Closed dmcguire81 closed 4 years ago
Merging #596 into master will not change coverage. The diff coverage is
100.00%
.
@@ Coverage Diff @@
## master #596 +/- ##
=======================================
Coverage 85.68% 85.68%
=======================================
Files 87 87
Lines 4994 4994
Branches 795 795
=======================================
Hits 4279 4279
Misses 578 578
Partials 137 137
Impacted Files | Coverage Δ | |
---|---|---|
petastorm/workers_pool/process_pool.py | 92.70% <ø> (ø) |
|
petastorm/reader.py | 90.24% <100.00%> (ø) |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update 1a617b6...9489b71. Read the comment docs.
I still owe you a test, here, so I'll get to this, today.
@selitvin let me know if this test works for you
@dmcguire81: looks good. Let's fix whatever flake8 complaints about and land it.
@selitvin can you restart the Travis CI build? It looks like a brittle test is timing out, which seems unrelated.
@selitvin can you restart the Travis CI build? It looks like a brittle test is timing out, which seems unrelated.
Hmmm. This is annoying. Did not see these timeouts before... Restarted.
These test timeouts appear to be in the similar place as the hanging you observed with s3. Do you think they are related?
These test timeouts appear to be in the similar place as the hanging you observed with s3. Do you think they are related?
Could be. I'll pull you into the ongoing discussion with pyarrow
and s3fs
maintainers, since they asked for a reproducer that you may be able to build more quickly than me, if you're more familiar. Needing to work around this deadlock is also related to our discussion of parameterizing the FileSystem
instance.
@selitvin https://github.com/dask/s3fs/issues/365
Addresses #595