浏览器是否没有正确遵循 HTTP 规范的缓存控制？

Question

I am somewhat new to web development and have noticed an issue, Browsers seem to not respect the Cache-Control header, I have it set to no-cache, no-store, must-revalidate but yet many of my clients have a cache to begin with (which no-store should prevent according to https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/Cache-Control#no-store ) and the cache is used rather than revalidating with the server leading to broken pages when I change a JS script referenced in a page, only after I tell them to refresh without cache does the browser then fetch the new file but for the browsers to be compliant with the HTTP protocol and spec, don't they need to respect the no-store policy or are none of the major browsers properly compliant with the HTTP protocol/spec and why haven't they been fixed so we don't need workaround solutions like query strings appended to files or using the file's hash or last modification date?

Answer 1

You initially served the resource without cache headers. In that case, the specification allows the client to choose the cache time itself:

Since origin servers do not always provide explicit expiration times, a cache MAY assign a heuristic expiration time when an explicit time is not specified, employing algorithms that use other header field values (such as the Last-Modified time) to estimate a plausible expiration time.

Different browsers will use different algorithms, but in any case it probably won't be very long. Your problem might have already resolved itself.

As for query strings, I think your confusion comes from conflating at least three distinct issues. One is the HTTP protocol mechanism for communicating cache policies. That is covered in RFC 7234 and mainly involves the proper use of the Cache-Control response header.

A separate issue is what cache strategy to use. That is, which resources should be cached and for how long? There are different ways to approach this, my suggestion would be to follow the best practices discussed here .

Finally, there's how to fix your mistake if you communicated the wrong cache policy and now need an already-cached resource to be ignored or invalidated. In that case, if possible, you could just use a different resource (ie change the name). Adding query strings is sometimes suggested here, but it's not a great solution since the standard does not forbid clients from caching resources with query strings.

Getting back to your question, you can temporarily fix your mistake (missing Cache-Control headers) by changing the name of the linked resource, or just by waiting a short time for the heuristic expiration time to pass. Longer term, you should decide how you want your different resources to be cached, and then use Cache-Control to communicate that intent to the browser.