Package Details: geant4 10.2.2-1

Git Clone URL: https://aur.archlinux.org/geant4.git (read-only)
Package Base: geant4
Description: A simulation toolkit for particle physics interactions.
Upstream URL: http://geant4.cern.ch/
Licenses: custom: http://geant4.cern.ch/license/
Replaces: geant4-deb
Submitter: Eothred
Maintainer: donpicoro
Last Packager: donpicoro
Votes: 11
Popularity: 0.000000
First Submitted: 2010-04-08 08:54
Last Updated: 2016-06-29 15:43

Dependencies (31)

Sources (2)

Latest Comments

donpicoro commented on 2015-01-22 13:54

I do not want to roll it to everybody just yet. Mine with Qt4 work just fine. I tried with Qt5 and it seems to work as well and you help me trying it out

In order to make it work with Qt5 please when compiling comment out the following line:
#patch -Np1 Geant4InterfaceOptions.cmake < ${srcdir}/Geant4InterfaceOptions.patch

and simply change QT_SELECT=4 by QT_SELECT=5. I would appreciate the feedback if people let me know it works for them as well.

donpicoro commented on 2015-01-22 13:53

I tried and it seems like it is working but I do not want to roll it to everybody just yet. Mine with Qt4 work just fine. I tried with Qt5 and its to work as well.

In order to make it work with Qt5 please when compiling comment out the following line:
#patch -Np1 Geant4InterfaceOptions.cmake < ${srcdir}/Geant4InterfaceOptions.patch

and simply change QT_SELECT=4 by QT_SELECT=5 and try it. I would appreciate the feedback if people let me know it works for them as well.

Stevenwalton commented on 2015-01-22 07:40

Since the new package has come out are you going to add the fix for qt5?

Hell-G commented on 2014-12-16 19:21

Anybody having issues with the new Geant4 version?
I can't run my code and the official examples I have tested. I always get an error concerning the G4EMLOW data files. I have version 6.41 installed, as it comes with this package. This is also the version reported on the Geant4 homepage for version 10.01. According to the error message, it looks like that Geant4 is looking for the data file from version 6.35 of G4EMLOW but that is of course not installed anymore. Here is a sample error message:

-------- EEEE ------- G4Exception-START -------- EEEE -------
*** G4Exception : em0003
issued by : G4SeltzerBergerModel::ReadData()
Bremsstrahlung data file </usr/share/geant4-ledata/G4EMLOW6.35/brem_SB/br1> is not opened!G4LEDATA version should be G4EMLOW6.23 or later.

*** Fatal Exception *** core dump ***
-------- EEEE -------- G4Exception-END --------- EEEE -------


*** G4Exception: Aborting execution ***

Thanks for your feedback!

donpicoro commented on 2014-10-27 09:53

The only reason to have the "internal" clhep as opposed the "external" one was to save the user from compiling something extra. Most of my applications work perfectly with the internal one but I guess in order to be more inclusive we can have it as a dependency until the point something breaks.

And as for the Qt, I do not justify a new release bump for it. As I said let's wait and see the actual support for Qt in the next release.

JP-Ellis commented on 2014-10-27 09:48

Regarding the cmake >= 3, I did that because cmake 3 introduces a new way of importing libraries and remember getting some issues at one stage... but I can't replicate them now so it probably isn't strictly necessary.

Also, regarding the CLHEP, I thought it better to include the full CLHEP (and I in fact created a clhep-git version on AUR); however, I was doing a long run just then and it appears that the latest CLHEP isn't as compatible as I had hoped as it (very rarely) generates an invalid floating point operation.

Having said that, the floating point error may be restricted to my computer, since on my computer I get "WARNING - FPE detection is activated" at the start of every Geant4 run, but on another computer I regularly use I do not. I guess it is up to you whether to make CLHEP a dependency or not.

PS: I also made a mistake with one of the building options. `-DQT_QMAKE_EXECUTABLE=ON` can be ignored (in fact, it is invalid as it stands and shows that Geant4 completely ignores your environment and manually looks for the files).

JP-Ellis commented on 2014-10-27 09:43

Regarding the cmake >= 3, I did that because cmake 3 introduces a new way of importing libraries and remember getting some issues at one stage... but I can't replicate them now so it probably isn't strictly necessary.

Also, regarding the CLHEP, I thought it better to include the full CLHEP (and I in fact created a clhep-git version on AUR); however, I was doing a long run just then and it appears that the latest CLHEP isn't as compatible as I had hoped as it (very rarely) generates an invalid floating point operation.

Having said that, the floating point error may be restricted to my computer, since on my computer I get "WARNING - FPE detection is activated" at the start of every Geant4 run, but on another computer I regularly use I do not. I guess it is up to you whether to make CLHEP a dependency or not.

donpicoro commented on 2014-10-26 22:52

I see, just commenting those lines out, eh?

I reverted to the stable version. People can use the version they prefer but I plan to 'maintain' only the stable one.

I kept the external clhep as you had it and the cmake>=3 dependency although not sure why. The important one is definitively NOT to have 2.6 ;-)

Thanks for the patch and I hope it works for everybody =P.
We'll see how it turns out on Dec. 05 when the new release comes out.

/Pico



JP-Ellis commented on 2014-10-26 11:37

This has issues with Qt5, and unfortunately, Geant4 by default searches for the Qt5 libraries first (despite Qt5 support being experimental at the moment). It also searches for the Qt5 libraries in such a way that using QT_SELECT=4 is ignored.

I have managed to fix this issue by patching the cmake file so that it ignores Qt5 entirely. Note that the pkgbuild has been update to the beta version (feel free to revert it).

### START OF PKGBUILD ###
# Maintainer: Luis Sarmiento < Luis.Sarmiento-ala-nuclear.lu.se >
pkgname='geant4'
pkgver=10.1beta1
_pkgver=10.01.b01
#pkgver=10.0.2
#_pkgver=10.00.p02
pkgrel=1
pkgdesc="A simulation toolkit for particle physics interactions."
depends=('cmake>=3' 'xerces-c' 'qt4' 'glu' 'soxt' 'mesa' 'clhep')
replaces=('geant4-deb')
optdepends=('java-environment: for histogram visualizations and analysis'
'tcsh: for C Shell support'
'python: for G4Python support'
'geant4-abladata: Data files for nuclear shell effects in INCL/ABLA hadronic mode'
'geant4-ledata: Data files for low energy electromagnetic processes'
'geant4-levelgammadata: Data files for photon evaporation'
'geant4-neutronhpdata: Neutron data files with thermal cross sections'
'geant4-neutronxsdata: Data files for evaluated neutron cross sections on natural composition of elements'
'geant4-piidata: Data files for shell ionisation cross sections'
'geant4-radioactivedata: Data files for radioactive decay hadronic processes'
'geant4-realsurfacedata: Data files for measured optical surface reflectance'
'geant4-saiddata: Data files from evaluated cross-sections in SAID data-base')
url="http://geant4.cern.ch/"
arch=('x86_64' 'i686')
license=('custom: http://geant4.cern.ch/license/')
options=('!emptydirs')
install="${pkgname}.install"
source=("http://geant4.cern.ch/support/source/${pkgname}.${_pkgver}.tar.gz"
"Geant4InterfaceOptions.patch"
"${pkgname}.install")
md5sums=('6e2276cf017a60f042f1edc271e0c172'
'1d282b540d671bad495b3d86c32e4324'
'c3652d79011cd3fab2e1081e22f9f513')

