1 files changed, 74 insertions, 0 deletions
diff --git a/internal/onionloc/onionloc.go b/internal/onionloc/onionloc.go
new file mode 100644
index 0000000..63e512a
--- /dev/null
+++ b/internal/onionloc/onionloc.go
@@ -0,0 +1,74 @@
+package onionloc
+
+import (
+	"net/http"
+	"strings"
+
+	"golang.org/x/net/html"
+)
+
+const (
+	HTTPHeaderName = "Onion-Location"
+)
+
+func HTTP(rsp *http.Response) (string, bool) {
+	v, ok := rsp.Header[HTTPHeaderName]
+	if !ok {
+		return "", false
+	}
+	if len(v) != 1 {
+		return "", false
+	}
+	return v[0], true
+}
+
+func HTML(rsp *http.Response) (string, bool) {
+	z := html.NewTokenizer(rsp.Body)
+	for {
+		tt := z.Next()
+		if tt == html.ErrorToken {
+			return "", false // EOF and other errors
+		}
+
+		switch tt {
+		case html.StartTagToken, html.SelfClosingTagToken:
+			t := z.Token()
+
+			// Looking for the html meta tag, see:
+			// https://www.w3schools.com/tags/tag_meta.asp
+			//
+			// We expect two attributes: "key" and "content"
+			if strings.ToLower(t.Data) != "meta" {
+				break
+			}
+			if len(t.Attr) != 2 {
+				break
+			}
+
+			// We're looking for the "http-equiv" key, see:
+			// https://www.w3schools.com/tags/att_meta_http_equiv.asp
+			//
+			// In particular with the value "onion-location", see:
+			// https://community.torproject.org/onion-services/advanced/onion-location/
+			//
+			// If we have all this and a following content
+			// attribute, that is the Onion-Location URL that this
+			// page advertises.  We make no attempt to validate
+			// whether the content value is really an onion address.
+			attr := t.Attr[0]
+			if strings.ToLower(attr.Key) != "http-equiv" {
+				break
+			}
+			if strings.ToLower(attr.Val) != "onion-location" {
+				break
+			}
+			attr = t.Attr[1]
+			if strings.ToLower(attr.Key) != "content" {
+				break
+			}
+
+			return attr.Val, true
+		default:
+		}
+	}
+}