Thanks for the answers and the patch. The patch worked :) The sequences of images is taken from the same vector of image. The method that takes the sequences is in the Sources/Library/ImageCanvas.cpp file, in the ImageCanvas::getCanvasForSave() method. If does a simple task: 1. Makes a copy of a vector which contains the sequences of the image 2. Resizes all images on the copied vector some specific size (Big, MediumBig, MediumSmall, Small) 3. Add an optional logo. 4. Returns the resized image sequence.
I think it should not be the source of the problem, because I tried to explicitly create an empty vector, and iterate over the sequences of images from the original vector to create a new Image variable (on the stack), and push manually to the vector; and it still crashed (without the patch). Btw, does the failing tests have to do with the drawing on the Images? Because I might be troubled again if it is, because the program does many drawing on images. I'll be doing my tests. Thanks a lot! :) On Thu, 2008-04-03 at 10:42 -0700, [EMAIL PROTECTED] wrote: > The patch we sent causes one of our regression tests to fail so it seems > likely a robust patch will require some deep thought. We need to disassociate > a I/O handler when the image is cloned so each thread has its own I/O > handler but in the mean-time it looks likely doing so will break other parts > of ImageMagick. Stand by while we do some testing. > > In the mean-time can you identify the code block in your program that > generates the 4 different image sequences? Somehow they are ending up with > the same I/O handler and we need to know why that is. _______________________________________________ Magick-users mailing list [email protected] http://studio.imagemagick.org/mailman/listinfo/magick-users
