better fix for inexistant widgets because of i18n bugs

(this will only result in no entry menu to switch that or that option)

also, it has the nice side effet to keep current options setting on exit even
for eg: wizard option wich can only be set when wizards are installed.
thus if one remove wizards, his wizard options are kept and will be there
as before when he'll install them back.

btw, we should really write a tool to check po so that menu entrys are always
translated the same way, whether they have the _ (menu declaration for key
shortcuts) or not (widget access by path), else it badly broke widget access.

and we should enforce its execution on commit to prevent having broken
translations in cvs.

1 files changed, 7 insertions, 7 deletions

diff --git a/control-center b/control-center
index a5d0eb9d..6d6af974 100755
--- a/control-center
+++ b/control-center
@@ -537,14 +537,14 @@ sub update_exp {
 
 $emb_box->set_size_request(-1, $index * 50);
 
-eval {
-    if (defined $show_log_check_box) {
-     $show_log_check_box->set_active($show_logs);
-     $embedded_check_box->set_active($embedded);
-     $expert_wizard_check_box->set_active($expert_wizard) if $::isWiz;
-    };
+foreach([ $show_log_check_box, $show_logs ], [ $embedded_check_box, $embedded ], if_($::isWiz, [ $expert_wizard_check_box, $expert_wizard ])) {
+    my ($widget, $value) = @$_;
+    if (defined $widget) {
+     $widget->set_active($value);
+    } else {
+     print STDERR "BUG with LANGUAGE $ENV{LANGUAGE}\n";
+    }
 };
-print STDERR "BUG with LANGUAGE $ENV{LANGUAGE}\n" if $@;
 
 hide_socket_and_clean();