Session.construct() wasn't properly deprecated since the deprecation
was only mentioned in the docstring. Proper deprecation requires use
of warnings/debtcollector and documentation.
bp deprecations
Change-Id: Ieff238aff9d39cfbbb80381b2392c33d0359acb3
client.HTTPClient wasn't properly deprecated since all it had was
a comment in the code. Proper deprecation requires use of warnings
and documentation.
bp deprecations
Change-Id: I1c50c1441b23a79831e6e1df749084130e4b9af7
The generated docs didn't provide enough information for a developer
to get started using the API. This change enhances the documentation
for the module so that a developer knows where to go to start (create
a Client).
Partial-Bug: #1330769
Change-Id: I907187d34ebf2c2e662ff7b9547b0ecaef008414
The argument to the :raises: directive is the class name. If the
class name is a valid reference it's rendered as a link to the
class. This change cleans up the :raises: directives to use the
reference correctly and use a valid class reference.
Change-Id: I84188b60de0ab4c6b5b2fb5a203c43bfde094707
We don't need vim modelines in each source file, it can be set in
user's vimrc.
Change-Id: Ic7a61430a0a320ce6b0c4518d9f5d988e35f8aae
Closes-Bug: #1229324
Session.construct will create a session based upon the kwargs that used
to be passed to a client __init__ function. This will allow clients an
easy path to providing compatibility with deprecated arguments.
Make use of the function throughout discovery. Discovery was initially
released prior to the session object being completed and was therefore
handled with the same arguments as a client. Instead we should use a
session object so use the conversion function to convert those kwargs
into a session object if one is not provided.
Change-Id: I8dc1e0810ea6ebc6ea648ec37d7881825c566676
This patch adjust import items and add missing blank lines acording
to http://docs.openstack.org/developer/hacking/#imports
{{stdlib imports in human alphabetical order}}
\n
{{third-party lib imports in human alphabetical order}}
\n
{{project imports in human alphabetical order}}
\n
\n
{{begin your code}}
hacking project also enforce some checks for import group.
Let make the change in keytoneclient
Change-Id: Ic83bd5ee426905588f4a2d555851a9a01fc69f02
Allow discovery of the API versions a server supports and create an
appropriate client based on this.
Implements: blueprint api-version-discovery
Change-Id: I63e6759889066a784dc47e35152c82e1ead7951d
Simply pull in HTTPClient from its new location. This also restores the
tests that were modified back to there original form to make sure
nothing has changed from the original.
Fixes: bug 1211998
Change-Id: Ia3241a4352eb6b37c6489cd4dd1bc373de107e12
The discoverable entry point is to be client.Client however adding this
functionality to the current client.py is impossible as we end up with
circular dependencies.
This patch simply renames the current client.py to httpclient.py to make
future patches that will modify client.py more readable.
Required for: blueprint api-version-discovery
Change-Id: Ibcea03f6e1df0ae05329297166a8b8117fc3ce7b
Client V3 is a completely different implementation of the API, if we
want to intermix the V2 and V3 APIs then we don't need to version our
client at all.
Change-Id: I80579cd41df73ccfc6c8feb76843772829afac4e
Added support for domain scoping.
Enhancement on AccessInfo to support reading v2/v3 token information.
Enhancement on ServiceCatalog for reading/filtering v2/v3 service
catalog information.
Change-Id: Ibb678b9933d3673e37d0fba857a152a3c5d2b4f4
Commit 06d9437e8388b369546d760607f17cb5022750e9 made using a keyring
optional by adding a function to handle the imports. The problem is that
import needs to be available at a global level for it to be usable in
other parts of the code.
The function is replaced by a top-level try/except.
Fixing the imports isn't enough to get keyring support working because
in get_auth_ref_from_keyring() the wrong token was being used when
evaluating token expiration. It was using the token already in the
HTTPClient object and not the one returned from the keyring.
Some simple tests were added to prevent future regressions. These
tests will be skipped if the keyring or pickle packages are not
installed.
Change-Id: I1fe2c9e5cdf275df2047018368da2e4b3d2d6de2
Fixes: bug #1183072
- F401: imported but unused
- F999: syntax error in doctest
Fixing a couple of the F401's simply required fixing the doctests syntax
where the imports were actually in use.
Change-Id: If78abbb143daf8d005a71c5ab52836df29c5e0cd
- E125: continuation line does not distinguish itself from next logical
line
- E126: continuation line over-indented for hanging indent
Change-Id: I626a6d5d57db927e8b239f90569b5601c772f28b
This corrects an odd problem where Horizon would stand up multiple
client objects, which would cause duplicate/triplicate/dozens of
repeated log lines in its log files, due to multiple identical
handlers being added to the logging object
Fixes Bug 1182678
Change-Id: I020b3999c9008b996286ccb74a7801c96c43e71c
Instead of implementing token retrieval twice, let's use the code provided
by the TokenManger to get a token in raw format from Keystone.
