test_utilities: Teach drake_py_unittest where the source lives #10969
Conversation
|
+@EricCousineau-TRI for feature review, please. |
There was a problem hiding this comment.
At some point, it would be nicer to make this extremely explicit - drake_py_unittest should have some logic to scribble in the main path to the Bazel invocation script. But eh, that's a relatively non-functional change for high-effort.
Thanks!
Reviewed 1 of 1 files at r1.
Reviewable status: all discussions resolved, platform LGTM from [ericcousineau-tri], labeled "do not merge"
|
CI found a few corner cases. I'll work on it. |
Previously, we were searching in "." for the source code. That is dangeous when running a test manually from bazel-bin (instead of via bazel test), and is ambiguous for reused basenames like lcm_test.
jwnimmer-tri
force-pushed
the
py-unittest-main
branch
from
83e9e4a
to
03a9950
Compare
There was a problem hiding this comment.
As it turns out, the prior find logic was "getting lucky" on the install_test. By making the rules more rigid, I had to make the install_test more explicit. PTAL.
Reviewable status: all discussions resolved, platform LGTM from [ericcousineau-tri], labeled "do not merge" (waiting on @EricCousineau-TRI)
There was a problem hiding this comment.
Reviewed 5 of 5 files at r2.
Reviewable status: all discussions resolved, platform LGTM from [ericcousineau-tri], labeled "do not merge"
|
+@ggould-tri for platform review per schedule, please. |
There was a problem hiding this comment.
Reviewed 5 of 5 files at r2.
Reviewable status: all discussions resolved, platform LGTM from [ericcousineau-tri, ggould-tri], labeled "do not merge"
Previously, we were searching in "." for the source code. That is dangerous when running a test manually from bazel-bin (instead of via bazel test), and is ambiguous for reused basenames like lcm_test.
Without this PR, running
bazel-bin/bindings/pydrake/py/lcm_testby hand fails.This change is