----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: http://git.reviewboard.kde.org/r/102671/#review6765 -----------------------------------------------------------
This has _basically_ been approved in http://lists.kde.org/?l=kde-core-devel&m=131616667329975&w=2 by Aaron - dunno if the patch has been applied yet or will be applied at all... :/ - Thomas On Sept. 20, 2011, 12:18 p.m., Thomas Gahr wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > http://git.reviewboard.kde.org/r/102671/ > ----------------------------------------------------------- > > (Updated Sept. 20, 2011, 12:18 p.m.) > > > Review request for kdelibs. > > > Summary > ------- > > This patch changes KDialog::screenRect to call > QDesktopWidget::screenGeometry(widget) instead of QDesktopWidget::geometry(), > which returns the wrong geometry if a second monitor has been connected and > disconnected. > This fixes a bug described here: > http://article.gmane.org/gmane.comp.kde.devel.core/71875 > The testcase and results can be found here: > http://article.gmane.org/gmane.comp.kde.devel.core/71911 > > Behaviour of QDesktopWidget::geometry() vs. > QDesktopWidget::screenGeometry(widget) are equivalent in case of non-faulty > ::geometry() and fix the problem in case of faulty ::geometry(). So I > wouldn't consider this a workaround that needs to be maintained/changed back > once (if) the bug in Qt has been fixed. On the contrary: using > ::screenGeometry(widget) is more verbose than ::geometry(), which make more > sense in the context of KDialog::screenRect(QWidget* widget,int screen) IMHO > > > Diffs > ----- > > kdeui/dialogs/kdialog.cpp 0cabb85 > > Diff: http://git.reviewboard.kde.org/r/102671/diff > > > Testing > ------- > > Tested successfully with 4.7.0 on Fedora 15 and current master > > > Thanks, > > Thomas > >
