[PATCH 0/1] MR2915: mshtml: Allow protocols with no ports when checking target origin.

[PATCH 1/1] mshtml: Allow protocols with no ports when checking target origin.

[Gabriel Ivăncescu]

From: Gabriel Ivăncescu

Signed-off-by: Gabriel Ivăncescu 
---
 dlls/mshtml/htmlwindow.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/dlls/mshtml/htmlwindow.c b/dlls/mshtml/htmlwindow.c
index 6c550172d68..e05dc143c08 100644
--- a/dlls/mshtml/htmlwindow.c
+++ b/dlls/mshtml/htmlwindow.c
@@ -3219,8 +3219,10 @@ static HRESULT check_target_origin(HTMLInnerWindow *window, const WCHAR *target_
         goto done;

hres = IUri_GetPort(uri, &port);
-    if(hres != S_OK)
+    if(hres != S_OK) {
+        hres = FAILED(hres) ? hres : S_OK;  /* some protocols don't have ports (e.g. res) */
         goto done;
+    }
     hres = IUri_GetPort(target, &port2);
     if(hres == S_OK && port != port2)
         hres = S_FALSE;

[Jacek Caban (＠jacek)]

Jacek Caban (@jacek) commented about dlls/mshtml/htmlwindow.c:
>      goto done;
>
>  hres = IUri_GetPort(uri, &port);
> - if(hres != S_OK)
> - if(hres != S_OK) {
> -
>    hres = FAILED(hres) ? hres : S_OK;  /* some protocols don't have ports (e.g. res) */

Should we make sure that target URI also doesn't have a port?

[Gabriel Ivăncescu (＠insn)]

On Tue May 30 11:30:40 2023 +0000, Jacek Caban wrote:
> Should we make sure that target URI also doesn't have a port?

I think that's unnecessary since we already tested if the protocols
match earlier, right? Unless I'm missing something.

[Jacek Caban (＠jacek)]

On Tue May 30 13:55:29 2023 +0000, Gabriel Ivăncescu wrote:
> I think that's unnecessary since we already tested if the protocols
> match earlier, right? Unless I'm missing something.

Port may be optional. In case of builtin protocols, urlmon is smart about it (see default_ports array), but that does not scale to arbitrary protocols.