Release Notes - v0.10.0¶
Reusable page-evidence capture with SPA-aware settling, plus a one-shot fetch
CLI command for grabbing a page without wiring up a session.
Highlights¶
pagecapture subpackage
A reusable package that captures normalized page evidence — screenshot, rendered
HTML, visible text, and structure (title, lang, links, headings) — for reuse
across accessibility, functional/journey, and internationalization consumers.
import "github.com/plexusone/w3pilot/pagecapture"
// Navigate, settle, and capture in one call.
ev, err := pagecapture.Capture(ctx, pilot, "https://example.com")
// Or capture the already-loaded page without navigating.
ev, err := pagecapture.CaptureCurrent(ctx, pilot)
// Or crawl same-host <a href> links.
site, err := pagecapture.CaptureSite(ctx, pilot, "https://example.com")
SPA-aware settling
The default SettleDOMStable polls the DOM until it stops changing and holds
real content, gated by a minimum-settle floor. Client-rendered apps never
reliably reach the load event or network-idle, and settling too early captures
a half-rendered shell — the floor plus consecutive-stable-sample check fixes
that.
cfg := pagecapture.Config{
SettleMode: pagecapture.SettleDOMStable, // default
MinSettle: 2500 * time.Millisecond, // floor before capture
}
fetch CLI command
Fetch a page's rendered text or HTML in one shot — launch, navigate, wait, and extract — without managing a session by hand.
w3pilot fetch https://example.com # visible text
w3pilot fetch https://example.com --html # full rendered HTML
w3pilot fetch https://spa.example.com --wait networkidle # wait for network to settle
w3pilot fetch https://example.com --format json # {url,title,content}
Upgrading¶
No breaking API changes. pagecapture is additive and depends only on the
w3pilot root package plus the standard library, so importing the root package is
not affected.
go get github.com/plexusone/[email protected]