<div class="gmail_quote">Hmm.. I've added these functions just recently: <a href="http://websvn.kde.org/?view=revision&revision=1099644">http://websvn.kde.org/?view=revision&revision=1099644</a><br><br>The logic was the following:<br>
<br>extent() - is the area where the node has an influence on the result image. In other words, it is an area, where we should update when the layer changes it's visibility.<br><br>KisMask::extent() - works right this way. It tells the area, where selection is. And this is right for most of the masks, except of the transparency mask. A transparency mask changes the entire image, except of the selection. It means that we should update everywhere, except of the selection. And if we don't have parent, we can't do anything.<br>
<br>So, i think, you should fix KisSelectionTest instead to confirm this logic. Or this logic is wrong somewhere?<br><br><br>On Wed, Mar 24, 2010 at 4:36 PM, Sven Langkamp <span dir="ltr"><<a href="mailto:sven.langkamp@gmail.com">sven.langkamp@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">Hi,<br><br>I have attached a patch to fix the KisSelectionTest. The problem was that the extent of a KisTransparencyMask without a parent is always an empty QRect,<br>
Now the code from KisMask is used that uses the selection extent when there is a selection.<br>
<br>Can you have a look if that is correct?<br><font color="#888888"><br><br>Sven<br>
</font><br>_______________________________________________<br>
kimageshop mailing list<br>
<a href="mailto:kimageshop@kde.org">kimageshop@kde.org</a><br>
<a href="https://mail.kde.org/mailman/listinfo/kimageshop" target="_blank">https://mail.kde.org/mailman/listinfo/kimageshop</a><br>
<br></blockquote></div><br><br clear="all"><br>-- <br>Dmitry Kazakov<br>