Keep a multi-line setext heading setext when re-rendering Markdown - #304

MarkdownRenderer renders every setext heading as an ATX heading. That only works when the heading is on one line; for a heading that spans lines, the lines after the first become a paragraph:

>>> md = marko.Markdown(renderer=MarkdownRenderer)
>>> md("Foo\nBar\n---\n")
'## Foo\nBar\n'

<!-- before -->
<h2>Foo
Bar</h2>
<!-- after re-rendering -->
<h2>Foo</h2>
<p>Bar</p>

With emphasis across the lines (Foo *bar\nbaz*\n====, spec example 81) the emphasis breaks too.

render_setext_heading() now keeps the setext form when the rendered heading contains a line break, with each line and the ===/--- underline prefixed for quotes and list items. One-line setext headings are still rendered as ATX headings, as before.

Added test_markdown_renderer_multiline_setext_heading with four cases (plain, emphasis across lines, in a quote, in a list item); they fail on master. Across the CommonMark spec examples, the number whose re-rendered Markdown gives different HTML goes from 27 to 24. pytest tests, black, isort and flake8 pass, and mypy reports the same 5 errors as on master.

Used Claude (Opus) to help find and patch this; I reviewed and tested the change