Angus Leeming <[EMAIL PROTECTED]> writes: > I don't think that the patch should be committed as it is, although it has > enabled Andreas to work out what was going wrong --- good detective work, > Andreas!
Thanks! :-) > > It seems to me that the global 'bool showing' should actually be a > one-per-dialog variable. ... > Do I miss anything? > The "showing" flag only protects the call to show() itself, therefore it should be named "opening". The dialog is shown between the call to show() and the subsequent hide(). IMO we don't need machinery to allow a show() inside a show(). Ciao Andreas