https://bugzilla.novell.com/show_bug.cgi?id=479874 User mkudlvasr@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=479874#c7 Martin Kudlvasr <mkudlvasr@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |michael.meeks@novell.com --- Comment #7 from Martin Kudlvasr <mkudlvasr@novell.com> 2009-03-25 08:51:25 MST --- Well, now it's split into 2 different problems: problem 1: Should the last dialog have F10 mapped to cancel? Argumentation of Lukas is very logical in this part and I agree, that user should be really warned before he wipes the disks. problem 2: Behaviour specified in YCPDialogParser::parseWidgetTreeTerm, YCPDialogParser::parsePushButton, YPushButton::setFunctionKey works like this: role of a PushButton is determined by Key shortcut, explicitely specified role, id string guess - in this order. Which means, that is you specify key shortcut, button role will be always assigned according to this. `PushButton (`id (`cancel), `opt(`cancelButton, `key_F10, `default) will have OkButton role regardless of the role you specify in opt. With this behaviour there is no way to specify in YCP a pushbutton with cancelButton role and F10 shortcut. Maybe this behaviour was programmed into yast to force equivalency of Ok,Cancel,Help button and F10,F9,F1 keys. Yast2-gtk only correctly behaves according to (previously decided) button role, so this is not yast2-gtk problem. I propose we agree on rules by which button roles and function keys are decided and then change them in code. Proposed rules: Button role is determined by (in this order): role option key option widget id After that key shortcut is determined by (in this order): key option button role -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.