> Additional thoughts how to tackle these?
Okay, I got this working too now -- thanks again, it works and I can try to
get further with the research problem we look at!
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this
Here is the latest updates:
1. After setting
# Location of system crti.o, in case we build our own gcc
export
SAGE_CRTI_DIR="/usr/lib/gcc/x86_64-redhat-linux/4.4.7/../../../../lib64"
in src/bin/sage-env-config.in by hand, the compilation of 8.8.beta7 with
#27907 merged now works. Many thanks
Following the setting of SAGE_CRTI_DIR in
"build/pkgs/gcc/spkg-configure.m4" seems to work:
(sage-sh) stumpcl2@dlp762:sage$ echo $CC
gcc
(sage-sh) stumpcl2@dlp762:sage$ echo $CRTI
(sage-sh) stumpcl2@dlp762:sage$ CRTI=`$CC -print-file-name=crti.o
2>/dev/null || true`
(sage-sh)
No, it hasn't. I just did "make distclean" and then "./configure" which
recreated src/bin/sage-env-config, but it says
export SAGE_CRTI_DIR="@SAGE_CRTI_DIR@"
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and
The bad symlink there was probably generated when you didn't regenerate
"configure".
Could you check that this time SAGE_CRTI_DIR has been correctly
substituted in src/bin/sage-env-config (and local/bin/sage-env-config)?
On Tuesday, June 4, 2019 at 9:44:21 PM UTC-7, Matthias Koeppe wrote:
>
>
Something went really wrong there with the symbolic linking script that I
added in
https://git.sagemath.org/sage.git/commit/?id=73be6cf53f30b53e5a8a3dedf8f30c2c9960475a
On Tuesday, June 4, 2019 at 11:30:22 AM UTC-7, Christian Stump wrote:
>
> [stumpcl2@dlp762 ~]$ cd progs/sage/
>
[stumpcl2@dlp762 ~]$ cd progs/sage/
[stumpcl2@dlp762 sage]$ ls -l local/lib/gcc/*/7.2.0/*crt*
-rw-r--r-- 1 stumpcl2 15301 2896 4. Jun 16:22
local/lib/gcc/x86_64-pc-linux-gnu/7.2.0/crtbegin.o
-rw-r--r-- 1 stumpcl2 15301 3224 4. Jun 16:22
local/lib/gcc/x86_64-pc-linux-gnu/7.2.0/crtbeginS.o
What does
ls -l local/lib/gcc/*/7.2.0/*crt*
say
On Tuesday, June 4, 2019 at 11:12:01 AM UTC-7, Christian Stump wrote:
>
> here comes
>
>
> /tmp/stumpcl2/sage/local/var/tmp/sage/build/mpir-3.0.0-644faf502c56f97d9accd301965fc57d6ec70868.p0/src/config.log
>
--
You received this message because
here comes
/tmp/stumpcl2/sage/local/var/tmp/sage/build/mpir-3.0.0-644faf502c56f97d9accd301965fc57d6ec70868.p0/src/config.log
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send
On Tuesday, June 4, 2019 at 7:32:01 AM UTC-7, Christian Stump wrote:
>
> I finally got
>
> [gcc-7.2.0] Successfully installed gcc-7.2.0
>
> using Matthias' solution. Unfortunately, then the next build of mpir-3.0.0
> fails with "no compiler found" (with and without the --with-mp=mpir flag
>
It appears that it is not an issue of this particular old system as
Matthias pointed out. I would appreciate further help at
https://trac.sagemath.org/ticket/27907 to figure out how to provide the
correct version of
[gcc-7.2.0] /usr/lib/../lib/crti.o: could not read symbols: File in
wrong
I've encountered similar, but different, problems on a different Linux
system. Opened https://trac.sagemath.org/ticket/27907 for this.
On Tuesday, May 28, 2019 at 10:15:08 AM UTC-7, Dima Pasechnik wrote:
>
> On Tue, May 28, 2019 at 6:43 PM Christian Stump
> > wrote:
> >
> > > well, building
assembly errors - means either broken toolchain (old `as` picked up ?), or
a failure to detect the correct architecture...
I guess sysadmins of this system see such errors every day, given how messy
that setup is.
On Wed, 29 May 2019 08:08 Christian Stump,
wrote:
> Okay, thank you for checking
Okay, thank you for checking -- I think that machine needs some upgrades!
In case anyone else runs into this issue, here is a partial fix: I first
compiled 8.6 from scratch and then upgraded to 8.8.beta6 and this worked.
Unfortunately, my own code on top of 8.8.beta6 again does not build with
On Tue, May 28, 2019 at 6:43 PM Christian Stump
wrote:
>
> > well, building gcc 7 with a 7-years old complier (gcc version 4.4.7
> > 20120313 (Red Hat 4.4.7-23), as the log says)
> > probably not possible.
> > Do they have something newer installed?
>
> They did install for me both gcc-7.2.0
> well, building gcc 7 with a 7-years old complier (gcc version 4.4.7
> 20120313 (Red Hat 4.4.7-23), as the log says)
> probably not possible.
> Do they have something newer installed?
They did install for me both gcc-7.2.0 (didn't work) and gcc-7.3.0 (didn't
work either).
Let me emphasize
well, building gcc 7 with a 7-years old complier (gcc version 4.4.7
20120313 (Red Hat 4.4.7-23), as the log says)
probably not possible.
Do they have something newer installed?
On Tue, May 28, 2019 at 4:17 PM Christian Stump
wrote:
>
> > ./configure --with-mp=mpir
> > make
>
> This did not work
> ./configure --with-mp=mpir
> make
This did not work -- but if I do not provide any gcc, sage compiles its own
and this works for gcc-7.2.0 on sage-8.6 while it does not work for the
same gcc-7.2.0 on sage-8.7. You find the log for the failing gcc-7.2.0
attached.
--
You received this
On Tue, May 28, 2019 at 1:50 PM Christian Stump
wrote:
>
> Thanks Dima! -- here is the requested information:
>
> It's a fresh install, config.log is attached. In just this moment, the system
> admin told me that he was able to compile sage-8.6 using the standard gcc
> compiler by RHEL6 (this
Thanks Dima! -- here is the requested information:
It's a fresh install, config.log is attached. In just this moment, the
system admin told me that he was able to compile sage-8.6 using the
standard gcc compiler by RHEL6 (this is copy-and-paste, I don't know what
that means), while the error
I suspect you don't have libgmpxx installed. What is the output of
$ find /usr/ -name libgmpxx.so*
and of
$ find /usr/ -name libgmp.so*
On Tue, May 28, 2019 at 9:54 AM Dima Pasechnik wrote:
>
> could you attach Sage's config.log ?
> I suspect that the system installation of GMP is a bit
could you attach Sage's config.log ?
I suspect that the system installation of GMP is a bit broken or
incomplete, or perhaps there are more than one, and they clash. (and our
recent GMP checking code fails to pick it up)
is it a fresh install, or an update?
--
anyhow, you can force the use
22 matches
Mail list logo