http_client: fixed support of runtime tests

[mabrarov]

Summary

Fixed support of runtime tests in HTTP client. Refer to #11686 (comment).

The enabled tests were successfully executed on CI within 099bb5c temporary commit using fluent/fluent-bit-ci#153 - refer to https://github.com/fluent/fluent-bit/actions/runs/24943722795?pr=11724.

---

Testing

• [N/A] Example configuration file for the change.
• [N/A] Debug log output from testing the change.
• Attached Valgrind output that shows no leaks or memory corruption was found - refer to flb_run_code_analysis.log for the output of command

  TEST_PRESET=valgrind ./run_code_analysis.sh

• [N/A] Run local packaging test showing all targets (including any new ones) build.
• [N/A] Set ok-package-test label to test for all targets (requires maintainer to do).

Documentation

• [N/A] Documentation required for this feature

Backporting

• [N/A] Backport to latest stable release.

---

Fluent Bit is licensed under Apache 2.0, by submitting this pull request I understand that this code will be released under the terms of that license.

Summary by CodeRabbit

• Bug Fixes

  • Added a null-pointer safety check to HTTP client connection setup.

• Tests

  • Reduced the default test skip list so Elasticsearch and forward outputs are included in default test runs.

• Chores

  • CI unit-test workflow updated to source reusable test scripts from a different repository reference.

[coderabbitai]

📝 Walkthrough

Walkthrough

Reduced default skipped-tests in the analysis script, added a null check when initializing HTTP client network pointers to avoid dereferencing a NULL connection, and changed CI unit-test jobs to checkout reusable CI scripts from a different repository/ref.

Changes

Cohort / File(s)	Summary
Build script run_code_analysis.sh	Changed default SKIP_TESTS assignment to remove flb-rt-out_elasticsearch and flb-rt-out_forward, resulting in fewer -DFLB_WITHOUT_* CMake flags when SKIP_TESTS is unset.
HTTP client src/flb_http_client.c	create_http_client() now checks u_conn != NULL before reading or reassigning u_conn->net, preventing a potential null dereference.
CI workflow .github/workflows/unit-tests.yaml	Unit-test jobs now checkout mabrarov/fluent-bit-ci at ref: feat/enable_tests into ci/ instead of the previous calyptia/fluent-bit-ci.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Suggested reviewers

• edsiper
• niedbalski
• patrick-stephens
• celalettin1286

Poem

🐇 I hopped through diffs at break of day,

pruned two skips that used to stray.
I fenced a NULL with gentle care,
pointed CI to fresh new air.
Tiny hops, a tidy lair.

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly identifies the main change: fixing HTTP client support for runtime tests, which directly corresponds to the primary modification in src/flb_http_client.c.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

🧹 Nitpick comments (1)

src/flb_http_client.c (1)

1020-1032: LGTM — null-guard correctly protects u_conn access in test paths.

The added if (u_conn != NULL) wrapper correctly prevents the NPE on u_conn->net / u_conn->upstream when create_http_client() is invoked in runtime-test mode with a NULL connection, while preserving the original behavior for the non-NULL path. When u_conn == NULL, c->original_net_setup stays NULL (from flb_calloc), which is consistent with the NULL checks in http_client_clamp_connection_io_timeout (Line 1499), http_client_update_connection_io_timeout (Line 1524), and http_client_unbind_connection (Line 1592).

Optional: this block now duplicates the net-setup selection logic already centralized in http_client_bind_connection() (Lines 1567–1588). Consider replacing the body with a call to http_client_bind_connection(c, u_conn) after the c->u_conn = u_conn; assignment is removed from Line 1019 — the helper already sets c->u_conn, selects original_net_setup, copies request_net_setup, and repoints u_conn->net. The only behavioral difference is that the helper additionally invokes http_client_update_connection_io_timeout(), which is a no-op here since response_timeout/read_idle_timeout are still 0 at this point, so it should be safe. Feel free to defer if you prefer to keep this PR minimally scoped.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@src/flb_http_client.c` around lines 1020 - 1032, The duplicated net-setup
logic in create_http_client (the if (u_conn != NULL) block that sets
c->original_net_setup, copies into c->request_net_setup and rewires u_conn->net)
should be replaced by a call to the existing helper
http_client_bind_connection(c, u_conn); remove the manual c->u_conn = u_conn
assignment that precedes it so the helper can set c->u_conn itself; keep in mind
http_client_bind_connection also calls http_client_update_connection_io_timeout
(which is harmless here because response_timeout/read_idle_timeout are 0), so
simply call http_client_bind_connection(c, u_conn) instead of duplicating the
selection and rewire code.

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Nitpick comments:
In `@src/flb_http_client.c`:
- Around line 1020-1032: The duplicated net-setup logic in create_http_client
(the if (u_conn != NULL) block that sets c->original_net_setup, copies into
c->request_net_setup and rewires u_conn->net) should be replaced by a call to
the existing helper http_client_bind_connection(c, u_conn); remove the manual
c->u_conn = u_conn assignment that precedes it so the helper can set c->u_conn
itself; keep in mind http_client_bind_connection also calls
http_client_update_connection_io_timeout (which is harmless here because
response_timeout/read_idle_timeout are 0), so simply call
http_client_bind_connection(c, u_conn) instead of duplicating the selection and
rewire code.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: f71e2bd1-e11f-444f-bf3d-d8ce25282e9e

📥 Commits

Reviewing files that changed from the base of the PR and between 856d762 and 2ebb0e7.

📒 Files selected for processing (2)

• run_code_analysis.sh
• src/flb_http_client.c

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Inline comments:
In @.github/workflows/unit-tests.yaml:
- Around line 116-117: In the unit-test workflow jobs that currently reference
the personal fork (the repository: mabrarov/fluent-bit-ci and ref:
feat/enable_tests entries in the unit-test job), revert those two fields to the
official CI repository and branch used elsewhere (e.g., repository:
fluent/fluent-bit-ci and the stable ref previously used) so both Linux and macOS
unit-test jobs match the other workflows (pr-perf-test.yaml and
call-run-integration-test.yaml); update the repository and ref values in the
unit-test job definitions to remove the personal fork and feature branch.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 434bfef2-66e6-4d3f-a46e-8e3cb76877ee

📥 Commits

Reviewing files that changed from the base of the PR and between 2ebb0e7 and cd068ee.

📒 Files selected for processing (2)

• .github/workflows/unit-tests.yaml
• run_code_analysis.sh

🚧 Files skipped from review as they are similar to previous changes (1)

• run_code_analysis.sh

[cosmo0920]

We need to use shorter than 80 characters commit messages:

❌ Commit cd068ee223 failed:
Commit subject too long (>80 chars): 'workflows: temporarily switched to CI scripts with enabled runtime tests for Forward and Elasticsearch output plugins, for Disk I/O metrics and Process metrics input plugins.'

❌ Commit 9664979ccd failed:
Commit subject too long (>80 chars): 'scripts: enabled runtime tests for Forward output plugin in dev code analysis script'


Commit prefix validation failed.
Error: Process completed with exit code 1.