100% Latest & Updated Microsoft GH-300 Practice Test Questions, Exam Dumps & Verified Answers!
30 Days Free Updates, Instant Download!
Question 360
An engineer is managing history and context in a long Copilot conversation. The developer wants a final answer using only facts shown in supplied logs. What should the developer do before accepting the proposed result?
- Identify the specific latest approved checkpoint and outstanding goal.
- Mark that assumption superseded and use verified results.
- Reintroduce a concise verified constraints summary before the next task.
- Constrain conclusions to those logs and identify unsupported gaps.
Explanation
Correct Answer
D:Source-bounded responses should not silently invent missing operational details.
Incorrect Answers
A:Continuation should preserve the correct scope and decisions. This addresses a different situation: A user says 'continue' but the prior answer contained three unrelated tasks.
B:Updated evidence should replace stale speculative context. This addresses a different situation: The conversation contains an earlier assumption later disproved by tests.
C:Explicit state refresh reduces accidental loss of requirements. This addresses a different situation: Context limits cause a key requirement to disappear from the active conversation.
Question 361
A developer asks Copilot to generate unit tests for a discount function. The current tests cover only a positive percentage discount. Which action is the most appropriate response to this situation?
- Replace them with synthetic deterministic fixtures.
- Add the expired-promotion scenario and verify the rejection path.
- Add zero, boundary, and invalid percentage cases with explicit expected outcomes.
- Inject or freeze the clock used in the unit test.
Explanation
Correct Answer
C:Boundary conditions reveal defects missed by happy-path tests.
Incorrect Answers
A:Test coverage does not require exposing personal customer data. This addresses a different situation: The suggested test uses production customer records as input fixtures.
B:Business-rule coverage must include invalid-state transitions. This addresses a different situation: The test suite passes but omits the case for an expired promotion.
D:Deterministic time makes behavior reproducible. This addresses a different situation: The discount function depends on current time and tests fail intermittently.
Question 362
A developer asks Copilot to generate unit tests for a discount function. A generated test asserts only that a function does not throw. What should the developer do before accepting the proposed result?
- Inject or freeze the clock used in the unit test.
- Assert the exact resulting price and rounding behavior.
- Add the expired-promotion scenario and verify the rejection path.
- Add zero, boundary, and invalid percentage cases with explicit expected outcomes.
Explanation
Correct Answer
B:A nonthrowing result does not establish the business rule.
Incorrect Answers
A:Deterministic time makes behavior reproducible. This addresses a different situation: The discount function depends on current time and tests fail intermittently.
C:Business-rule coverage must include invalid-state transitions. This addresses a different situation: The test suite passes but omits the case for an expired promotion.
D:Boundary conditions reveal defects missed by happy-path tests. This addresses a different situation: The current tests cover only a positive percentage discount.
Question 363
A developer asks Copilot to generate unit tests for a discount function. The suggested test uses production customer records as input fixtures. Which response best addresses the stated risk or requirement?
- Add the expired-promotion scenario and verify the rejection path.
- Replace them with synthetic deterministic fixtures.
- Assert the exact resulting price and rounding behavior.
- Add zero, boundary, and invalid percentage cases with explicit expected outcomes.
Explanation
Correct Answer
B:Test coverage does not require exposing personal customer data.
Incorrect Answers
A:Business-rule coverage must include invalid-state transitions. This addresses a different situation: The test suite passes but omits the case for an expired promotion.
C:A nonthrowing result does not establish the business rule. This addresses a different situation: A generated test asserts only that a function does not throw.
D:Boundary conditions reveal defects missed by happy-path tests. This addresses a different situation: The current tests cover only a positive percentage discount.
Question 364
A developer asks Copilot to generate unit tests for a discount function. The discount function depends on current time and tests fail intermittently. What is the most defensible next step under responsible development practices?
- Assert the exact resulting price and rounding behavior.
- Add zero, boundary, and invalid percentage cases with explicit expected outcomes.
- Replace them with synthetic deterministic fixtures.
- Inject or freeze the clock used in the unit test.
Explanation
Correct Answer
D:Deterministic time makes behavior reproducible.
Incorrect Answers
A:A nonthrowing result does not establish the business rule. This addresses a different situation: A generated test asserts only that a function does not throw.
B:Boundary conditions reveal defects missed by happy-path tests. This addresses a different situation: The current tests cover only a positive percentage discount.
C:Test coverage does not require exposing personal customer data. This addresses a different situation: The suggested test uses production customer records as input fixtures.
Question 365
A developer asks Copilot to generate unit tests for a discount function. The test suite passes but omits the case for an expired promotion. Which recommendation directly resolves this issue?
- Replace them with synthetic deterministic fixtures.
- Add the expired-promotion scenario and verify the rejection path.
- Assert the exact resulting price and rounding behavior.
- Inject or freeze the clock used in the unit test.
Explanation
Correct Answer
B:Business-rule coverage must include invalid-state transitions.
