one test segfaults on 32-bit s390

from build.log
...
test 1
    Start 1: test_main
1: Test command: /builddir/build/BUILD/librepo/tests/run_tests.sh
1: Test timeout computed to be: 9.99988e+06
1: Tests using directory: /tmp/librepog8fbJH
1: Running suite(s): cheksum
1:  gpg
1:  handle
1:  internal_mirrorlist
1:  metalink
1:  mirrorlist
1:  package_downloader
1:  repoconf
1: 
1: (process:29021): GLib-WARNING **: GError set over the top of a previous GError or uninitialized memory.
1: This indicates a bug in someone's code. You must ensure an error is NULL before it's set.
1: The overwriting error message was: Value of option 20 is not set
1: 
1: (process:29023): GLib-WARNING **: GError set over the top of a previous GError or uninitialized memory.
1: This indicates a bug in someone's code. You must ensure an error is NULL before it's set.
1: The overwriting error message was: Value of option 20 is not set
1:  repomd
1:  url_substitution
1:  util
1:  version
1: 95%: Checks: 43, Failures: 0, Errors: 2
1: /builddir/build/BUILD/librepo/tests/test_repoconf.c:56:E:Main:test_parse_repoconf_minimal:0: (after this point) Received signal 11 (Segmentation fault)
1: /builddir/build/BUILD/librepo/tests/test_repoconf.c:56:E:Main:test_write_repoconf:0: (after this point) Received signal 11 (Segmentation fault)
1/2 Test #1: test_main ........................***Failed    1.14 sec

Version-Release number of selected component (if applicable):
librepo-1.7.14-1.fc23 and newer


The test passes on all other arches (s390x, ppc64*, aarch64), so it might be a 32-bit big endian issue, or related to size_t being defined as unsigned long as opposed to the common unsigned int.