Comment 7 for bug 1880999

The python-boto3 unit and functional tests seem to do a pretty good job exercising python-botocore and python-s3transfer. I ran those tests against botocore/updates and s3transfer/updates for each suite, and compared those results to the same tests run against botocore/proposed and s3transfer/proposed for each suite. The raw results are available here, alongside the script used to run the tests:

https://git.launchpad.net/~davidkrauser/+git/logs/tree/LP1875485/20200714?id=17b08c1f3ea389dd3f49f2a511f6ed6ed45c6ac4

The summary is as follows:

==> bionic-proposed-py.log <==
OK (SKIP=1)

==> bionic-proposed-py3.log <==
OK

==> bionic-updates-py.log <==
OK (SKIP=1)

==> bionic-updates-py3.log <==
OK

==> eoan-proposed-py.log <==
OK (SKIP=1)

==> eoan-proposed-py3.log <==
OK

==> eoan-updates-py.log <==
OK (SKIP=1)

==> eoan-updates-py3.log <==
OK

==> xenial-proposed-py.log <==
FAILED (SKIP=1, errors=10, failures=1)

==> xenial-proposed-py3.log <==
FAILED (errors=10, failures=1)

==> xenial-updates-py.log <==
FAILED (SKIP=1, errors=1, failures=1)

==> xenial-updates-py3.log <==
FAILED (errors=1, failures=1)

Seeing that the awscli verification for xenial against this package failed, and the test results are worse here against proposed for xenial, I'm marking this as failed verification for xenial. The other suites look to be in good order.

Also: the documentation related python-boto3 tests were skipped for these runs. I couldn't get them to pass reliably, even with the dependencies from updates.