Tokenization-related test tweaks (backport from #8185)#8499
Tokenization-related test tweaks (backport from #8185)#8499hendrikmakait merged 1 commit intodask:mainfrom
Conversation
| second.wait() | ||
|
|
||
| fs = c.map(func, [first] * 5, [second] * 5) | ||
| fs = c.map(func, [first] * 5, [second] * 5, key=[f"x{i}" for i in range(5)]) |
There was a problem hiding this comment.
With dask/dask#10883, func no longer tokenizes to 5 different random values, so we ended up with a single key.
Unit Test ResultsSee test report for an extended history of previous test failures. This is useful for diagnosing flaky tests. 27 files + 1 27 suites +1 10h 15m 49s ⏱️ + 40m 16s For more details on these failures, see this check. Results for commit e9fd17a. ± Comparison against base commit 9a9468c. ♻️ This comment has been updated with latest results. |
328ae70 to
ed0ded1
Compare
4889ab2 to
e9fd17a
Compare
|
|
||
| @gen_cluster(client=True, nthreads=[("127.0.0.1", 1), ("127.0.0.1", 2)]) | ||
| async def test_dont_steal_worker_restrictions(c, s, a, b): | ||
| future = c.submit(slowinc, 1, delay=0.10, workers=a.address) |
There was a problem hiding this comment.
It looks like this was used to seed the task duration. I think that by removing this, this test will be vulnerable to changes to the default task duration and could potentially fail to raise.
There was a problem hiding this comment.
I see that we assert that (almost) the same scenario does indeed balance in test_work_steal_no_kwargs. Maybe it makes sense to adjust the delay there and add a comment that references that this tests a precondition of the others?
There was a problem hiding this comment.
I expanded on the cleanup on test_steal and, since it's unrelated to tokenization, moved it to a separate PR
e9fd17a to
d5cf4de
Compare
Backport from #8185, which may not happen in the short term future.
This unblocks dask/dask#10883.