prepare() {
cd ${srcdir}/${pkgname}.${_pkgver}
patch -Np1 -i ../Geant4InterfaceOptions.patch
}

build() {
[ -d ${srcdir}/build ] || mkdir ${srcdir}/build
cd ${srcdir}/build
env -i \
QT_SELECT=4 \
PATH=/usr/bin \
cmake -Wno-dev \
-DCMAKE_BUILD_TYPE=Debug \
-DCMAKE_INSTALL_PREFIX=/usr \
-DCMAKE_INSTALL_LIBDIR=lib \
-DGEANT4_USE_GDML=ON \
-DGEANT4_USE_G3TOG4=ON \
-DGEANT4_USE_QT=ON \
-DQT_QMAKE_EXECUTABLE=ON \
-DGEANT4_USE_XM=ON \
-DGEANT4_BUILD_MULTITHREADED=ON \
-DGEANT4_USE_SYSTEM_CLHEP=ON \
-DGEANT4_USE_OPENGL_X11=ON \
-DGEANT4_USE_INVENTOR=ON \
-DGEANT4_USE_RAYTRACER_X11=ON \
-DGEANT4_USE_SYSTEM_EXPAT=ON \
-DGEANT4_USE_SYSTEM_ZLIB=ON \
-DGEANT4_BUILD_CXXSTD=c++11 \
../${pkgname}.${_pkgver}

make
}

package() {

#Since the basic package does not include the data files, their
#configuration should be removed from the configuration file. Data
#files are also available on the AUR and the environment variables are
#set automatically for you from the packages.

msg "Removing wrongly set environment variables"

variables=("G4LEDATA" \
"G4LEVELGAMMADATA" \
"G4NEUTRONHPDATA" \
"G4NEUTRONXSDATA" \
"G4PIIDATA" \
"G4RADIOACTIVEDATA" \
"G4REALSURFACEDATA" \
"G4SAIDXSDATA" \
"G4ABLADATA" )


for _varname in ${variables[*]}
do
sed -i "/${_varname}/d" ${srcdir}/build/InstallTreeFiles/geant4.sh
sed -i "/${_varname}/d" ${srcdir}/build/InstallTreeFiles/geant4.csh
done

cd ${srcdir}/build
make DESTDIR="${pkgdir}" install

echo 'pushd /usr/bin &> /dev/null && source geant4.sh && popd &>/dev/null' > ${srcdir}/geant4.profile.sh
echo 'pushd /usr/bin &> /dev/null && source geant4.csh && popd &>/dev/null' > ${srcdir}/geant4.profile.csh
install -d ${pkgdir}/etc/profile.d
install -m755 ${srcdir}/geant4.profile.sh ${pkgdir}/etc/profile.d/geant4.sh
install -m755 ${srcdir}/geant4.profile.csh ${pkgdir}/etc/profile.d/geant4.csh
}
### END OF PKGBUILD ###

### START OF Geant4InterfaceOptions.patch ###
diff --git a/cmake/Modules/Geant4InterfaceOptions.cmake b/cmake/Modules/Geant4InterfaceOptions.cmake.new
index 256bc91..94c2de6 100644
--- a/cmake/Modules/Geant4InterfaceOptions.cmake
+++ b/cmake/Modules/Geant4InterfaceOptions.cmake
@@ -88,29 +88,29 @@ if(GEANT4_USE_QT)
find_package(Qt5OpenGL QUIET)
find_package(Qt5PrintSupport QUIET)

- if(Qt5Core_FOUND
- AND Qt5Gui_FOUND
- AND Qt5Widgets_FOUND
- AND Qt5OpenGL_FOUND
- AND Qt5PrintSupport_FOUND)
- # Compatibility
- macro(qt4_wrap_cpp)
- qt5_wrap_cpp(${ARGN})
- endmacro()
- set(Qt5_USE_FILE_IN "${PROJECT_SOURCE_DIR}/cmake/Templates/Geant4UseQt5.cmake.in")
- set(QT_USE_FILE "${PROJECT_BINARY_DIR}/Geant4UseQt5.cmake")
- configure_file("${Qt5_USE_FILE_IN}" "${QT_USE_FILE}" @ONLY)
- get_target_property(QT_QMAKE_EXECUTABLE ${Qt5Core_QMAKE_EXECUTABLE} IMPORTED_LOCATION)
- set(G4QTLIBLIST "-lQt5PrintSupport -lQt5Widgets -lQt5Gui -lQt5Core")
- set(G4GLQTLIBLIST "-lQt5OpenGL ${G4QTLIBLIST}")
- else()
+ # if(Qt5Core_FOUND
+ # AND Qt5Gui_FOUND
+ # AND Qt5Widgets_FOUND
+ # AND Qt5OpenGL_FOUND
+ # AND Qt5PrintSupport_FOUND)
+ # # Compatibility
+ # macro(qt4_wrap_cpp)
+ # qt5_wrap_cpp(${ARGN})
+ # endmacro()
+ # set(Qt5_USE_FILE_IN "${PROJECT_SOURCE_DIR}/cmake/Templates/Geant4UseQt5.cmake.in")
+ # set(QT_USE_FILE "${PROJECT_BINARY_DIR}/Geant4UseQt5.cmake")
+ # configure_file("${Qt5_USE_FILE_IN}" "${QT_USE_FILE}" @ONLY)
+ # get_target_property(QT_QMAKE_EXECUTABLE ${Qt5Core_QMAKE_EXECUTABLE} IMPORTED_LOCATION)
+ # set(G4QTLIBLIST "-lQt5PrintSupport -lQt5Widgets -lQt5Gui -lQt5Core")
+ # set(G4GLQTLIBLIST "-lQt5OpenGL ${G4QTLIBLIST}")
+ # else()
unset(Qt5Core_DIR CACHE)
unset(Qt5Gui_DIR CACHE)
unset(Qt5Widgets_DIR CACHE)
unset(Qt5OpenGL_DIR CACHE)
unset(Qt5PrintSupport_DIR CACHE)
find_package(Qt4 REQUIRED COMPONENTS QtCore QtGui QtOpenGL)
- endif()
+ # endif()

find_package(OpenGL REQUIRED)
### END OF Geant4InterfaceOptions.patch ###

Hell-G commented on 2014-05-21 20:26

No worries! Thank you very much for your help!
I don't think it has anything to do with you preparing the package. I guess it is Geant4 and Qt5 related. I will see, if I can get help from the Geant4 community.

donpicoro commented on 2014-05-21 08:05

Sorry for not being more helpful here but without proper reproducibility of the problem it becomes almost impossible for me (also non Qt expert) to diagnose anything.

