Closed sloriot closed 1 month ago
Do we have to boil it down for a gcc bug report?
I create this issue to track a problem observed recently with
g++ (GCC) 14.1.1 20240507 (Red Hat 14.1.1-1)
In this test result and probably this one.
I can only be reproduce with optimizers with
DNDEBUG
defined.gdb
says:
That is the compiler I have on my own laptop (Fedora Linux 40):
[lrineau@fernand]~% rpm -q gcc
gcc-14.1.1-1.fc40.x86_64
[lrineau@fernand]~% gcc --version
gcc (GCC) 14.1.1 20240507 (Red Hat 14.1.1-1)
Copyright (C) 2024 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
So... to reproduce: use that compiler, -DNDEBUG
, -O3
... and which CMake target have you compiled?
ERROR: execution of test_region_growing_on_point_set_2 in 0.400651 s : SEGFAULT
I can reproduce, with -O2 -DNDEBUG
, with that command:
[lrineau@fernand]~/Git/cgal-master% cmake -DCGAL_ROOT=$PWD -DCMAKE_CXX_FLAGS='-O2 -g -DNDEBUG' -S Shape_detection/test/Shape_detection -B build-Shape_detection && cmake --build build-Shape_detection --clean-first -t test_region_growing_on_point_set_2 -v && /home/lrineau/Git/cgal-master/build-Shape_detection/test_region_growing_on_point_set_2
This can no longer be reproduced with gcc 14.2.1 from Fedora 40. It was probably a compiler bug. Closing this issue.
I create this issue to track a problem observed recently with
g++ (GCC) 14.1.1 20240507 (Red Hat 14.1.1-1)
In this test result and probably this one.
I can only be reproduce with optimizers with
DNDEBUG
defined.gdb
says:This happens in a non concurrent context, only accessing a FT (Lazy_exact_nt) stored in an array (but copied).