| 1 |
> On Tue, 12 Oct 2010 21:28:55 +1300 |
| 2 |
> |
| 3 |
> François Bissey <f.r.bissey@×××××××××.nz> wrote: |
| 4 |
> > Yes you have the typical configuration that was leading to trouble and |
| 5 |
> > that we have dealt with. If you are using the latest linbox ebuild from |
| 6 |
> > the overlay and it still doesn't work, send us the output of |
| 7 |
> > ldd /usr/lib64/liblinboxsage.so |
| 8 |
> |
| 9 |
> I'm using the |
| 10 |
> '/var/lib/layman/sage-on-gentoo/sci-libs/linbox/linbox-1.1.6-r3.ebuild' |
| 11 |
> from yesterday-synced layman 'sage-on-gentoo' overlay. |
| 12 |
> |
| 13 |
> ldd /usr/lib64/liblinboxsage.so gives the following: |
| 14 |
> |
| 15 |
> linux-vdso.so.1 => (0x00007fffd2f75000) |
| 16 |
> libntl-5.5.2.so => /usr/lib/libntl-5.5.2.so (0x00007ff6a00a6000) |
| 17 |
> liblinbox.so.0 => /usr/lib/liblinbox.so.0 (0x00007ff69fe95000) |
| 18 |
> libblas.so.0 => /usr/lib/libblas.so.0 (0x00007ff69fc72000) |
| 19 |
> libgmp.so.10 => /usr/lib/libgmp.so.10 (0x00007ff69fa03000) |
| 20 |
> libgivaro.so.0 => /usr/lib/libgivaro.so.0 (0x00007ff69f7b2000) |
| 21 |
> libstdc++.so.6 => /usr/lib/gcc/x86_64-pc-linux-gnu/4.4.4/libstdc |
| 22 |
> ++.so.6 (0x00007ff69f49e000) libm.so.6 => /lib/libm.so.6 |
| 23 |
> (0x00007ff69f21d000) libc.so.6 => /lib/libc.so.6 (0x00007ff69eeb7000) |
| 24 |
> libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x00007ff69ec9f000) |
| 25 |
> libgmp.so.3 => /usr/lib/libgmp.so.3 (0x00007ff69ea48000) |
| 26 |
> libgf2x.so.1 => /usr/lib/libgf2x.so.1 (0x00007ff69e83a000) |
| 27 |
> libgfortran.so.3 |
| 28 |
> => /usr/lib/gcc/x86_64-pc-linux-gnu/4.4.4/libgfortran.so.3 |
| 29 |
> (0x00007ff69e54b000) libatlas.so.0 => /usr/lib/libatlas.so.0 |
| 30 |
> (0x00007ff69df07000) /lib64/ld-linux-x86-64.so.2 (0x00007ff6a0889000) |
| 31 |
> |
| 32 |
> > And also how recently you have synced the main tree, has the last fix |
| 33 |
> > required an eclass bug to be fixed. |
| 34 |
> |
| 35 |
> I have synced the main tree yesterday, just before updating Sage and |
| 36 |
> the packages it depends on. |
| 37 |
> I'm not sure I understand you right about 'has the last fix required |
| 38 |
> an eclass bug to be fixed'. |
| 39 |
> |
| 40 |
I meant an "as an eclass bug had to be fixed", the fix developed by me |
| 41 |
Martin von Gagern wouldn't take because the autotool eclass missed some |
| 42 |
double quotes see bug #339227. |
| 43 |
|
| 44 |
Basically as far as we understand that's a problem when using gsl for your |
| 45 |
cblas. There are two workarounds while we work out why it doesn't work for |
| 46 |
you: |
| 47 |
1) switch to atlas cblas |
| 48 |
or |
| 49 |
2) compile gsl with the "static-libs" useflag. |
| 50 |
Then recompile linbox. |
| 51 |
|
| 52 |
Before you do that, it would be helpful if you sent us a build log of |
| 53 |
lnbox-1.1.6-r3 and if you know how the configuration log of linbox. |
| 54 |
|
| 55 |
Cheers, |
| 56 |
Francois |
Archives