search

jshttp / fresh

HTTP request freshness testing
MIT License
161 stars 28 forks source link

Add a max-age=0 patch for some versions of Safari. #9

Closed chaaz closed 10 years ago

chaaz commented 10 years ago

Some versions of Safari (at least my version, 7.0.3 OSX) will send Cache-Control:max-age=0 instead of the usual Cache-Control:no-cache when refreshing a page, after deleting the cache. It expects the server to respond with up-to-date data. However, if the server response with a 304, then Safari will display a blank page to the user.

This very simple patch treats "max-age=0" the same as it does "no-cache": it automatically treats the request as fresh, regardless of other parameters.

btmdave commented 10 years ago

Any update if this will get added in? It's pretty breaking.

imyller commented 10 years ago

I agree with @btmdave; this needs to be added in.

Just tested with Safari 7.0.4, node-fresh 0.2.2, express 4.4.3:

On page reload, Safari sends Cache-Control: max-age=0 instead of Cache-Control: no-cache

Express then incorrectly responds with 304 resulting in empty page rendered as Safari no longer has a cached copy.

imyller commented 10 years ago

Safari is complying to RFC-2616 when it sends Cache-Control: max-age=0. Safari requests end-to-end revalidation when reloading the page manually.

http://www.w3.org/Protocols/rfc2616/rfc2616-sec14.html#sec14.9.3

Especially: http://www.w3.org/Protocols/rfc2616/rfc2616-sec14.html#sec14.9.4

Specific end-to-end revalidation The request includes a "max-age=0" cache-control directive, which forces each cache along the path to the origin server to revalidate its own entry, if any, with the next cache or server. The initial request includes a cache-validating conditional with the client's current validator.

node-fresh module should react to max-age=0 by preferring fresh content instead of allowing 304 response.

jonathanong commented 10 years ago

okay wtf this was added before but reverted. @dougwilson do you know what's up? i'm just not going to push a release until we solve this

dougwilson commented 10 years ago

@jonathanong idk, but it clearly seems like max-age=0 should be forcing a refresh, i.e. I believe this PR is valid.

dougwilson commented 10 years ago

Nevermind, it may not be. I'm reading more into RFC 7234 sec 4 still, but yes, max-age=0 is not exactly equivalent to no-cache, but I'm still determining what the exact behavior should be (but the current behavior is not correct, but I think the behavior in the PR is also not correct).