Created attachment 174965 [details] gsl19test patch gsl19test does not work properly when /bin/sh is /bin/dash. There were 2 issues: 1) The output redirect operators seem to be bash specific. All output went to the terminal..... 2) When invoking valgrind the 'eval' needs to precede GSL_TEST_VERBOSE=1. Otherwise that environment variable is not seen by valgrind. Two tweaks: 1) Determine the number of available processors and use that in make -j when building gsl-1.9 2) Replace -v with -q in the valgrind invocation. The file out-V is already quite large (approx 600 MB). It would be enormous with -v.
Thanks, applied with one small tweak, use command -v instead of which to find nproc. which isn't POSIX and might not be available itself. commit 849dca0b4bef48e1f3640751b24d9e3e7f2d4aaf Author: Florian Krohm <flo2030@eich-krohm.de> Date: Fri Nov 1 00:31:27 2024 +0100 gsl19test does not work properly when /bin/sh is /bin/dash There were 2 issues: 1) The output redirect operators seem to be bash specific. All output went to the terminal..... 2) When invoking valgrind the 'eval' needs to precede GSL_TEST_VERBOSE=1. Otherwise that environment variable is not seen by valgrind. Two tweaks: 1) Determine the number of available processors and use that in make -j when building gsl-1.9 2) Replace -v with -q in the valgrind invocation. The file out-V is already quite large (approx 600 MB). It would be enormous with -v. https://bugs.kde.org/show_bug.cgi?id=494960 Do note that on "most" systems the gsltest (make auxchecks) fail with: +FAIL: random32-bsd, 10000 steps (626977143 observed vs 1663114331 expected) +FAIL: random64-bsd, 10000 steps (1262850638 observed vs 864469165 expected) +FAIL: random32-libc5, 10000 steps (827587933 observed vs 1967452027 expected) +FAIL: random64-libc5, 10000 steps (963916554 observed vs 2106639801 expected)