Incorrect Answers
A:Test coverage does not require exposing personal customer data. This addresses a different situation: The suggested test uses production customer records as input fixtures.
C:A nonthrowing result does not establish the business rule. This addresses a different situation: A generated test asserts only that a function does not throw.
D:Deterministic time makes behavior reproducible. This addresses a different situation: The discount function depends on current time and tests fail intermittently.
Question 366
A service owner requests integration tests for an order API. The team has these independent requirements: (1) A test mocks every dependency and never exercises the database boundary. (2) A failed payment should not leave an order in a paid state. Which TWO actions together satisfy all requirements? Choose TWO.
- Integrate the tests into the supported CI workflow and observe results.
- Add a controlled integration test using the actual persistence interface.
- Use approved test doubles or a sanctioned sandbox endpoint.
- Assert the rollback or compensation behavior across service boundaries.
- Reset or isolate the test data before each scenario.
Explanation
Correct Answers
B:Integration testing checks interactions that isolated mocks can hide.
D:Integration correctness includes transactional failure handling.
Incorrect Answers
A:Repeatable automation makes integration failures visible to reviewers. This addresses a different situation: The test suite runs only locally and nobody knows its CI status.
C:External production dependencies can make test suites slow, costly, and unsafe. This addresses a different situation: A generated test relies on a paid external API endpoint during CI.
E:Test isolation prevents order-dependent failures. This addresses a different situation: The test starts with leftover records from earlier runs.
Question 367
A service owner requests integration tests for an order API. The test starts with leftover records from earlier runs. Which response best addresses the stated risk or requirement?
- Add a controlled integration test using the actual persistence interface.
- Assert the rollback or compensation behavior across service boundaries.
- Integrate the tests into the supported CI workflow and observe results.
- Reset or isolate the test data before each scenario.
Explanation
Correct Answer
D:Test isolation prevents order-dependent failures.
Incorrect Answers
A:Integration testing checks interactions that isolated mocks can hide. This addresses a different situation: A test mocks every dependency and never exercises the database boundary.
B:Integration correctness includes transactional failure handling. This addresses a different situation: A failed payment should not leave an order in a paid state.
C:Repeatable automation makes integration failures visible to reviewers. This addresses a different situation: The test suite runs only locally and nobody knows its CI status.
Question 368
A service owner requests integration tests for an order API. A generated test relies on a paid external API endpoint during CI. What is the most defensible next step under responsible development practices?
- Use approved test doubles or a sanctioned sandbox endpoint.
- Reset or isolate the test data before each scenario.
- Add a controlled integration test using the actual persistence interface.
- Integrate the tests into the supported CI workflow and observe results.
Explanation
Correct Answer
A:External production dependencies can make test suites slow, costly, and unsafe.
Incorrect Answers
B:Test isolation prevents order-dependent failures. This addresses a different situation: The test starts with leftover records from earlier runs.
C:Integration testing checks interactions that isolated mocks can hide. This addresses a different situation: A test mocks every dependency and never exercises the database boundary.
D:Repeatable automation makes integration failures visible to reviewers. This addresses a different situation: The test suite runs only locally and nobody knows its CI status.
Question 369
A service owner requests integration tests for an order API. A failed payment should not leave an order in a paid state. Which recommendation directly resolves this issue?
- Add a controlled integration test using the actual persistence interface.
- Reset or isolate the test data before each scenario.
- Assert the rollback or compensation behavior across service boundaries.
- Use approved test doubles or a sanctioned sandbox endpoint.
Explanation
Correct Answer
C:Integration correctness includes transactional failure handling.
Incorrect Answers
A:Integration testing checks interactions that isolated mocks can hide. This addresses a different situation: A test mocks every dependency and never exercises the database boundary.
B:Test isolation prevents order-dependent failures. This addresses a different situation: The test starts with leftover records from earlier runs.
D:External production dependencies can make test suites slow, costly, and unsafe. This addresses a different situation: A generated test relies on a paid external API endpoint during CI.
Question 370
A service owner requests integration tests for an order API. The test suite runs only locally and nobody knows its CI status. Which action is the most appropriate response to this situation?
- Reset or isolate the test data before each scenario.
- Assert the rollback or compensation behavior across service boundaries.
- Use approved test doubles or a sanctioned sandbox endpoint.
- Integrate the tests into the supported CI workflow and observe results.
Explanation
Correct Answer
D:Repeatable automation makes integration failures visible to reviewers.
Incorrect Answers
A:Test isolation prevents order-dependent failures. This addresses a different situation: The test starts with leftover records from earlier runs.
B:Integration correctness includes transactional failure handling. This addresses a different situation: A failed payment should not leave an order in a paid state.
C:External production dependencies can make test suites slow, costly, and unsafe. This addresses a different situation: A generated test relies on a paid external API endpoint during CI.
GitHub Copilot (GH-300) Practice Tests
