Verification

VerificationTest(input, expected, …) — @enumeratio/aestimatio's reference-example primitive (design/aestimatio.md §4). It evaluates input under the given constraints, compares the result with expected (default: structural sameness), and returns a TestResultObject: the outcome (Success, Failure, Error or Aborted), the input and actual output, and how long it took. <notatio-test-result-object> draws it as an In/Out pair with an outcome badge.

#Success

No mismatch, no message: a passing test just shows its In/Out.

source · editable

<notatio-cell value="VerificationTest(PowerModList(3, 1/2, 11), [5, 6])" />

#Failure

The badge names what was expected; the Out line still shows what actually came back.

source · editable

<notatio-cell value="VerificationTest(PowerModList(3, 1/2, 11), [1, 2])" />

#Aborted, under a time constraint

A 30-digit semiprime's factorization outruns a 0.05s TimeConstraint: the badge reads Aborted, and there is no Out line to show.

source · editable

<notatio-cell value="VerificationTest(FactorInteger(100000001000039100000310002511), [[100000000000031, 1], [1000000010000081, 1]], TimeConstraint -> 0.05)" />

Options

SameTest, TimeConstraint, MemoryConstraint and TestID are trailing rules, Wolfram's way. SameTest takes any predicate of two arguments, applied instead of structural sameness. MemoryConstraint is enforced only inside the isolated (worker) evaluator — in-process, asking for one turns the test itself into an Error rather than silently skipping the bound (see @enumeratio/aestimatio's own comment on MemoryConstrained).

#TestID

source · editable

<notatio-cell value='VerificationTest(2 + 2, 4, TestID -> "arithmetic-sanity")' />

In the browser: evaluateInWorker

@enumeratio/aestimatio/browser runs one evaluation in a plain dedicated Worker — the browser counterpart of @enumeratio/aestimatio/node's evaluateIsolated. A hard time limit is a real terminate(); a memory limit is best-effort only, since a browser gives a worker no memory cap to set — the host polls performance .measureUserAgentSpecificMemory() (or Chromium's performance.memory where that isn't available) and terminates once it reads past the bound. That reading is the page's memory, not the worker's alone, so it is a guardrail, not the real per-process cap evaluateIsolated gets from worker_threads' resourceLimits.

Wiring a live demo into this page (spinning up an actual Worker from the docs bundle) is deferred past this PR — see packages/aestimatio/tests/browser.test.ts for evaluateInWorker's host-side logic instead, exercised with a fake Worker so it runs without a browser at all: a normal resolution, a timeMs kill, a signal abort, and a memoryBytes bound tripped by an injected measureMemory.

As a Vue component

#The Vue wrapper

source

<TestResultObject outcome="Success" input="1 + 1" actual="2" time="0" />