You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Describe the bug run-tests -m '/failures/' doesn't seem to be working as expected. The test should attempt to build all targets under //tests/failures/... that are marked with the manual tag and expect them to fail. However, if I add the manual tag to a suceeding target, e.g. //tests/failures/transitive-deps:lib-c, then the overall test still suceeds.
To Reproduce
Apply the following patch and run bazel build //tests/run-tests && bazel-bin/tests/run-tests -m '/failures/'.
diff --git a/tests/RunTests.hs b/tests/RunTests.hs
index 9cc83b6..a589457 100644
--- a/tests/RunTests.hs
+++ b/tests/RunTests.hs
@@ -63,7 +63,7 @@ main = hspec $ do
for_ all_failure_tests $ \test -> do
it test $ do
- assertFailure (bazel ["build", "test"])
+ assertFailure (bazel ["build", test])
-- Test that the repl still works if we shadow some Prelude functions
it "repl name shadowing" $ do
However, it uncovers another failure
Failures:
tests/RunTests.hs:65:7:
1) failures //tests/failures/transitive-deps:lib-dFailure
uncaught exception: IOException of type InvalidArgument
fd:15: hGetContents: invalid argument (invalid byte sequence)
To rerun use: --match "/failures///tests/failures/transitive-deps:lib-dFailure/"
tests/RunTests.hs:65:7:
2) failures //tests/failures/transitive-deps:lib-cFailure
uncaught exception: IOException of type InvalidArgument
fd:15: hGetContents: invalid argument (invalid byte sequence)
To rerun use: --match "/failures///tests/failures/transitive-deps:lib-cFailure/"
The text was updated successfully, but these errors were encountered:
Problem description by Andreas Herrmann:
> run-tests -m '/failures/' doesn't seem to be working as expected.
> The test should attempt to build all targets under
> //tests/failures/... that are marked with the manual tag and expect
> them to fail. However, if I add the manual tag to a suceeding target
> e.g. //tests/failures/transitive-deps:lib-c, then the overall test
> still suceeds.
Manually tested via description that a succeeding test now actually
fails.
Fixes#943
Problem description by Andreas Herrmann:
> run-tests -m '/failures/' doesn't seem to be working as expected.
> The test should attempt to build all targets under
> //tests/failures/... that are marked with the manual tag and expect
> them to fail. However, if I add the manual tag to a suceeding target
> e.g. //tests/failures/transitive-deps:lib-c, then the overall test
> still suceeds.
Manually tested via description that a succeeding test now actually
fails.
Fixes#943
Describe the bug
run-tests -m '/failures/'
doesn't seem to be working as expected. The test should attempt to build all targets under//tests/failures/...
that are marked with themanual
tag and expect them to fail. However, if I add themanual
tag to a suceeding target, e.g.//tests/failures/transitive-deps:lib-c
, then the overall test still suceeds.To Reproduce
Apply the following patch and run
bazel build //tests/run-tests && bazel-bin/tests/run-tests -m '/failures/'
.Expected behavior
The test should fail with the above patch applied.
Environment
shell.nix
)Additional context
The following patch provides additional output
The following patch resolves the initial problem.
However, it uncovers another failure
The text was updated successfully, but these errors were encountered: