This introduces a linter for PEP257 to avoid trivial nitpicking of
docstrings in code reviews.
Change-Id: I01ebad7b70cf61dd80d3c06c6808d8178fbdd634
Related-Bug: 1501544
Depends-On: I60adf0dca4aa32f4ef6bca61250b375c8a3703c6
The subprocess_without_shell_equals_true test checks that subprocess
is called because it can easily be used incorrectly. The current use
is correct since it passes a list rather than a command string.
Change-Id: Ia31b1911547560e245cd1ae0c91cf7789146424f
Use oslo.cache now that it's been made available.
Note that we only deprecate the cache backends, but not remove
them, they are public APIs and subject to deprecation.
Also remove most of the caching code, but keep the portions
that register backends, since they are still available.
Note that the tests that were deleted were testing cache
functionality, they now exist in oslo.cache.
Co-Authored-By: Brant Knudson <bknudson@us.ibm.com>
Change-Id: I39d3ff1dd2edad79effa56450b444174b09f971e
Updated the docs to better reflect v3 deployments. Essentially, matched
the [app:service_v3] entry with the value in the ini file.
Closes-Bug: 1504891
Change-Id: I57f6a2a5287b2f6aa108586c1ad07b9a6235d53d
The OpenStack standard log format string contains fields for
user ID, project ID, etc., that will be displayed if the
appropriate fields are set in the RequestContext. This change sets
the fields in the RequestContext.
Closes-Bug: 1500222
Change-Id: I4ae6da483cfae98f910a2905e80b6af606ae6468
There are files hanging around the top level directory that only
handle the versioning routes of keystone (/v2.0 and /v3).
These should be moved to their own package to further isolate
these APIs.
Closes-Bug: #1504892
DocImpact
Change-Id: Ica0ddcbeb6f7fc00a4ad3919fa16bf135637a607
oslo.policy 0.12.0 contains a change to use requests to do the http
check rather than urllib. This change caused keystone tests to fail
because the keystone tests were mocking urllib, making assumptions
about how oslo.policy is implemented. Keystone doesn't need to test
internal features of oslo.policy, so these tests are removed.
Change-Id: I9d6e4950b9fe75cbb94100c8effdcec002642027
Closes-Bug: 1505374
Change I952cac73a9713bde4ad757371ca8b4ded93f207e refactored the keystone test
cases to use the six.moves.http_client for HTTP status codes instead of
integers. This change refactors the method names to follow the same pattern.
Change-Id: I90b17a7196075c164fe8bbd0f43af13a118e4c7e
Change I952cac73a9713bde4ad757371ca8b4ded93f207e refactored the keystone test
cases to use the six.moves.http_client for HTTP status codes instead of
integers. This change refactors the method names to follow the same pattern.
Change-Id: I6f0cc05d9bf45a32d50e83151141796ba76325ec
Change I952cac73a9713bde4ad757371ca8b4ded93f207e refactored the keystone test
cases to use the six.moves.http_client for HTTP status codes instead of
integers. This change refactors the method names to follow the same pattern.
Change-Id: Ic830ba42200ea39d0a0bd5f3355b8b32b292ae83
Change I952cac73a9713bde4ad757371ca8b4ded93f207e refactored the keystone test
cases to use the six.moves.http_client for HTTP status codes instead of
integers. This change refactors the method names to follow the same pattern.
Change-Id: I5b48598afa55a7c15ca21e79f9c572001e17b069
currently notifications are emitted for v2 calls, but the
initiator field is not filled in.
Co-authored-by: sam leong <chio-fai-sam.leong@hpe.com>
Change-Id: Ie2c3fe8d105d59ab89b7f6625e159d4eb6e923b0
Closes-Bug: #1485035
A string inside an exception was not being passed to
i18n translation so it will only show up as English if it
was ever thrown.
Here is a grep I used to make sure there were no more
exception strings that were not being passed to translations
before they were thown.
grep -r 'exception' . | grep '("' | grep -v '_'
Change-Id: I51a0d6d5b1d4053c380c8be5a0e6ac4e61985b81
