[MERGED] check all definitions of a function before suggesting 'const'

[YashasSamaga]

What this PR does / why we need it:

const qualifier suggestions were being given based solely on the function definition that was currently being looked at. This caused wrong suggestions when one state of a function modifies the argument while another state did not.

The commit in this PR adds code to share usage details across multiple definitions of the functions. Hence, when the const qualifier is suggested, it takes into account of all the function states.

The commit is just a proof-of-concept for the time being. The commit only shares uWRITTEN status only but it could be generalized to share other usage flags as well which could solve #371 (comment)

Which issue(s) this PR fixes:

Fixes #371

What kind of pull this is:

• A Bug Fix
• A New Feature
• Some repository meta (documentation, etc)
• Other

Additional Documentation:

[YashasSamaga]

Ready.

[stale]

This issue has been automatically marked as stale because it has not had recent activity.

[Daniel-Cortez]

This seems like a useful bugfix, as it saves from the need to manually apply #pragma unused as a workaround in all function states that don't modify the argument.
@YashasSamaga Why did you close this PR?

[YashasSamaga]

I lost touch with the code and I planned on writing a generic system for any flag instead of just uWRITTEN but never got time to do it. Also got annoyed with passing arguments to functions in this and the {...} PR. I think a better solution is required than having to pass int written everywhere or maybe I am thinking too much.

[Daniel-Cortez]

Hmm... I see what you mean; a few functions in sc1.c indeed do have an unusually high number of arguments. But implementing a flag system might complicate the code even more. And even if it won't, does it really have to be done in this PR, and not in a separate one later (after this and the {...} PRs are merged)?