Build/Test Tools: Trim the routine PHPUnit database matrix

[lancewillett]

Trac ticket: https://core.trac.wordpress.org/ticket/66069

Reduce the database versions tested on pushes, pull requests, and manual runs. Keep the complete database lists on the weekly schedule, following the existing PHP matrix pattern.

• MySQL: retain 5.7, 8.0, and 8.4 on routine runs; test 9.7 weekly.
• MariaDB: retain 5.5, 10.6, 10.11, 11.4, and 11.8 on routine runs; test 10.3 and 10.5 weekly.

Routine runs drop from 86 to 68 jobs (21% fewer); weekly runs remain at 182. These totals include shared Gutenberg preparation and exclude notification and failure-handling jobs. Memcached, alternate-domain, reporting, HTML API, innovation-release, and PHP prerelease configurations are preserved.

Label-triggered full-matrix runs on PRs are tracked separately in https://core.trac.wordpress.org/ticket/66071.

Validation

• actionlint .github/workflows/phpunit-tests.yml
• uvx --offline zizmor@1.24.1 --persona=regular --offline .github/workflows/phpunit-tests.yml
• Expanded the before/after matrices for push, pull request, manual, and scheduled events. Confirmed 86 → 68 for each routine event and 182 → 182 weekly; every retained job configuration matches exactly, including reporting and other additional configurations.

Use of AI tools

AI assistance: Yes
Tool(s): Codex
Model(s): GPT-6, GPT-5.6
Used for: Workflow edits, matrix validation, code review, and drafting this description. I take responsibility for the submitted change.

---

This pull request is for code review only. Please keep other discussion in the Trac ticket. Do not merge this pull request.

[github-actions]

The following accounts have interacted with this PR and/or linked issues. I will continue to update these lists as activity occurs. You can also manually ask me to refresh this list by adding the props-bot label.

Core Committers: Use this line as a base for the props when committing in SVN:

Props lancewillett, lucatume.

To understand the WordPress project's expectations around crediting contributors, please review the Contributor Attribution page in the Core Handbook.

[Copilot]

🟢 Approval recommended

The workflow logic change is straightforward and consistent with the existing scheduled-vs-routine matrix pattern, with only a minor comment wording clarification suggested.

Pull request overview

This PR updates the PHPUnit GitHub Actions workflow to reduce the routine database-version test matrix for MySQL and MariaDB while keeping the full matrix on the weekly scheduled run, aligning database coverage with the existing “full on schedule, trimmed otherwise” PHP matrix pattern.

Changes:

• Switch MySQL db-version to a scheduled full list vs. trimmed list for push/PR/manual runs.
• Switch MariaDB db-version to a scheduled full list vs. trimmed list for push/PR/manual runs.
• Minor comment reflow/clarification around the MariaDB matrix documentation.

File summaries

File	Description
.github/workflows/phpunit-tests.yml	Trims routine MySQL/MariaDB DB-version matrices while preserving the full weekly scheduled matrix.

Review details

• Files reviewed: 1/1 changed files
• Comments generated: 1
• Review effort level: Lite

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[lancewillett]

Updated the comment and routine MySQL selection in 6edd6f8. Routine runs now use 5.7, 8.0, and 8.4. One factual correction: MySQL 9.7 is an LTS release; the revised comment explains the coverage choice without misclassifying it.

---

_{Adversarial review · gpt-6}

[lucatume]

I would retouch the versions to match current usage, the MariaDB version suggestion is not blocking.

[lucatume]

MySQL 9.7 is 0.022% of reported installs and it isn't an LTS, so the new comment doesn't quite describe it either. 8.0 is 14.07%, the most used MySQL version there is; it's the one going weekly.

Swapping the two is free, same job count, and routine coverage goes from 63.8% to 77.9%.

Suggested change

	# The scheduled run tests the full database matrix. Other events test the oldest listed version and supported LTS releases.
	db-version: ${{ github.event_name == 'schedule' && fromJSON('["5.7","8.0","8.4","9.7"]') || fromJSON('["5.7","8.4","9.7"]') }}
	# The scheduled run tests the full database matrix. Other events test the oldest listed version, the most used release and the most recent LTS.
	db-version: ${{ github.event_name == 'schedule' && fromJSON('["5.7","8.0","8.4","9.7"]') || fromJSON('["5.7","8.0","8.4"]') }}

[lancewillett]

Applied the swap in 6edd6f8: routine MySQL runs use 5.7, 8.0, and 8.4, with 9.7 retained in the weekly matrix. The comment now describes the compatibility and deployment-coverage rationale.

---

_{Adversarial review · gpt-6}

[lucatume]

10.6 is 9.67% of installs, second only to 10.11 among the versions this job runs. It moves to the weekly run while 5.5 at 0.311% stays per-push.

This one isn't free: +6 jobs, so 86 -> 68 rather than 86 -> 62. Routine coverage goes to 87.6% and a 10.6 only regression doesn't sit on trunk until Sunday.

Suggested change

	# The scheduled run tests the full database matrix. Other events test the oldest listed version and supported LTS releases.
	db-version: ${{ github.event_name == 'schedule' && fromJSON('["5.5","10.3","10.5","10.6","10.11","11.4","11.8"]') || fromJSON('["5.5","10.11","11.4","11.8"]') }}
	# The scheduled run tests the full database matrix. Other events test the oldest listed version and the most used releases.
	db-version: ${{ github.event_name == 'schedule' && fromJSON('["5.5","10.3","10.5","10.6","10.11","11.4","11.8"]') || fromJSON('["5.5","10.6","10.11","11.4","11.8"]') }}

[lancewillett]

Added MariaDB 10.6 to routine runs in 6edd6f8. Matrix expansion confirms 68 routine jobs, down from 86, and the same 182 weekly jobs. Reporting, memcached, alternate-domain, innovation, and PHP prerelease configurations are unchanged. Updated the PR description with the new lists and counts.

---

_{Adversarial review · gpt-6}