Maybe pull HTML detection up from rss-fetcher?

[philbudne]

The rss-fetcher tries to generate a clearer error Exception for HTML by catching ExpatError exceptions and sniffing the top of the document:

    p = sitemap_parser.XMLSitemapParser(url, text)
    try:
        sitemap = p.sitemap()
    except ExpatError:          # try translateing ExpatError
        top = response.content[:1024].lower()
        if top.find(b'<!doctype') or top.find(b'<html'):
            raise Exception("html?")
        raise

Maybe pull this up into XMLSitemapParser.sitemap so that mc_sitemap_tools.discover (etc) can also benefit?