum, 0)
            self.suppressionStats[ruleNum] += 1
            return
        else:
        .......

With this the output become as expected:

>python misra/misra.py src/boilerplate.c.dump
Checking src/boilerplate.c.dump...
Checking src/boilerplate.c.dump, config ...

No output, which means 'no violations'.

I'm sure it is not the best solution, but I want to share it with you just to let you know about this issue.

Thanks!

[Daniel Marjamäki]

hmm looks pretty reasonable to me.