[ajeya-krishna]

What does this PR do?

Problem

An empty CI variable such as RENDER_GIT_COMMIT prevents release autodetection
from reaching a populated GITHUB_SHA, leaving the detected release empty.

Solution

Skip empty CI values, preserving the order and contents of nonempty values.
Add helper regressions and a real-client initialization test.

How we tested it

The six new cases had 3 failures before the fix and all pass after it.
41 targeted tests passed; full unit suite: 969 passed, 2 skipped.

Type of change

• Bug fix

Verification

uv run --frozen pytest tests/unit/test_environment.py tests/unit/test_initialization.py tests/unit/test_resource_manager.py
uv run --frozen ruff check .
uv run --frozen ruff format --check .
uv run --frozen mypy langfuse --no-error-summary
LANGFUSE_PUBLIC_KEY=test-public-key LANGFUSE_SECRET_KEY=test-secret-key \
LANGFUSE_BASE_URL=http://localhost:3000 \
uv run --frozen pytest -q -n 4 --dist worksteal tests/unit

Python 3.12. No e2e/server or live-provider tests run; this is a local configuration
helper. The full unit run uses dummy credentials, not a live account.

Checklist

• Diff reviewed using code_review.md.
• Tests added for the behavior change.
• No docs/examples/environment-template change needed.
• No generated files edited.
• No credentials committed.

AI-assisted implementation and regression testing with Codex.

Confidence Score: 5/5

The PR appears safe to merge.

Summary

Skips empty CI release values so release detection can reach the next populated variable.

• Langfuse finds the first nonempty release value from CI.

No actionable issues found. Tests were not run during this review.

_{Reviews (1) · Last reviewed commit: "fix(config): skip empty CI release varia..." · Reviewed by Greptile}

[claude]

Claude Code Review

This pull request is from a fork — automated review is disabled. A repository maintainer can comment @claude review to run a one-time review.

[CLAassistant]

All committers have signed the CLA.

[chrikrah]

@ajeya-krishna From my perspective this can be merged at d9c85a2. Both callers, client.py:335 and resource_manager.py:631, already skip an empty LANGFUSE_RELEASE with or, and only the helper stopped on "".

$ pytest tests/unit/test_environment.py tests/unit/test_initialization.py -q -p no:cacheprovider   # Python 3.12
26 passed in 2.92s                                    # d9c85a2
3 failed, 23 passed in 2.53s                          # environment.py from d2222a5, tests kept
FAILED tests/unit/test_environment.py::test_get_common_release_envs[all-empty]
FAILED tests/unit/test_environment.py::test_get_common_release_envs[skip-empty]
FAILED tests/unit/test_initialization.py::TestClientInitialization::test_empty_ci_release_falls_back_to_next_variable

non-blocking: when every CI variable is empty, the base helper returned "". The if v is not None filter at resource_manager.py:638 kept it, so the resource had an empty release attribute. At head the attribute goes away. Does a server-side filter or the UI read that empty release?

[ajeya-krishna]

@chrikrah Thanks for checking this. Omitting the release attribute when all CI variables are empty is intentional: it now matches the existing behavior when those variables are unset, instead of exporting an empty release. I verified the actual _init_tracer_provider() resource for unset, all-empty, and empty-then-populated inputs; the first two omit the attribute and the last preserves github-sha. All 41 environment, initialization, and resource-manager tests pass at d9c85a2.

I haven't verified server/UI handling of an explicitly empty release, so I can't claim those filters are equivalent. No additional code change proposed; the all-empty helper case is already covered by the PR.