h_raw depends on -fsanitize=bounds in order to function. The removal of that option in r274978 broke the test helper, which in turn broke the testcase. This wasn't visible until clang 3.5.0 hit the tree.
A commit references this bug: Author: ngie Date: Fri Jan 2 05:40:04 UTC 2015 New revision: 276527 URL: https://svnweb.freebsd.org/changeset/base/276527 Log: Don't install h_raw if dealing with clang 3.5.0+ to unbreak the tests2 Jenkins job The h_raw application doesn't do proper bounds checking without the option being supplied via the build, which means that it doesn't throw signals and fail as expected PR: 196430 X-MFC with: r276479 Changes: head/contrib/netbsd-tests/lib/libc/ssp/t_ssp.sh head/lib/libc/tests/ssp/Makefile
To originators/assignees of this PR: A commit to the tree references this PR, however the PR is still in a non-closed state. Please review this PR and close as appropriate, or if closing the PR requires a merge to stable/10, please let re@ know as soon as possible. Thank you. Glen
Garrett, since we're not going to merge a newer version of clang to stable/10 soon, I don't think this test case needs a quick merge either? Maybe just close this bug for now?
(In reply to Dimitry Andric from comment #3) Sounds reasonable. I wasn't sure about that... Closing -- thanks!
Wait. Now that I re-read this, it was a placeholder to fix the issue on CURRENT (or wherever the commit was merged to). Reopening, but explicitly setting all of the mfc flags to -.
A commit references this bug: Author: ngie Date: Fri Oct 30 07:48:04 UTC 2015 New revision: 290183 URL: https://svnweb.freebsd.org/changeset/base/290183 Log: - Re-enable h_raw with clang 3.7.0+ - Fix the compiler check to allow the test to be compiled for gcc PR: 196430 MFC after: never Sponsored by: EMC / Isilon Storage Division Changes: head/lib/libc/tests/ssp/Makefile
A commit references this bug: Author: ngie Date: Thu Feb 9 08:25:31 UTC 2017 New revision: 313468 URL: https://svnweb.freebsd.org/changeset/base/313468 Log: MFC r276527,r277648: r276527: Don't install h_raw if dealing with clang 3.5.0+ to unbreak the tests2 Jenkins job The h_raw application doesn't do proper bounds checking without the option being supplied via the build, which means that it doesn't throw signals and fail as expected PR: 196430 r277648 (by jilles): Enable utimensat tests from NetBSD. As with other tests from c063, a required #include <sys/stat.h> was missing. Changes: _U projects/stable-10-backport-test-changes/ projects/stable-10-backport-test-changes/contrib/netbsd-tests/lib/libc/c063/t_utimensat.c projects/stable-10-backport-test-changes/contrib/netbsd-tests/lib/libc/ssp/t_ssp.sh projects/stable-10-backport-test-changes/lib/libc/tests/c063/Makefile projects/stable-10-backport-test-changes/lib/libc/tests/ssp/Makefile
A commit references this bug: Author: ngie Date: Thu Feb 9 08:39:08 UTC 2017 New revision: 313469 URL: https://svnweb.freebsd.org/changeset/base/313469 Log: MFC r276527,r277648,r279154,r279397,r283584: r276527: Don't install h_raw if dealing with clang 3.5.0+ to unbreak the tests2 Jenkins job The h_raw application doesn't do proper bounds checking without the option being supplied via the build, which means that it doesn't throw signals and fail as expected PR: 196430 r277648 (by jilles): Enable utimensat tests from NetBSD. As with other tests from c063, a required #include <sys/stat.h> was missing. r279154 (by jilles): nice(): Correct return value and [EPERM] error. PR: 189821 Obtained from: NetBSD Relnotes: yes r279397 (by jilles): nice(): Put back old return value, keeping [EPERM] error. Commit r279154 changed the API and ABI significantly, and {NZERO} is still wrong. Also, preserve errno on success instead of setting it to 0. PR: 189821 Relnotes: yes r283584 (by emaste): memmem(3): empty little string matches the beginning of the big string This function originated in glibc, and this matches their behaviour (and NetBSD, OpenBSD, and musl). An empty big string (arg "l") is handled by the existing l_len < s_len test. Changes: _U projects/stable-10-backport-test-changes/ projects/stable-10-backport-test-changes/contrib/netbsd-tests/lib/libc/gen/t_nice.c projects/stable-10-backport-test-changes/contrib/netbsd-tests/lib/libc/string/t_memmem.c projects/stable-10-backport-test-changes/lib/libc/gen/nice.3 projects/stable-10-backport-test-changes/lib/libc/gen/nice.c projects/stable-10-backport-test-changes/lib/libc/string/memmem.3 projects/stable-10-backport-test-changes/lib/libc/string/memmem.c
A commit references this bug: Author: ngie Date: Thu Feb 9 21:29:19 UTC 2017 New revision: 313482 URL: https://svnweb.freebsd.org/changeset/base/313482 Log: MFC r276527: Don't install h_raw if dealing with clang 3.5.0+ to unbreak the tests2 Jenkins job The h_raw application doesn't do proper bounds checking without the option being supplied via the build, which means that it doesn't throw signals and fail as expected PR: 196430 Changes: _U stable/10/ stable/10/contrib/netbsd-tests/lib/libc/ssp/t_ssp.sh stable/10/lib/libc/tests/ssp/Makefile