Blender file causing segfault when loaded

While scripting some some low-level mesh stuff I ended up with a blend file that can no longer be loaded without segfaulting. Is there some way to at least extract the script I was working on from within the file? And should I file a bug report for the segfault? The backtrace doesn’t show much, looks like something related to triangle normals, but I can at least provide the file.

# Blender 2.92.0, Commit date: 2021-02-24 16:25, Hash 02948a2cab44

# backtrace
blender(BLI_system_backtrace+0x34) [0x562b5c4e27e4]
blender(+0xe14cfd) [0x562b59fb9cfd]
/usr/lib/libc.so.6(+0x3cf80) [0x7f1ed929ff80]
blender(normal_tri_v3+0) [0x562b5c47cc80]
blender(+0x334181a) [0x562b5c4e681a]
/usr/lib/libtbb.so.2(_ZN3tbb10interface78internal20isolate_within_arenaERNS1_13delegate_baseEl+0x46) [0x7f1ee2c660a6]
blender(+0x3341f33) [0x562b5c4e6f33]
blender(+0x334215e) [0x562b5c4e715e]
/usr/lib/libtbb.so.2(+0x2c105) [0x7f1ee2c6c105]
/usr/lib/libtbb.so.2(+0x2c43c) [0x7f1ee2c6c43c]
/usr/lib/libtbb.so.2(+0x25d97) [0x7f1ee2c65d97]
/usr/lib/libtbb.so.2(+0x243e1) [0x7f1ee2c643e1]
/usr/lib/libtbb.so.2(+0x2081c) [0x7f1ee2c6081c]
/usr/lib/libtbb.so.2(+0x20a8a) [0x7f1ee2c60a8a]
/usr/lib/libpthread.so.0(+0x9299) [0x7f1ee2986299]
/usr/lib/libc.so.6(clone+0x43) [0x7f1ed9362053]

# Python backtrace

To get to the script you may try File->Link in a new .blend file and see if you can get to it like that.

I had cases like this 2 or 3 times, where my startup file would segfault. I could never figure out what causes this. I did realize however, that I could load it successfully in like 1/10 attempts. Maybe worth a try.

Good tip, didn’t think of that. It works, thanks.