[Rivet] rivet 2 on Mac OSX

Hannes Jung hannes.jung at cern.ch
Tue Jun 10 18:38:01 BST 2014


Hi Andy

that a lot.... that was it,... it compiles and works...
Puhhhhh


So the point is:

having in the PATH first everything from OSX and then /sw/bin
(before I had it the other way around, and then it complained about python from fink)
export PATH=$PATH:/sw/bin
export CFLAGS=-Qunused-arguments
export CPPFLAGS=-Qunused-arguments
INSTALL_PREFIX=/Users/jung/jung/cvs/rivet2 CC=clang CXX=clang++ ./rivet-bootstrap

and then of course your changes for fastjet and the Makefiles.
Another thing, what I had to change in the bootstrap script was
(including the --no-check-certifiate):

function wget_untar { wget --no-check-certificate $1 -O- | tar xz; }


Thanks so much... this is really great....

Have a nice evening

Cheers




On 10.06.2014, at 16:59, Andy Buckley <andy.buckley at cern.ch<mailto:andy.buckley at cern.ch>> wrote:

On 10/06/14 16:07, Hannes Jung wrote:
Dear Andy

thanks for not giving up ...

my build dir is:
~/jung/cvs/rivet2/build> ls
HepMC-2.06.09  README-hannes.bck  boost_1_55_0
boost_1_55_0.tar.gz.1  gsl-1.16         rivet-bootstrap-old
README-hannes  YODA-1.1.0         boost_1_55_0.tar.gz  fastjet-3.0.6
    rivet-bootstrap  rivet-bootstrap.bck

so I do not yet have Rivet..... are you sure the problem comes from the
rivet completion ?

Aha, I read more carefully: we have the same bug in YODA's
bin/Makefile.am, with the same suggested solution.

So make that change in YODA and run "autoreconf -i" in the YODA-1.1.0
dir; then go back to your build dir and run

wget http://www.hepforge.org/archive/rivet/Rivet-2.1.2.tar.gz -O - | tar xf

to get Rivet, and cd Rivet-2.1.2/bin and make the same change +
autoreconf there. Finally, run the bootstrap one last (?) time; it
should run the configure scripts again but not do any more downloading,
nor any building until it gets to Rivet.

I hope this will work, but it relies on your system autotools working.
If it doesn't succeed let me know.

Since there are no other fixes so far, no-one else has hit this issue,
and it doesn't change the behaviour of a successful build, I am tempted
to just "silently" remake the Rivet and YODA tarballs with this fix
included. A bit naughty, but it's not worth a full release and ~everyone
except Hannes will not notice at all! A more proper way is to make e.g.
Rivet-2.1.2.post1, cf. Python's version scheme
http://legacy.python.org/dev/peps/pep-0440/#post-releases ... but I find
it a bit ugly and confusing. Preferences, developers?

Andy


On 10.06.2014, at 15:58, Andy Buckley <andy.buckley at cern.ch<mailto:andy.buckley at cern.ch>
<mailto:andy.buckley at cern.ch>> wrote:

Hi Hannes. In answer to the last question... yes, but we're now very
close! YODA and Rivet are fully building, including the Python modules
so it's really the very last install step where something has gone wrong.

The issue now is a bug on our side, but a very minor and strange one
that I never heard of before. The installation of the bash_completion
script only happens if $prefix/etc/bash_completion.d exists, and I guess
that it has not existed for most Mac users so far. If it does exist then
in bin/Makefile.am we install it like this:

install --mode 644 rivet-completion $(bashcomp_dir)/; fi

which works just fine on Linux. But on Macs the install command seems to
take different arguments (argh!), so the installation falls over with
the message that you got. I'm guessing here, but I think it will work
for you if you replace that line in Makefile.am with this:

$(install_sh_DATA) rivet-completion $(bashcomp_dir)/; fi

Can you try this? Since the build has all completed, no need to re-run
the bootstrap script: just cd into the Rivet-2.1.2 build dir, make the
change, and call "make install". Cross fingers again ;-)

Or you could delete your $prefix/etc/bash_completion.d directory, but
that has downsides... and I'd like to know if this works. If so, it'll
be fixed in the next release. Thanks for helping, and persevering...
really close to success now!

