2 - Possibly but it’s going to be REALLY small and depend on how large your cookies are. If you only have a few cookies and they are relatively short it’s probably not worth the effort (webpagetest.org cookies get applied to my static content as well because it’s not really worth the effort for how few there are).
I’m confused by this. There’s only one page on my site that sets cookies, and after reading this post I changed the domain from .datesphere.com to .www.datesphere.com. However, I’m still failing all the cookie tests when you look at my results.
So am I just screwed forever since I already set cookies on the naked domain? This “you can’t fix it” explanation seems flat out bizarre, as I’m used to being able to fix / undo things on the internet, especially since it sounds like Marvin was able to change the cookie domain and it worked.
Any ideas on what I did wrong / why this didn’t work for me?
Well, you did half of the solution by setting the cookie to the fqdn, but all of the static resources are still being served from www.datesphere.com. You need to serve them from a different fqdn otherwise they will still carry the cookies (static.datesphere.com for example).
That said, cookies is probably one of the last things you should focus on. Once you reduce the number of requests you’ll be serving less than 10 requests total and it looks like there’s only the one cookie so it’s really not going to make a measurable difference.
Yep - dynamically may be more difficult though (is it dynamically selected from a reasonable set of possible images or a huge pool of images)? Are they all dynamic or are some of them static (like the flags)?
All the images are generated based on php variables, so I guess I’d say they’re all dynamic. I understand sprites in theory (having one big image containing all your images and showing only a specific part using css rules for each individual image), but I’ll have to read up on how you make it work when images are being pulled in dynamically. I’m sure there’s a way to do it.