@v8areu: I'm sorry but I'm unable to reproduce your issue. Do you have any special configuration linked to gfortran?
Search Criteria
Package Details: silo 4.10.2-8
Git Clone URL: | https://aur.archlinux.org/silo.git (read-only) |
---|---|
Package Base: | silo |
Description: | A Mesh and Field I/O Library and Scientific Database |
Upstream URL: | https://wci.llnl.gov/simulation/computer-codes/silo |
Licenses: | |
Submitter: | jedbrown |
Maintainer: | eolianoe |
Last Packager: | eolianoe |
Votes: | 5 |
Popularity: | 0.000000 |
First Submitted: | 2009-02-13 15:59 |
Last Updated: | 2018-01-22 19:11 |
Dependencies (4)
Required by (1)
Sources (4)
Latest Comments
pychuang commented on 2018-04-23 17:45
eolianoe commented on 2017-03-05 13:19
v8areu commented on 2017-03-05 03:36
Hello, I got an error while building the package from 'makepkg -sri' command. Here are the following build and error message:
https://gist.github.com/v8areu/824ca1164aefc2d989ed4e1a8572c5bb
Thank you.
I have both the latest gcc and gcc-fortran from official package. I don't know why I got an error "checking whether we are using the GNU Fortran compiler... no"
eolianoe commented on 2016-09-22 14:53
@muon, @MartinDiehl: I was able to reproduce the bug but randomly... As proposed by @fallenrgh, downgrading to c++98 seems to fix the problem, but I do not know why...
Let me know if you still have some errors.
muon commented on 2016-09-20 16:03
@eolianoe: Thanks for the reply. I have both hdf5 and hdf5_18 installed from the official repos. I just built with a commented out check function, and it built fine.
eolianoe commented on 2016-09-20 11:51
@muon, @MartinDiehl: Which package are you using for hdf5? In order to build the package you can disable the check function and I'll try to investigate more.
muon commented on 2016-09-19 19:59
I just got the same result as @MartinDiehl on x86_64. (silo-4.10.2-5). Output from `makepkg` command:
https://gist.github.com/raybsmith/1e4e6a05fb670de385e07b35eb6be50a
Let me know if any other info would help.
eolianoe commented on 2016-09-16 16:50
@MartinDiehl: Are you sure, that you have the latest version of the PKGBUILD? Because it seems to be linked to bug reported by smradlev a few months ago and I'm able to build it in a clean i686 chroot.
MartinDiehl commented on 2016-09-16 10:37
For me, the build fails (on i686). It seems that a shell script is processed by the preprocessor
In file included from /usr/include/c++/6.2.1/tuple:39:0,
from /usr/include/c++/6.2.1/bits/stl_map.h:63,
from /usr/include/c++/6.2.1/map:61,
from rocket.cxx:66:
./array:1:2: error: invalid preprocessing directive #!
#! /bin/sh
eolianoe commented on 2016-08-24 09:09
@fallenrgh: building fine here for now, I will add the `-std=c++98` flags if there is more trouble
[Update] OK, I found the problem is on my side: a previous admin installed
f77
on my machine. And nowadays not many people havef77
in their system, so this shouldn't happen to most users. But I'll keep the original message just in case.[Original message] Not sure why, but I had to set F77=gfortran in order to build some Fortran 77 tests. Without using
gfortran
, it will try to usef77
as Fortran 77 compiler, and some compilations failed.For example, the following compilation was executed when doing
make check
in thecheck()
section of PKGBUILD:And it produced the following error:
After using
gfortran
as the Fortran 77 compiler, everything worked fine.