Andy




On 10/06/14 15:42, Hannes Jung wrote:
Dear Andy

thanks a lot for the updates, now this works, but another error appears:
.....
running install_egg_info
Removing build/yoda-1.1.0-py2.7.egg-info
Writing build/yoda-1.1.0-py2.7.egg-info
CC="clang" CXX="clang++" CXXFLAGS="-Wno-sign-compare
-Wno-strict-prototypes" ARCHFLAGS=""  /usr/bin/python setup.py install
--prefix=/Users/jung/software/rivet --force
running install
running build
running build_py
running build_ext
running install_lib
copying build/lib.macosx-10.9-intel-2.7/yoda/__init__.py ->
/Users/jung/software/rivet/lib/python2.7/site-packages/yoda
copying build/lib.macosx-10.9-intel-2.7/yoda/core.so ->
/Users/jung/software/rivet/lib/python2.7/site-packages/yoda
copying build/lib.macosx-10.9-intel-2.7/yoda/script_helpers.py ->
/Users/jung/software/rivet/lib/python2.7/site-packages/yoda
copying build/lib.macosx-10.9-intel-2.7/yoda/util.so ->
/Users/jung/software/rivet/lib/python2.7/site-packages/yoda
byte-compiling
/Users/jung/software/rivet/lib/python2.7/site-packages/yoda/__init__.py
to __init__.pyc
byte-compiling
/Users/jung/software/rivet/lib/python2.7/site-packages/yoda/script_helpers.py
to script_helpers.pyc
running install_egg_info
Removing
/Users/jung/software/rivet/lib/python2.7/site-packages/yoda-1.1.0-py2.7.egg-info
Writing
/Users/jung/software/rivet/lib/python2.7/site-packages/yoda-1.1.0-py2.7.egg-info
make[3]: Nothing to be done for `install-data-am'.
Making install in tests
make[2]: Nothing to be done for `install-exec-am'.
make[2]: Nothing to be done for `install-data-am'.
Making install in bin
.././install-sh -c -d '/Users/jung/software/rivet/bin'
/usr/bin/install -c yoda-config yodacnv yodamerge yodadiff yodahist
yoda2aida yoda2flat aida2yoda aida2flat flat2yoda
'/Users/jung/software/rivet/bin'
if [[ -d "/Users/jung/software/rivet/etc/bash_completion.d" && -w
"/Users/jung/software/rivet/etc/bash_completion.d" ]]; then \
    install --mode 644 yoda-completion
/Users/jung/software/rivet/etc/bash_completion.d/; fi
install: illegal option -- -
usage: install [-bCcpSsv] [-B suffix] [-f flags] [-g group] [-m mode]
             [-o owner] file1 file2
     install [-bCcpSsv] [-B suffix] [-f flags] [-g group] [-m mode]
             [-o owner] file1 ... fileN directory
     install -d [-v] [-g group] [-m mode] [-o owner] directory ...
make[2]: *** [install-data-local] Error 64
make[1]: *** [install-am] Error 2
make: *** [install-recursive] Error 1
~/jung/cvs/rivet2/build>


Hm.... is it only me, who has this enourmous problems with Rivet on a
Mac ?

Cheers
Hannes



On 10.06.2014, at 11:10, Andy Buckley <andy.buckley at cern.ch<mailto:andy.buckley at cern.ch>
<mailto:andy.buckley at cern.ch>
<mailto:andy.buckley at cern.ch>> wrote:

Hi Hannes,

Looks like this is a FastJet bug in OS X Mavericks, which was fixed last
year. From http://fastjet.fr/all-releases.html:

29 October 2013: fastjet-3.0.6 (manual, doxygen, fjcore) is a minor
maintenance release. Main changes:
fixed compilation issues for two plugins using clang (clang-500.2.79,
Apple LLVM version 5.0 based on LLVM 3.3svn) under OS X 10.9 (Mavericks)

We hadn't updated the fastjet version in our bootstrap script, because
until recently LCG had not installed releases later than 3.0.3, but I've
done it now. Re-download the bootstrap and try again... hopefully it'll
work this time.

By the way, on Mavericks I think it is still necessary to set

export CFLAGS=-Qunused-arguments
export CPPFLAGS=-Qunused-arguments

before running the bootstrap script, due to an OS X clang/Python bug:
see http://stackoverflow.com/a/22322645/91808 if you want details.
Otherwise it will fail late in the build as it tries to build the Rivet
or YODA Python modules.

