SIGSEGV in BoxStateEstimator

Hi folks,

I’m getting this occasional SIGSEGV playing back an SVO file with object detection enabled:

PC@       7070f005f548 BoxStateEstimator<BoxMeasurement>::update(BoxMeasurement*)
PC@       7070f005fcd3 box_track_state::BoxTrackState<BoxStateEstimator<BoxMeasurement> >::update()
PC@       7070eff3044e tracksUpdateThread<BoxTrack<BoxStateEstimator<BoxMeasurement> > >::operator()(slutils::cv::Range const&) const
PC@       7070ef1aade2 tbb::interface9::internal::start_for<tbb::blocked_range<int>, (anonymous namespace)::ProxyLoopBody, tbb::auto_partitioner const>::execute()
PC@       7070f043637d tbb::internal::custom_scheduler<tbb::internal::IntelSchedulerTraits>::local_wait_for_all(tbb::task&, tbb::task*)
PC@       7070f041fc5a tbb::internal::arena::process(tbb::internal::generic_scheduler&)
PC@       7070f04299a8 tbb::internal::market::process(rml::job&)
PC@       7070f042ee7c tbb::internal::rml::private_worker::run()
PC@       7070f042f0cc tbb::internal::rml::private_worker::thread_routine(void*)
PC@       706ffbc94ac2 start_thread at ./nptl/pthread_create.c:442
PC@       706ffbd268bf __GI___clone3
PC@   ffffffffffffffff (unknown)

Looks like an unchecked nullptr somewhere in box tracking

Hi @Waelyn,

Thank you for contacting us and for the backtrace.

Could you send us the SVO at support@stereolabs.com so that we can replicate the issue and ensure it’s fixed in the next release?

Also, is it fairly reproducible ? If you could point out any details to help us reproduce (HW setup, other apps running in parallel, voltage, load at the moment of the crash, …), it will be helpful to help us reproduce and release a proper fix

Thanks!