Why are these functions preserved?

[AlexGuteniev]

I see that other preserved functions have _CRTIMP2 or similar, but not these.
Should these really be preserved?

STL/stl/src/winapisupp.cpp

Lines 157 to 175 in ed8150e

	// TRANSITION, ABI: preserved for binary compatibility
	extern "C" DWORD __cdecl __crtFlsAlloc(_In_opt_ PFLS_CALLBACK_FUNCTION const lpCallback) {
	return FlsAlloc(lpCallback);
	}
	// TRANSITION, ABI: preserved for binary compatibility
	extern "C" BOOL __cdecl __crtFlsFree(_In_ DWORD const dwFlsIndex) {
	return FlsFree(dwFlsIndex);
	}
	// TRANSITION, ABI: preserved for binary compatibility
	extern "C" PVOID __cdecl __crtFlsGetValue(_In_ DWORD const dwFlsIndex) {
	return FlsGetValue(dwFlsIndex);
	}
	// TRANSITION, ABI: preserved for binary compatibility
	extern "C" BOOL __cdecl __crtFlsSetValue(_In_ DWORD const dwFlsIndex, _In_opt_ PVOID const lpFlsData) {
	return FlsSetValue(dwFlsIndex, lpFlsData);
	}

[frederick-vs-ja]

I guess they don't actually need to be preserved. They don't appear in the result of dumpbin /exports, so perhaps we can remove them now.

[StephanTLavavej]

We can't remove things that were ever needed from the import lib, but since these weren't being injected there, it does seem quite plausible that they can be removed, and that the "preserved for ABI" comments were mistakenly added too mechanically.

[AlexGuteniev]

The header declarations were removed in #1194 .

Looks like it was my decision to keep these, after failure with removing _CRTIMP2, #1194 (comment):

Actually probably need to keep not only _CRTIMP2, but every function visible from header

And I wasn't corrected. Probably the PR was scary enough to do more removal.

[StephanTLavavej]

Yeah, there was a ton going on in that PR, and I didn't notice that they were never declared as exported. We've gotten better at noticing this as the codebase has gotten simpler (a lot of the weirdness in winapisupp was because it had organically grown so huge).