| Summary: | mask origin is messed up on saving/loading | ||
|---|---|---|---|
| Product: | [Applications] krita | Reporter: | Halla Rempt <halla> |
| Component: | General | Assignee: | Dmitry Kazakov <dimula73> |
| Status: | RESOLVED FIXED | ||
| Severity: | critical | ||
| Priority: | NOR | ||
| Version: | unspecified | ||
| Target Milestone: | --- | ||
| Platform: | unspecified | ||
| OS: | Linux | ||
| Latest Commit: | http://commits.kde.org/calligra/992c5c23ea09b525dd90e1aed1e4ada7406d75dd | Version Fixed In: | |
| Sentry Crash Report: | |||
| Attachments: | image with mask to play with | ||
Created attachment 72866 [details]
image with mask to play with
Was already present in 2.4 Git commit 348ee8b634ffd41bc59997e1dee841d6df2da259 by Dmitry Kazakov. Committed on 26/10/2012 at 13:23. Pushed by dkazakov into branch 'master'. Fixed loading of masks with the offset That patch introduces the following changes: 1) Removed lazy initialization of masks' selection. Now you always must call initSelection after the creation of the mask (probably, we can move it into c-tor). 2) KisMask::setSelection must only be used when we need to switch the underlying selection explicitly (like, e.g. in KisImage::setGlobalSelection). In the rest of the cases you need to work with already existent selection. M +1 -0 krita/image/kis_image.cc M +6 -23 krita/image/kis_mask.cc M +3 -0 krita/image/tests/kis_filter_mask_test.cpp M +26 -17 krita/image/tests/kis_mask_test.cpp M +2 -0 krita/image/tests/kis_transparency_mask_test.cpp M +24 -0 krita/sdk/tests/testutil.h M +11 -14 krita/ui/kra/kis_kra_load_visitor.cpp M +1 -1 krita/ui/kra/kis_kra_load_visitor.h M +11 -8 krita/ui/kra/kis_kra_loader.cpp M +4 -4 krita/ui/kra/kis_kra_loader.h http://commits.kde.org/calligra/348ee8b634ffd41bc59997e1dee841d6df2da259 Git commit 7da5237740f207f41227a5e12dfca246fa37497b by Dmitry Kazakov. Committed on 26/10/2012 at 13:23. Pushed by dkazakov into branch 'vector_compositioning_kazakov'. Fixed loading of masks with the offset That patch introduces the following changes: 1) Removed lazy initialization of masks' selection. Now you always must call initSelection after the creation of the mask (probably, we can move it into c-tor). 2) KisMask::setSelection must only be used when we need to switch the underlying selection explicitly (like, e.g. in KisImage::setGlobalSelection). In the rest of the cases you need to work with already existent selection. M +1 -0 krita/image/kis_image.cc M +6 -23 krita/image/kis_mask.cc M +3 -0 krita/image/tests/kis_filter_mask_test.cpp M +26 -17 krita/image/tests/kis_mask_test.cpp M +2 -0 krita/image/tests/kis_transparency_mask_test.cpp M +24 -0 krita/sdk/tests/testutil.h M +11 -14 krita/ui/kra/kis_kra_load_visitor.cpp M +1 -1 krita/ui/kra/kis_kra_load_visitor.h M +11 -8 krita/ui/kra/kis_kra_loader.cpp M +4 -4 krita/ui/kra/kis_kra_loader.h http://commits.kde.org/calligra/7da5237740f207f41227a5e12dfca246fa37497b Git commit 992c5c23ea09b525dd90e1aed1e4ada7406d75dd by Jaroslaw Staniek, on behalf of Dmitry Kazakov. Committed on 26/10/2012 at 13:23. Pushed by staniek into branch 'kexi-altertable-staniek'. Fixed loading of masks with the offset That patch introduces the following changes: 1) Removed lazy initialization of masks' selection. Now you always must call initSelection after the creation of the mask (probably, we can move it into c-tor). 2) KisMask::setSelection must only be used when we need to switch the underlying selection explicitly (like, e.g. in KisImage::setGlobalSelection). In the rest of the cases you need to work with already existent selection. M +1 -0 krita/image/kis_image.cc M +6 -23 krita/image/kis_mask.cc M +3 -0 krita/image/tests/kis_filter_mask_test.cpp M +26 -17 krita/image/tests/kis_mask_test.cpp M +2 -0 krita/image/tests/kis_transparency_mask_test.cpp M +24 -0 krita/sdk/tests/testutil.h M +11 -14 krita/ui/kra/kis_kra_load_visitor.cpp M +1 -1 krita/ui/kra/kis_kra_load_visitor.h M +11 -8 krita/ui/kra/kis_kra_loader.cpp M +4 -4 krita/ui/kra/kis_kra_loader.h http://commits.kde.org/calligra/992c5c23ea09b525dd90e1aed1e4ada7406d75dd |
The origin of a mask is wrong after saving and loading an image. Reproducible: Always Steps to Reproduce: 1. create an image 2. create a (transparency mask) 3. crop or move the mask around 4. save 5. load Actual Results: The mask is shown offset to the bottom-right. Expected Results: It should be in the correct location 06:06:01 * PovAddict runs in circles screaming 06:08:45 * eages gets back and is mildly confused. 06:09:23 < PovAddict> krita messed up my handiwork when I saved it 06:09:55 < PovAddict> I fixed it, saved file, closed krita, reopened the file and it's messed up in the same way again 06:09:57 * PovAddict is not amused 08:32:40 < boud> PovAddict: that sounds serious. What happened exactly? 08:32:52 < PovAddict> I'll try to repro with a simpler case later 08:32:57 < PovAddict> I made a large canvas 08:33:06 < PovAddict> pasted two images as layers (both smaller than the canvas) 08:33:17 < PovAddict> changed their positions 08:33:33 < PovAddict> carefully and timeconsumingly created a transparency mask on the top layer 08:34:17 < PovAddict> finally I cropped the canvas to remove the large empty space around the actual images (I made the canvas larger than necessary to let me move the layers around more freely) 08:34:46 < PovAddict> I saved the image, and when I reopened it, the transparency mask was shifted from its position quite a lot 08:35:03 < boud> hm, that's not good. Which version of krita are you using? 08:35:25 < PovAddict> I think the offset from the correct position is related to how much I cropped 08:35:39 < PovAddict> anyway, I used the move tool to put the transparency mask layer back in place 08:35:41 < PovAddict> saved the image 08:35:43 < PovAddict> reopened it 08:35:47 < PovAddict> and it was out of place again 08:36:13 < PovAddict> krita 2.4.3 08:36:22 < boud> ok, let me try to reproduce 08:37:01 < PovAddict> I'll try tomorrow again, using simpler images (I can't give you my .kra, due to privacy and due to 50MB+) 08:38:09 < boud> yes, I can easily reproduce 08:38:29 < boud> the mask is more to the bottom-right than it should be 08:38:36 < PovAddict> oh perfect 08:39:32 < boud> now I'll check 2.5rc1 08:40:12 < boud> still in 2.5, okay, I'll create a bug report