Description of problem: tests failed during build (make check) of findutils in ppc64le arch Version-Release number of selected component (if applicable): findutils-4.5.12-2.fc21 How reproducible: rpmbuild --target ppc64le Actual results: ../build-aux/test-driver: line 107: 25009 Aborted "$@" > $log_file 2>&1 FAIL: test-isinf PASS: test-isnand-nolibm PASS: test-isnanf-nolibm ../build-aux/test-driver: line 107: 25027 Aborted "$@" > $log_file 2>&1 FAIL: test-isnanl-nolibm Additional info: floating-point test cases do not work correctly with the PowerPC long double ("double double")format. These tests "accidentally" succeeded anyway in the big-endian variant, but are now actually failing in little-endian mode. A patch to fix the tests for ppc64le has been submitted to upstream to the bug-gnulib mailing list. http://lists.gnu.org/archive/html/bug-gnulib/2013-12/msg00107.html note that this bug affects packages like gnulib, including coreutils, findutils, grep, and libunistring.
Are there any concerns that prevent the patch from being applied upstream?
This patch have been written originally by Ulrich Weigand (see mail archive url I put in description) and it concerns sevreal packages. Finally I didn't found what have been done in upstream for findutils. In my point of view, it fixes failing tests and I didn't see side effects but I am not the best person to validate it in upstream.
It needs to be accepted by _gnulib_ upstream. Then findutils will automatically get the fix on a next import of gnulib...
the gnulib fix is applied in upstream as http://git.savannah.gnu.org/gitweb/?p=gnulib.git;a=commitdiff;h=e962c07c0 Kamil, please apply the fix also to findutils package, it is currently blocking the merge of our bootstrap ppc64 little endian koji with the existing ppc koji.
fixed in indutils-4.5.12-4.fc21 Thank you for getting the patch upstream!
Already fixed in Fedora 21, closing...