Eliminate pynac's handle_factor method

[gh-DaveWitteMorris]

After #31554, the handle_factor method will be used in only 3 places in the pynac source code:

expairseq.cpp:994:  ex newfactor = mf.handle_factor(elem, _ex1);
expairseq.cpp:1003: ex newfactor = mf.handle_factor(elem, _ex1);
expairseq.cpp:1049: combine_overall_coeff(ex_to<numeric>(mf.handle_factor(elem.rest, _ex1)));

It is completely unnecessary in those places, because handle_factor(x, 1) is equal to x. So we should refactor the code to get rid of it (and the class make_flat_inserter that contains it).

Here is the definition:

class make_flat_inserter
{
	public:
		make_flat_inserter(const epvector &epv, bool b)
		{}
		make_flat_inserter(const exvector &v, bool b)
                {}
		ex handle_factor(const ex &x, const ex &coeff)
		{
			if (is_exactly_a<numeric>(coeff) and coeff.is_zero())
				return coeff;
		        return x;
		}
};

[chapoton]

tentative commit

---

New commits:

​b2d3661

remove handle_factor

[chapoton]

seems to go smoothly..

[mkoeppe]

Upstream has slightly more substantial code there ​https://www.ginac.de/ginac.git/?p=ginac.git;a=blob;f=ginac/expairseq.h;h=0618e2d777e8ecc29876f643bbe262425486d903;hb=HEAD