I need to set up a specific expire header to json files, much much lower than the rest of files. Can I do this in W3TV? I couldn't find a way.
The default of 31536000 seconds is ok for all other file types. But I use the JSON REST API to deliver data to an AngularJS+Cordova App, and was having a problem with content not being updated. We figured out that It was the json expire header when we manually configured 300 seconds, problem is W3 TC constantly overrides this change.
Is there a way to tell W3 TC to use a lower expire header for json files? Or a way manually enter a value that's not overridden by W3TC?
The only idea I've come up with is to rewrite the json expire header rule at the bottom of .htaccess, but I don't know if this is going to prevent W3TC to edit or erase it. Also, having a repeated rule seams just plain wrong from the begging.
Or is there any way to tell Angular to re download the json file even if the cache header tells it to keep it for a year?
What do you think?
Thanks!
FG
Add a random property to the end of the URL to fetch the JSON file. This is what jQuery does to ensure that the cache is not used for JSON requests.
Assume your file URL is http://example.com/myfile.json, then you would fetch it with http://example.com/myfile.json?__random=1 the first time and http://example.com/myfile.json?__random=2 the second time etc. of course you should use totally random numbers instead of 1, 2 etc.
W3TC offers 3 groups of header policies to manage for user agent (browser) caching. The third group on the browser cache settings page has an "Other" section where the headers available there would be applied to the json response from WP unless there's a bug or implementation that prevents this behavior from occurring. W3TC sets the directives for the nginx or apache web server to apply the headers specified on this page and there are ways that this policy could fail to be applied, however the intent is as indicated.
Related
i am looking for a solution for an auto-complete dropdown box which needs to load entries from a huge json (json file is also being updated / generated every second.)
I tried "typeahead.js" but by default it caches the json file in browser and was not able to display new entries added to json file.
is there a solution for an auto-complete text box which can load entries from the server as fast as possible ?
please suggest.
thanks
In your case, you can take advantage of Bloodhound, the typeahead.js suggestion engine. It provides two options. Prefetch and Remote.
In Prefetch, data is fetched and processed on initialization. If the browser supports local storage, the processed data will be cached there to prevent additional network requests on subsequent page loads.
In Remote, it will fetch data from remote source when ever you need it. But remember, in order to prevent an obscene number of requests being made to the remote endpoint, requests are rate-limited.
I think you should user Remote option in your situation.
Reference : Link
There are two ways which are mentioned in the documentation for typeahead.js
You can make the TTL value = 1 for prefetch ( this did;t work for me )
https://github.com/twitter/typeahead.js/blob/master/doc/bloodhound.md
Or you can use clearPrefetchCache on click page load or click of a button .
category.clearPrefetchCache();
I just came across a website pagesource and saw this in the header:
<link href="../css/style.css?V1" rel="stylesheet" type="text/css" />
Could we actually pass GET data to css? I tried searching but found no results apart from using PHP. Could anyone help make meaning of the ?V1 after the .css
I know this forum is for asking programming problems, however I decided to ask this since I have found no results in my searches
First of all, no you can't pass GET parameters to CSS. Sorry. That would have been great though.
As for the example url. It can either be a CSS page generated by any web server (doesn't have to be PHP). In this case the server can serve different pages or versions of the same page which might explain the meaning of V1, Version 1. The server can also dynamically generate the page with a server-side template. This is an example from the Jade documentaion:
http://cssdeck.com/labs/learning-the-jade-templating-engine-syntax
It can also just be used as cache buster, for versioning purposes. Whenever you enter a url the browser will try to fetch it only if it doesn't already have a cached copy which is specific to that URL. If you have made a change in your content (in this instance the css file) and you want the browser to use it and not the cached version you can change the url and trick the browser to think it's a new resource that is not cached, so it'll fetch the new content from the server. V1 can then have a symantic meaning to the developer serving as a note (ie I've changed this file once...twice..etc) but not actually do anything but break the cache. This question addresses cache busting.
There are different concepts.
At first, it only is a link - it has a name, it might have an extension, but this is just a convention for humans, and nothing more than a resource identifier for the server. Once the browser requests it, it becomes a server request for a resource. The server then decides how to handle this request. It might be a simple file it just has to return, it might be a server side script, which has to be executed by a server side scripting interpreter, or basically anything else you can imagine.
