| FazBrowse GitHub Viewer | Trending | | Home |
| Tools: [Download Repo ZIP] [Original HTTPS Page] |
Sorry, something went wrong.
|
I suppose I should let @tseaver take the lead here |
Sorry, something went wrong.
| ... "robots-pubsub", | ||
| ... filter='log:apache-access AND textPayload:robot') | ||
| >>> sink.pubsub_topic = 'projects/my-project/topics/my-topic' | ||
| ... 'log:apache-access AND textPayload:robot', |
logName in the v2 API expects the fully qualified name. The colon operator is “contains” so it works since the fully qualified name contains itself, but it is not indexed and can be extremely slow.
|
Ok, addressed all review comments, changed commit messages/PR name to reflect that logging changes are optimizations rather then bugfixes,assuming Travis passes it's ready for another pass. |
Sorry, something went wrong.
|
@waprin Thanks for the research, as well as the patch! |
Sorry, something went wrong.
… the MWID/X.509 cert sources detected (#1848) The Python SDK will use a hybrid approach for mTLS enablement: - If the GOOGLE_API_USE_CLIENT_CERTIFICATE environment variable is set (either true or false), the SDK will respect that setting. This is necessary for test scenarios and users who need to explicitly control mTLS behavior. - If the GOOGLE_API_USE_CLIENT_CERTIFICATE environment variable is not set, the SDK will automatically enable mTLS only if it detects Managed Workload Identity (MWID) or X.509 Workforce Identity Federation (WIF) certificate sources. In other cases where the variable is not set, mTLS will remain disabled. ** This change also adds the helper method `check_use_client_cert` and it's unit test, which will be used for checking the criteria for setting the mTLS to true ** This change is only for Auth-Library, other changes will be created for Client-Library use-cases. --------- Signed-off-by: Radhika Agrawal <agrawalradhika@google.com> Co-authored-by: Daniel Sanche <d.sanche14@gmail.com>
… enables mTLS if GOOGLE_API_USE_CLIENT_CERTIFICATE is not set, when the MWID/X.509 cert sources detected (#1859) Add public wrapper for check_use_client_cert which enables mTLS if GOOGLE_API_USE_CLIENT_CERTIFICATE is not set, when the MWID/X.509 cert sources detected. Also, fix check_use_client_cert to return boolean value. Change #1848 added the check_use_client_cert method that helps know if client cert should be used for mTLS connection. However, that was in a private class, thus, created a public wrapper of the same function so that it can be used by python Client Libraries. Also, updated check_use_client_cert to return a boolean value instead of existing string value for better readability and future scope. --------- Signed-off-by: Radhika Agrawal <agrawalradhika@google.com> Co-authored-by: Daniel Sanche <d.sanche14@gmail.com>
Librarian Version: v0.5.0 Language Image: us-central1-docker.pkg.dev/cloud-sdk-librarian-prod/images-prod/python-librarian-generator:latest <details><summary>google-auth: 2.43.0</summary> ## [2.43.0](googleapis/google-auth-library-python@v2.42.1...v2.43.0) (2025-11-05) ### Features * Add public wrapper for _mtls_helper.check_use_client_cert which enables mTLS if GOOGLE_API_USE_CLIENT_CERTIFICATE is not set, when the MWID/X.509 cert sources detected (#1859) ([1535eccb](googleapis/google-auth-library-python@1535eccb)) * Enable mTLS if GOOGLE_API_USE_CLIENT_CERTIFICATE is not set, if the MWID/X.509 cert sources detected (#1848) ([395e405b](googleapis/google-auth-library-python@395e405b)) * onboard `google-auth` to librarian (#1838) ([c503eaa5](googleapis/google-auth-library-python@c503eaa5)) </details>
Correct the way we check whether `self._done_timeout` is an instance of `object` class or not. Fixes #1838 🦕
Fixes internal issue 427306238
| Back | FazBrowse Home | New Git URL |
Two changes here:
logger.list_entries() was using just the logger name rather than the fully qualified name, you can see an example on the docs that from v1 to v2 the change from log to logName also changed to expect fully qualified name.
The query was also using the colon operator, which is a "has" operator. Problem 1 is that this is a slower query because it's not indexed, problem 2 is that I'm pretty sure there's a bug because it's not working.
Second set of changes is the docs, just some random small discrepancies between them and the actual API. The biggest discrepancy was in the Sink constructor, the samples all use keyword arguments and special names for different sink types, the actual code seems to take these all as general positional arguments.