adding few more c++17 operators override.

[devnexen]

No description provided.

[nwf-msr]

Discarding the alignment requirement seems dubious? While it's generally true that snmalloc prefer to align things much more strongly than generally required, it's still possible to ask for yet stronger alignment. I think the correct thing to do is to use

snmalloc/src/mem/sizeclasstable.h

Line 131 in d0df2d0

aligned_size(size_t alignment, size_t size)

as done with

snmalloc/src/override/malloc.cc

Lines 158 to 174 in d0df2d0

	SNMALLOC_EXPORT void*
	SNMALLOC_NAME_MANGLE(memalign)(size_t alignment, size_t size)
	{
	if ((alignment == 0) || (alignment == size_t(-1)))
	{
	errno = EINVAL;
	return nullptr;
	}
	if ((size + alignment) < size)
	{
	errno = ENOMEM;
	return nullptr;
	}
	return SNMALLOC_NAME_MANGLE(malloc)(aligned_size(alignment, size));
	}

I do not know to what extent those checks are also required by C++ .

[nwf-msr]

Suggested change

	void operator delete(void* p, std::align_val_t)EXCEPTSPEC
	void operator delete(void* p, std::align_val_t) EXCEPTSPEC

I'm surprised clang-format was OK with that without the whitespace.

[nwf-msr]

... oh wow, it in fact requires that it not be there. That... that sure seems like a bug, but I guess it's best to do what it wants. Sorry!

[mjp41]

@nwf-msr sorry I am confused now. I think @devnexen has now added the space, but are you saying the space shouldn't be there?

[devnexen]

by the way is clang-format9 still recommended ? Thankfully I still have it somewhere.

[mjp41]

It is, but I might move this and Verona forward to a newer version. It is getting a bit long in the tooth now.

[mjp41]

@devnexen would be able to run clang-format to unblock the CI?

Changes have been addressed from this review and @nwf-msr is OOF at the moment.