Since this doesn't look like a packaging issue I suggest you to post a question at the Geant4 forums:
http://hypernews.slac.stanford.edu/HyperNews/geant4/cindex
possibly under the category "(Graphical) User Interfaces" or wherever you see fit.

keep me posted should you find some insight about this.

Hell-G commented on 2014-05-21 00:35

Thank you very much for the additional suggestions! Unfortunately, it didn't work either. I'm out of ideas at the moment. Maybe an update will help at some point.

donpicoro commented on 2014-05-16 08:54

Or... this may or may not be the answer... but have you had a look at qtchooser? from
extra/qtchooser
?
and then at the PKGBUILD try to run (in the same line)

QT_SELECT=qt4 cmake ...

Another option is for example, to set Qt4 symlink /etc/xdg/qtchooser/4.conf to ~/.config/qtchooser/default.conf.

let me know how it goes

Hell-G commented on 2014-05-15 22:09

Unfortunately, it doesn't work for me. Even with
cmake -DGEANT4_USE_QT=ON -DQT_QMAKE_EXECUTABLE=/usr/bin/qmake-qt4
cmake still chooses Qt5 over Qt4. After this configure step, ccmake shows me:

Qt5Core_DIR /usr/lib64/cmake/Qt5Core
Qt5Gui_DIR /usr/lib64/cmake/Qt5Gui
Qt5OpenGL_DIR /usr/lib64/cmake/Qt5OpenGL
Qt5PrintSupport_DIR /usr/lib64/cmake/Qt5PrintSuppor
Qt5Widgets_DIR /usr/lib64/cmake/Qt5Widgets

This is confusing me, as most of the similar problems I found online were solved by this.

I also tried "export QT_SELECT=4" before cmake with the same result. This is surprising as this changes the output of "qmake --version" from
QMake version 3.0
Using Qt version 5.2.1 in /usr/lib
to
QMake version 2.01a
Using Qt version 4.8.6 in /usr/lib
but cmake still uses Qt5.

My next try was using
find_package( Qt4 REQUIRED )
set(QT_QMAKE_EXECUTABLE /usr/bin/qmake-qt4)
in the CMakeLists.txt of Geant4, but with no luck as well.

Any other ideas? Thanks!

Hell-G commented on 2014-05-15 00:53

Thanks donpicoro for your quick reply!
I'm using QT5 because I have both versions installed and as you say, Geant4 looks first for QT5. I don't really have a preference and to be honest, before your post, I was not even aware of the differences between QT4 and QT5. Now that I checked the features of QT5, I don't see why QT4 should not be fine for running Geant4 visualizations. Apart from the intrinsic urge that higher version numbers are better of course ;-)
Apparently, they are more specific in the installation guide than in the release notes. From the release notes it looked like that QT5 should be just fine. I will try to compile Geant4 against QT4 and see, if it works better.
Thanks a lot!

donpicoro commented on 2014-05-14 20:40

According to the documentation:
http://geant4.web.cern.ch/geant4/UserDocumentation/UsersGuides/InstallationGuide/html/ch02s03.html
it says:

"Please note that Qt5 support is still experimental."

If you have both installed Qt5 takes precedence over Qt4. You can try forcing the compilation against Qt4 byt setting the QT_QMAKE_EXECUTABLE variable to /usr/bin/qmake-qt4.
Of course that is if you want to try to run it with QT4. I do not use Qt5 and I have no issues whatsoever. Is there a reason to need QT5?

Hell-G commented on 2014-05-14 20:25

Hi guys,

I'm having the same issue with QT5 since I upgraded to Geant4 version 10. The error just before the segmentation fault is not related to the problem. "Qt Warning - invalid keysym: dead_actute" is just a typo in the libX11 package. See for instance here:
http://unix.stackexchange.com/questions/111181/how-to-find-out-whats-wrong-on-all-of-my-qt-programs
The typo is still there on my system, although I have libX11 version 1.6.2-1.

When I try to run the basic examples, the segmentation fault does not occur all the time. Sometimes the OpenGL viewer opens, but most of the time it crashes. I'm not really experienced with gdb, but I was able to get these hints:

Available UI session types: [ Qt, Xm, GAG, tcsh, csh ]
[New Thread 0x7fffe48f9700 (LWP 3107)]
[New Thread 0x7fffd7ac0700 (LWP 3108)]

Program received signal SIGSEGV, Segmentation fault.
0x00007fffefeb81ca in QWidgetPrivate::showChildren(bool) ()
from /usr/lib/libQt5Widgets.so.5

According to the release notes of Geant4 10, it sounds like QT5 should be supported. What do you think about these statements in the release notes?
- Provided scripts for support of Qt5 external libraries.
- Added cmake configuration for Wt and Qt5 in "common" and "basic" modules.
- Fixes for support of Qt5.

I looked through the Geant4 Hypernews. For version 10 they sometimes recommend to go to QT5 instead of QT4 to avoid problems.

Any new insights here?
Thanks!

xzy3186 commented on 2014-05-13 12:01

@Eothred, donpicoro

Thank you all for the nice job!

Eothred commented on 2014-05-13 10:58

Hi,

I keep backups of my pkgbuilds on gitorious:
https://gitorious.org/eothred/aur/source/209595bc240542ff0d29f4beb485971d1523738d:geant4
(this is from back when I maintained this package, quite a while ago now!)

The newer version for Chakra might save you:
https://gitorious.org/eothred/ccr/source/99dc59ba953531b0b1300807db50bb12b02312ff:geant4

donpicoro commented on 2014-05-13 10:00

Hi xzy3186,
unfortunately I do not have a copy of it. Sorry.

If you want to make your own PKGBUILD you should remember that the cmake procedure was on alpha stage at that time so your mileage may vary. However it is not very well documented.

Ideally it should be completely possible to do it with cmake but probably it would be easier to do it using autotools. Here is the installation guide from 9.4:
http://geant4.slac.stanford.edu/tutorial/installation/Geant4.9.4/Linux/Geant4_9_4_Linux_Installation.htm#_Doing_the_Initial

I hope it helps somehow.

/Pico

xzy3186 commented on 2014-05-13 09:48

Hi, thanks again for maintaining this package. Do you have any backup PKGBUILD for the version 9.4?

donpicoro commented on 2014-01-22 09:20

1) I do not have a segfault with gv. The important part is that it creates successfully the prim file.

2) as it seems your geant4 installation was compiled against qt5 which for geant4 is still experimental. I do not use qt5 and perhaps that is why I cannot reproduce the error. What qt5 packages do you have so I try to check for that?

3) Fine, I would blame it in on Qt5 for now.

MarkH commented on 2014-01-22 00:00

1) dawn successfully created an eps file and gv opened. Segfault occurs upon closing gv. Generated prim file: http://pastebin.com/BPRmbNUS

2) ldd output: http://pastebin.com/rV8fs0wm

3) Everything is en_US.UTF-8

donpicoro commented on 2014-01-21 15:35

Grrr... I'm flying blind without being able to reproduce the error myself.

