[Rivet] Python, Rivet 1.20, CERN AFS, LXBATCH

[Andy Buckley]

On 29/03/10 11:07, Ben Waugh wrote:
> On 26/03/10 18:17, Andy Buckley wrote:
>> Ah, so it seems to be a problem with GCC 4 in general. Which is really
>> strange since I've built Rivet several times on SLC5 GCC43 using the
>> bootstrap script. Maybe it should be reported to LCG AA.
>>
>> I did notice this line:
>>
>> checking for g++...
>> /afs/cern.ch/sw/lcg/contrib/gcc/4.3.2p1/x86_64-slc5-gcc34-opt/bin/g++
>>
>> I don't quite understand these combinations: this looks like a copy of
>> GCC4.3.2, built via GCC3.4... huh? I always use the gcc43 LCG tag
>> version of the GCC43 setup script... not sure if that makes a
>> difference, but I've not seen this error before on GCC4.3.
> 
> That does look odd. Which script do you mean by "the gcc43 LCG tag
> version of the GCC43 setup script"?

Before running rivet-bootstrap on a 64 bit SLC5 machine like lxplus5, I
source this gcc43 setup script (since the LCG fastjet, HepMC, etc. are
all built with gcc43):

. /afs/cern.ch/sw/lcg/contrib/gcc/4.3/x86_64-slc5-gcc43-opt/setup.sh

i.e. set up gcc 4.3 with an LCG tag that also says "gcc43". I'm not sure
what the other combinations of these setup scripts for gcc43 mean, but
this is the once that looked most sensible to me!

Anyway, this has worked for me before... can you see if it works for you?

> Maybe we should report something to
> LCG AA, but at the moment I'm not sure exactly what.

If this setup script solves the problem, then maybe we should
report/query the other GCC setup scripts. If it doesn't solve the
problem, then maybe we should ask for this setup to also provide an up
to date SWIG... in fact, for all I know *this* setup script is already
doing just that!

Andy

-- 
Dr Andy Buckley
SUPA Advanced Research Fellow
Particle Physics Experiment Group, University of Edinburgh