Summary: | dev-python/pandas-0.23.4 - ImportError: libgfortran.so.4: cannot open shared object file | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Juergen Rose <rose> |
Component: | Current packages | Assignee: | Gentoo Science Related Packages <sci> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | frp.bissey, gentoo, jstein |
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
See Also: | https://bugs.gentoo.org/show_bug.cgi?id=646090 | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
build.log
build.log for sage |
Description
Juergen Rose
2018-08-30 14:31:35 UTC
gcc-8 is your culprit. There was a bump from libgfortran.so.3 to libgfortran.so.4 in gcc-7, we now have a bump to libgfortran.so.5 in gcc-8. All the bugs we had opened before because blas/lapack where linked to libgfortran.so.3 and gcc-7 was trying to link libgfortran.so.4 will appear once more with an increment. Try revdep-rebuild --library 'libgfortran.so.4' -- --exclude gcc before trying to merge pandas again. (In reply to François Bissey from comment #1) > gcc-8 is your culprit. There was a bump from libgfortran.so.3 to > libgfortran.so.4 in gcc-7, we now have a bump to libgfortran.so.5 in gcc-8. > > All the bugs we had opened before because blas/lapack where linked to > libgfortran.so.3 and gcc-7 was trying to link libgfortran.so.4 will appear > once more with an increment. > > Try > revdep-rebuild --library 'libgfortran.so.4' -- --exclude gcc > > before trying to merge pandas again. 'revdep-rebuild --library 'libgfortran.so.4' -- --exclude gcc' seems not to be enough to solve the problem. I have here the next similar problem, 'emerge sage' fails, but a subsequent revdep-rebuild does not detect ImportError: libgfortran.so.4: cannot open shared object file: No such file or directory ************************************************************************ Error building the Sage library ... >>> Failed to emerge sci-mathematics/sage-8.3-r1, Log file: >>> '/var/log/portage/sci-mathematics:sage-8.3-r1:20180902-163052.log' root@lynxold:/usr/local/portage/sci-geosciences(38)# revdep-rebuild --library 'libgfortran.so.4' -- --exclude gcc * This is the new python coded version * Please report any bugs found using it. * The original revdep-rebuild script is installed as revdep-rebuild.sh * Please file bugs at: https://bugs.gentoo.org/ * Collecting system binaries and libraries * Checking dynamic linking consistency Your system is consistent root@lynxold:/usr/local/portage/sci-geosciences(39)# If you give me a bit more of the sage build log I may be able to figure out what's been missed. (In reply to François Bissey from comment #3) > If you give me a bit more of the sage build log I may be able to figure out > what's been missed. I assume that atlas is responsible for this issue: root@lynxold:/usr/local/portage/sci-geosciences(40)# genlop -t atlas | tail Mon Jan 1 15:26:46 2018 >>> sci-libs/atlas-3.11.39 merge time: 2 hours, 10 minutes and 44 seconds. Wed Jan 3 06:00:13 2018 >>> sci-libs/atlas-3.11.39 merge time: 2 hours, 39 minutes and 32 seconds. Sun Aug 12 15:27:11 2018 >>> sci-libs/atlas-3.11.39 merge time: 3 hours, 8 minutes and 43 seconds. root@lynxold:/usr/local/portage/sci-geosciences(41)# genlop -t gcc | tail Sun May 20 08:44:23 2018 >>> sys-devel/gcc-7.3.0-r3 merge time: 57 minutes and 29 seconds. Mon Jun 25 11:52:29 2018 >>> sys-devel/gcc-7.3.0-r3 merge time: 59 minutes and 16 seconds. Sun Aug 19 21:21:38 2018 >>> sys-devel/gcc-8.2.0-r2 merge time: 1 hour, 31 minutes and 12 seconds. Created attachment 545940 [details]
build.log for sage
Right, that would fit since f77blas and lapack from atlas link to libgfortran. And the log you give basically say that numpy tries to open libgfortran.so.4 and the only way that should happen is through lapack. If revdep-rebuild doesn't work you'll have to do is 1) rebuild you blas implementations. 2) rebuild the matching lapack implementation (if not covered by (1)). 3) rebuild all the dependencies has given by `equery d blas lapack` I think this bug has been fixed by bug 646090. |