There are some things I would like to try/know with the same example you used
1) try using the following lines as your vis.mac file:
/vis/open DAWNFILE
/vis/drawVolume
/vis/scene/create
/vis/viewer/flush

2) can you please give me the output of "ldd -d -r exampleB1" from the folder where you build the application.

3) Are you using any non "en_US.UTF-8" locale on your system?

MarkH commented on 2014-01-19 11:20

I've posted the full output of running example B1 on a native installation (not a virtual machine) to http://pastebin.com/SZNfrvMV. The relevant error messages occur at the bottom:

"Qt Warning - invalid keysym: dead_actute"
Segmentation fault (core dumped)

donpicoro commented on 2014-01-18 16:45

mmm that is extrange, would it be possibe that there are missing dependencies? that could be it.

Can you put the full error? As I told you it works for me so I'm guessing is a dependency issue.

MarkH commented on 2014-01-17 21:05

I have all of the data files installed and I've rebooted to source them. The example (and my own code) still crashes in the same place.

$ env | grep G4
G4LEVELGAMMADATA=/usr/share/geant4-levelgammadata/PhotonEvaporation3.0
G4NEUTRONXSDATA=/usr/share/geant4-neutronxsdata/G4NEUTRONXS1.4
G4LEDATA=/usr/share/geant4-ledata/G4EMLOW6.35
G4NEUTRONHPDATA=/usr/share/geant4-neutronhpdata/G4NDL4.4
G4RADIOACTIVEDATA=/usr/share/geant4-radioactivedata/RadioactiveDecay4.0
G4ABLADATA=/usr/share/geant4-abladata/G4ABLA3.0
G4PIIDATA=/usr/share/geant4-piidata/G4PII1.3
G4SAIDXSDATA=/usr/share/geant4-saiddata/G4SAIDDATA1.1
G4REALSURFACEDATA=/usr/share/geant4-realsurfacedata/RealSurface1.0

donpicoro commented on 2014-01-17 20:18

ok, that was easy and I am lucky...

fortunately I also got the segmentation fault otherwise I wouldn't have noticed that (as it seems) your problem is that you are missing the package geant4-saiddata which is an optional dependency and of geant4 also availabel in the AUR.

I installed myself and it worked right away (do not forget to run "source /etc/profile.d/geant4.sh", log out and log in again, or reboot. Your choice.

I hope that solves your problems. Please let me know how it goes.

MarkH commented on 2014-01-17 20:01

After building /usr/share/Geant4-10.0.0/examples/basic/B1/ and running ./exampleB1, the program crashes at the line 106 in exampleB1.cc.

G4UIExecutive* ui = new G4UIExecutive(argc, argv);

I'm running this in VirtualBox, so I get the error message:

!!Assertion Failed!!
Expression: RT_SUCCESS_NP(rc)
Location : /build/virtualbox/src/VirtualBox-4.3.6/src/VBox/Runtime/common/misc/thread.cpp(227) void rtThreadLockRW()
VERR_INVALID_HANDLE (-4) - Invalid loader handle.

However, I have run the same program on a native installation and I get a normal "Segmentation Fault (core dumped)".


Commands run:
$ cmake /usr/share/Geant4-10.0.0/examples/basic/B1/
$ make
$ ./exampleB1

donpicoro commented on 2014-01-17 10:51

Sorry... I am writing a paper this week and hadn't had time to have a look on whether even my own simulations are crashing. I will have a look over the weekend.

Can you give me the details of the exact commands you ran so I can try to reproduce the crash and potentially fix it or file a corresponding bug report somewhere?

Thanks.

/Pico

MarkH commented on 2014-01-15 00:14

OK, now I think I've found a real problem. After upgrading to 4.10, any program with visualizations crashes with a segfault. For example, the first basic example (/usr/share/Geant4-10.0.0/examples/basic/B1/) compiles fine, but running it results in a segfault when calling

UIManager->ApplyCommand("/control/execute init.mac");

Pure text output simulations work fine.

donpicoro commented on 2014-01-03 03:37

Hehe,
new environment variables only require you to log out and log in again =P

Rebooting, of course, also does the trick ;-)

I am glad it is working. I am currently on holidays and I haven't had time to properly test it myself.

Thanks for the feedback.

MarkH commented on 2014-01-03 01:37

Nevermind. Restarting my computer loaded the correct environment variables. All is well.

Thanks for the hard work.

MarkH commented on 2014-01-03 01:27

After updating to 10.0, I'm having trouble running programs that ran under the previous version of geant4. I get the following error message:

-------- EEEE ------- G4Exception-START -------- EEEE -------
*** G4Exception : em0003
issued by : G4SeltzerBergerModel::ReadData()
Bremsstrahlung data file </usr/share/geant4-ledata/G4EMLOW6.32/brem_SB/br6> is not opened!G4LEDATA version should be G4EMLOW6.23 or later.

*** Fatal Exception *** core dump ***
-------- EEEE -------- G4Exception-END --------- EEEE -------


*** G4Exception: Aborting execution ***
Aborted (core dumped)


I have the latest geant4-ledata installed (6.35-1), so it seems like some definition in geant4 didn't get updated by the geant4-ledata package.

donpicoro commented on 2013-03-28 14:49

I figured out how to handle/keep the location of the data files automatically.

From now on everything should "just work". It should not matter whether you update/install/remove just the package or just the data sets.

If you encounter any errors, please let me know.

donpicoro commented on 2013-03-23 19:33

@carbolymer:

Mmm that is strange. I just rebuilt it with no problems on x86_64 machine.

Is your system up to date? Maybe is a 32bits machine? I haven't tested the package for that architecture in a while.

Anonymous comment on 2013-03-23 18:38

I've got a problem with QT during installation:
http://pastebin.com/raw.php?i=LyCjx6f2
Any ideas? Package is not compiling properly.

donpicoro commented on 2013-02-13 15:40

When you updrade the package, the file /etc/profile.d/geant4.sh will be overwritten. I am not sure how to handle this for now.
So please, before updating take note of the data set you have defined there and make sure you place them back, for example in my case.

pushd /usr/bin &> /dev/null && source geant4.sh && popd &>/dev/null
export G4LEDATA=/usr/share/geant4-ledata/G4EMLOW6.32
export G4LEVELGAMMADATA=/usr/share/geant4-levelgammadata/PhotonEvaporation2.3
export G4RADIOACTIVEDATA=/usr/share/geant4-radioactivedata/RadioactiveDecay3.6
export G4REALSURFACEDATA=/usr/share/geant4-realsurfacedata/RealSurface1.0

or simply reinstall the datasets you want. At installation time the environment variables are automatically added to that file.

Sorry for the minor inconvenience.

donpicoro commented on 2012-04-05 20:37

@exaos: Yes and No...
The packages are not actually listed anywhere... For that I would have to actually create different packages instead of simply shipping them split into this one...
For now I think I will stay with this naming scheme... it goes with the environment variables in geant4. I am glad it works for you now.

