Re: [overpass] Segmentation fault

[Igor Brejc <>]

No problem, I'm going to bed, too :-)

On Jul 30, 2017 10:46 PM, "Roland Olbricht" <> wrote:

Hi,

thank you for the debug data. The object in which later the crash happens is at least properly constructed. I would like to check now whether the object is still valid before deconstruction. This is again a gdb session.

Delete the .lock file, then start gdb.

break file_blocks_index.h:296

Then

run </tmp/crash_input.osc

Once the break point is reached the first time: there is an address after th (this=) parameter. Please use that address as follows:

print *this
print *(File_Blocks_Index<Uint32_Index>*)0x77d750

Both commands are supposed to do the same. But I have seen the first one to fail from time to time.

I will go to bed after this mail. Hence, I'm sorry that the next round cannot continue before tomorrow morning.

Best regards,

Roland