AJ @ WpFASTER.org

Hi stellamaris5,

Try disabling database and object caching.

Best,
AJ

Hi Gemini23,

It looks like you’ve solved the render blocking JavaScript issue. Your CSS file is very small, so, as Frank alluded to, you should just use “inline all CSS” which will eliminate that warning. We have found that so long as the weight of the HTML ends up under 50kb there is no additional performance benefit to inlining only critical CSS as opposed to all of it.

The other warning is for the Google Web font you are using, which you can load asynchronously with this plugin. Bear in mind, however, that asyncing the font may or may not remove the warning in Google PSI and may in fact lower your site’s PSI grades. This is a bug with PSI (which, to be honest, just isn’t a particularly useful tool). To garner actual, useful performance metrics, we recommend WebPageTest.org (as does Google, in fact).

Best,
AJ

…I spoke too soon lol: https://developers.google.com/speed/pagespeed/insights/?url=https%3A%2F%2Fwww.wpfaster.org&tab=desktop

(As an aside, note that we’re also asyncing the webfont and PSI is not picking up on that either for Mobile or Desktop.)

We, ourselves, don’t really care about this as PSI is a ‘Best Practices’ tool not a measurement tool yielding actionable, best performance metrics. But, I thought you might want to know anyway given the common misconception that PSI is the be-all-end-all arbiter of whether or not a website is performant.

Best,
AJ

Great, Frank, thanks. You’re always such a big help. ??

AJ

That does clarify, thanks!

After that you can simple remove the one of the comment-sequences (double-slash) to activate one (or more) of the functions in there

I understand that this activates the filter, but I’m not clear as to how to then specify exactly what we want the filter doing and where that specific code is to go…

After un-commenting the generic filter in the “autoptimize helper” plugin so as to activate that filter,
A.) are we, other than un-commenting it, otherwise not touching the generic filter?
B.) Where do we put the specific filter code i.e. the code with specific instructions (I’m guessing at the bottom of “autoptimize helper”?)

For Example:

.
.
.
.
/* autoptimize_filter_css_datauri_exclude: exclude background images from being turned into data uri's;

@param $imageexcl: default images excluded (empty)
@return: comma-seperated list of images to exclude */
add_filter('autoptimize_filter_css_datauri_exclude','my_ao_exclude_image',10,1);
function my_ao_exclude_image($imageexcl) {
	return "adfreebutton.jpg, otherimage.png";
	}

.
.
.
.
.

(End/bottom of autoptimze helper plugin template)

add_filter('autoptimize_filter_css_datauri_exclude','my_ao_exclude_image',10,1);
function my_ao_exclude_image($imageexcl) {
	return "1specific.jpg, 2specific.jpg,, 3specific.jpg, 4specific.jpg,";
	}

^ Would that be correct?

Hope I’m making sense,
AJ

@frank,

No, just a dabbler in languages and a sucker for a good witticism in any tongue lol.

@celso,

I see what you’re saying. ??

Best,
AJ

don’t ever be bloated as W3TC

Not sure we’re disagreeing, mate(?).

Were there an Autoptimize “pro” (version) that also handled caching (e.g. page, database, object, fragment) in the manner in which W3TC does, (in addition to the optimization features AO already offers), such a plugin would utterly dominate the WordPress WPO space. Beyond a doubt. Completely. — For reasons that are immediately obvious to any and all that have dealt with W3 Total Cache…

Autoptimize is in a unique position to do precisely this; it would be nice to see the benefits go to Frank.

(although i love to use [W3tc] when not using WPengine)

And you do so because it is an awesome plugin.

bloated

Lot’s of people accuse W3TC of being “bloated” and yet none offer an actual argument as to why they’re saying it is.

The reason? Because it isn’t:
As far as code streamlining goes, the current version of W3 Total Cache is — relative to what is does — as streamlined (i.e. un-bloated) as is technically possible.

The onus, of course, is on you to prove the assertion that W3TC is “bloated”. I won’t hold my breath ??

At any rate, none of this has much to do with my premise(s), which is/are that Frank (and AO) are uniquely positioned to do wonderful things.

For Frank: Als alle gekken konden vliegen hadden we een permanente zonsverduistering. ??

AJ

The PSI update causing this issue has been retracted, Frank. You’re still awesome ??

AJ

*raises hand for beta testing*

Exciting! What will be in the paid-for version, Frank? To what extent will it differ from the free version?

I still think you should do an “Autoptimize PRO” (whatever), which would include the ability right from within the UI to implement the extended functionality offered by the helper and sell it on CodeCanyon… add white-labeling and caching features comparable to W3TC and, you know, corner the market and get rich lol.

I may be a dreamer, but I’m no the only one ??
AJ

Will do.

Ultimately, I suppose, the problem is that we mere mortals awaken each day to brand new Google-imposed conundrums, the solution to which only Google knows and impishly does not tell. ??

I’ll keep you updated on what, if anything, I discover.

AJ

Okay, Frank.

Thanks again.

Okay Frank, thanks. I believe what was used was indeed the path, not the full URL, but I’ll have to double check that.

AJ

Sorry, I should have been more specific.

An order form (URL B) is being loaded on a page (URL A) by way of an iframe. Both URLs are on WordPress, both on the same site/domain.

The order form page (URL B) has a template, JavaScript and CSS exclusive to that page, so it ‘breaks’ when attempting to apply the global block of critical CSS to it. Ergo, we need to exclude that page being framed (URL B) from being Autoptimized.

This URL (URL B) is what we’re trying to apply the filter to but it won’t take. If it should, then it should. I just need to make sure we’re doing it in a way compliant with AO (i.e. the above code).

Given my elaboration, would you say the above code should be working?

Thanks as always,
AJ

Do you think it could be the fact that it’s a Windows Server? That’s what the server company are saying.

MySQL and PHP have not been designed for Windows servers. So yes, if you are able to switch it would probably save a lot of current and future headaches. In other words, Windows servers *can* be configured correctly for WordPress (and W3TC, therefore), but are very easy to mess up.

AJ

It is a server configuration issue, therefore.

Let’s recap what we’ve done:
1.) Exhibit A: We have, in increasing orders of magnitude, lessened W3TC’s interaction with the site’s server, up to and now including complete removal of the plugin. Once we were at the point where W3TC was merely caching to basic disk, things were as good as they could get. Thus indicating that W3TC was attempting to overcome server issues which it could not, problems increasing in severity the more W3TC attempted interaction with the server. In short, W3TC was doing the best job it could given the server (configuration) it had to work with.

2.) Exhibit B: The site, with no caching mechanisms in place whatsoever, should be faster than it was when W3TC was implemented and caching to basic disk (where we reached pinnacle performance.) But it isn’t, it is worse. Thus also indicating the server is misconfigured in itself.

Short version: the site’s server needs tuned-up, mate.

Best,
AJ