Remove FPP dependencies on native int types

[bocchino]

Change Description

This PR makes changes necessary to remove the dependencies of the FPP model and generated C++ on the native int types NATIVE_INT_TYPE and NATIVE_UINT_TYPE.

Incidental changes:

• I reformatted a few files via clang-format. We want to do this reformatting anyway on a larger scale. Reformatting is not the primary purpose of this PR, but the reformatting helped me understand the code. Since reformatting obscures change/delete info, I reformatted only hpp files, so you can still see the change/delete info in the corresponding cpp files.

• I added some comments in FpConfig.h explaining what the configurations do.

• I renamed FwBuffSizeType to FwSizeStoreType as discussed. I added a type alias for FwBuffSizeType for backwards compatibility.

Rationale

This is a first step in removing NATIVE_INT_TYPE and NATIVE_UINT_TYPE from the code base.

Next Steps

1. Make an alpha release of FPP that eliminates references to NATIVE_INT_TYPE and NATIVE_UINT_TYPE. At this time we can replace NATIVE_INT_TYPE with FwIndexType in the portNum arguments of the port handler implementations.
2. Scrub NATIVE_INT_TYPE and NATIVE_UINT_TYPE from the code in this repo.
3. In FpConfig.h, replace configurable types except for PlatformInt/Uint with fixed-width types, to eliminate static analysis warnings.

Related Issues

nasa/fpp#391

[github-advanced-security]

CodeQL found more than 20 potential problems in the proposed changes. Check the Files changed tab for more details.

[bocchino]

Note: There are many false-positive warnings due to (1) problems in F Prime that this PR is addressing or (2) problems with the static analysis. Dismissing these warnings one-by-one in the GUI is too painful. I gave up. (You have to click dismiss, wait a few seconds until it resets to the top of the changed file list, and scroll back manually to where you were.)

[bocchino]

There are many false-positive warnings due to (1) problems in F Prime that this PR is addressing or (2) problems with the static analysis.

It turns out that the static analysis does not work as expected with regard to fixed-width types. It seems that when you declare a variable

Foo bar;

It wants Foo to be a literal fixed-width numeric type like int32_t or maybe I32. It won't accept a general type alias, even if the aliased type has a fixed width. This seems like a bug in the static analysis.

[LeStarch]

1. Review need for pointer cast in model. If needed, switch to PlatformPointerCase
2. Review type event counter derived types

[LeStarch]

Make them all enum store type, but keep queue size as it is.

[bocchino]

The event counters are for counting events emitted at runtime, for purposes of throttling. So I think they should be FwIndexType.

[bocchino]

With the change of FwSizeType to I64, there are a lot of implicit casts from I64 to NATIVE_UINT_TYPE and/or U32. I for this reason, I had to disable -Wconversion in the FPP compilation checks (too many warnings). This issue will be fixed in due course as we replace NATIVE_UINT_TYPE and U32 with FwSizeType, e.g., in the Fw::Buffer interface.

[bocchino]

@LeStarch I made the requested changes. I also had to introduce a type alias into Fw::Serializable, to resolve the discrepancy between the declared types NATIVE_UINT_TYPE and the eventual types FwSizeType. We'll have FPP refer to the alias and change the alias later. That way we decouple the FPP changes from the F Prime changes.

[LeStarch]

I see the expected changes fixed!