GripMock

Appearance

Verify API ​

Asserts that an endpoint was called an exact number of times. It reads the same store as the History API, so it needs HISTORY_ENABLED (the default); with recording off every check reports 0 actual calls.

Request ​

  • Method: POST
  • URL: /api/verify
bash
curl -X POST http://127.0.0.1:4771/api/verify \
  -H 'Content-Type: application/json' \
  -d '{"service":"helloworld.Greeter","method":"SayHello","expectedCount":3}'

Response ​

200 when the count matches:

json
{"message": "ok", "time": "2026-08-19T09:20:55Z"}

400 when it does not — the body carries both numbers, so a script can report the difference without re-reading the history:

json
{
  "expected": 3,
  "actual": 7,
  "message": "expected helloworld.Greeter/SayHello to be called 3 times, got 7"
}

The comparison is exact: expectedCount: 3 fails on two calls and on four. Use GET /api/history?service=…&method=… when you need a range or just the evidence.

Session scope ​

X-Gripmock-Session counts that session's calls plus the global ones — the same scope its stubs match in:

bash
curl -X POST http://127.0.0.1:4771/api/verify \
  -H 'Content-Type: application/json' \
  -H 'X-Gripmock-Session: team-a' \
  -d '{"service":"helloworld.Greeter","method":"SayHello","expectedCount":2}'

Without the header every call is counted, including calls made under a session. On a shared server that makes an unscoped assertion sensitive to a colleague's traffic: give each test its own session and the count becomes stable.

Counting model ​

Verify counts per service and method. The embedded SDK's ExpectationsWereMet counts per stub, which is why two stubs on one method (say Once() and Twice()) are checked independently there but add up here.