Problem compiling REST with Garfield

RESTFramework
1

REST version : v2.2.13
REST commit : 72e9d452

I am having troubles to make REST to recognise the Garfield installation. These are the steps I followed up to now, referencing the installation tutorials:

  1. I installed Garfield with the ./installGarfield.sh script. I also used the other installation scripts to be sure all the required software was installed. I checked that the required paths are in my .bashrc.
  2. I compiled REST as usual, with the following options for the cmake: -DREST_G4=ON and -DREST_GARFIELD=ON

But now, when I try to run the REST_DrawCombinedGasCurves macro (for example) that requires garfield to obtain the parameters of the gases, I get the following messages:

This REST is not compiled with garfield, Do not use Longitudinal Diffusion from TRestGas!
Please define the Longitudinal Diffusion in each process!

Am I doing something wrong?
Thank you very much!

2

Hi Elisa,

when you execute rest-config --flags what is the output you are getting?

Make sure you are using the version with Garfield installed.

rest-config --welcome

should give you that hint.

When you execute make install it is installing it at that location?

Just as additional comment …

The REST_DrawCombinedGasCurves macro is quite old, and might be obsolete. Only documented macros should be trusted.

For the moment, not many of them are documented …

Use: restManager ListMacros command to get a list of macros.

3

Additionally I believe if you run once cmake with Garfield disabled, and then try to enable it next time using cmake again, then it will not change in the final installation.

I believe this happened to me right now.

This a small bug, it should be fixed. But in the meantime clearing your build directory should do the job.

Be careful here :slight_smile:

cd build
rm -r *
4

Hi Javier,

Thank you for the tips. I am sorry this is going to be all over the place but I am attempting two installations of REST: on one hand, by my own at my own home, and on the other hand, the administrator of the server is attempting another where I don’t have admin permits. So:

  • The administrator achieved to fix the problem I mentioned before but I don’t know exactly what he did. With the installation he made, I can run REST correctly (not entirely true, but matter for another topic). However, when I type restManager ListMacros, I get the following

    RandomEngineRoot:
    Generator type: TRandom3
    Seed: 1569142573
    Heed:
    Database path: /home/common/software/garfieldpp/Heed/heed++/database
    [== Initializing ListMacros ==]
    REST ERROR. Task : ListMacros not found!!

    Also, how can I add new macros to be started with restRoot?

  • In my own installation, when I attempt make -j4, I get this error:

    [ 20%] Building CXX object source/metadata/CMakeFiles/RestMetadata.dir/__/__/rootdict/CINT_TRestGeometry.cxx.o
In file included from /home/eruizchl/REST_v2/source/metadata/general/inc/TRestGas.h:50:0,
             from /home/eruizchl/REST_v2/source/metadata/general/inc/TRestGeometry.h:27,
             from /home/eruizchl/REST_v2/source/metadata/general/src/TRestGeometry.cxx:19:
/home/common/software/garfieldpp/Include/TrackHeed.hh:9:38: fatal error: heed++/code/HeedParticle.h: No such file or directory
 #include "heed++/code/HeedParticle.h"
                                  ^
compilation terminated.
In file included from /home/eruizchl/REST_v2/source/metadata/general/inc/TRestGas.h:50:0,
             from /home/eruizchl/REST_v2/build/rootdict/CINT_TRestGas.cxx:40:
/home/common/software/garfieldpp/Include/TrackHeed.hh:9:38: fatal error: heed++/code/HeedParticle.h: No such file or directory
 #include "heed++/code/HeedParticle.h"
                                  ^
compilation terminated.
In file included from /home/eruizchl/REST_v2/source/metadata/general/inc/TRestGas.h:50:0,
             from /home/eruizchl/REST_v2/source/metadata/general/src/TRestGas.cxx:216:
/home/common/software/garfieldpp/Include/TrackHeed.hh:9:38: fatal error: heed++/code/HeedParticle.h: No such file or directory
 #include "heed++/code/HeedParticle.h"
                                  ^
compilation terminated.
In file included from /home/eruizchl/REST_v2/source/metadata/general/inc/TRestGas.h:50:0,
             from /home/eruizchl/REST_v2/source/metadata/general/inc/TRestGeometry.h:27,
             from /home/eruizchl/REST_v2/build/rootdict/CINT_TRestGeometry.cxx:40:
/home/common/software/garfieldpp/Include/TrackHeed.hh:9:38: fatal error: heed++/code/HeedParticle.h: No such file or directory
 #include "heed++/code/HeedParticle.h"
                                  ^
compilation terminated.

    And I don’t understand why because I can perfectly find the files in the folders they are supposed to be.

Thank you very much!!

