From 7531bb0b9374dbb669d8e0e88a444593198066fe Mon Sep 17 00:00:00 2001
From: Antoine Denommee Pigeon <adenommee@summit-tech.ca>
Date: Wed, 16 Mar 2022 13:11:09 -0400
Subject: [PATCH] Correctly handle badly formatted protocols

Remove the bad protocol from the URL instead of parsing that protocol inside the domain.

Change-Id: Ibde25c59914236c67c72f3a882e66b03645f8789
---
 src/hackney_url.erl | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)

diff --git a/src/hackney_url.erl b/src/hackney_url.erl
index bd9c063c..2ebe19a5 100644
--- a/src/hackney_url.erl
+++ b/src/hackney_url.erl
@@ -52,7 +52,8 @@ parse_url(URL) ->
   parse_url(URL, #hackney_url{transport=hackney_tcp, scheme=http}).
 
 parse_url(URL, S) ->
-  {URL1, Fragment} =  parse_fragment(URL),
+  URL0 = parse_protocol(URL),
+  {URL1, Fragment} =  parse_fragment(URL0),
   case binary:split(URL1, <<"/">>) of
     [URL1] ->
       parse_addr1(URL1, S#hackney_url{raw_path = raw_fragment(Fragment),
@@ -277,6 +278,14 @@ parse_fragment(S) ->
       {S1, F}
   end.
 
+parse_protocol(S) ->
+  case binary:split(S, <<"://">>) of
+    [Url] ->
+      Url;
+    [_InvalidProtocol, Url] ->
+      Url
+  end.
+
 
 %% @doc Decode an URL encoded binary.
 %% @equiv urldecode(Bin, crash)