Checkpoint Lab — Logs, Reports, output.xml, Rebot, and Result Post-Processing

1. Checkpoint scenario

You own a release-evidence pipeline for a synthetic product. Two independent smoke suites must be reported together. A separate release-gate suite intentionally fails once, is rerun with repaired synthetic input, and must be merged so the final view reflects the corrected test while retaining the original failure. Your deliverable is not merely “all green”: it is a reproducible evidence packet explaining every artifact and transformation.

2. Setup and preflight

python --version
python -m robot --version
python -m robot.rebot --version

• Use Robot Framework 7.4.2 stable for the chapter baseline and Python 3.8+.
• Use a new disposable directory such as rf14-checkpoint.
• No network service, external library, Pabot, browser, database, SSH, CI provider, container, or paid platform is required.
• Use only synthetic values. Do not introduce real credentials or personal data.
• Create sources/, tools/, and evidence/; do not point --outputdir at an existing shared results directory.

3. Predict before execution

4. Create the checkpoint sources

*** Test Cases ***
API Contract Fixture Is Coherent
    [Tags]    checkpoint    independent
    Should Be Equal    schema-v1    schema-v1

Save as sources/a.robot.

*** Test Cases ***
Package Fixture Is Coherent
    [Tags]    checkpoint    independent
    Should Be Equal As Integers    7    7

Save as sources/b.robot.

*** Variables ***
${STATE}    degraded

*** Test Cases ***
Release Gate
    [Tags]    checkpoint    rerun
    Should Be Equal    ${STATE}    healthy    msg=Synthetic release state must be healthy

Stable Invariant
    [Tags]    checkpoint    rerun
    Should Be Equal    stable    stable

Save as sources/release.robot. Reuse the read-only inspector/checksum/privacy scripts from Lesson 2, or recreate them from that lesson.

5. Execute and combine the independent evidence

python -m robot --outputdir evidence/a sources/a.robot
python -m robot --outputdir evidence/b sources/b.robot
python -m robot.rebot   --name "Checkpoint Independent Bundle"   --output combined.xml   --outputdir evidence/combined   evidence/a/output.xml evidence/b/output.xml

Verify the raw inputs still exist. Inspect combined.xml and confirm two tests beneath a new aggregate root. Record the exact command and process return code.

6. Capture the controlled failure before rerunning

python -m robot   --variable STATE:degraded   --outputdir evidence/release/original   sources/release.robot
# Preserve the expected non-zero status and first-failure artifacts.

python -m robot   --rerunfailed evidence/release/original/output.xml   --variable STATE:healthy   --outputdir evidence/release/rerun   sources/release.robot

Before merging, independently inspect both XML files. The original must show Release Gate FAIL and Stable Invariant PASS. The rerun should contain only Release Gate and it should PASS. If those observations differ, stop and diagnose selection/input identity rather than forcing a merge.

7. Intentionally make—and diagnose—the wrong aggregation choice

# Diagnostic exercise only: combining a rerun is semantically wrong for the final view.
python -m robot.rebot   --name "WRONG Rerun Bundle"   --output wrong-combined.xml   --outputdir evidence/release/wrong   evidence/release/original/output.xml   evidence/release/rerun/output.xml

Inspect the result. You have preserved both roots instead of replacing the failed test in the same logical suite. This can make a reviewer see both the old failure and the rerun as independent evidence. Keep this artifact only as the diagnostic example, then use --merge for the production-correct derivative.

8. Produce and verify the corrected merged result

python -m robot.rebot   --merge   --output merged.xml   --outputdir evidence/release/merged   evidence/release/original/output.xml   evidence/release/rerun/output.xml

Verify: same logical top-level suite, two total tests, both final PASS, replacement note for the rerun test, and the original failure still preserved under evidence/release/original. A green merged report is acceptable only because the evidence packet also explains the earlier failure and rerun.

9. Generate publication derivatives

python -m robot.rebot   --xunit xunit.xml   --outputdir evidence/release/publish   evidence/release/merged/merged.xml

This creates default log/report plus xUnit under the publication directory. For a restricted CI system you may publish all three. For a broader audience, publish only artifacts allowed by your privacy policy; do not assume HTML is inherently safe.

10. Verification checklist

• Record generator, generated, schemaversion, top suite, test count, and statuses for each raw/combined/merged XML.
• Confirm schema version is expected for the pinned Robot Framework version.
• Confirm independent combination created a new parent rather than replacing tests.
• Confirm the rerun executed only the original failed test.
• Confirm merge restored the full logical suite and final test statuses.
• Confirm original failure artifacts still exist unchanged.
• Record sizes and SHA-256 checksums for raw and derived artifacts.
• Open the log/report and verify hierarchy/status visually.
• Confirm xUnit exists and its test count matches the intended final view.
• Run the synthetic forbidden-marker scan over publishable artifacts.

11. Write the CI retention/redaction policy

12. Cleanup / rollback

Cleanup is intentionally boring: after the evidence review, remove only the run-owned rf14-checkpoint directory. Do not delete parent directories, shared CI workspaces, or repository source. In a real CI pipeline, cleanup happens only after required raw and derivative artifacts have been validated and uploaded to their correct retention tier.

13. What Chapter 14 adds to the production operating model

Chapter 13 made status truthful; Chapter 14 makes that truth durable. Your operating model now includes raw machine-readable evidence, explicit presentation derivatives, schema/version checks, combine-versus-merge identity rules, first-failure retention, artifact checksums, privacy/access tiers, and reproducible Rebot transformations. This is the foundation Chapter 15 will use when selecting tests, rerunning failures, and defining execution profiles.

14. Knowledge check

15. Checkpoint complete

You have built a complete local evidence pipeline: separate raw runs, XML inspection, correct independent combination, preserved controlled failure, failed-test rerun, diagnosed wrong aggregation, correct merge, report/log/xUnit derivatives, checksums, privacy checks, and retention policy. The result is auditable because every transformation is explicit and the original evidence remains attributable.

Further reading

• Robot Framework 7.4.2 User Guide — Post-processing outputs — Rebot, combine, merge, and JSON/XML result processing.
• Robot Framework 7.4.2 User Guide — Different output files — output, log, report, xUnit, and output-directory behavior.
• Robot Framework 7.4.2 User Guide — Removing and flattening keywords — output-size and evidence trade-offs.
• Robot Framework 7.4.2 result.xsd — authoritative XML schema version and root attributes.
• Robot Framework 7.4.2 release notes — stable release details and built-in Testdoc deprecation.
• Robot Framework 7.4.2 on PyPI — pinned package metadata and Python requirement.
• Robot Framework releases — re-check stable/pre-release status when updating this chapter.