I confirm, but I think it's a problem of correctly updating the projection
(like a dirty rect isn't set correctly).

On Thursday 28 September 2006 02:18, Moritz Göbelbecker wrote:
> When you draw on a layer while there is a layer mask, the screen will not
> display the image correctly (Edit Mask was disabled, of course): 

All of it confirmed, I'll have a better look at what happens tonight. Thanks 
for reporting this :-)

SVN commit 589923 by coppens:

Ok, I think I found the problem. Basically, reverting r587841 on kis_painter.cc fixes it (took me a while to realize the problem wasn't with my code, but with KisPainter, bah :-( ). The issue is that using oldRawData uses the data from _before_ starting to paint on the mask, which obviously isn't what we want.
Cyrille, this was introduced by you with commit message 'better fix, with optimization this time :)', so I'm hoping this doesn't break anything (Regression testing for 2.0, we _need_ it!)? Also, I'd like to know what this was supposed to optimize (well I can guess at it it, but I'd like to know for sure :]).

BUG:134770
CCMAIL:cberger@cberger.net

 M  +1 -1      kis_painter.cc  


--- branches/koffice/1.6/koffice/krita/core/kis_painter.cc #589922:589923
@@ -336,7 +336,7 @@
             //const Q_UINT8 *srcData = srcdev->pixel(srcX, srcY);
             Q_INT32 srcRowStride = srcdev->rowStride(srcX, srcY);
             KisHLineIteratorPixel srcIt = srcdev->createHLineIterator(srcX, srcY, columns, false);
-            const Q_UINT8 *srcData = srcIt.oldRawData();
+            const Q_UINT8 *srcData = srcIt.rawData();
 
             //const Q_UINT8 *selData = seldev->pixel(dstX, dstY);
             Q_INT32 selRowStride = seldev->rowStride(dstX, dstY);