5

Hi Elisa, I would open a new post for the questions you are doing since they do not seem related to Garfield anymore…

Anyway

Basically, if they are hosted at /macros/ directory they will be installed when doing make install, and then available when loading restRoot.

It exists the following file?
$REST_PATH/macros/REST_ListMacros.C

Please, make sure you do not have local changes on your cmake files, and re-build. Please, share the output of cmake, and make.

git reset --hard
cd build
rm -r *
cmake -DREST_GARFIELD=ON ../
make

what is the output of those commands?

6

Sorry!! since the problem was with garfield and heed++ I though it was…

No it does not. In the $REST_PATH/macros/ folder I only find CMakeLists.txt, and several folders for each type of macros.

I did this, and I got the very same result.

I asked to the admin of the server, and he told me that he had to make several changes to the cmakelists and installation scripts to make REST work. I created a new branch called “mainz” so i can push the version he modified and see the changes. Is this ok? I’ll wait until I know your thougths, its been a while and I don’t know the work flow right now. We used to have our own branches before, but I see it is not the case anymore.

7

No, macros have nothing to be with Garfield …

Make sure you are at the right installation. What is the result of make install?

When I do make install it appears how it installs the macro files, for example

make install output

– Up-to-date: /programas/REST/v2.2.14/./macros/geant4/REST_ListIsotopes.C

– Up-to-date: /programas/REST/v2.2.14/./macros/geant4/REST_MeanTrackLeng.C

– Up-to-date: /programas/REST/v2.2.14/./macros/geant4/REST_FindIsotopes.C

– Up-to-date: /programas/REST/v2.2.14/./macros/geant4/REST_ReadNEvents_inROI.C

– Up-to-date: /programas/REST/v2.2.14/./macros/geant4/REST_ReadNEvents.C

– Up-to-date: /programas/REST/v2.2.14/./macros/CMakeLists.txt

– Up-to-date: /programas/REST/v2.2.14/./macros/REST_ListMacros.C

– Up-to-date: /programas/REST/v2.2.14/./macros/tools

– Up-to-date: /programas/REST/v2.2.14/./macros/tools/REST_ValidateGeometry.C

– Up-to-date: /programas/REST/v2.2.14/./macros/tools/REST_MergeFiles.C

– Up-to-date: /programas/REST/v2.2.14/./macros/tools/REST_DrawCombinedGasCurves.C

– Up-to-date: /programas/REST/v2.2.14/./macros/tools/REST_ProduceReponseSignal.C

– Up-to-date: /programas/REST/v2.2.14/./macros/tools/REST_CheckReadout.C

– Up-to-date: /programas/REST/v2.2.14/./macros/tools/REST_CheckRunFileList.C

– Up-to-date: /programas/REST/v2.2.14/./macros/tools/REST_GenerateGasFile.C

– Up-to-date: /programas/REST/v2.2.14/./macros/tools/REST_DrawResponseSignal.C

– Up-to-date: /programas/REST/v2.2.14/scripts/restv2_1ToXml.py

– Up-to-date: /programas/REST/v2.2.14/./data

– Installing: /programas/REST/v2.2.14/./data/dataURL

– Up-to-date: /programas/REST/v2.2.14/./data/definitions

– Up-to-date: /programas/REST/v2.2.14/./data/definitions/processes.rml

– Up-to-date: /programas/REST/v2.2.14/./data/definitions/materials.xml

– Up-to-date: /programas/REST/v2.2.14/./data/definitions/gases.rml

– Up-to-date: /programas/REST/v2.2.14/./data/definitions/readouts.rml

8

The operation with branches is well described at the contribution guide.

It is ok, I will test those changes on our server, and then merge with the development branch v2.2.14_dev. We should never diverge branches, or produce results with local modifications, so that we are able to track/reproduce results produced with REST using git and the REST versioning system.

We should get to a merged version that can be installed on any system. Perhaps you could invite the administrator at Mainz to join this discussion.

9

This is the commit of Garfield that works for me c84002f. I know past versions changed the file structure of include files, and perhaps also the naming. That was usually the problem.

Thats the Garfield commit that is installed with the ./instalGarfield.sh script.

echo checkout the code...

#svn co http://svn.cern.ch/guest/garfield/tags/v1r0 $GARFIELD_HOME

git clone https://gitlab.cern.ch/garfield/garfieldpp.git $GARFIELD_HOME

cd $GARFIELD_HOME

git checkout c84002f1
10

I crosschecked the modifications that the admin made, and he only had to modify a little the CMakeLists.txt where Garfield is searched. With these modifications, the compilation was compleated without errors.

Since they are small changes, I deleted the mainz branch, checked into the 2.2.14_dev and did a commit following the contribution guide.

11

Thanks @elisa, it seems to me that the commit fixes a small bug not appearing in all conditions, or Garfield installations. Looks harmless for other systems.