Just to be clear I have *read* http://code.google.com/chrome/whitepapers/prerender.html So my understanding of pretender is based on this whitepaper. :)
Le 13 août 2014 à 10:09, Jonas Sicking <jo...@sicking.cc> a écrit : >> And what would be the fallback for non JS user agents (cue accessibility and >> HTTP scripting here)? > > The same as any other time when prerendering is used. The fallback > would simply be that the next page is loaded at normal speed, rather > than being prerendered. The only thing that the page would have to do > special here when the initial page attempts to send a message to the > prerendered page about which bug's data to load, if the API for > talking to the message channel is missing, don't do anything. <!doctype html> <html> <title>bug</title> <link rel='prerender' href='http://bugzilla.mozilla.org/show_bug.cgi?justLoadTemplate=1'/> <a href="/bug0001">my next bug</a> and <a href="/bug0002">my next next bug</a> </html> In your scenario, this would not work without scripting. I can see how useful it could be with Firefox OS and transitions between different states of an application. On the Web, it means that we get yet another dependency on JS for this to be working for capturing the click event and sending a XHR with Accept: application/json. On the other hand if no JS, the server receives the HTTP GET requests on the resource, but then without pre-rendering. :) I guess it's why I'm asking questions to be sure how it would be working. I was asking questions about the caching because once http://bugzilla.mozilla.org/show_bug.cgi?justLoadTemplate=1 is downloaded, rendered and cached. It could be reused for all subsequent requests with the same pattern. But in return it means we create an issue for Web developers for developing and testing, or just modifying the cache. I guess people can rely to cache busting techniques such as timestamp parameters in URIs. > I'm definitely getting the impression that we're still talking past > each other. I guess maybe me or you are misunderstanding how the > prerendering spec is defined. I understand what you are saying. :) But maybe we are assuming different things for the Web and/or have read a different spec. It's ok. Thanks Jonas. No more questions. :) -- Karl Dubost, Mozilla http://www.la-grange.net/karl/moz _______________________________________________ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
