|
|
[Rivet] It's a rivet bug ... 90% sure
Christopher Lester
lester at hep.phy.cam.ac.uk
Sat Nov 12 00:37:37 GMT 2011
Hmm.
I thought I had browsed pretty much everywhere on the rivet pages ...
but in truth I hadn't found (or even noticed!) "troubleshooting".
I guess - like sheep - you wear an established path in the landscape
and then cease to see other possibilities.
However, I'm interested in the comment thereon:
"If you have a personalised LaTeX setup, and are not setting
up your shell environment from the installation-generated rivetenv.sh
script, then ..."
since I certainly *was* using the installation-generated rivetenv.sh
(and still am) ... and these were not setting my TEXINPUTS env var.
The autogenerated script I was using was this one:
----------------------------------------------
# These variables need to exist
prefix=/atlas/lester/usr/local
exec_prefix=${prefix}
datarootdir=${prefix}/share
export PATH="$exec_prefix/bin:$PATH"
export
LD_LIBRARY_PATH="${exec_prefix}/lib:/afs/cern.ch/sw/lcg/external/HepMC/2.06.04/x86_64-slc5-gcc43-opt/lib:/afs/cern.ch/sw/lcg/external/fastjet/2.4.3/x86_64-slc5-gcc43-opt/lib:$LD_LIBRARY_PATH"
export
PYTHONPATH="/atlas/lester/usr/local/lib/python2.6/site-packages:$PYTHONPATH"
export TEXMFHOME="${datarootdir}/Rivet/texmf:$TEXMFHOME"
export HOMETEXMF="${datarootdir}/Rivet/texmf:$HOMETEXMF"
export TEXMFCNF="${datarootdir}/Rivet/texmf/cnf:$TEXMFCNF"
if (complete &> /dev/null); then
test -e "${datarootdir}/Rivet/rivet-completion" && source
"${datarootdir}/Rivet/rivet-completion"
fi
----------------------------------------------------
which sets TEXMFHOME and a number of similar things, but no TEXINPUTS
(or LATEXINPUTS).
So the text on
http://projects.hepforge.org/rivet/trac/wiki/TroubleShooting
still might be a bit misleading in the sense that the installation
generated script is (at least sometimes) not sufficient to set up the
env ... or at least that seems to have been the case for me.
On Fri, 11 Nov 2011, Andy Buckley wrote:
> Hi Chris,
>
> Glad it works (and that you think the plots are beautiful!).
>
> I've added an item on this phenomenon to
> http://projects.hepforge.org/rivet/trac/wiki/TroubleShooting , which due to
> the link in the left navigation menu is hopefully one of the first places
> that people will look... did you?
>
> I think we're ready for a 1.7.0 release now...
>
> Andy
>
>
> On 11/11/11 09:55, Christopher Lester wrote:
>>
>>
>> Having done:
>>
>> bash-3.2$ export
>> TEXINPUTS=$TEXINPUTS:/atlas/lester/usr/local/share/Rivet/texmf/tex/:
>> bash-3.2$ echo $TEXINPUTS
>> .:/usr/share/texmf/tex//:/usr/share/texmf/bibtex//:/usera/lester/LESTERHOME/proj/latex:/usera/lester/LESTERHOME/proj/papers/STYLES//:/atlas/lester/usr/local/share/Rivet/texmf/tex/:
>>
>> bash-3.2$
>>
>>
>> to the env as previously submitted to you, then
>> I find that both
>>
>> rivet-mkhtml Rivet.aida
>>
>> and
>>
>> rivet-mkhtml Rivet_moo.aida
>>
>> perform equally successfully - both make non-blank plots.
>>
>> ---
>>
>> So that's good!
>>
>> So did I do something wrong in my installation? I've had another look at
>>
>> http://projects.hepforge.org/rivet/trac/wiki/GettingStarted
>>
>> and I don't see any bits that I missed that would have made me
>> check/alter my TEXINPUTS path .... is there some missing documentation
>> there? Given that Rivet.aida make beautiful plots, it would have taken
>> me a long time (without your expert help) to track down the fault to
>> missing TEXINPUTS path ...
>>
>> Thanks for your help!
>>
>> On Thu, 10 Nov 2011, Frank Siegert wrote:
>>
>>> Hi Chris,
>>>
>>> On 10/11/11 19:20, Christopher Lester wrote:
>>>>
>>>> $LATEXINPUTS is blank, but I do have a semi-custom $TEXINPUTS
>>>>
>>>> bash-3.2$ echo $TEXINPUTS
>>>> .:/usr/share/texmf/tex//:/usr/share/texmf/bibtex//:/usera/lester/LESTERHOME/proj/latex:/usera/lester/LESTERHOME/proj/papers/STYLES//
>>>>
>>>
>>> If I'm not mistaken, this should include
>>> ${prefix}/share/Rivet/texmf/tex/ and end with a ":". Could you try
>>> whether that helps?
>>>
>>> Best,
>>> Frank
>>>
>>>
>>>
>>>
>>>>
>>>>
>>>> My full env is attached as moo.txt.
>>>>
>>>> Also attached are success.tgz containing Rivet.aida and the plots dir
>>>> made from it, and failure.tgz containing Rivet_moo.aida, and the
>>>> (failed, blank) plots made from it.
>>>>
>>>> (rm -rf plots was done between runs)
>>>>
>>>>
>>>>> I just did a similar test with some output from another test:
>>>>>
>>>>> cp Rivet.aida with_underscore.aida
>>>>> compare-histos with_underscore.aida
>>>>> make-plots *.dat
>>>>>
>>>>> and everything worked fine. The path element /with_underscore.aida/...
