Re: diff from 2022-08-24 12:41 to 2022-08-27 02:29

["Dr. Thomas Orgis" <thomas.orgis%uni-hamburg.de@localhost>]

Am Sun, 28 Aug 2022 08:52:41 +0000
schrieb Makoto Fujiwara <makoto%ki.nu@localhost>:

> reporting URL:
>     http://www.ki.nu/pkgsrc/reports/current/NetBSD-9.0/20220827.0229/meta/report.html

> --- 1. Previous success but  error at this time ---- (6)
>  http://www.ki.nu/pkgsrc/reports/current/NetBSD-9.0/20220827.0229/mpich-3.4.3?C=M;O=D

I'd like some input on that one. It fails to get a correct Fortran
setup. The change causing this is that the f90 option is on by default
now, and hence USE_LANGUAGES+=fortran.


=> Bootstrap dependency digest>=20211023: found digest-20220214
=> Checksum BLAKE2s OK for mpich-3.4.3.tar.gz
=> Checksum SHA512 OK for mpich-3.4.3.tar.gz
===> Installing dependencies for mpich-3.4.3
==========================================================================
The supported build options for mpich are:

	debug f90

The currently selected options are:

	f90

You can select which build options to use by setting PKG_DEFAULT_OPTIONS
or the following variable.  Its current value is shown:

	PKG_OPTIONS.mpich (not defined)

==========================================================================
=> Tool dependency libtool-base>=2.4.2nb9: found libtool-base-2.4.7nb1
=> Tool dependency mktools-[0-9]*: found mktools-20220614
=> Tool dependency cwrappers>=20150314: found cwrappers-20220403
=> Tool dependency checkperms>=1.1: found checkperms-1.12
=> Build dependency libtool-fortran>=2.4.2nb9: found libtool-fortran-2.4.7nb1
=> Full dependency bash-[0-9]*: found bash-5.1.16
=> Full dependency perl>=5.0: found perl-5.36.0
=> Full dependency openpa>=1.0.1: found openpa-1.0.4
=> Full dependency gcc7>=7.5.0nb6: found gcc7-7.5.0nb6

[…]

sourcing /tmp/parallel/mpi-ch/work/mpich-3.4.3/src/pm/hydra/mpichprereq
checking whether the compiler defines __func__... yes
checking whether the compiler defines __FUNC__... no
checking whether the compiler sets __FUNCTION__... yes
checking whether C compiler accepts option -O2... yes
checking whether routines compiled with -O2 can be linked with ones compiled without -O2... yes
checking for shared library (esp. rpath) characteristics of CC... done (results in src/env/cc_shlib.conf)
checking whether Fortran 77 compiler accepts option -O2... no
checking how to get verbose linking output from gfortran... configure: WARNING: compilation failed

checking for Fortran 77 libraries of gfortran... 
checking whether gfortran accepts the FLIBS found by autoconf... no
checking for valid entries in FLIBS... 
checking whether gcc links with FLIBS found by autoconf... yes
checking whether Fortran 77 and C objects are compatible... no
checking for file... file
configure: error: ****  Incompatible Fortran and C Object File Types!  ****
F77 Object File Type produced by "gfortran -O -fallow-argument-mismatch " is : : cannot open `' (No such file or directory).
 C  Object File Type produced by "gcc -O2 -fPIC -D_FORTIFY_SOURCE=2 -I/usr/pkg/include -I/usr/include   -O2" is : : ELF 64-bit LSB relocatable, x86-64, version 1 (SYSV), not stripped.
*** Error code 1

The build is not happy with Fortran. I also spot a test == syntax error
there … not yet able to locate its source. Also not with set -x.

Can anyone on NetBSD 9 confirm those issues (and ideally, debug)?


Alrighty then,

Thomas

-- 
Dr. Thomas Orgis
HPC @ Universität Hamburg