| Summary: | failing test case: lib.libarchive.functional_test.test_write_filter_zstd | ||
|---|---|---|---|
| Product: | Base System | Reporter: | Li-Wen Hsu <lwhsu> |
| Component: | tests | Assignee: | Baptiste Daroussin <bapt> |
| Status: | Open --- | ||
| Severity: | Affects Only Me | CC: | arichardson, bapt, emaste, mm |
| Priority: | --- | ||
| Version: | CURRENT | ||
| Hardware: | Any | ||
| OS: | Any | ||
|
Description
Li-Wen Hsu
2019-09-19 12:11:11 UTC
A commit references this bug: Author: lwhsu Date: Thu Sep 19 13:23:25 UTC 2019 New revision: 352515 URL: https://svnweb.freebsd.org/changeset/base/352515 Log: Temporarily add test_write_filter_zstd BROKEN_TESTS as it always fails in CI There is no trivial way to mark single libarchive test skip currently so just add it to BROKEN_TESTS for now. PR: 240683 Sponsored by: The FreeBSD Foundation Changes: head/lib/libarchive/tests/Makefile What is weird is yes I can reproduce this bug but only under bhyve, not on regular amd64 machines. I am adding mm@ to see if he has more insight I can't reproduce this either natively or on CHERI RISC-V. Maybe we should try removing it from BROKEN_TESTS? (In reply to Alex Richardson from comment #3) I haven't checked this for a while, maybe it was fixed silently. Please check it again in bhyve and feel free to enable it. I think we need a "try broken" CI job to monitor if the skipped/disabled tests get fixed. (In reply to Li-Wen Hsu from comment #4) > I think we need a "try broken" CI job to monitor if the skipped/disabled tests > get fixed. Ideally the failing tests would just be XFAIL and not fail the whole test / generate email, but absent that having a regular job that tries the broken/excluded tests indeed sounds like a good idea. We could check the test case name in functional_test.sh? |