Skip to content

Incident Review Guide

Guide Maps

graph LR
  family["Reproducible Research"]
  program["Deep Dive Make"]
  guide["Capstone docs"]
  section["INCIDENT_REVIEW_GUIDE"]
  page["Incident Review Guide"]
  proof["Proof route"]

  family --> program --> guide --> section --> page
  page -.checks against.-> proof
flowchart LR
  case["Run incident-audit"] --> symptom["Read the captured failure"]
  symptom --> cause["Map it to the defect class"]
  cause --> repair["Name the repair rule"]
  repair --> capstone["Return to the main capstone boundary"]

Use this guide when you want one executed failure bundle instead of a list of repro file names. The point is to study a defect class with evidence, not to admire a broken Makefile in isolation.


Reading Order

Read an incident bundle in this order:

  1. route.txt
  2. README.md
  3. repro-case.mk
  4. command.txt
  5. output.txt
  6. summary.txt
  7. PROOF_GUIDE.md
  8. review-questions.txt

That route keeps the failure specimen ahead of the explanation, so you do not mistake the summary for the evidence.

Back to top


What Each Surface Gives You

File Why it exists
repro-case.mk the smallest preserved defect specimen
command.txt the exact command used to reproduce the incident
output.txt the concrete failure output or race symptom
summary.txt the named defect class and the repair lesson
PROOF_GUIDE.md the matching proof route in the real capstone
review-questions.txt prompts that force a learner to name cause and repair

Back to top


What A Good Incident Review Should Answer

By the end of one incident review, you should be able to say:

  • which output or edge was modeled dishonestly
  • whether the failure depends on -j or is visible in serial mode
  • which repair belongs in the main capstone rather than in the repro itself
  • which target in the real capstone would prove the repaired property

Back to top


Best Incident Sequence

If you are new to the repro pack, use this order:

  1. shared mutable output
  2. directory creation race
  3. order-only misuse
  4. generated-header modeling

That sequence moves from visibly broken outcomes to more subtle graph lies.

Back to top


Companion Surfaces

Use these when the incident bundle raises a bigger question:

  • REPRO_GUIDE.md for the broader failure-class catalog
  • PROOF_GUIDE.md for the corresponding production proof routes
  • ARCHITECTURE.md when the repair depends on ownership boundaries

Back to top