That installation of attr seems to come from conda, I'd probably drop out of 
the conda environment to do the installation...if that is possible.

This issue looks similar and that's what they did: 
https://stackoverflow.com/questions/74078492/relocation-error-after-export-ld-library-path

On 11 August 2023 17:33:17 CEST, Steen Lysgaard <[email protected]> wrote:
>Hi Alan,
>
>adding -ffree-line-length-512 as you suggested makes the build work! Many 
>thanks.
>
>One additional question though:
>
>On a different server we have the attr package installed by the package 
>manager:
>$ yum list 'attr*'
>Installed Packages
>attr.x86_64                                                          
>2.4.46-13.el7                                                    @anaconda
>
>BigDFT has the attr v. 2.5.1 package as a dependency. Thus when building 
>BigDFT it uses attr v. 2.5.1 and this makes it not recognize any commands e.g. 
>tar. The error message is:
>
>  List of modules to be treated: I: unknown keys defined in configuration 
> file: addpath
> tar: relocation error: /lib64/libacl.so.1: symbol getxattr, version ATTR_1.0 
> not defined in file libattr.so.1 with link time reference
> jhbuild build: failed to unpack 
> /home/scratch3/stly/modules/skylake-modules-test/build/BigDFT/1.9.1/foss-2021b/bigdft-suite-1.9.1/yaml-0.1.6.tar.gz
>
>Commenting out the attr dependency in the easyconfig file leads to a undefined 
>reference errors:
>
>D_potrf_double: error: undefined reference to 'dpotrf_'
> /home/scratch3/stly/modules/skylake-modules-test/build/BigDFT/1.9.1/foss-2021b/bigdft-suite-1.9.1/bigdft_builddir/install/lib/libfutile-1.a(linalg.o):linalg.f90:function
>  __wrapper_linalg_MO D_potrf_simple: error: undefined reference to 'spotrf_'
> collect2: error: ld returned 1 exit status
> make[3]: *** [chess_toolbox] Error 1
>
>Any advice on this issue?
>
>Best regards,
>Steen
>
>________________________________
>From: Alan O'Cais
>Sent: Thursday, August 10, 2023 17:09
>To: Steen Lysgaard
>Cc: easybuild
>Subject: Re: [easybuild] BigDFT build issue: BigDFT-1.9.1-foss-2021b.eb
>
>
>If the below works, can you let us know and I can PR an update the easyconfig.
>
>
>On 10-Aug-23 4:09 PM, Alan O'Cais wrote:
>
>Dear Steen,
>
>
>It _looks_ like this could be resolved with giving the option 
>`-ffree-line-length-512` to GFortran. I suspect that __FILE__ is a very long 
>string and it is pushing the line length over the limit (which is why others 
>may not have seen it). The second error is (I think) a consequence of the 
>truncation.
>
>
>Edit the relevant easyconfig line to be
>
>local_extra_flags = '-fallow-argument-mismatch -ffree-line-length-512'
>
>
>Alan
>
>
>On 10-Aug-23 3:41 PM, Steen Lysgaard wrote:
>Hello,
>
>I am trying to install BigDFT with the easyconfig file 
>BigDFT-1.9.1-foss-2021b.eb (eb BigDFT-1.9.1-foss-2021b.eb -r 
>--detect-loaded-modules=unload)
>
>However, the installation fails, with what seems to be an error with the 
>compilation of a Fortran file of libAbinit which is included in the BigDFT 
>package.
>
>The following error message is present in the log file:
>
>mpifort -DHAVE_LIBPAW_BIGDFT -DHAVE_DFT_LIBXC -DHAVE_MPI -DHAVE_MPI1 -I. [... 
>Removed Include instructions ...] -O2 -ftree-vectorize -march=native 
>-fno-math-errno -fallow-argument-mismatch -fallow-argument-mismatch
>  -std=legacy -c -o m_libpaw_tools.o `test -f 'libpaw/m_libpaw_tools.F90' || 
> echo 
> '/home/scratch3/stly/modules/icelake-modules-test/build/BigDFT/1.9.1/foss-2021b/bigd
>  ft-suite-1.9.1/libABINIT/src/'`libpaw/m_libpaw_tools.F90
>  
> /home/scratch3/stly/modules/icelake-modules-test/build/BigDFT/1.9.1/foss-2021b/bigdft-suite-1.9.1/libABINIT/src/libpaw/m_libpaw_tools.F90:408:132:
>
>    408 | &  __FILE__,__LINE__)
>        |                                                                      
>                                                               1
>  Error: Line truncated at (1) [-Werror=line-truncation]
>  
> /home/scratch3/stly/modules/icelake-modules-test/build/BigDFT/1.9.1/foss-2021b/bigdft-suite-1.9.1/libABINIT/src/libpaw/m_libpaw_tools.F90:408:5:
>
>    408 | &  __FILE__,__LINE__)
>        |     1
>  Error: Unterminated character constant beginning at (1)
>  f951: some warnings being treated as errors
>  make[1]: *** [Makefile:1908: m_libpaw_tools.o] Error 1
>  make[1]: *** Waiting for unfinished jobs....
>  make[1]: Leaving directory 
> '/home/scratch3/stly/modules/icelake-modules-test/build/BigDFT/1.9.1/foss-2021b/bigdft-suite-1.9.1/bigdft_builddir/libABINIT/src'
>  make: *** [Makefile:367: all-recursive] Error 1
>
>  *** Building libABINIT *** [8/11]^O
>  make -j 113
>  *** Error during phase build of libABINIT: ########## Error running make -j 
> 113  *** [8/11]^O
>
>
>I can get around the first error with -Wno-error=line-truncation, but not the 
>second one, and I don't really understand it.
>
>I have tried the newer 1.9.4 version of BigDFT, where the problem seems to 
>persist. It also happens with foss-2022a
>
>Some info:
>$ uname -r
>4.18.0-477.15.1.el8_8.x86_64
>
>$ lsb_release -a
>LSB Version:    :core-4.1-amd64:core-4.1-noarch
>Distributor ID: AlmaLinux
>Description:    AlmaLinux release 8.8 (Sapphire Caracal)
>Release:        8.8
>Codename:       SapphireCaracal
>
>If anybody can help it would be much appreciated.
>
>Best regards,
>Steen
>
>
>
>
>
>[http://www.dtu.dk/-/media/DTU_Generelt/Andet/mail-signature-logo.png]
>Steen Lysgaard
>Scientific Software Developer
>DTU Energy
>
>Mob. +45 51 94 48 86
>[email protected]<mailto:[email protected]>
>Anker Engelunds Vej
>Building 301
>2800 Kgs. Lyngby
>www.dtu.dk/english<https://www.dtu.dk/english>

Reply via email to