I'm tempted to think we should bail if GNOME_ACCESSIBILITY is set since we should bail after this if GNOME_ACCESSIBILITY is set, otherwise nobody will ever check the return message.
>+ if (!bus)
>+ return;
>+ dbus_connection_set_exit_on_disconnect(bus, FALSE);
nit, blank line.
>+ if (a11yPendingCall) {
nit,if ( !a11yPendingCall) goto dbus_done;
also we don't usually put a11y in local variables.
Comment on attachment 581750
Updated patch.
>+PreInit()
>+{
>+#ifdef MOZ_ENABLE_DBUS
>+ static bool sChecked = FALSE;
>+ if (sChecked)
>+ return;
>+ sChecked = TRUE;
nit, blank line before sChecked = true
I'm tempted to think we should bail if GNOME_ACCESSIBILITY is set since we should bail after this if GNOME_ACCESSIBILITY is set, otherwise nobody will ever check the return message.
>+ if (!bus) _set_exit_ on_disconnect( bus, FALSE);
>+ return;
>+ dbus_connection
nit, blank line.
>+ if (a11yPendingCall) {
nit,if ( !a11yPendingCall) goto dbus_done;
also we don't usually put a11y in local variables.
>+ dbusSuccess = true;
>+ }
>+
>+ break;
nit, put break in block above.
>diff --git a/widget/ src/gtk2/ nsWindow. cpp b/widget/ src/gtk2/ nsWindow. cpp tyService. h" yService. h" ocument. h"
> #ifdef ACCESSIBILITY
>-#include "nsIAccessibili
>+#include "nsAccessibilit
> #include "nsIAccessibleD
>-#include "prenv.h"
>-#include "stdlib.h"
>
> using namespace mozilla;
> using namespace mozilla::widget;
btw while I suppose ti doesn't really happen since these are present later unconditionally why are these here?