openstack
/
keystone
Code Issues Proposed changes

pkiz String conversion 

When the PKIZ token provider signs a token, force conversion of the
token to a string as the keystone client cms code produces unicode.

Without this change, Keystone errors out when validating tokens.  It
was only visible in tempest tests.

bp: compress-tokens
Change-Id: Ic932240fbbe5bd9da6197b37cf86092bde510f40
This commit is contained in:
Morgan Fainberg 2014-06-17 09:22:47 -04:00
parent e3d52279d7
commit b384e6de86
1 changed files with 7 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
keystone/token/providers/pkiz.py
View File

@ -32,9 +32,13 @@ ERROR_MESSAGE = _('Unable to sign token.')
class Provider(common.BaseProvider):
def _get_token_id(self, token_data):
try:
token_id = cms.pkiz_sign(jsonutils.dumps(token_data),
CONF.signing.certfile,
CONF.signing.keyfile)
# force conversion to a string as the keystone client cms code
# produces unicode. This can be removed if the client returns
# str()
# TODO(ayoung): Make to a byte_str for Python3
token_id = str(cms.pkiz_sign(jsonutils.dumps(token_data),
CONF.signing.certfile,
CONF.signing.keyfile))
return token_id
except environment.subprocess.CalledProcessError:
LOG.exception(ERROR_MESSAGE)