Cheers,
Andy



On 10/06/14 09:24, Hannes Jung wrote:
Dear Riveters

still having problems to get rivet 2 installed on Mac OSX Mavericks

following the advice of Andy at our CMS meeting on Friday, I tried
to compile everything using clang and clang++ with the command:

INSTALL_PREFIX=$HOME/software/rivet CC=clang CXX=clang++ MAKE="make
-j8"
./rivet-bootstrap


but then I got an error on fastjet:

libtool: compile:  clang++ -DHAVE_CONFIG_H -I.
-I../../../include/fastjet -O3 -Wall -g -Woverloaded-virtual
-DDROP_CGAL
-I. -I./../../../include -MT libCDFcode_la-JetCluAlgorithm.lo -MD -MP
-MF .deps/libCDFcode_la-JetCluAlgorithm.Tpo -c JetCluAlgorithm.cc<http://JetCluAlgorithm.cc>
<http://JetCluAlgorithm.cc>
<http://JetCluAlgorithm.cc>
<http://JetCluAlgorithm.cc>  -fno-common -DPIC -o
.libs/libCDFcode_la-JetCluAlgorithm.o
JetCluAlgorithm.cc<http://JetCluAlgorithm.cc> <http://JetCluAlgorithm.cc>
<http://JetCluAlgorithm.cc>
<http://JetCluAlgorithm.cc>:81:16: error: use of
undeclared identifier 'abs'; did you
   mean 'fabs'?
       int dIEta = abs(iEtaSeedTower - iEtaPreClusterTower);
                   ^~~
                   fabs
/usr/include/math.h:431:15: note: 'fabs' declared here
extern double fabs(double);
           ^
JetCluAlgorithm.cc<http://JetCluAlgorithm.cc> <http://JetCluAlgorithm.cc>
<http://JetCluAlgorithm.cc>
<http://JetCluAlgorithm.cc>:82:16: error: use of
undeclared identifier 'abs'; did you
   mean 'fabs'?
       int dIPhi = abs(iPhiSeedTower - iPhiPreClusterTower);
                   ^~~
                   fabs
/usr/include/math.h:431:15: note: 'fabs' declared here
extern double fabs(double);
           ^
2 errors generated.
make[3]: *** [libCDFcode_la-JetCluAlgorithm.lo] Error 1


Any idea to solve this ?
Is this just a Mac problem ?

Thanks a lot
Cheers
Hannes



***********************************************************************
Hannes Jung
Email: Hannes.Jung at desy.de<mailto:Hannes.Jung at desy.de> <mailto:Hannes.Jung at desy.de>
<mailto:Hannes.Jung at desy.de>
<mailto:Hannes.Jung at desy.de>
mobile :+49 40 8998 93741
http://www.desy.de/~jung
Tel: +49 (0) 40 8998 3741
Fax: +49 (0) 40 8994 3741
DESY, CMS 01B/02.213
Notkestr.85, 22603 Hamburg, FRG
***********************************************************************






_______________________________________________
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 Andy Buckley, Royal Society University Research Fellow
Particle Physics Expt Group, University of Glasgow / PH Dept, CERN

***********************************************************************
Hannes Jung
Email: Hannes.Jung at desy.de<mailto:Hannes.Jung at desy.de> <mailto:Hannes.Jung at desy.de>
<mailto:Hannes.Jung at desy.de>
mobile :+49 40 8998 93741
http://www.desy.de/~jung
Tel: +49 (0) 40 8998 3741
Fax: +49 (0) 40 8994 3741
DESY, CMS 01B/02.213
Notkestr.85, 22603 Hamburg, FRG
***********************************************************************






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

***********************************************************************
Hannes Jung
Email: Hannes.Jung at desy.de<mailto:Hannes.Jung at desy.de> <mailto:Hannes.Jung at desy.de>
mobile :+49 40 8998 93741
http://www.desy.de/~jung
Tel: +49 (0) 40 8998 3741
Fax: +49 (0) 40 8994 3741
DESY, CMS 01B/02.213
Notkestr.85, 22603 Hamburg, FRG
***********************************************************************






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

***********************************************************************
Hannes Jung
Email: Hannes.Jung at desy.de<mailto:Hannes.Jung at desy.de>
mobile :+49 40 8998 93741
http://www.desy.de/~jung
Tel: +49 (0) 40 8998 3741
Fax: +49 (0) 40 8994 3741
DESY, CMS 01B/02.213
Notkestr.85, 22603 Hamburg, FRG
***********************************************************************




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.hepforge.org/lists-archive/rivet/attachments/20140610/fae0b515/attachment.html>


More information about the Rivet mailing list