Fix MarkdownRenderer

[jeffpatton1971]

CrefToMarkdown was not pulling in the entire namespace, so it was using the text after the first '.' and thus generating invalid filenames "xml2doc.md"

[jeffpatton1971]

What changed and why

• Fixed per-type member links in CrefToMarkdown: previously used id.Split('.')[0] which grabbed only the first namespace segment (e.g., “Xml2Doc”), producing links like Xml2Doc.md#....
• Now we extract the containing type using the last dot before the method parameter list (or the last dot overall), yielding the full type ID (e.g., Xml2Doc.Core.MarkdownRenderer). This generates correct links such as:
• Docs/Xml2Doc.Core.MarkdownRenderer.md#method-rendertodirectorystring

After change

• Rebuild and regenerate docs. The link in Xml2Doc.Core.MarkdownRenderer.md for “RenderToDirectory(string)” should point to: Docs/Xml2Doc.Core.MarkdownRenderer.md#method-rendertodirectorystring

[jeffpatton1971]

• The failing assertion expected a link label “Transform(...)”, but ShortLabelFromCref didn’t convert method-level generic arity tokens (``n) in the method name. It produced “Transform(...)” instead, so your regex didn’t match.
• The change adds a replacement that maps method generic arity to “<T1,...>” (e.g., Transform``1 → Transform), matching the pattern used in your test.
• Links already point to the correct per-type file and anchor (fixed earlier by selecting the containing type via the last dot before “(”); this update aligns the link label with your expectations.