test-admin-deploy-var: Don't rely on OSTREE_FEATURES #3184
Conversation
|
Skipping CI for Draft Pull Request. |
|
(testing this next, will undraft when tested) |
smcv
force-pushed
the
issue3183
branch
2 times, most recently
from
24ea77e
to
f6f378f
Compare
|
This was more involved than I had expected! |
Signed-off-by: Simon McVittie <[email protected]>
This avoids false negatives from `ostree --version | grep -q ...` exiting with failure under `set -o pipefail` because `grep -q` can exit as soon as it sees the desired string, leaving `ostree --version` to be terminated by `SIGPIPE` next time it writes to stdout. Signed-off-by: Simon McVittie <[email protected]>
This is set during build-time testing, but unset during "as-installed" tests. Resolves: ostreedev#3183 Signed-off-by: Simon McVittie <[email protected]>
| # avoid ostree --version being killed with SIGPIPE and exiting with a | ||
| # nonzero status under `set -o pipefail`. |
There was a problem hiding this comment.
As I said, more involved than I expected. This PR brought to you by the dept. of Unix arcana.
There was a problem hiding this comment.
Even though it's eminently useful, I stopped using pipefail years ago after being burned by the same issue in tar. It might be nice to install a SIGPIPE handler in the version printing code since this is likely to affect other ostree consumers.
There was a problem hiding this comment.
One thing I've said many times with coworkers in live video chats, but rarely externally is that a nontrivial component of why I still (20+ years later) wake up and log into my computer for this job is that I just really enjoy collaborating with lots of smart people across the internet. Also there's the continual "surprise" factor in computers as in "what did someone figure out how to make them do this year" (and right now it's AI).
Anyways this little bit combines those two - I just want to say thanks @smcv as one of those smart people on the Internet for everything you do 🙏 - and this little bit of pipefail is one of those surprises 😄
Signed-off-by: Simon McVittie <[email protected]> Forwarded: ostreedev/ostree#3184 Gbp-Pq: Name tests-Generalize-has_gpgme-has_sign_ed25519-into-has_ostr.patch
This avoids false negatives from `ostree --version | grep -q ...` exiting with failure under `set -o pipefail` because `grep -q` can exit as soon as it sees the desired string, leaving `ostree --version` to be terminated by `SIGPIPE` next time it writes to stdout. Signed-off-by: Simon McVittie <[email protected]> Forwarded: ostreedev/ostree#3184 Gbp-Pq: Name tests-Use-skip_without_ostree_feature-to-detect-libarchiv.patch
This is set during build-time testing, but unset during "as-installed" tests. Bug: ostreedev/ostree#3183 Signed-off-by: Simon McVittie <[email protected]> Forwarded: ostreedev/ostree#3184 Gbp-Pq: Name test-admin-deploy-var-Don-t-rely-on-OSTREE_FEATURES.patch
tests: Generalize has_gpgme, has_sign_ed25519 into has_ostree_feature
tests: Use skip_without_ostree_feature to detect libarchive, composefs
This avoids false negatives from
ostree --version | grep -q ...exiting with failure underset -o pipefailbecausegrep -qcan exit as soon as it sees the desired string, leavingostree --versionto be terminated bySIGPIPEnext time it writes to stdout.test-admin-deploy-var: Don't rely on OSTREE_FEATURES
This is set during build-time testing, but unset during "as-installed" tests.
Resolves: installed-tests fail with 2024.3: test-admin-deploy-var.sh: line 24: OSTREE_FEATURES: unbound variable #3183