Anonymous comment on 2012-04-05 13:40

At last, I successfully got the packages like -- :-D

geant4-9.5.1-2 geant4-abladata-3.0-1 geant4-ledata-6.23-1
geant4-levelgammadata-2.2-1 geant4-neutronhpdata-4.0-1
geant4-neutronxsdata-1.1-1 geant4-piidata-1.3-1 geant4-radioactivedata-3.4-1
geant4-realsurfacedata-1.0-1

I thinks, it's better to rename all the data packages as "geant4-data-XXX-<ver>", e.g. geant4-data-abla, geant4-data-le, etc. It will help user to search by names.

donpicoro commented on 2012-04-04 23:47

Ok, uploaded again. It is now split :)
when you install the geant4 package you are actually installing geant4 and each one of the data sets as an independent package so people who do not really want to use (some or all of) them can easily and clean uninstall them.

I am currently in testing stage to check whether all the different environment variables are correctly set but please report back if for some reason it doesn't work for you.

@exaos: By using this approach I managed to lower the temporal folder usage down to a max of about 2.9GB so I hope it now works for you... it was one hell of a change but I think it was worth it.

donpicoro commented on 2012-04-04 16:25

@exaos: You are right...

I monitored the temporal disk consumption and you are a some MB too short :(
The temporal folder takes about 1GB before decompressing the data files.
It suddenly jumps to 2.4GB when decompressed
And since creating the package requires to 'copy the data files' to the package folder in the end it ends up being 4442MB(!!!) big.
Not to mention that to install it you still need on your root partition (/) another 1.5GB that the final package actually uses.

I know is VERY stupid to take all that space at once and I am working on creating a group of packages instead of this unholy mess.

In the meantime if you need to install/upgrade desperately, there is a workaround to this... if you are using yaourt as your AUR helper go to the file /etc/yaourtrc uncomment and change the line
#TMPDIR="/tmp"
to a location with enough space.

Sorry for the inconvenience... Please check back later today to check whether I manage to tear this apart into a group of packages.

Anonymous comment on 2012-04-04 09:46

Though my "/tmp" is 4GB, I failed on "out of space" ERROR.

donpicoro commented on 2012-04-03 21:47

updated to version 9.5.1 (patch01).
I gave it an attempt to split the package to no success :(
so, again, if you do not need to download the data files just edit the PKGBUILD accordingly.
Sorry.

NOTE: Make sure you have enough space to build it... remember that it is about 1.5GB -yes, GB(!)-
Uninstalling the previous version might help you to liberate some space or configure your building directory in a partition with enough free space.

If someone knows how to split a package with different source arrays in the PKGBUILD to avoid the hassle of redownloading the data files... suggestions are welcome.

There is a third option that is to properly create two PKGBUILD files... one for geant4 and another for geant4-data. Any preference?

donpicoro commented on 2012-04-03 21:23

updated to version 9.5.1 (patch01).
I gave it an attempt to split the package to no success :(
so, again, if you do not need to download the data files just edit the PKGBUILD accordingly.
Sorry.

donpicoro commented on 2012-02-17 14:11

I will definitively consider doing it. Maybe even just for the sake of learning. For now until I get more complaints it won't be my priority to split the packages -at least for a couple of weeks-. Please remember I am writing my PhD thesis and that is my priority.

Nevertheless it looks easy enough to give a go really soon. However I do not want to just edit that example. Is it documented somewhere?
Thank you very much.

Eothred commented on 2012-02-17 09:14

What you could consider is using split package to split this in geant4/geant4-data, where the latter would hold the data files. You find an example on how to split here: https://aur.archlinux.org/packages.php?ID=47680

donpicoro commented on 2012-02-12 03:37

Well. If too many people face the issue then yes I would consider disabling it but I prefer to have by default the most complete posible "package" built.

for now I think having documented the alternative is enough.
please report back if you encounter any other issue. :)

Anonymous comment on 2012-02-12 03:00

OK. Thanks.
I think, it's better to let the user download the data manually, because the data archives are big and the network status is quite different.

donpicoro commented on 2012-02-12 00:15

Then download yourself the data files you need from the following site:

http://geant4.cern.ch/support/download.shtml

then unpack them somewhere, for example to: /usr/local/GEANT4/data/ and then define the variables somewhere in your global scripts

export G4ABLADATA=/usr/local/GEANT4/data/G4ABLA3.0
export G4LEDATA=/usr/local/GEANT4/data/G4EMLOW6.23
export G4LEVELGAMMADATA=/usr/local/GEANT4/data/PhotonEvaporation2.2
export G4NEUTRONHPDATA=/usr/local/GEANT4/data/G4NDL4.0
export G4NEUTRONXSDATA=/usr/local/GEANT4/data/G4NEUTRONXS1.1
export G4PIIDATA=/usr/local/GEANT4/data/G4PII1.3
export G4RADIOACTIVEDATA=/usr/local/GEANT4/data/RadioactiveDecay3.4
export G4REALSURFACEDATA=/usr/local/GEANT4/data/RealSurface1.0

and tell me how it goes.

Anonymous comment on 2012-02-12 00:08

I succeeded in building the package by turning off the GEANT4_INSTALL_DATA.

donpicoro commented on 2012-02-11 22:47

Mmm that is VERY odd. I would suggest you to try two things:

1) Please try again to install the package, just for the sake of it.
2) Try deactivating the feature that downloads the data files since the error you reported seems to be related to it.
To do so just change in the PKGBUILD form
-DGEANT4_INSTALL_DATA=ON \
to
-DGEANT4_INSTALL_DATA=OFF \

and please let me know how it goes. If by doing solution two it works I strongly suggest you trying solution 1) since apparently the Photon-evaporation files you downloaded got corrupted.

donpicoro commented on 2012-02-11 22:46

Mmm that is VERY odd. I would suggest you to try two things:

1) Please try again to install the package, just for the sake of it.
2) Try deactivating the feature that downloads the data files since the error you reported seems to be related to it.
To do so just change in the PKGBUILD form
-DGEANT4_INSTALL_DATA=ON \
to
-DGEANT4_INSTALL_DATA=OFF \

and please let me know how it goes.

Anonymous comment on 2012-02-11 22:36

Build error:

CMake Error: Problem with archive_read_next_header(): Truncated input file (needed 11264 bytes, only 0 available)
CMake Error: Problem with archive_write_header(): Invalid empty pathname
CMake Error: Current file:
CMake Error: Problem with archive_read_next_header(): Truncated input file (needed 11264 bytes, only 0 available)make[2]: ^C*** [Externals/PhotonEvaporation-2.2/src/PhotonEvaporation-stamp/PhotonEvaporation-download] 中断
make[1]: *** [CMakeFiles/PhotonEvaporation.dir/all] 中断
make: *** [all] 中断

Eothred commented on 2012-02-08 10:41

Yep, I'll leave the geant4-deb package there for the time being, since it is precompiled. However I guess at some point (already perhaps?) it will break and then it should be removed from AUR.

donpicoro commented on 2012-02-04 00:47

This package replaces the other one called geant4-deb since there is no need for that since I have Eothred 'blessing' to keep the Geant4 project going on Arch.

