| Commit message (Collapse) | Author | Age | Files | Lines |
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
For mysql ForeignKey constraints were removed but the fields stayed as indexes.
This migration drops them.
bp db-sync-models-with-migrations
Change-Id: I3baeac4047cd65ac5d7733ba909c45d0874f17d8
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | | |
The test_mask_password test doesn't have to run 4 times.
Change-Id: If82e8f22d2594dbdf237e9cbe9eba5b944fb2663
|
|\ \ \ |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
When adding a user to a project, if the default role is missing, create it.
Bug 1176270
Change-Id: Id972ccf9c132c362a0b85049d248530dc2d56d54
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
DocImpact
Change-Id: I1b1de8f7e07afe8af8a5cbb83de7f935cea04670
|
|\ \ \ \ |
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Allow each of the extensions to have their own
sql migration repository instead of mixing them into
the common repo. db_sync must be called explicitly on the extension.
In the past, it was assumed that only migrations for backends backed in
sql would be run. In practice, however, all of the migrations were run
every time. The code has been modified to reflect this.
Adds parameter --extension to the cli for db_sync and db_version
to test out the migrations
bin/keystone-manage db_sync --extension example
will migrate to version 1 and
bin/keystone-manage db_sync --extension example 0
will migrate it back to 0
to check the version
bin/keystone-manage db_version --extension example
blueprint multiple-sql-migrate-repos
DocImpact
Change-Id: I6852d75bde6506c535fa3d74537e3c1bbd6578d8
|
|\ \ \ \ \ |
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
In middleware/s3_token.py, here only use swift for a logger and path
split functionality. We should remove swift dependency by using new
codes.
fixes bug #1178738
Change-Id: Icc2648720e220a873d1fb8e9961d777ceabef70b
|
|\ \ \ \ \ \ |
|
| | |_|_|_|/
| |/| | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Fixes bug 1206257
Change-Id: Ia522e023a2d66ec25bc909db12d358f7c0ee6952
|
|\ \ \ \ \ \ |
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
This patch removes the use of kwargs from manager calls where not
required. Dogpile.cache (the targeted caching library) does not
support kwargs out of the box with its cache-key-generator. This
change allows us to support the default cache-key-generator; while
it is possible to create a new cache-key-generator function, there
are many possible edge-cases to deal with when making cache
invalidation calls (ensuring the arguments are the same) as well as
possible performance implications (depending on the depth of method
introspection needed to determine how to invalidate the cache).
As an added bonus, this change brings the code touched more in-line
with the rest of keystone where most manager/driver calls do not
use kwargs unless absolutley required.
blueprint: caching-layer-for-driver-calls
Change-Id: I035c976314fb48f657661f681f7c1760d3c547a6
|
|\ \ \ \ \ \ \
| |_|_|_|/ / /
|/| | | | | | |
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
With 19081b834991d263d84c761dcf422a8c9faf40a1, the token_provider
is created in load_backends() so it doesn't have to be created by
the tests.
This makes the tests using token_provider_api consistent with how
other backends are used in the tests.
This change only affects unit tests.
Change-Id: I9c05c7876bd14fb28e77fe1e49dd509a2a1cb788
|
|\ \ \ \ \ \ \ |
|
| | |_|_|_|/ /
| |/| | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
When creating items like project, role through V3 API, if name field
is missing, the server will return 500 internal error or 409 conflict
error, shall return 400 bad request error instead
Change-Id: I9deddcd4cf801306c16555b59bba7fdfab6556a7
Fixes: Bug 1136342
|
|\ \ \ \ \ \ \ |
|
| | |_|_|/ / /
| |/| | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
This patch syncs models with migrations for:
-Domain
-GroupDomainGrant
-GroupProjectGrant
-Role
-UserDomainGrant
-UserProjectGrant
No actual schema change is taking place, this patch just corrects errors
in the model definitions.
partially implements bp db-sync-models-with-migrations
Change-Id: Ib3f22f56a9d00c725f21036ab4bfe4df5fad6876
|
|\ \ \ \ \ \ \ |
|
| | |/ / / / /
| |/| | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Bug 1183874
Change-Id: I3881e360485c4c1ff6d67aec12d47130c6b0eb70
|
|\ \ \ \ \ \ \ |
|
| | |_|/ / / /
| |/| | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
The dependency injection code doesn't handle circular
dependencies. This change makes it so that the
dependency injection code allows circular dependencies.
Part of fix for bug 1204605
Change-Id: I8de166a352ac727c7ddf27bae420b7c7ab22415f
|
|\| | | | | |
| |_|_|_|_|/
|/| | | | | |
|
| |/ / / /
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
As part of the process during several tests setUp
where the backends are reloaded, automatic dependency
injection takes place. The REGISTRY is being updated
with new providers and it's also looking up the
required dependencies. Some of the providers for the
requirements may not have been updated yet with the
new provider object, so it loads an object that
was created from a previous test run rather than the
current one. This can cause tests to
fail when one class gets a ref to the old one
(it depends on the order that the tests are run).
