http_utils update using std::span

[devnexen]

No description provided.

[nunojpg]

We are storing a copy of the data at m_body, not a reference/pointer.

With std::span you just store a pointer, to you need to make sure the original passed data remains available.

Think of it as string vs string_view.

[reneme]

Suggested change

	DataSource_Memory src(vec.data(), vec.size());
	DataSource_Memory src(vec);

DataSource_Memory already learned to deal with std::span<>

[lieser]

We are storing a copy of the data at m_body, not a reference/pointer.

I agree that we should be careful about this. And make really sure that not storing a copy is what we want and brings us enough performance that we care about that it is worth the potential lifetime issue.

If we stay with the current storing a copy of data in Response, we should change the constructor to takes a std::vector<uint8_t> (and not a reference). If the caller no longer needs the body this would allow to avoid creating a copy of the body without the lifetime issue just storing a pointer to it has.