SEGV in rtengine::HaldCLUT::correct()

[innir]

Hi,

another report - I'm not sure if it's still reproducible, I tried a little and was not able to:

Dear Maintainer,

Film simulation is a great tool to instantly get your colors right. That's why I'm experimenting a lot with it. Unfortunately there seems to be a problem:

• Entering a directory with RAWs using film simulation in their pp3s, or
• changing the simulated film in before-after view, or
• going from editor back to file list sometimes crashes the application. In that case, it will often crash again, as soon as the thumbnail for the affected image becomes visible or the image itself is double-clicked.

Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fff99ffb700 (LWP 1673)] 0x0000555555c427b4 in rtengine::HaldCLUT::correct(rtengine::Imagefloat&, int, float, float, float, float&, float&, float&) () (gdb) bt

0 0x0000555555c427b4 in rtengine::HaldCLUT::correct(rtengine::Imagefloat&, int, float, float, float, float&, float&, float&) ()

1 0x0000555555ad923c in ?? ()

2 0x00007ffff363f716 in ?? () from /usr/lib/x86_64-linux-gnu/libgomp.so.1

3 0x00007ffff32070a4 in start_thread (arg=0x7fff99ffb700) at pthread_create.c:309

4 0x00007ffff2f3bcbd in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:111

Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fffc67fc700 (LWP 1779)] 0x0000555555c427b4 in rtengine::HaldCLUT::correct(rtengine::Imagefloat&, int, float, float, float, float&, float&, float&) () (gdb) bt

0 0x0000555555c427b4 in rtengine::HaldCLUT::correct(rtengine::Imagefloat&, int, float, float, float, float&, float&, float&) ()

1 0x0000555555ad923c in ?? ()

2 0x00007ffff363b7df in GOMP_parallel () from /usr/lib/x86_64-linux-gnu/libgomp.so.1

3 0x0000555555ac93e4 in rtengine::ImProcFunctions::rgbProc(rtengine::Imagefloat, rtengine::LabImage, rtengine::EditBuffer, LUT&, LUT&, LUT&, rtengine::SHMap, int, LUT&, LUT&, LUT&, float, float, rtengine::ColorGradientCurve const&, rtengine::OpacityCurve const&, bool, LUT&, LUT&, rtengine::ToneCurve const&, rtengine::ToneCurve const&, rtengine::ToneCurve const&, rtengine::ToneCurve const&, double&, double&, double&, float&, float&, float&, double, int, int) ()

4 0x0000555555ba27dc in rtengine::Thumbnail::processImage(rtengine::procparams::ProcParams const&, int, rtengine::TypeInterpolation, std::string, double, double, float, float, float, float, std::string, double&) ()

5 0x00005555559ce7f6 in Thumbnail::processThumbImage(rtengine::procparams::ProcParams const&, int, double&) ()

6 0x000055555586db80 in ThumbImageUpdater::Impl::processNextJob() ()

7 0x00007ffff6d4bb02 in ?? () from /usr/lib/x86_64-linux-gnu/libglibmm-2.4.so.1

8 0x00007ffff6fdd2b8 in ?? () from /lib/x86_64-linux-gnu/libglib-2.0.so.0

9 0x00007ffff6fdc925 in ?? () from /lib/x86_64-linux-gnu/libglib-2.0.so.0

10 0x00007ffff32070a4 in start_thread (arg=0x7fffc67fc700) at pthread_create.c:309

11 0x00007ffff2f3bcbd in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:111

It might be also related to cropping, as it only started to happen after a crop was applied.

I already removed the whole cache, but that made it effectivly crash more often.

I hope this has already been fixed upstream, so it won't take to long to find the culprit.

Best, Björn

See: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=769028

Best, Philip

[Beep6581]

Hey Please see http://rawpedia.rawtherapee.com/How_to_write_useful_bug_reports and supply:

• the raw file + PP3 + HALD CLUT (or its name, if its shipped in the RawTherapee Film Simulation Collection,
• the AboutThisBuild contents
• details of your system and RAM
• a stack backtrace using a debug build, not a release type build,
• anything else stated there.

[innir]

I guess it's a duplicate of #2558. So it's already fixed.