Comparison
Client-side routingvsSPA fallback
Client-side routing
you intercepted the link click, pushed a new URL and rendered a different component in place.
Mapping URLs to components in the browser, updating history without a document request. It is what makes an application feel instant between screens, and it is the source of the whole category of bugs where the browser and the application disagree about where you are. The server still needs to serve the application for every routable URL, or a direct visit or refresh returns a 404.
Full entry →SPA fallback
the deep link worked from inside the app and returned a 404 when you refreshed the page.
A server rule serving the application's HTML for any path the router might own, so a direct request to a client-side route reaches the application instead of a missing file. Every static host has its own way to configure it, and forgetting it is the classic first deploy of a single-page application. Done bluntly it also turns genuine 404s into a 200 with an empty page, which is bad for crawlers and worse for monitoring.
Full entry →