Suppression file name cache.

[MartinBP]

Implement a file name cache per suppression to avoid calling PathMatch::match again and again for the same suppression and file name combination.

Local benchmark running cppcheck with misra c-2012 has been performed.

Before
10min 34sec
After
1min 36sec

In this test it was a few header files producing the majority of the error messages, which is why the cache speeds it up by a lot.

[glankk]

Looks good

[firewave]

As this is a global cache it might have side effects when the object is used for more than one analysis. This applies to the tests and might also apply to the GUI. As the suppressions are always mutable if both of them have usage issues they have been pre-existing. We should check and fix any issues before we merge this to make sure things don't get worse.

[firewave]

We should also add a performance test for this in Python.

[firewave]

Also needs a ticket.

[MartinBP]

Thanks for the feedback.
I have added a few comments below.

As this is a global cache it might have side effects when the object is used for more than one analysis. This applies to the tests and might also apply to the GUI. As the suppressions are always mutable if both of them have usage issues they have been pre-existing. We should check and fix any issues before we merge this to make sure things don't get worse.

I'm afraid that this expands a little beyond the knowledge I currently have of this project.

We should also add a performance test for this in Python.

Is this something you (or someone else) can assist with?
Or can you point me to some performance test that has already been implemented to get some inspiration.

Also needs a ticket.

Should I create one? In that case I will just need to request access first.