[Rivet] Segmentation Fault when running Rivet 2.5.0

[Andy Buckley]

Not obviously -- we can't "easily" tell from a compiled plugin library 
which Rivet library version it was compiled against, let alone whether 
it's compatible.

The easiest way to tell is by trying and getting a crash! ;-) Best to 
assume that a change of library version will always require a 
recompile... often it won't, but better to be safe and sometimes there 
will be non-critical improvements, or even important non-crash-inducing 
changes, maybe in inline functions.

Andy


On 14/07/16 17:14, Christian Johnson wrote:
> Is there a possibility of doing some sort of compatibility check of
> custom analyses?
> Maybe give the user a warning that the custom analyses in the path are
> old and that they need to be re-compiled?
>
> Cheers,
> Christian
>
>
>> On Jul 14, 2016, at 16:33, Christian Gutschow <chris.g at cern.ch
>> <mailto:chris.g at cern.ch>> wrote:
>>
>> Don’t think so, RIVET_ANALYSIS_PATH wasn’t even set for me at the time
>> :-/
>>
>> Cheers,
>> Chris
>>
>>> On 14 Jul 2016, at 16:28, Andy Buckley <a.g.buckley at gmail.com
>>> <mailto:a.g.buckley at gmail.com>> wrote:
>>>
>>> Aha, thanks for letting us know.
>>>
>>> In general you should rebuild your personal analysis codes when
>>> upgrading Rivet, since they've been compiled to expect the same
>>> function interfaces in the Rivet library. Let us know if rebuilding
>>> them doesn't solve the problem...
>>>
>>> Chris G, is this maybe your problem, too? Maybe we should remind
>>> people of this in each release email, or at least in the non-patch
>>> release announcements.
>>>
>>> Andy
>>>
>>>
>>> On 14/07/16 14:46, Christian Johnson wrote:
>>>> Through some more digging, it looks like the problem was due to Rivet
>>>> 2.5.0 not liking one of the analyses that I add to RIVET_ANALYSIS_PATH.
>>>> Removing the analysis from the path solved the issue. Sorry for the
>>>> trouble guys. I will check what Rivet 2.5.0 doesn't like about this
>>>> analysis and get back to you.
>>>>
>>>> Cheers,
>>>> Christian
>>>>
>>>>> On Jul 14, 2016, at 15:35, Christian Johnson
>>>>> <christian.johnson at cern.ch <mailto:christian.johnson at cern.ch>
>>>>> <mailto:christian.johnson at cern.ch>> wrote:
>>>>>
>>>>> Another update...
>>>>>
>>>>> I checked with another mac/homebrew user and his Rivet worked just
>>>>> fine. The only difference between our environments is that;
>>>>> He uses:
>>>>>
>>>>>    Apple LLVM version 7.3.0 (clang-703.0.31)
>>>>>    Target: x86_64-apple-darwin15.4.0
>>>>>    Thread model: posix
>>>>>    InstalledDir: /Library/Developer/CommandLineTools/usr/bin
>>>>>
>>>>> And I use:
>>>>>
>>>>>    Apple LLVM version 7.3.0 (clang-703.0.31)
>>>>>    Target: x86_64-apple-darwin15.5.0
>>>>>    Thread model: posix
>>>>>    InstalledDir: /Library/Developer/CommandLineTools/usr/bin
>>>>>
>>>>>
>>>>> I think that is OS X 10.11.4 vs 10.11.5.
>>>>>
>>>>> Christian
>>>>>
>>>>>> On Jul 14, 2016, at 15:06, Christian Johnson
>>>>>> <christian.johnson at cern.ch <mailto:christian.johnson at cern.ch>
>>>>>> <mailto:christian.johnson at cern.ch>> wrote:
>>>>>>
>>>>>> So I have tested the RIVETDEV installation as well and I get the same
>>>>>> result.
>>>>>> By the way, the cython-0.23.5 link in the script appears to be broken
>>>>>> at the moment.
>>>>>>
>>>>>> I'll try installing from home-brew again but ensuring that it builds
>>>>>> against gcc.
>>>>>> Which version are you guys building against?
>>>>>>
>>>>>> Cheers,
>>>>>> Christian
>>>>>>
>>>>>>
>>>>>>> On Jul 14, 2016, at 14:27, Christian Johnson
>>>>>>> <christian.johnson at cern.ch <mailto:christian.johnson at cern.ch>
>>>>>>> <mailto:christian.johnson at cern.ch>> wrote:
>>>>>>>
>>>>>>> Hi Chris and Andy,
>>>>>>>
>>>>>>> Thanks for the info!
>>>>>>>
>>>>>>> So through some more testing, I found that I was not properly
>>>>>>> linking gcc to begin with so Rivet was building against clang.
>>>>>>> So I think the issue it with clang.
