QA TEMPLATE / SAMPLE

Bug Report

Defect report | Sequence overwrite, impact and disposition

Synthetic Order Book teaching scenario, not a real exchange test. Based on this site's stale-sequence Fail Demo. Prices, quantities and times are supplemental teaching assumptions. Runs, defects and evidence names are synthetic; there was no real product execution, attached evidence or customer approval.

What this document answers

Record triggers, expected / actual differences, impact and disposition. Keep defect severity separate from the overall exit decision.

Blank template

[Required] applies to every case. [Conditional] is required when applicable; explain when it is not. Copy the text into your case documents and replace the brackets. This static page has no editing or saving function.

[Required] Defect ID / title: [trigger + observed impact]
[Required] Version / environment / time / reporter: [ ]
[Required] QA-proposed Severity: [S1 / S2 / S3 / S4]; impact evidence: [ ]
[Required] Client decision: [pending / severity, rationale, decision owner, date]
[Conditional] Fix Priority: [separate field, pending client decision]
[Required] Trace: requirement / oracle version, risk, Scenario, Case / version, Run, data row, failed Step / CP
[Required] Preconditions / Test Data: [ ]
[Required] Reproduction steps: [numbered; minimal and complete]
[Required] Expected Result / source: [ ]
[Required] Actual Result: [observable difference, not merely Fail]
[Required] Impact / known scope / reproduction rate: [successful reproductions / actual attempts]
[Required] Evidence: [raw artifact location / time / sanitization; a summary alone is insufficient]
[Required] Cause: [product / test / environment / data / unknown; do not guess without evidence]
[Required] Defect status / owner / next action: [ ]
[Conditional] Fix version, retest Run, regression scope and risk rationale: [ ]
[Conditional] Deferral acceptance: [authority, reason, residual risk, owner, review date]
Overall results and exit belong in the Quality Report. This defect references relevant Runs and evidence only.

Filled example

DEF-MDQ-004 | Old sequence 103 overwrites an Order Book already at 104

Retains the synthetic defect ID from this site's Fail Demo. Status Triaged · demo; QA-proposed S1. No customer decision; root cause Unknown.

Field	Content
Version / environment / report	build-demo-1 / DEMO-BOOK; DEMO-TESTER (fictional). For time and evidence, see DEMO-RUN-FAIL-001.
Case / data / checkpoint	TC-MDQ-017 v1.2 / STALE-103 / attempt 1 / S2～S3、CP2～CP3.
Reproduction steps	Load SNAP-104; send E-103 once; read complete state after event completion. Baseline and values follow DEMO-DC-MDQ v1.0.
Expected / actual	Expect rejection with sequence=104, bid 100×2, ask 101×3 unchanged. Synthetic actual: event accepted, sequence=103, bid 100×9, ask 101×3.
Impact and reproduction	Market-data state regresses and may mislead downstream decisions; no real trade or loss is claimed. This document expands one synthetic attempt only. Synthetic reproduction statements elsewhere on the site are not a real reproduction rate.
Ownership / disposition	DEMO-ENGINEER must investigate; speculation is not root cause. Stop the case, preserve snapshots and append a Run after a fix. No fix or retest evidence exists.

Severity and lifecycle

Field	Rule
Severity	S1: core-state correctness or integrity compromised; S1 is proposed here. S2: important function fails. S3: limited impact with a workaround. S4: minor issue. In real work, the designated authority decides and records the rationale.
Priority	Case P0 determines test order. The owner assigns defect fix Priority separately; P0 cannot substitute for Severity.
Status	New → Triaged → In Progress → Fixed → Verified → Closed; a failed retest returns to Reopened. Fixed is a fix claim requiring a new Run as evidence. This example remains Triaged.

Retest and regression

After a fix, append a Run using the original TC-MDQ-017 input and check CP2 / CP3. Regress TC-MDQ-016 to ensure normal updates are not rejected, and TC-MDQ-018 to protect full reconnect replacement. Confirm oracles before adding equal sequences, gaps or reordering. Do not mark unexecuted plans Pass.

Related records

The original failure remains in the Run Log; the Quality Report cites this defect and batch evidence. See the Quality Report.