fix: use .read() instead of .content.read() in aiohttp transport #1899
Conversation
Using .read() is more robust when working with caching libraries like aiohttp-client-cache, which may cause .content.read() to return empty bytes on subsequent reads. .read() ensures the full body is returned. Fixes #1248
|
👋 Jules, reporting for duty! I'm here to lend a hand with this pull request. When you start a review, I'll add a 👀 emoji to each comment to let you know I've read it. I'll focus on feedback directed at me and will do my best to stay out of conversations between you and other bots or reviewers to keep the noise down. I'll push a commit with your requested changes shortly after. Please note there might be a delay between these steps, but rest assured I'm on the job! For more direct control, you can switch me to Reactive Mode. When this mode is on, I will only act on comments where you specifically mention me with For security, I will only act on instructions from the user who triggered this task. New to Jules? Learn more at jules.google/docs. |
Using .read() is more robust when working with caching libraries like aiohttp-client-cache, which may cause .content.read() to return empty bytes on subsequent reads. .read() ensures the full body is returned. Fixes #1248
chalmerlowe
added
the
kokoro:run
yoshi-kokoro
removed
the
kokoro:run
chalmerlowe
added
the
kokoro:run
yoshi-kokoro
removed
the
kokoro:run
|
Temporarily removing the requirement for system-3.10 to pass due to a known issue with that CI/CD check that is being addressed in an separate PR #1901 |
Acknowledged. |
chalmerlowe
deleted the
fix-aiohttp-cache-empty-response-3355868403646689815
branch
PR created by the Librarian CLI to initialize a release. Merging this PR will auto trigger a release. Librarian Version: v1.0.1 Language Image: us-central1-docker.pkg.dev/cloud-sdk-librarian-prod/images-prod/python-librarian-generator@sha256:b8058df4c45e9a6e07f6b4d65b458d0d059241dd34c814f151c8bf6b89211209 <details><summary>google-auth: 2.46.0</summary> ## [2.46.0](v2.45.0...v2.46.0) (2026-01-05) ### Features * Recognize workload certificate config in has_default_client_cert_source for mTLS for Agentic Identities (#1907) ([0b9107d](0b9107d5)) ### Bug Fixes * Fix test coverage for mtls_helper (#1886) ([02e7163](02e71631)) * use .read() instead of .content.read() in aiohttp transport (#1899) ([12f4470](12f4470f)) * add types to default and verify_token and Request __init__ based on comments in the source code. (#1588) ([59a5f58](59a5f588)) * fix the document of secure_authorized_session (#1536) ([5d00147](5d001470)) * raise RefreshError for missing token in impersonated credentials (#1897) ([94d04e0](94d04e09)) * remove setup.cfg configuration for creating universal wheels (#1693) ([c767531](c767531c)) ### Documentation * update urllib3 docstrings for v2 compatibility (#1903) ([3f1aeea](3f1aeea2)) </details>
3 participants
Fixes issue where response content is empty when using
aiohttp-client-cacheby usingresponse.read()instead ofresponse.content.read().Rationale:
In
aiohttp,.read()is the standard, high-level method for reading the full response body. It correctly handles internal state (like caching the body for subsequent calls), which is essential for compatibility with wrappers likeaiohttp-client-cache. Accessing.content.read()directly bypasses this logic, leading to empty responses when the stream has already been consumed or intercepted by the caching layer.Benefits:
google-authis used in conjunction withaiohttpwrappers or middleware.aiohttpbest practices for consuming response bodies.Risk Mitigation:
_CombinedResponseis designed to load the full response into memory. Changing to.read()preserves this behavior exactly; it just uses a safer accessor.aiohttp's.read()returns a reference to the bytes, there is no additional memory overhead compared to the previous implementation.PR created automatically by Jules for task 3355868403646689815 started by @chalmerlowe