And Nginx delivers its own headers (like the rules in the. It would completely ignore the dynamic nature introduced by the execution of that PHP. I’m sure you can see that this assessment would be wrong. If it considering just the final HTML and Javascript, it would treat pages as cacheable even when they contained a PHP snippet. My reading of this feature is that Nginx is deciding if a page can be served from cache by looking at its contents after processing by the CDN (WordPress). Siteground have decided to implement something called “Static Cache to NGINX Direct Delivery”. Here’s the messy bit that’s probably causing the problem. If they did that transfer for my website when I had SG Optimizer installed, they would have copied all those caching rules left behind by their plugin! Siteground don’t permit clients to view or modify their nf file. To do that, they would have had to automate the conversion of the. My guess is that Siteground automated their move from Linux to Nginx. When I deleted SG Optimizer, it left those rules in the. Little did I know that SG Optimizer added dozens of rules to the. So I deleted SG Optimizer thinking that I had returned to my previous situation. So I decided to try SG Optimizer to see if I could manage caching a bit better. I knew right away that it was caused by caching. Then I started having problems with Woody Snippets not being executed on each load. Under Linux they recommended the use of a plugin called SG Optimizer that controlled the caching. No doubt it was buried behind some changes to T&C that I did not read. Any help you can supply would be most welcome. After that, they only run if I click on Refresh ( or circular arrow on browser). They only run the first time the page is accessed. I have used PHP snippets on several other pages. I have tried other combinations of meta tags with no success. I’ve checked the URL with and it shows that the cached page is usable for 180 days despite my meta tags in the head section as follows: htaccess and there is no caching plugin installed. I have checked that “Cloudfare” is Inactive, SuperCacher is “off”. They claim that they are not doing any special kind of caching. I have tried various header changes and removed all. However when displaying the page later using the same browser, the cached page is displayed (showing the old date/time). When displaying the page the first time, the display shows the correct (current) time. It runs near the top of the page (it appears with text asking the visitor to click on refresh if the date/time are old). PHP snippet simply contains: date_default_timezone_set('Australia/Perth') I’m pulling my hair out! Any ideas would be appreciated. I have tried using a Caching plugin (SG Optimizer) that provides an “exclude” feature and I’ve nominated the pages in question to be “excluded” but that hasn’t worked either. I have tried my best to turn off caching on the page but the browser is ignoring my header tags of While this is a workaround, it is not a solution since my clients won’t realise that they need to click on “refresh” to see the correct version. I can fix this manually by clicking on the “refresh” button () then an up-to-date version of the page is displayed. It is clearly out of date and the PHP snippet is not being run. Recently (probably started a few months back but I can’t pinpoint a date) I noticed that in the morning when I first view these pages, my browser (both Firefox and Chrome) displays an old cached version of the page. Location.reload(bool) might not be officially defined, but it seems to be quite well defined de facto, except from your recent change.I have a number of pages that use PHP snippets (all are “private” so can’t give you a link). Hack works, but I think it's sad that I can not rely on simply location.reload(true). I can't even redirect to the login page myself since it's a general SSO-solution in front of the servers that provides an authentication form on unauthenticated requests.
I thought I could use location.reload(true) to force a reload that would redirect to the login page, but since it does not work in chrome I have to implement some other hack to force the browser to actually reload the resource.
My specific use case taking me here is an authenticated rich web application where I have to handle expired sessions. The behavior I see now is that Ctrl-F5 disregards the cache, while reload(true) does not.
Location.reload() and that Ctrl-F5 reloads like location.reload(true)? Because That you should stick to two types of reloads: One that use the cache and one that It, wasn't the idea of the discussion in !topic/loading-dev/gD-MPRcfwVA/discussion