Test Case 2: Deflake Success

[oxve]

Verifying flaky test retry success.

Bug: 551996294

[github-actions]

🤖 Gemini Suggested Commit Message

---

ci: Mock CI steps to verify flaky test retries

This change replaces standard build and test actions with mocked
implementations using a new helper script and configuration. These
mocks allow the CI to simulate transient failures on specific attempts
to verify that the pipeline correctly handles and recovers from flaky
test scenarios.

By using controlled mocks instead of actual builds, the infrastructure
retry logic can be validated quickly and reliably without depending
on natural test flakiness.

Bug: 551996294

---

💡 Pro Tips for a Better Commit Message:

1. Influence the Result: Want to change the output? You can write custom prompts or instructions directly in the Pull Request description. The model uses that text to generate the message.
2. Re-run the Generator: Post a comment with: /generate-commit-message

[github-actions]

Context	State	Description
CI / android	✅ PASS	Passed
CI / linux	✅ PASS	Passed

[gemini-code-assist]

Code Review

This pull request mocks several GitHub Actions steps—including GN generation, Ninja builds, browser tests, on-host tests, and artifact archiving—by replacing their actual execution with a Python helper script (mock_step.py) and a configuration file (mock_config.json) to simulate build and test outcomes. The review feedback focuses on improving the robustness of the helper script, specifically by resolving the configuration path relative to the script location, safely handling empty directory paths during directory creation, and defensively casting configuration values to prevent type errors.

[gemini-code-assist]

Using os.environ.get('GITHUB_WORKSPACE', '.') to resolve the configuration file path can fail if the script is executed from a different working directory (such as cobalt/src as configured in the build action) and GITHUB_WORKSPACE is not set or points elsewhere. Resolving the path relative to the script's own location (__file__) is more robust and ensures it works consistently across different execution environments.

Suggested change

	config_path = os.path.join(
	os.environ.get('GITHUB_WORKSPACE', '.'),
	'.github/config/mock_config.json')
	script_dir = os.path.dirname(os.path.abspath(__file__))
	repo_root = os.path.abspath(os.path.join(script_dir, '..', '..'))
	config_path = os.path.join(repo_root, '.github/config/mock_config.json')

[gemini-code-assist]

If filepath is a simple filename without any directory components (e.g., "result.xml"), os.path.dirname(filepath) will return an empty string "". Calling os.makedirs("", exist_ok=True) will raise a FileNotFoundError in Python. Guarding the directory creation prevents this crash.

Suggested change

	os.makedirs(os.path.dirname(filepath), exist_ok=True)
	dirname = os.path.dirname(filepath)
	if dirname:
	os.makedirs(dirname, exist_ok=True)

[gemini-code-assist]

Defensive programming: rule.get('fail_until_attempt', 0) might return a non-integer type (such as a string or None if configured incorrectly or set to null in JSON). Comparing an integer attempt with a non-integer using <= will raise a TypeError in Python 3. Casting and handling potential conversion errors ensures the script is resilient to malformed configuration.

    try:
      fail_until = int(rule.get('fail_until_attempt', 0))
    except (TypeError, ValueError):
      fail_until = 0
    if attempt <= fail_until: