Jonas Sicking wrote: > On Dec 5, 2012 3:52 PM, "Dietrich Ayala" <[email protected]> > wrote: > > 3) What's the Web platform plan for fixing this? Is there one? > > Unfortunately the appcache suffers from a.other problem than a slow > gecko implementation. > > It's also hated by authors. > > So my plan was to come up with an alternative spec proposal and then > kill two birds in one rewrite. > > But I'be started to think that an alternative spec will take too long > to come up with even a draft for.
The thing I hate about AppCache is that it abuses HTTP, redefining HTTP semantics, and is more complex than necessary. The subtle, non-obvious differences between normal HTTP semantics and AppCache semantics has resulted in lots of bugs found (and fixed) in our HTTP implementation, and AppCache is probably the single biggest feature that contributes to complexity in our HTTP implementation. AFAICT, all that complexity is (or will become) superfluous. In particular, I think that it is *highly* likely that an AppCache-like thing is best implemented efficiently on top of the other (existing) Web Platform features + SPDY features like SPDY server push. For example, with SPDY (HTTP 2.0) server push, we can send a request to the home page of an app, and based on (e.g. the ETag of) that single request, the server can immediately tell us "Hey, you are going to need to download the following resources right now; better better get started." AFAICT, that's already much better than what AppCache provides. So, I hope that the Web API team collaborates with the networking team to make sure that we don't send the W3C or WHATWG down a path that is redundant or at odds with what the networking team is doing with SPDY. > So what I think we should do is to do a rewrite which implements the > current spec, but that has better performance. We can fix some other > problems while we are at it, such as integrate with the soon-to-land > quota manager. This would allow us to get rid of the prompt and > provide better guarantees for apps. Given the benefits that SPDY provides, and the benefits that packaging apps provides, I'm pretty doubtful there is room left for AppCache to have a meaningful existence, beyond being a compatibility burden for us. Cheers, Brian _______________________________________________ dev-b2g mailing list [email protected] https://lists.mozilla.org/listinfo/dev-b2g