>>>>> does exist in the .dat files, but make-plots is apparently escaping it
>>>>> as it should.
>>>>>
>>>>> The only things I can think of are that you have an old version of
>>>>> Rivet from before when we fixed this (1.6.0 should be safe; my
>>>>> trunk-about-to-become-1.7.0 version certainly seems to be), or perhaps
>>>>> that you have a local custom setting of LATEXINPUTS or similar that
>>>>> means that underscore.sty is not visible to LaTeX? Rivet's setup
>>>>> script installs some LaTeX packages like that which sometimes aren't
>>>>> available on crappy SLC5, but if your LaTeX setup overwrites the Rivet
>>>>> part of the TeX path then you might be out of luck. This is fixable,
>>>>> though!
>>>>>
>>>>> Hope that helps,
>>>>> Andy
>>>>>
>>>>>
>>>>> On 04/11/11 14:54, Christopher Lester wrote:
>>>>>>
>>>>>> I'm now 90% sure its a rivet bug ...
>>>>>>
>>>>>> At first I thought the problem was that Rivet didn't like the aida
>>>>>> file
>>>>>> to have the name [ANALYSIS].aida ... but I disproved that by
>>>>>> checking I
>>>>>> could rename other analyses' aida files without problem.
>>>>>>
>>>>>> What I have now found, is that the problem seems only to occur when I
>>>>>> rename the aida file to have an underscore "_" in it.
>>>>>>
>>>>>> So:
>>>>>>
>>>>>> rivet-mkhtml Rivet.aida ; # Works!
>>>>>>
>>>>>> cp Rivet.aida Moo.aida
>>>>>> rivet-mkhtml Moo.aia ; # Works!
>>>>>>
>>>>>> cp Rivet.aida Moo_moo.aida
>>>>>> rivet-mkhtml Moo_moo.aia ; # Fails!
>>>>>>
>>>>>> #(blank plots)
>>>>>>
>>>>>> I hope that helps.
>>>>>>
>>>>>> Anyway, for the moment I now have a work around ...
>>>>>>
>>>>>> On Fri, 4 Nov 2011, Christopher Lester wrote:
>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> Attached are three identical files differing only in name. Rivet.aida
>>>>>>> was generated from a rivet analysis called LESTER_TEST1. THe other
>>>>>>> two
>>>>>>> files are just copies of the same Aida file, but with different
>>>>>>> names:
>>>>>>> Moo.aida and LESTER_TEST1.aida.
>>>>>>>
>>>>>>> I find it very odd, then, that
>>>>>>>
>>>>>>> rivet-mkhtml Rivet.aida
>>>>>>> rivet-mkhtml Moo.aida
>>>>>>>
>>>>>>> both work as expected (i.e. make sensible plots) but
>>>>>>>
>>>>>>> rivet-mkhtml LESTER_TEST1.aida
>>>>>>>
>>>>>>> makes "blank" plots ... i.e. goes through all the motions of
>>>>>>> producing
>>>>>>> png and pdf files and a web page and all ... but makes all the plots
>>>>>>> as uniform expanses of white canvas with nothing on.
>>>>>>>
>>>>>>> Is this a bug or a feature?
>>>>>>>
>>>>>>
>>>>>> -------------------------------------
>>>>>> Christopher Lester 01223 690509
>>>>>> 59 Mulberry Close, Cambridge, CB4 2AS
>>>>>> -----------------------------------------------------
>>>>>> Dr Christopher Lester Tel: +44 (0)1223 337232
>>>>>> Room 952, Department of Physics, Cavendish Laboratory
>>>>>> JJ Thomson Avenue, Cambridge, CB3 0HE, United Kingdom
>>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Dr Andy Buckley
>>>>> SUPA Advanced Research Fellow
>>>>> Particle Physics Experiment Group, University of Edinburgh
>>>>>
>>>>> The University of Edinburgh is a charitable body, registered in
>>>>> Scotland, with registration number SC005336.
>>>>>
>>>>>
>>>>
>>>> -------------------------------------
>>>> Christopher Lester 01223 690509
>>>> 59 Mulberry Close, Cambridge, CB4 2AS
>>>> -----------------------------------------------------
>>>> Dr Christopher Lester Tel: +44 (0)1223 337232
>>>> Room 952, Department of Physics, Cavendish Laboratory
>>>> JJ Thomson Avenue, Cambridge, CB3 0HE, United Kingdom
>>>>
>>>>
>>>> _______________________________________________
>>>> Rivet mailing list
>>>> Rivet at projects.hepforge.org
>>>> http://www.hepforge.org/lists/listinfo/rivet
>>>
>>
>> -------------------------------------
>> Christopher Lester 01223 690509
>> 59 Mulberry Close, Cambridge, CB4 2AS
>> -----------------------------------------------------
>> Dr Christopher Lester Tel: +44 (0)1223 337232
>> Room 952, Department of Physics, Cavendish Laboratory
>> JJ Thomson Avenue, Cambridge, CB3 0HE, United Kingdom
>>
>
>
> --
> Dr Andy Buckley
> SUPA Advanced Research Fellow
> Particle Physics Experiment Group, University of Edinburgh
>
> The University of Edinburgh is a charitable body, registered in
> Scotland, with registration number SC005336.
>
>
-------------------------------------
Christopher Lester 01223 690509
59 Mulberry Close, Cambridge, CB4 2AS
-----------------------------------------------------
Dr Christopher Lester Tel: +44 (0)1223 337232
Room 952, Department of Physics, Cavendish Laboratory
JJ Thomson Avenue, Cambridge, CB3 0HE, United Kingdom
More information about the Rivet
mailing list
|