Again, do not trick yourself in thinking "this is a CSS file", just because it has a css extension, or is called style.
Whatever runs at the server, and actually answers the request, will return something. And this something then is given a meaning. It might be CSS, it might be HTML, it might be JavaScript, or an image or just a binary download. To help the browser to understand what it is, the server returns a Content-Type header.
If no content type is given, the browser has to guess what it is. Or the nice web author gave a hint on what to expect as response - in this case he gave the hint of text/css. Again, this is how the returned content should be interpreted by the client/browser, not how that content is supposed to created on the server side.
And about the ?V1? This could mean different things. Maybe the user can configure a style (theme) for the website and this method is used to dispatch different styles. Or it can be used for something called "cache busting" (look it up).
You can pass whatever you want; the server decides what to do with the data.
After all, PHP isn't your only option for creating a server. If i wrote a server in Node.js, set up a route for /css/style.css and made it return different things depending on what query was given, neither the server nor browser will bat an eyelid.
I've got some issues regarding images and secure access to the cloud (S3, in this case).
I do want to have direct access to avoid stressing my server so I do have two options:
1.) Generate a signed url on my server and send it back to client to load from
2.) Generate an Authorization-Header sent within the request from the client
Now, 1. & 2. work just fine, however, I am in trouble regarding images in html.
For 1.), images will be re-requested each time as the url changes every time and thus renders caching useless.
For 2.), caching would work as the url'd stay constant though I've seeked the net to figure how to add a custom http request header to image request including my security token without luck.
So, what options are left for me? How do others resolve that issue?
thanks
Alex
I've recently added HTTP headers to my site to inform the browser to check with the server every time it comes across a given JS/CSS URL. I've tested it and it works perfectly; all browsers now make conditional GET requests.
Here's the problem though -- people still have the old headers cached; headers which more or less told the browser "cache this forever; don't bother asking for an update!". This can be busted with a hard refresh. I don't want to have to communicate to everyone to please hit F5 on any buggy pages after we push out code.
Are there any HTTP header(s)/HTML meta tag(s) I could put on the HTML document itself to say "Browser, ignore the headers you have on the JS/CSS files and download the latest version of all the included files on this page"?
Eventually this problem will work itself out as more and more people clear their cache or learn to refresh on their own. But, I'd rather fix it now. Then in a month or so, I'll remove the HTML-level headers to get caching where I want -- on a per resource basis.
EDIT: I do not want to rename the resources or add on query parameters. That's what we used to use (?v=18, ?v=19, etc.) and it was a chore to increment that number every time we updated resources. Even doing that programmatically isn't the ideal solution; especially now that our server is configured correctly. It makes more sense to do it on the HTTP level so it works regardless of how you're accessing it -- included on a page, directly from the address bar, or otherwise.
pass a parameter to on the script source which will force a reload of the script... in fact you could do it by version or similiar
<script src="/test/script/myawesomescript.js?ver=1.0&pwn=yes" ...>
that would work and be seemless to the other users... when you feel like it has been long enough go back to the old way. but this will work if you want to force a refresh from users.
This method is utilized to prevent caching of webpages by some frameworks. Let me know if you were successful
http://css-tricks.com/can-we-prevent-css-caching/ -- here is a link to the concept for css (should work in js too) -- the biggest difference is you dont want it to never cache, so dont use a time stamp, use my style like from above :) enjoy!
Basically the only way is to get the browser not to use the cached URL.
One method is to use a cache-busting dummy parameter on the end of the URL.
some-name.css?q=1
That will force the browser to reload that file (because that URL isn't in the cache), and the downloaded file won't be cached because of your new headers. However: you may need to use this new name indefinitely, because you can't guarantee that once you leave off the dummy parameter again the cached version may still be used.
The other method is to completely rename the file.
my-new-name.css
Is there a way to retrieve only part of a page? I need to access a certain number of characters after the sexdecillionth character, and I know that it is possible. I can't request the page, because it returns response body too large. What should I do?
The platform and language don't matter to me, because I can access them all.
Retrieve it from what?
Generally speaking, you'll need to open an HTTP connection and use the Range: header to request the start location of the page download. I don't know that every web server supports it, though, so your mileage may vary.