Got latest alpha build to see if the multi-monitor resolution offset problem was fixed, or to see if at least manually changing the resolution still results in total pressure sensitivity loss but now simply performing brush strokes on a new canvas will crash Krita with 100% reproducibility. I can usually make one brush stroke, but a second one will cause a crash. I then ran it in debug mode to collect the print out from Dr. Mingw. The contents are pasted in the details section. It is a three monitor setup with a Cintiq 21UX as the third monitor on Windows 7. Reproducible: Always Steps to Reproduce: 1. Open Krita 2. Create new document. 3. Touch and drag stylus across tablet surface. Lift. 4. Touch stylus to surface again. 5. Crash. Actual Results: 100% crash. Expected Results: Performed a second brush stroke. Here is the print out from Dr. Mingw: krita.exe caused an Access Violation at location 000000006754BB3A in module libkritapigment.dll Reading from location FFFFFFFFFFFFFFFF. AddrPC Params 000000006754BB3A 0000000000000002 0000000000000020 000000002321F8B0 libkritapigment.dll!KoStreamedMath<(Vc_1::Implementation)6>::genericComposite<false, false, OverCompositor32<unsigned char, unsigned int, false, true>, 4> 000000006763885D 0000000013727960 000000006F63E635 000000002321FAF0 libkritapigment.dll!KoOptimizedCompositeOpOver32<(Vc_1::Implementation)6>::composite 00000000674DCCBB 00000200000004C0 0000023F000004FF 0000000000000000 libkritapigment.dll!KoColorSpace::bitBlt 000000006F63197B 000000002321FC10 0000000000000440 0000000000000180 libkritaimage.dll!KisPainter::bitBltImpl<false> 000000006F423DBF 0000000700000001 0000000000000000 0000000018A3BD20 libkritaimage.dll!KisPainter::bitBlt 000000006F678713 0000000016E1E570 0000000016E1E5A8 00000000182C36A0 libkritaimage.dll!KisIndirectPaintingSupport::mergeToLayerImpl<KisPostExecutionUndoAdapter> 000000006F4FE2C9 0000000018A43378 0000000000000005 0000000000000000 libkritaimage.dll!KisIndirectPaintingSupport::mergeToLayer 0000000066509DE6 0000000000000000 0000000000CF0D6F 0000000016C2D500 libkritaui.dll!KisPainterBasedStrokeStrategy::finishStrokeCallback 000000006F64DE93 0000000000000000 0000000000ABC9EE 0000000000000000 libkritaimage.dll!KisUpdateJobItem::run 0000000000ABE085 0000000016C2D500 0000000018BC6A30 0000000016F18410 Qt5Core.dll!QThreadPool::clear 0000000000AC050C 0000000016F18410 0000000000000000 0000000000000000 Qt5Core.dll!QThread::setTerminationEnabled 000007FEFF04415F 000007FEFF0D1EA0 0000000016F18410 0000000000000000 msvcrt.dll!_endthreadex 000007FEFF046EBD 0000000000000000 0000000000000000 0000000000000000 msvcrt.dll!_endthreadex 0000000076EA59BD 0000000000000000 0000000000000000 0000000000000000 kernel32.dll!BaseThreadInitThunk 00000000770DA2E1 0000000000000000 0000000000000000 0000000000000000 ntdll.dll!RtlUserThreadStart
Hi Andy, Thanks for your report. I wondered how I missed this, but I guess that the laptop I tested on yesterday doesn't have the right vectorization extensions. I tested on my SP3, and there I confirm the crash. I'll try to fix it first thing.
I've updated the post with a new build which doesn't crash, but which also doesn't use the vectorization extensions.
Hi Boudewijn. Thanks for the update. This was happening on my work machine so I'll check it out on Monday when I'm back in the office. I know a bug report probably isn't the right place for this but I just wanted to say I think you guys are doing great work on Krita. I'm a really big fan. :)
Confirming crash is no longer occurring for me on posted updated build.