The package is fully functional for my simulations so far with the 'internal' clhep library that comes with the geant4.9.5. Most users will just have the package that works 'out-of-the-box'.
If that is not your case just install clhep (also available from the AUR) and modify the geant4 PKGBUILD to be build against it.
Just modify the line GEANT4_USE_SYSTEM_CLHEP and eventually you might have to also include CLHEP_ROOT_DIR. In case you need the documentation can be found here:

http://geant4.web.cern.ch/geant4/UserDocumentation/UsersGuides/InstallationGuide/html/ch02s03.html

BEWARE that the package downloads ALL the data files by default. Its ~1.5GB are because of it. Consider deactivating GEANT4_INSTALL_DATA if you are ABSOLUTELY sure you do not need them or if you plan to add them yourself on a later stage.

Please test it and vote for it if you find it useful. You can also find DAWN and DAVID on the AUR and eventually I can also add AIDAJNI, AIDA, JAIDA, JAS3 or WIRED to the AUR if you ask me ;-)
Enjoy.

donpicoro commented on 2012-02-04 00:42

This package replaces the other one called geant4-deb since there is no need for that I have Eothred 'blessing' to keep the Geant4 project going.

It is fully functional for my simulations so far with the 'internal' clhep library that comes with the geant4.9.5. Most users will just have the package that works 'out-of-the-box'.
If that is not your case just install clhep (also available from the AUR) and modify the geant4 PKGBUILD to be build against it.
Just modify the line GEANT4_USE_SYSTEM_CLHEP and eventually you might have to also include CLHEP_ROOT_DIR. In case you need the documentation can be found here:

http://geant4.web.cern.ch/geant4/UserDocumentation/UsersGuides/InstallationGuide/html/ch02s03.html

Please test it and vote for it if you find it useful. You can also find DAWN and DAVID on the AUR and eventually I can also add AIDAJNI, AIDA, JAIDA, JAS3 or WIRED to the AUR if you ask me ;-)
Enjoy.

Eothred commented on 2012-02-02 18:38

Ah, excellent. I'll keep myself notified of comments to this package for a while.. Good luck!

I took the script from the Debian package I maintained back in the days: http://wiki.debian.org/DebianScience/Geant4

I'm not sure it's documented anywhere. Essentially it is just a wrapper script which sets up the Geant4 environment variables. You can either source it or put it before your normal command, e.g. "g4env make" instead of just "make".

That said, I think the script is deprecated, as the cmake build system is auto-generating geant4.sh. If you find that geant4.sh is sufficient I wouldn't keep the g4env script anymore...

donpicoro commented on 2012-02-02 06:10

@Eothred:
ok, I think I got it now. I am struggling with converting my applications to the new cmake compilation style but other than that I think I got it now.

Feel free to leave this baby on my hands.
I will provably take over this whole simulation business for arch. Dawn, dawncut, david and so on... any suggestions? you seem to [have] work[ed] on this topics.

Cheers.

donpicoro commented on 2012-02-02 02:25

I just adopted clhep and I am interested in putting my two cents for arch. I already managed to install this new version of geant4 on my own so I think I can take it from now.
HOWEVER,
A little help regarding the usage of the g4env script will come handy specially for application developers. I cannot compile my programs and they are not -yet- migrated to the new cmake so I need to know how to use the backwards compatibility :P.

If you don't mind explaining a bit. I do not mind adopting :D

Cheers.

donpicoro commented on 2012-02-01 20:38

I just adopted clhep and I am interested in putting my two cents for arch. I already managed to install this new version of geant4 on my own so I think I can take it from now.

Eothred commented on 2012-02-01 09:05

I'm not using Arch at the moment. Could anyone volunteer to take over this package?

donpicoro commented on 2012-01-11 02:12

@Eothred:
Perhaps something here can be useful to you?
http://hypernews.slac.stanford.edu/HyperNews/geant4/get/installconfig/1394.html?inline=-1

Eothred commented on 2011-12-17 10:12

Ah, sorry, you said you had it installed. Then I will investigate further. Thanks for reporting.

Eothred commented on 2011-12-17 10:12

Try to remove this line:
-DGEANT4_USE_QT=ON \

If you don't have Qt installed then you should probably exclude that one. That said, I should add it to list of dependencies. Thanks for the warning.

donpicoro commented on 2011-12-17 01:30

I found the following error and I really can't figure out why. Any clues? (qt 4.7.4-3 installed)

CMake Error at /usr/share/cmake-2.8/Modules/FindPackageHandleStandardArgs.cmake:91 (MESSAGE):
Could NOT find Qt4 (missing: QT_QMAKE_EXECUTABLE QT_MOC_EXECUTABLE
QT_RCC_EXECUTABLE QT_INCLUDE_DIR QT_LIBRARY_DIR QT_QTCORE_INCLUDE_DIR
QT_QTCORE_LIBRARY QT_QTGUI_INCLUDE_DIR QT_QTGUI_LIBRARY
QT_QTOPENGL_INCLUDE_DIR QT_QTOPENGL_LIBRARY QT_UIC_EXECUTABLE)
Call Stack (most recent call first):
/usr/share/cmake-2.8/Modules/FindPackageHandleStandardArgs.cmake:252 (_FPHSA_FAILURE_MESSAGE)
/usr/share/cmake-2.8/Modules/FindQt4.cmake:1163 (FIND_PACKAGE_HANDLE_STANDARD_ARGS)
cmake/Modules/Geant4InterfaceOptions.cmake:76 (find_package)
CMakeLists.txt:84 (include)

Eothred commented on 2011-12-08 15:23

New release.

Be warned, this package currently includes the data files which are 1.6 GB of ascii data installed to /usr/share/!

If you don't want that remove the line "-DGEANT4_INSTALL_DATA=ON \"

I'm thinking that one should instead do a split-build but I don't have time just now. Patches to make this build better are welcome!

DarioP commented on 2011-08-04 13:22

I managed to solve the lG3toG4 error disabling G3 support in the PKGBUILD and recompiling. That should be ok as I'll never will have to face something written in fortran..

I've also been able to compile the first example using make instead of the recommended gmake, many people recommend to symlink gmake to make.

Thank you for maintaining this package!

DarioP commented on 2011-08-04 09:41

I've installed the package and try to learn this enviornment. Compilations and all was fine, I also appreciated a lot the environment setting scripts.

But when I tried to compile the first novice example using the make command, I got the error:

/usr/bin/ld: cannot find -lG3toG4
collect2: ld returned 1 exit status

and I'm not able to fix it.

I'm also confused about the gmake command which is not present in my installation and it's reported in the official G4 guide..

DarioP commented on 2011-08-04 09:36

I've installed the package and try to learn this enviornment. Compilations and all was fine, I also appreciated a lot the environment setting scripts.