>>>>>>>
>>>>>>> I verified this by uninstalling everything in Homebrew to start
>>>>>>> clean and then used the rivet-bootstrap to install everything.
>>>>>>> I am still running into the same error as before. Let me try running
>>>>>>> the RIVETDEV option.
>>>>>>>
>>>>>>> Cheers,
>>>>>>> Christian
>>>>>>>
>>>>>>>> On Jul 14, 2016, at 12:57, Andy Buckley <a.g.buckley at gmail.com
>>>>>>>> <mailto:a.g.buckley at gmail.com>
>>>>>>>> <mailto:a.g.buckley at gmail.com>> wrote:
>>>>>>>>
>>>>>>>> We'll look into this as soon as there's a GCC6 system within
>>>>>>>> developer reach. I'm guessing it's due to something like Cython (or
>>>>>>>> at least the version that we used in our Rivet tarball build) not
>>>>>>>> yet supporting the GCC6 binary interface... but I don't know the
>>>>>>>> details.
>>>>>>>>
>>>>>>>> It would be interesting to know if you can build Rivet using the
>>>>>>>> "developer mode" of the bootstrap script and GCC6 -- that will
>>>>>>>> regenerate the Python interfacing code rather than relying on the
>>>>>>>> version that we bundled. Use like this: INSTALL_RIVETDEV=1
>>>>>>>> ./rivet-bootstrap
>>>>>>>>
>>>>>>>> Andy
>>>>>>>>
>>>>>>>>
>>>>>>>> On 14/07/16 10:11, Christian Gutschow wrote:
>>>>>>>>> Hi Christian,
>>>>>>>>>
>>>>>>>>> I had the same problem last week, mainly because I updated gcc
>>>>>>>>> at the
>>>>>>>>> same time I updated Rivet — big mistake. Ended up having to
>>>>>>>>> revert to
>>>>>>>>> the old gcc, reinstalling Python and only then it all worked
>>>>>>>>> fine...
>>>>>>>>>
>>>>>>>>> Cheers,
>>>>>>>>> Chris
>>>>>>>>>
>>>>>>>>>> On 14 Jul 2016, at 10:51, Christian Johnson
>>>>>>>>>> <christian.johnson at cern.ch <mailto:christian.johnson at cern.ch>
>>>>>>>>>> <mailto:christian.johnson at cern.ch>
>>>>>>>>>> <mailto:christian.johnson at cern.ch>> wrote:
>>>>>>>>>>
>>>>>>>>>> Hi Team,
>>>>>>>>>>
>>>>>>>>>> I have recently updated to Rivet 2.5.0 and Yoda 1.6.2. Though I am
>>>>>>>>>> running into a "Segmentation Fault 11" when I run "rivet
>>>>>>>>>> --list-analyses."
>>>>>>>>>> It looks like it is complaining at this line "all_analyses =
>>>>>>>>>> rivet.AnalysisLoader.analysisNames()" in the rivet python script.
>>>>>>>>>>
>>>>>>>>>> I am running it on Mac OS X 10.11.5, using gcc 6.1 and python
>>>>>>>>>> 2.7.12
>>>>>>>>>> installed using Homebrew.
>>>>>>>>>>
>>>>>>>>>> Cheers,
>>>>>>>>>> Christian_______________________________________________
>>>>>>>>>> Rivet mailing list
>>>>>>>>>> Rivet at projects.hepforge.org
>>>>>>>>>> <mailto:Rivet at projects.hepforge.org>
>>>>>>>>>> <mailto:Rivet at projects.hepforge.org>
>>>>>>>>>> <mailto:Rivet at projects.hepforge.org>
>>>>>>>>>> https://www.hepforge.org/lists/listinfo/rivet
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> —
>>>>>>>>>
>>>>>>>>> Dr. Christian Gütschow
>>>>>>>>>
>>>>>>>>> University College London
>>>>>>>>> Department of Physics and Astronomy
>>>>>>>>> Gower Street
>>>>>>>>> London WC1E 6BT
>>>>>>>>>
>>>>>>>>>> D10 Physics Building
>>>>>>>>>> +44 (0)20 7679 3775
>>>>>>>>>> chris.g at cern.ch <mailto:chris.g at cern.ch>
>>>>>>>>>> <mailto:chris.g at cern.ch> <mailto:chris.g at cern.ch>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> _______________________________________________
>>>>>>>>> Rivet mailing list
>>>>>>>>> Rivet at projects.hepforge.org
>>>>>>>>> <mailto:Rivet at projects.hepforge.org>
>>>>>>>>> <mailto: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
>>>>>>>
>>>>>>
>>>>>
>>>>
>>>
>>>
>>> --
>>> Dr Andy Buckley, Lecturer / Royal Society University Research Fellow
>>> Particle Physics Expt Group, University of Glasgow
>>
>>
>>
>>
>>  —
>>
>>  Dr. Christian Gütschow
>>
>>  University College London
>>  Department of Physics and Astronomy
>>  Gower Street
>>  London WC1E 6BT
>>
>>  > D10 Physics Building
>>  > +44 (0)20 7679 3775
>>  > chris.g at cern.ch <mailto:chris.g at cern.ch>
>>
>>
>>
>>
>


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