Change-Id: I769be118ee137580cabd5cabcf7843e7afe1e456
Signed-off-by: Julien Danjou <julien@danjou.info>
We should not try loading keyrings if we dont' have to. We
certainly should not print messages if keyrings are not even
needed
Fix for LP# 1145071
Change-Id: I700a2b40620359f6f1fbe03c6a6239d162ddea61
The requests module dropped all configuration with the 1.0.0 release.
There's no danger_mode and no 'verbose'' mode. The former
shouldn't be necessary anymore and the latter can be done by setting
a different log handler for the request.logging root logger.
Change-Id: I41bfaf2574f6d7fc21f86e0124ceae7df6481eee
Signed-off-by: Chuck Short <chuck.short@canonical.com>
This implements handling of token expiration. Once the token is expired,
this will request automatically for a new one.
A special case is introduced if the user specified a token when the client
is initialized: this is the auth_token_from_user. In this case, we can't
know the expiration date, so we just assume it will never expire and don't
handle it ourself.
Change-Id: I3771ff5d669da015d4aa259de422c5d81aed3eb4
Signed-off-by: Julien Danjou <julien@danjou.info>
body (string rep) is at 'data' in the kwargs dict. 'body' was deleted prior to this call.
Change-Id: Ieea3d11f5246ee785e41df3bc54ac61200808354
Fixes: bug #1104313
- There's no need to call parent init function since that's the default
behaviour.
- The token attribute is nor used nor updated anywhere.
Change-Id: Ib0b2729a396a2d761931ce0e178c49c49814eb21
Signed-off-by: Julien Danjou <julien@danjou.info>
Add a new cli argument (--timeout) which
is by default 600 seconds which will be
set in the requests library so that timeouts
can occur correctly.
Change-Id: I845c55dfb6f6b8345663ccdb5b150a2655f20026
* Implement correct certificate verification
* Add requests to tools/pip-requires
* Fix OS_CACERT env var help text
* Add info to README
* Rework tests to use requests
Pinned requests module to < 1.0 as 1.0.2 is now current in pipi
as of 17Dec2012.
Change-Id: I120d2c12d6f20ebe2fd7182ec8988cc73f623b80
Updates keystoneclient so that it uses a print statement to
stderr instead of a logger.warn if stderr is a tty.
This works around problems caused by the fact that logging isn't always
initialized when this module import runs (and fails) thus causing:
No handlers could be found for logger "keystoneclient.client"
instead of the intended log message.
Fixes LP Bug #1090396
Change-Id: I94e2c45eec14edfe3c2f356af6907aa827808a13
Updates the use_keyring option for the Keystone Client class
to default to False by default. This provides for a much
easier upgrade path the the most recent keystone-client code
and is backwards compatible with the previous release.
This does not change the default of the no_cache keystone CLI
which will still enable use_keyring by default.
Fixes LP Bug #1087434.
Change-Id: Iaafadde660b0542a9c0a9c1bb742cb369b523a68
User can optionally turn off keyring by specifying the --no-cache option.
It can also be disabled with environment variable OS-NO-CACHE.
Change-Id: I8935260bf7fd6befa14798da9b4d02c81e65c417
bug 1078589
* allow client values to be overridden, but use auth_ref if none
available
* added tests to match this flow
* refactored tokens into test_fixtures.py file
Change-Id: I771a2dee6dedf31d883417d9b4e6e64bbb620f14
blueprint solidify-python-api
* extended and updated documentation strings
* updated README.rst with latest options
* made debug a pass-through value, optionally set on client (instead of
just being pulled from environment variable)
* adding AccessInfo object and associated tests
(access.AccessInfo meant to be a cacheable object external to client
and ultimately to replace service_catalog and it's existing functionality)
* extending authtoken to support lists of endpoints
* maintaining a single entity for client.management_url with first from
list of possible endpoints
* create project_name and project_id synonyms to match tenant_name and
tenant_id
* replacing authenticate call to a pure method, not overloading the
resource/manager path that confuses base URL concepts.
* throw AuthorizationFailure if client attempts to access keystone
resources before it has a management url
* special case listing tenant using auth_url for unscoped tokens authorized
through client
* special case listing tokens.authenticate for Dashboard to allow unscoped
tokens to hand back parity information to dashboard
Change-Id: I4bb3a1b6a5ce2c4b3fbcebeb59116286cac8b2e3
Formally, OpenStack Keystone implements the OpenStack Identity API, and
this is a client to the API, not to Keystone itself.
Change-Id: If568866221a29ba041f0f2cd56dc81deeb9ebc00
The original IP is useful in cases where keystoneclient is used by a
different openstack component and we need to know who made the original
request. Otherwise it gets overwritten by e.g. Dashboard's host's IP.
bug 1046837
Change-Id: Ic22c565e92010afd89c8573c375919215b70d73d
This patch moves the json.loads(body) call in the
HTTP response handling to after the check for non-
200-300 return codes. This gets rid of the
ValueError exception raise when you hit, for instance,
a 400 or 404.
Also changes a number of logger.exception() calls to
logger.debug() calls, since some exceptions are expected
and should not be logged as exceptions per-se.
fixes LP bug#1067512
Change-Id: If66fb1846ddc19da5bc2f15c6e0dd09019a56932
