Turbo Drive cached HTML served without cache-control invalidation on server mutation

Turbo Drive caches GET responses to avoid full page reloads, but the framework does not automatically invalidate these caches when server-side data changes without corresponding cache-control directives. This can result in browser-reused stale HTML being presented during Turbo Drive navigation, even after a successful request. The interaction between Turbo’s request headers and standard browser caching mechanisms determines whether a fresh response is fetched or a prior response is served.

The research brief highlights version-dependent cache-control behavior, but exact differences across Turbo releases and browser engines are not verified.

• Under what conditions does Turbo Drive issue conditional requests versus plain GET navigations when cached responses are available?
• How do server-supplied Cache-Control or ETag headers influence browser cache hit decisions in conjunction with Turbo’s X-Turbo-Drive headers?
• Is there a header or meta-tag configuration that forces Turbo Drive to bypass cache without appending query parameters to every link?