-
Notifications
You must be signed in to change notification settings - Fork 3.9k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
sql/tests: TestRandomSyntaxSQLSmith failed [no locations to init GEOS] #110780
Comments
This failure included be0bb2c, so there must still be something missing for this test. |
Reproducible at commit
|
|
@cockroachdb/test-eng could we have your help on this again to figure out why the random syntax generator tests aren't getting libgeos? Thanks in advance! |
On master, via the above reproduction, the test actually passes, but the same issue remains,
Despite running with the same seed, the panic on 12a0fdf is caused by In any case, I am sending a PR to use a runfile, this is how other unit tests with libgeos dependency work. The change in be0bb2c builds libgeos in a directory that's not reachable from a test sandbox. |
110850: codeowners: add `test-eng-prs` to `pkg/clusterversion` r=healthy-pod a=healthy-pod part of: DEVINF-509 Release note: None 110926: sql: add libgeos runfile dependency for RSG tests r=rickystewart,herkolategan a=srosenberg In [1], we updated CI (nightly) to require a build step for `//c-deps:libgeos`. That didn't fix `rsg_test`'s implicit dependency on libgeos, since the produced artifacts are not reachable from the test sandbox. Instead, we use a runfile dependency--same way other test packages specify libgeos dependency. This enables dynamic loading via `bazel.Runfile`. [1] #110129 Epic: none Fixes: #110780 Release note: None Co-authored-by: healthy-pod <[email protected]> Co-authored-by: Stan Rosenberg <[email protected]>
In [1], we updated CI (nightly) to require a build step for `//c-deps:libgeos`. That didn't fix `rsg_test`'s implicit dependency on libgeos, since the produced artifacts are not reachable from the test sandbox. Instead, we use a runfile dependency--same way other test packages specify libgeos dependency. This enables dynamic loading via `bazel.Runfile`. [1] cockroachdb#110129 Epic: none Fixes: cockroachdb#110780 Release note: None
In [1], we updated CI (nightly) to require a build step for `//c-deps:libgeos`. That didn't fix `rsg_test`'s implicit dependency on libgeos, since the produced artifacts are not reachable from the test sandbox. Instead, we use a runfile dependency--same way other test packages specify libgeos dependency. This enables dynamic loading via `bazel.Runfile`. [1] cockroachdb#110129 Epic: none Fixes: cockroachdb#110780 Release note: None
Did you try the same seed on master? I'm wary of relying on random seeds to repro issues at different commits, since intervening code changes could change the number of calls to the RNG. I found that at the commit the issue was reported it was reproducible. |
Yep.
That's most likely the case. The two (RNG) sequences are not in 1-1 correspondence, hence the sequence on master didn't produce
Yep, the issue was reproducible at the commit. |
sql/tests.TestRandomSyntaxSQLSmith failed with artifacts on release-23.1 @ 12a0fdf76785787a3a7e83198f1adfd7184ea910:
Help
See also: How To Investigate a Go Test Failure (internal)
This test on roachdash | Improve this report!
Jira issue: CRDB-31609
The text was updated successfully, but these errors were encountered: