Re: segmentation faults

From: Axel Naumann <Axel.Naumann_at_cern.ch>
Date: Tue, 03 Jul 2007 15:11:53 +0200

Hi Thomas,

funcname==0xffffffffffffffff is just fine. Use GDB if you cannot send us

a stripped down version of your code.

Cheers, Axel.

Thomas Lauf wrote:
> Hi Axel,
>

>> this just means that there is some function being called via the
>> interpreter.

> Hm, but
>
> funcname=0xffffffffffffffff <Address DEFANGED_0xffffffffffffffff
> DEFANGED_out DEFANGED_of DEFANGED_bounds>
>
> in belows error message seems to me that something already goes wrong
> before that function is called, doesnt it?
>
> As I said, these error messages occur on different parts of the code,
> even if I rerun the same example. Unfortunately I am analyzing big data
> files, a small example is not at hand just now. Will see what I can do...
>
> Axel Naumann wrote:

>> this just means that there is some function being called via the
>> interpreter. Most probably, something goes wrong inside that function.
>> You could use GDB to find out what, or send (preferably only) the
>> relevant code such that I can reproduce the problem.
>>>
>>> in my program I get segmentations faults from time to time which I can't
>>> exactly pinpoint. I am a bit clueless where to search for the reason. A
>>> common thing in the backtrace is the following message:
>>>
>>> #14 0x000000000041fc92 in G__G__AnalysisDict_508_0_14
>>> (result7=0x7fffce058100, funcname=0xffffffffffffffff <Address
>>> DEFANGED_0xffffffffffffffff DEFANGED_out DEFANGED_of DEFANGED_bounds>,
>>> libp=0x7fffce04c150, hash=-105) at base/src/G__AnalysisDict.cxx:1518
>>>
>>> Is there a usual suspect for this?

>
Received on Tue Jul 03 2007 - 15:12:00 CEST