Work around reparenting issues in #40

I'm still really curious why this double reparent would be needed
though. Everything works fine on any other i3 config I've tried, and
even the exact same config in a VM works fine for me.

CHANGELOG.md

@@ -30,6 +30,11 @@ Versioning](https://semver.org/spec/v2.0.0.html).
   the suggestion to enable the `hack_reaper_update_display` workaround for
   REAPER when it is not already enabled.
 
+### Fixed
+
+- Added a workaround for reparenting issues with the plugin editor GUI on a
+  [specific i3 setup](https://github.com/robbert-vdh/yabridge/issues/40).
+
 ## [1.6.1] - 2020-09-28
 
 ### Fixed

src/wine-host/editor.cpp

@@ -179,13 +179,20 @@ Editor::Editor(const Configuration& config,
     // child window ourselves. This is a hack to work around the issue's
     // described in `Editor`'s docstring'.
     xcb_reparent_window(x11_connection.get(), wine_window, parent_window, 0, 0);
-    xcb_map_window(x11_connection.get(), wine_window);
     xcb_flush(x11_connection.get());
 
     ShowWindow(win32_handle.get(), SW_SHOWNORMAL);
     if (win32_child_handle) {
         ShowWindow(win32_child_handle->get(), SW_SHOWNORMAL);
     }
+
+    // HACK: I can't seem to figure why the initial reparent would fail on this
+    //       particular i3 config in a way that I'm unable to reproduce, but if
+    //       it doesn't work the first time, just keep trying!
+    //
+    //       https://github.com/robot-vdh/yabridge/issues/40
+    xcb_reparent_window(x11_connection.get(), wine_window, parent_window, 0, 0);
+    xcb_flush(x11_connection.get());
 }
 
 Editor::~Editor() {