@marginallyuseful/soll-playwright

Playwright integration for soll — scope E2E test runs to affected files.

Usage

Compose into your existing fixtures:

import { sollFixtures } from "@marginallyuseful/soll-playwright";
import { test as base } from "@playwright/test";

export const test = base.extend(sollFixtures);

Or use the convenience export (no custom fixtures):

import { test, expect } from "@marginallyuseful/soll-playwright";

How it works

The fixture handles everything automatically:

1. Selection: loads test→chunk edges from the edge store, checks chunk content hashes for validity, delegates to soll affected for graph traversal
2. Skip: calls testInfo.skip() for unaffected tests
3. Recording: instruments page.on('response') to capture in-scope URLs
4. Persistence: writes test→chunk edges with source file annotations after each test

Configuration

export const test = base.extend({
  ...sollFixtures,
  sollConfig: [
    {
      alwaysRun: ["**/visual-*.spec.ts"],
      runAllTriggers: ["playwright.config.*", ".env*"],
      inScopeOrigins: ["http://api.local:4000"],
    },
    { option: true },
  ],
});

Requirements

• A bundler plugin (@marginallyuseful/soll-vite, etc.) must be configured to produce the chunk overlay. Without it, the fixture has no chunk→source data and runs all tests.
• First run is always a full run (cold start) — it builds the test→chunk edge data that subsequent runs use for selection.