fix: the crawler and the Chrome guard keep to the site's port

Fetcher.Allowed and Key, the rendered comparison and the Chrome request
guard accept an address only on the site's effective port (explicit,
else 443 for https and 80 for http), so page code cannot make headless
Chrome reach other services on the site host's address and a link to
another port is not reported as a page of the site.

classes/crawl/checks.go

@@ -731,7 +731,7 @@ func checkRenderedPage(res *Result, p *Page, rec *Record) []Finding {
 	if p.Rendered == nil || res.fetcher == nil {
 		return nil
 	}
-	out := compareRendered(p.Facts, p.Rendered, res.fetcher.Hosts())
+	out := compareRendered(p.Facts, p.Rendered, res.fetcher.Key)
 	for i := range out {
 		out[i].URL, out[i].Locale, out[i].Record = p.Key, locale(rec), rec
 	}