This change clears out the registry of providers before
loading backends. It only affects testing.
Part of fix for bug 1204605
Change-Id: Ib845493fa13531225e4be7e3b6cc315b9d19a0f4
|
|\ \ \ \ \
| |/ / / /
|/| | | | |
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Keystone defines a custom Request class that's used in many
places (middleware, etc.). BaseApplication says that subclasses
should typically set Request as the custom RequestClass,
but for some reason it's not used in keystone.wsgi.Application.
This doesn't cause any problems at this point because Keystone's
custom Request is the same as webob.Request. bp user-locale-api
requires the custom Request to implement calculating the requested
locale, and if Application doesn't set this custom RequestClass then
most requests don't using the custom Request.
Part of changes for bp user-locale-api
Change-Id: If20ee9000aba89a5a2c94ed8a3dda7382142038e
|
|\ \ \ \ \
| |_|/ / /
|/| | | | |
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
The backends should be loaded before the app is deployed for
the client tests, otherwise the wrong backends may be loaded
when the app is deployed.
Part of fix for bug 1204605
Change-Id: I70680c3daea3a863bcbde07b1ee7e062cd150e51
|
|\ \ \ \ \
| |_|/ / /
|/| | | | |
|
| | |_|/
| |/| |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The Keystone server would print a warning when both the token
format and provider were set to the default.
Also, the Keystone server would not start if the format was
commented out and the provider was set to the uuid.Provider.
Fixes: bug 1204314
Change-Id: Id7db33a1f27c4986af153efc73b22db8c6a8942e
|
|\ \ \ \ |
|
| | |/ /
| |/| |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The doc string in exception.py of Keystone will be
returned with __doc__ method, but cannot realize the
internationalization.Change exception module to enable
i18n support.
Changes in the patch are:
1, useing class variable msg_fmt to replace class __doc__
2, modify wsgi.render_exception function using unicode
function to replace str function
3, modify/add UT test cases
Fixes: bug # 1179425
Change-Id: I75c1229c905a2625d2f6961d1a8dd3958eac51a5
|
|\ \ \ \
| |_|_|/
|/| | | |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The keystone.common.sql.core.Base class cached the global database
engine when get_session() was called. When the global database engine
changed to a new instance, the cached copy was used in subsequent
calls to get_session(), leading to using the old engine and tests
failing to run by themselves.
This change makes it so that when the global database engine is
changed, Base will use the new engine rather than the invalid one.
Change-Id: I75aa3c230d9b4fd666ab8d478c9e9a27669905e8
Fixes: Bug #1179259
|
|\ \ \ \ |
|
| | |/ /
| |/| |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Replaced `raise nose.exc.SkipTest()` statement to self.skipTest().
Removed unused nose.exc imports.
Fixes bug 1172794
Change-Id: Ieb353864acadef43508d185156c7fa1667baa845
|
|/ / /
| | |
| | |
| | |
| | |
| | | |
Bug 1204995
Change-Id: Ife92041ffc386b8f9629096ebd65020f3cc7ae26
|
| |/
|/|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Credential table has foreign key constraint
referring to tenant table which is dropped.
Since sqlite does not support alter table
drop constraint, the foreign key constraint
was not dropped. When we try to load credential
table using sqlite backend it fails because tenant
table does not exist. Fix is provided such that
the credential table is recreated without foreign
key constraint and the data is moved from old
credential table to the new credential table.
Fixes Bug #1190383
Change-Id: I3afb04254f33e12fccb7da84c8674feba36622c8
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
When you load an old version of keystoneclient doing legacy tests and
then start the test app the ec2 work that depends on keystoneclient sees
the recently loaded legacy one which doesn't have the required modules
in it.
If we load the app first the ec2 modules are resolved and finds the
original keystone client. The ec2 module then uses
'from keystoneclient.contrib.ec2 import utils as ec2_utils' so the
ec2_utils reference is kept in the file scope so isn't affected by the
tests changing keystoneclient.
It is not a long term fix for bug 1178532 but it solves the immediate
symptoms when running client tests independently.
Change-Id: I7267ca0d4740f037884fae95f8a6562ee86584b9
|
|\ \ \ |
|
| |/ /
| | |
| | |
| | |
| | |
| | | |
Modified one spelling mistake in tests/test_middleware
Change-Id: I4beffa77c38321a44b44d1893d2335319c23b5a5
|
|\ \ \ |
|
| |/ /
| | |
| | |
| | |
| | |
| | |
| | | |
Refactor tests/test_backend.py, remove a useless enumerate call in
for loop.
Change-Id: I5d1914ae3490f715437711d594ec903a2c6632f5
|
|\ \ \ |
|
| |/ /
| | |
| | |
| | |
| | |
| | | |
This also adds i18n to a few related strings and updates doc.
Change-Id: Icba582a085939f58581fa909b63a36cbad3b4e69
|
|\ \ \
| |/ /
|/| | |
|