Remove invalid chunk validation

[garye]

Fixes#2980

[lukesneeringer]

@dhermes Can you glance at this one? I lack sufficient context.

[dhermes]

@garye I'd guess there would be broken tests, but it's hard to say right now because there is a bad release of protobuf that is breaking everything

[lukesneeringer]

@dhermes Update on the protobuf thing: googleapis/api-client-staging#159

[dhermes]

I restarted the CircleCI build: https://circleci.com/gh/GoogleCloudPlatform/google-cloud-python/865

Hopefully it works?

[garye]

Can someone help me finish this off? The test failures don't really seem related. Getting this fix pushed and released will unblock a number of users.

[dhermes]

@garye Just re-started the CircleCI build: https://circleci.com/gh/GoogleCloudPlatform/google-cloud-python/876

Hopefully the bad imports are fully resolved. I'll follow-up if not.

[dhermes]

@garye So we've got a unit test failing that is actually caused by this PR:

def test_invalid_empty_second_chunk(self):
from google.cloud.bigtable.row_data import InvalidChunk
chunks = _generate_cell_chunks(['', ''])
first = chunks[0]
first.row_key = b'RK'
first.family_name.value = 'A'
first.qualifier.value = b'C'
response = _ReadRowsResponseV2(chunks)
iterator = _MockCancellableIterator(response)
prd = self._make_one(iterator)
with self.assertRaises(InvalidChunk):
> prd.consume_next()
E AssertionError: InvalidChunk not raised

ISTM the right move is just to delete this test case since it no longer applies?

[garye]

yeah, i think that's an invalid test case

[dhermes]

Ping me when this PR is updated with the test case removed? You can ignore the speech test failures. It looks like the underlying protobuf libraries are (now correctly) using single instead of double precision for message fields with single precision

[garye]

Will do, thanks

…

On Mon, Feb 13, 2017, 1:25 PM Danny Hermes ***@***.***> wrote: Ping me when this PR is updated with the test case removed? You can ignore the speech test failures. It looks like the underlying protobuf libraries are (now correctly) using single instead of double precision for message fields with single precision — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#2992 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AACTr3aTAOjGxQ5hgBk6h-nz-2JJq70Pks5rcKAcgaJpZM4L8Ql5> .

[garye]

Yeah, thanks I noticed that and (hopefully) pushed a fix.

[tseaver]

@garye Ugh, you need to be able to run tox -e lint locally:

bigtable/unit_tests/test_row_data.py:419:1: W293blanklinecontainswhitespace

[garye]

Ugh, OK. That will take a bit of time to set up.

…

On Mon, Feb 13, 2017, 4:11 PM Tres Seaver ***@***.***> wrote: @garye <https://github.com/garye> Ugh, you need to be able to run tox -e lint locally: bigtable/unit_tests/test_row_data.py:419:1: W293 blank line contains whitespace — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#2992 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AACTr05fXHOLT3PHG9DwvC9OaiCJnV7Gks5rcMbkgaJpZM4L8Ql5> .

[tseaver]

I think it should just work to run tox -e pylint. You probably want to export the following environment variables so it only runs on your changes:

$ export GOOGLE_CLOUD_TESTING_REMOTE=upstream # the name of the remote for this repo
$ export GOOGLE_CLOUD_TESTING_BRANCH=master
$ tox -e lint

[dhermes]

LGTM

[garye]

Thanks guys!

[dhermes]

@garye We hope to cut a release in the next 2-3 days as well