0

I'm developing a C application running on a remote embedded board - Intel x86 platform - running LUbuntu 13.10.
This application is launched by a PHP script (but eventually I can run it from the command line for debugging purpose).

After a day the board shows a window that reports an internal error: 

ExecutablePath: MyApp
ProblemType: Crash
Architecture: i386
CodeDump: (binary data)
..
SegvAnalysis: Skipped: missing required fields "Disassembly"
Signal: 11
..

This should be a Segmentation Error exception and of course I'd like to know where.

I'm not able to find the coredump file on the disk: do I need to enable the generation of it?

In the case, is it possible to understand which code raises the exception - possibly off-line?

regards

Tamara Koliada
  • 1,200
  • 2
  • 14
  • 31
SteMMo
  • 392
  • 1
  • 4
  • 23
  • If you can run it from command line, can you also run it with GDB? – Gerhardh Jan 12 '19 at 12:14
  • It might indeed be necessary to enable generation of core files. See https://stackoverflow.com/q/17965/2402272. – John Bollinger Jan 12 '19 at 12:18
  • 1
    and you can also under _valgrind_ – bruno Jan 12 '19 at 12:18
  • to complete my previous remark about _valgrind_ you can think it is crazy to run under _valgrind_ because out of it that already need one day. You see the effect after one day but may be the problem occurs long time before, for instance to destruct the heap can have a visible effect long time after, or even never. So if you can, try under _valgrind_ – bruno Jan 12 '19 at 12:28
  • Core dumps are actually already enabled in stock *Ubuntu. See e.g. [Where does ubuntu 14.04 drop core files?](https://superuser.com/q/849099/87525). – Ruslan Jan 12 '19 at 13:44

0 Answers0