Previous Results¶

The prev reference gives your script access to the result of the previous run. This lets you compare current behaviour against past behaviour --- detecting regressions, tracking trends, or suppressing duplicate notifications.

Providing Previous Results¶

Previous results are supplied per run, from a JSON file passed on the CLI (--prev is a short alias):

lacelang-executor run script.lace --vars vars.json --prev-results last_result.json

The file must contain a result JSON from a prior run of the same script (see the result structure). The library APIs of the reference executors can also feed each probe's last result back in automatically.

Access Syntax¶

prev mirrors the result structure. Access fields with dot notation and array indexing:

// Top-level outcome
prev.outcome                         // "success", "failure", "timeout", or null

// Run-scope variables from the previous run
prev.runVars.token
prev.runVars.count_before

// Individual call results
prev.calls[0].outcome                // "success", "failure", etc.
prev.calls[0].response.status        // HTTP status code
prev.calls[0].response.responseTimeMs

// Assertion results from the previous run
prev.calls[0].assertions[0].outcome  // "passed", "failed", "indeterminate"
prev.calls[0].assertions[0].actualLhs

Null When Missing¶

When no previous results are provided, prev resolves to null. Any field access on null propagates null --- it does not throw an error.

// If --prev-results was not provided:
prev.outcome              // null
prev.calls[0].response    // null (null propagation)
prev.runVars.token        // null (null propagation)

This means you can safely reference prev without checking whether it exists. The null will propagate through to your assertion, where it follows the standard null semantics:

prev.outcome eq null --- true (no previous results)
prev.outcome eq "success" --- false (null neq any value)
prev.calls[0].response.status lt 300 --- indeterminate (null in ordered comparison)

Practical Examples¶

Compare response time against previous run¶

get("$BASE_URL/api/search")
.expect(status: 200)
.assert({
  check: [
    // Warn if response time is more than double the previous run
    this.responseTime lt (prev.calls[0].response.responseTimeMs * 2)
  ]
})

Check if the previous run failed¶

get("$BASE_URL/health")
.expect(status: 200)
.assert({
  check: [
    // Soft-check: was the previous run also successful?
    prev.outcome eq "success"
  ]
})

Use a previous run-scope variable¶

get("$BASE_URL/api/metrics")
.expect(status: 200)
.assert({
  check: [
    // Verify count hasn't decreased since last run
    this.body.count gte prev.runVars.current_count
  ]
})
.store({ "$$current_count": this.body.count })

Validator warning

The validator emits a warning (PREV_WITHOUT_RESULTS) if your script uses prev references but --prev-results was not provided. The script will still run --- prev will just be null.