[Rivet] Rivet 2.5.1 segfault on debian jessie

[Andy Buckley]

Done! Also now with a hint about the "segfault after upgrade" 
semi-common issue.

https://rivet.hepforge.org/trac/wiki/TroubleShooting

Thanks,
Andy


On 05/09/16 14:42, David Bjergaard wrote:
> Hi,
>
> I refuse to give up my old environment, and I figured it out:
>> strace rivet
>> [...]
>> mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f0ae3f9b000
>> read(3, "00400000-0071f000 r-xp 00000000 "..., 1024) = 1024
>> read(3, "home/dave/rivet/local/lib/libBOO"..., 1024) = 1024
>> read(3, "fd3f000-7f0adfd43000 rw-p 000140"..., 1024) = 1024
>> read(3, "ibssl.so.1.0.0\n7f0ae045a000-7f0a"..., 1024) = 1024
>> read(3, "  /home/dave/rivet/local/lib/lib"..., 1024) = 1024
>> read(3, "ib/libfastjetplugins.so.0.0.0\n7f"..., 1024) = 1024
>> read(3, "922327                     /usr/"..., 1024) = 1024
>> close(3)                                = 0
> As soon as I saw libBOOSTFastJets.so I remembered that I had put a custom
> wrapper around the fastjet contrib functions.  Removing it fixed the segfault.
>
> All's well that ends well, I've got a new version of Rivet working and I can be
> on my merry way.
>
> Also, can someone put a gdb prescription on the troubleshooting wiki?  I could
> write a quick and dirty one.  This worked for me (finally):
> gdb python
> (gdb) run /path/to/rivet [rivet args]
>
> Its obvious in retrospect, but not at all when your not familiar with running
> C++ in python.
>
>
>     David
>
>
>
> Andy Buckley <andy.buckley at cern.ch> writes:
>
>> Not sure what exactly went wrong for you, but my guess is that a
>> Rivet*.so file from the old version was left in place somehow. And
>> since it wasn't rebuilt with the new API, etc. its contents segfaulted
>> on being dlopen()d / run.
>>
>> Best to install into a completely fresh area if possible, but a manual
>> clear-out of Rivet*.so and .info etc. files from $prefix/lib and
>> $prefix/share/Rivet is a good plan, as is removing the rivet directory
>> and any .egg files in the $prefix/lib*/python*/site-packages
>>
>> Andy
>>
>>
>> On 02/09/16 20:34, David Bjergaard wrote:
>>> Hi Holger,
>>>
>>> As it turns out... I just launched rivetenv in a clean environment, and it
>>> works.   This came about because I tried to upgrade from 2.4.something to 2.5.1,
>>> and I didn't want to blow away all the generators I had installed, so I just
>>> ran the bootstrap script without cleaning out the old install of rivet.
>>>
>>> I know that's a bit of a no-no, but the question is: can I keep my old
>>> rivet/local prefix, or should I just re-install my old generators fresh wit the
>>> new rivet?  I'm guessing do it all fresh since I know that the Rivet extensions
>>> probably won't work anymore since they were linked with an older version of
>>> rivet.
>>>
>>> Anyway, sorry for the noise, happy Friday,
>>>
>>>     David
>>>
>>> Holger Schulz <holger.schulz at durham.ac.uk> writes:
>>>
>>>> Hi David,
>>>>
>>>> could I ask you to dump the output of env after sourcing rivetenv to
>>>> this thread?
>>>>
>>>>
>>>> Thanks,
>>>> Holger
>>>>
>>>>
>>>> On 02/09/16 19:07, David Bjergaard wrote:
>>>>> Hi,
>>>>>
>>>>> I'm trying to install the latest version of rivet and I'm having no luck. The
>>>>> bootstrap script works great, it installs, and when I run rivetenv.sh and rivet
>>>>> I get 'Segmentation Fault'.  I didn't dig too far into why (I still can't get
>>>>> gdb to attach to rivet for the life of me).  Anyway, here's the relevant info:
>>>>> wget http://rivet.hepforge.org/hg/bootstrap/raw-file/2.5.1/rivet-bootstrap
>>>>> ./rivet-bootstrap
>>>>>
>>>>> dave at calypso:~/rivet-2.5/Rivet-2.5.1$ uname -a
>>>>> Linux calypso 3.16.0-4-amd64 #1 SMP Debian 3.16.7-ckt25-2+deb8u3 (2016-07-02) x86_64 GNU/Linux
>>>>> dave at calypso:~/rivet-2.5/Rivet-2.5.1$ gcc --version
>>>>> gcc (Debian 4.9.2-10) 4.9.2
>>>>> Copyright (C) 2014 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.
>>>>>
>>>>> I would have assumed that this would have been sufficient (I know the compiler
>>>>> is a bit old, but its still newer than lxplus).  I'm at your service for further
>>>>> debugging.
>>>>>
>>>>> Cheers,
>>>>>
>>>>>     David
>>>>> _______________________________________________
>>>>> Rivet mailing list
>>>>> Rivet at projects.hepforge.org
>>>>> https://www.hepforge.org/lists/listinfo/rivet
>>> _______________________________________________
>>> Rivet mailing list
>>> Rivet at projects.hepforge.org
>>> https://www.hepforge.org/lists/listinfo/rivet
>>>


-- 
Dr Andy Buckley, Lecturer / Royal Society University Research Fellow
Particle Physics Expt Group, University of Glasgow