* feat(javascript): compare static endpoints with HAR captures * test(javascript): preserve original HAR fixture bytes during formatting * test(javascript): load the byte-preserved HAR fixture
Static endpoint and capture fixture
This source-owned loopback application was recorded on Linux with Playwright
1.64.1 and Chromium headless shell 153.0.8010.12. recording.har.txt is the original
HAR 1.2 from browser.newContext({recordHar: {mode: "full", content: "embed"}});
its SHA-256 is f0e7d2c869c811d4272a7b2916c6531c8c6cc7938336a35575e2a6b7b6346714.
The .txt suffix preserves the recorder bytes without JSON formatting; the test
copies it to capture.har before production decoding.
The recorded origin is http://127.0.0.1:42845. Do not reserialize the HAR.
The server returned the two adjacent text files as /app.js and /, and JSON
for the API paths. It streamed each response in two chunks 40 ms apart. The
browser awaited all three fetch response bodies before context close. The five
recorded requests are /, /app.js, /api/users?ordinary=kept-source-only,
/users/123 and /observed-only. The script's unused function bodies were
never called. Transport authentication headers and cookies are absent.
The public acceptance test copies these sources into a temporary application and runs the production static analyzer, HAR decoder, comparison and retained views through CLI/MCP. It replays this recording offline and needs no browser or network. The recording proves only this small application's selected run.