But when I tried to compile the first novice example using the make command, I got the error:

/usr/bin/ld: cannot find -lG3toG4
collect2: ld returned 1 exit status

and I'm not able to fix it.

I'm also confused about the gmake command which is not present in my installation and it's reported in the official G4 guide..

Eothred commented on 2011-07-06 07:30

As said so done. I tried to write a g4env wrapper script to invoke the geant4 environment, but I don't think it works 100% yet. Still, since the PKGBUILD didn't work yet in any case (debian files were out of date), I decided to update immediately. Please also see the package "geant4-deb" if you don't want to go through the process of compiling for a couple of hours and still ending up with a beta version of geant4..

This package will not use beta versions forever, but since I decided to use cmake for configuring, I went with it for now since it is supposed to be fully implemented in 4.9.5. People are more than welcome to complain, and propose alternative PKGBUILD which uses the old configuration system and version 4.9.4.

Eothred commented on 2011-07-04 16:36

good news, I have the 64bit packages of the latest geant4 ready. further, I have a new PKGBUILD which builds the 9.4.5 beta using cmake (the old config scripts are deprecated when the new release is out). When I have everything in order, I will upload the PKGBUILD for the new beta version with the name "geant4", and then this package will be renamed "geant4-deb". I think that is the best solution. That means that your votes might be out of order, but that is a minor issue I find. You can already find the new scripts here if you are curious:
https://gitorious.org/eothred/aur/trees/master/geant4
https://gitorious.org/eothred/aur/trees/master/geant4-deb
When the new Debian files are available I will update everything.

Eothred commented on 2011-04-13 09:31

heh, it takes four days from me finally managing to get the package updated and then there is a new release (9.4, 25. of February). and now they already released the first patch as well. I'll update when I'll update, currently I have little spare time on my hands. sorry about that, hope this version is decent enough for most of you.

Eothred commented on 2011-02-21 08:19

upgraded.

Eothred commented on 2011-02-17 17:13

geant4 for 64bit will be available tomorrow, but I go away on holidays for the weekend so I cannot update aur before after the weekend. Please use the PKGBUILD in the previous pastebin link, run "makepkg -g" and update the md5 array before you run makepkg. That should work fine. Or you can wait until next week when I update aur.

Eothred commented on 2011-02-11 19:31

I notice it is taking MUCH longer than anticipated to update the package, so I strongly recommend annyone willing to write a pkgbuild that simply pulls from source and compiles. Contact me if you have something working and I will remove myself as the owner of this package. The new patchlevel for i686 is available and you can use the PKGBUILD here: http://aur.pastebin.com/dQNe513S

For x86_64 the packages are available tonight, but I am waiting for the CLHEP maintainer to upload the CLHEP for that architecture. It was linked against 2.0.4.7 this time, so I cannot use the one from aur. Will update the aur package here as soon as everything is in place.

I have a PhD to write this year, so unless it is much less painful to update the next version I most likely wont do it. Hence, the recommendation above.

Eothred commented on 2010-12-08 16:42

I am aware that the package is out of date yes.

May I ask if people that successfully installed this package and find it useful vote for it? It gives me a hint about how much the package is used, so I know how urgent the updates are... At the moment I get the impression it is not at all urgent. Thanks.

donpicoro commented on 2010-12-02 21:20

FOUND IT !!!
Acording to the Geant4 FAQ web page (http://geant4.cern.ch/UserDocumentation/UsersGuides/ForApplicationDeveloper/html/ch10.html) and I quote

'cernlib' is a command from the CERN program library (cernlib) returning a list of libraries needed to link a cernlib application. This command is only used in the 'g3tog4' module, however, if you do not make use of the 'g3tog4' tool, it's harmless.

while talking about the error:
gmake[1]: cernlib: Command not found.

So unless you are using some tools that MUST use the tool g3tog4 (Geant3 is written in Fortran) is safe to ignore it. It should go as an optional dependence and perhaps the package should be build with the g4tog4 module by default (is harmless to have it there I guess)

By the way. I will flag the package as out-of-date since there is a new patch (p02) since 24th September. I hope you don't mind.

Eothred commented on 2010-11-30 18:46

So I checked on the Geant4 mailing list, and I don't see any reason why CERNLIB should be a dependency. There might be some specifics about how packages are split in Debian, but I could not see anything in the patches that are applied to Geant4 before it is built. As a conclusion I moved it to optdepends for now, just in case. If anyone finds anything not working without CERNLIB, please let me know!

I plan on updating CLHEP and Geant4 on Debian as soon as I have time to spare, and I will bump this package as soon as I have finished that.

Eothred commented on 2010-11-27 11:45

Looking at the Geant4 webpage I see that they only list gcc and clhep as dependencies. I will investigate this further and drop the dependency if it is not required.

Eothred commented on 2010-11-27 11:34

Hi donpicoro!

Not sure why I used ==, probably just a bug that managed to get through (perhaps one = got in there after I had tested the script..). Thank you for pointing it out to me.

The reason why I use DEBIAN is that I myself maintain the DEBIAN package as well, so I know how they are built. I think I mentioned before that if anyone wants to write a new script that compiles the source from scratch, then I am more than happy to hand over the ownership of this package and rename mine geant4-deb. The reason why you might want the debian version as well is that they are precompiled, so they are quicker to install.

CERNLIB is from what I can see only available FOR x86_64 on AUR, but (still) not i686. I see that I sent piggy1983 an updated script in the comments about half a year ago, but it has still not been taken up for consideration. Try the script that is linked to in the comments instead.

I am not extremely familiar with Geant4, but I do think it is used for some low level mathematical operations etc. It is a dependency for the Geant4 package, and I am quite sure it should be there.

The reason that g4make,g4run and g4env are used is that this way you do not have to set up the environment. These are wrapper script that sets up your variables at run-time.
Usage:
- Compile a project: "g4make"
- Run a binary: "g4run ./binary"
- Check your environment: "g4run g4env"
These scripts were written by the previous Debian maintainer, and I think it is a nicer way to use Geant4 than the more default. If you want a more traditional setup, you can essentially set the variables listed with the command "g4run g4env" in ~/.bashrc.


You say that none of the examples work? They don't compile or they don't execute? What error messages do you get?

donpicoro commented on 2010-11-26 00:13

Hey, fellow archers !
I do not why nodoby else got away with it. But I HAD to modify the PKGBUILD from:
"clhep==${_clhepver}"
to
"clhep=${_clhepver}"
(just remove one '=' sign) in order to make it work. Is it normal?

I 'installed' the package but I have absolutely no idea how debian works and I am still unable to compile and run ANY example.
I have a couple of questions:
-> CERNLIB is only available for i686 on AUR that means there is no way to install this package properly for x86_64 using this PKGBUILD? Then I sugggest you to remove the arch i686 and only allow it for x86_64 unless I am wrong. Am I?
-> Why do we need CERNLIB? What does it do?

I skipped the CERNLIB installation and tried to compile and run ANY of the examples but honestly I have absolutely no idea how to proceed. Why would I use g4make? What exactly does it do? I doesn't work anyway !

I have LOTS of experience compiling and running Geant4 since many years ago but I've always done it the "./Configure -build" way. And I am totally lost right now. Any pointers, hits, rants?

Thanks in advance !!!

p.s: Why do we need to use the debian packages? I am sure we can manage to do it a little bit more 'vanilla'. :D

donpicoro commented on 2010-11-26 00:11

Hey, fellow archers !
I just 'installed' the package but I have absolutely no idea how debian works and I am still unable to compile and run ANY example.

I have a couple of questions:
-> CERNLIB is only available for i686 on AUR that means there is no way to install this package properly for x86_64 using this PKGBUILD? Then I sugggest you to remove the arch i686 and only allow it for x86_64 unless I am wrong. Am I?
-> Why do we need CERNLIB? What does it do?

I skipped the CERNLIB installation and tried to compile and run ANY of the examples but honestly I have absolutely no idea how to proceed. Why would I use g4make? What exactly does it do? I doesn't work anyway !

I have LOTS of experience compiling and running Geant4 since many years ago but I've always done it the "./Configure -build" way. And I am totally lost right now. Any pointers, hits, rants?

Thanks in advance !!!

p.s: Why do we need to use the debian packages? I am sure we can manage to do it a little bit more 'vanilla'. :D

donpicoro commented on 2010-11-26 00:10

Hey, fellow archers !
I just 'installed' the package but I have absolutely no idea how debian works and I am still unable to compile and run ANY example.

I have a couple of questions:
-> CERNLIB is only available for i686 on AUR that means there is no way to install this package properly for x86_64 using this PKGBUILD? Then I sugggest you to remove the arch i686 and only allow it for x86_64 unless I am wrong. Am I?
-> Why do we need CERNLIB? What does it do?

I skipped the CERNLIB installation and tried to compile and run ANY of the examples but honestly I have absolutely no idea how to proceed. Why would I use g4make? What exactly does it do? I doesn't work anyway !

I have LOTS of experience compiling and running Geant4 since many years ago but I've always done it the "./Configure -build" way. And I am totally lost right now. Any pointers, hits, rants?

Thanks in advance !!!

p.s: Why do we need to use the debian packages? I am sure we can manage to do it a little bit more 'vanilla'. :D

Eothred commented on 2010-09-02 17:32

Oops, my bad. Fixed now. Thanks!

Anonymous comment on 2010-09-02 17:26

I think there is a mistake in the "arch" line of the PKGBUILD. Instead of i386 it should be i686. After changing it, I could install the package.

Eothred commented on 2010-09-02 17:25

Good question. I thought so? Perhaps it is only needed during compilation. At the moment I basically set the same dependencies as I have in the debian package which I adopted by a more skilled packager than me. I can have a look more in detail. If anyone can confirm the dependency one way or the other please inform me in the comments here.

I would not be surprised if this package has issues in certain modules. I am not a heavy user myself, so I have not tested it as thorough as I would've liked (I also have a PhD to write so my spare time is not limitless ;) ). General testing is most welcome!

Anonymous comment on 2010-09-02 16:28

Hello
thanks for the packages. I am also using Geant4 at work and will give the packages a try. Hopefully saves some time . :) If you need help with anything feel free to contact me. I am willing to contribute.

Btw. I had it installed without cernlib. Is it really necessary as dependency?

Eothred commented on 2010-07-13 11:26

OK, thank you very much for your information. I see that there is another package called DAWN in AUR, which I added as optional dependency. http://aur.archlinux.org/packages.php?ID=26848

A few things you can help me with:
-Extend the line "optdepends" in the PKGBUILD file and post the changes here or send them to me. OpenGL is already defined as a dependency (libgl), I couldn't find JAS or wired3 in the Arch build system or in AUR (they might have a different name).
-Test the packages, and report any problems.

The debian build is based on a few patches (not written by me), and requires an automated build script. I will see what I can do about heprep (it isn't included in the current version?)

donpicoro commented on 2010-07-12 19:27

When you create your geometries, you can view them using many drivers as they call them. Different drivers work for different purposes. The HepRep is probably the preferred one when you are developing your geometry because it allows you to rotate it, zoom and and the like.
But if you want to show your already constructed geometry, let's say in a conference, it is better to use another: for example DAWN.

DAWN renders very nicely ghostscripts (.ps) files -hence the gv dependence- with textures and shadows... and if you use it properly in conjunction with the 'david' software you can find non trivial overlaps in your geometry.

Maybe I did not use the correct word... it is not 'additional dependencies'. What I wanted to say was 'optional dependencies'.
There are many more optional dependencies (openGL, wired3, JAS, root, etc), I just wanted to point out the fact that it will be nice to include some of the 'visualization' dependencies.

Eothred commented on 2010-07-11 08:57

Cleaned the PKGBUILD as requested.

What is dawn and gv required for? I am not using geant4 extensively myself, so I do not actually know the code very well. I'll be happy to add them as dependencies if they are needed for the "full experience".

Eothred commented on 2010-07-11 08:49

I use the debian package because it is easy for me (I am anyway maintaining the debian packages as well), and it pre-build installs much quicker (geant4 compilation is easily a couple of hours). I guess in principle this package should be named geant4-deb or something, and then another person could take care of creating a PKGBUILD which directly installs from source. In addition, this debian package uses some wrapper scripts so that you don't have to mess up your environment with all the geant4-variables.

Yes, I solved the clhep issue. Temporary fixes are commented out, but should remove them completely as you say. I just wanted to wait until this version was confirmed working.

donpicoro commented on 2010-07-08 20:26

WOW this is an interesting package... I would love to do something about it. How can I help? did you already solved the clhep issues? according to you last post it seems like it, but then I suggest you to have a clean PKGBUILD file... or explain a little bit more those temporary fixes.

Why do we have to rely on the debian package?... I have geant4 up an running with my arch installation with no problems at all.
How about adding dawn and gv as additional dependencies. I do not how, but I am willing to learn how to create maybe the wired and HepRApp packages.
What do you think?

Thanks !!!

Eothred commented on 2010-05-11 15:13

Sorry for my delay (too much to do at work these days..). The updated geant version is linked against clhep 2.0.4.6, and clhep in aur has updated package version. I checked 64bit and it seems to work fine. If someone could check 32bit version and comment here how it went, I'd be happy :)

Eothred commented on 2010-04-08 09:00

I apologize, but at the moment the Geant4 on our CERN Debian server is compiled against CLHEP version 2.0.4.0, which is conflicting with the current CLHEP in AUR.

I have checked a few of the Novice examples, and everything seems to run fine. Obviously testers are most welcome! Examples are installed in /usr/share/doc/geant4/examples/

There are some wrapper scripts to get the Geant4 environment variables set. In order to see your environment variables, use "g4run g4env". In order to compile your code to be ran with geant4, use "g4make". In order to run your binary, use "g4run" in front (though this one does not seem to be necessary at least for the novice examples I've tried).

If you find dependencies I forgot or other errors, please report!