Category Archives: wordpress

Autoptimize 2.2 coming your way, care to test?

So work on Autoptimize 2.2 is almost finished and I need your help testing this version before releasing (targeting May, but that depends on you!). The more people I have testing, the faster I might be able to push this thing out and there’s a lot to look forward to;

  • New option: enable/ disable AO for logged in users for all you pagebuilders out there
  • New minification/ caching system, significantly speeding up your site for non-cached pages (previously part of a power-up)
  • Switched to rel=preload + Filamentgroup’s loadCSS for CSS deferring
  • Additional support for HTTP/2 setups (no GUI, you might need to have a look at the API to see/ use all possibilities)
  • Important improvements to the logic of which JS/ CSS can be optimized (getPath function) increasing reliability of the aggregation process
  • Updated to a newer version of the CSS Minification component (albeit not the 3.x one, which seems a tad too fresh and which would require me to drop support for PHP 5.2 which will come but just not yet)
  • API: Lots of extra filters, making AO (even) more flexible.
  • Lots of bugfixes and smaller improvements (see GitHub commit log)

So if you want to help:

  1. Download the zip-file from Github
  2. Overwrite the contents of wp-content/plugins/autoptimize with the contents of autoptimize-master from the zip
  3. Test and if any bug (regression) create an issue in GitHub (if it doesn’t exist already).

Very much looking forward to your feedback!

Autoptimize CSS defer switching to loadCSS (soon)

Historically Autoptimize used its own JS-implementation to defer the loading of the main CSS, hooking into the domContentLoaded event and this has worked fine. I knew about Filament Group’s loadCSS, but saw no urgent reason to implement it as I saw no big advantages vs. my homegrown solution. That changed when criticalcss.com’s Jonas contacted me, pointing out that the best way to load CSS is now using the rel="preload" approach, which as of loadCSS 1.3 is also the way loadCSS works;

<link rel="preload" href="path/to/mystylesheet.css" as="style" onload="this.rel='stylesheet'">

As rel="preload" currently is only supported by Chrome & Opera (both Blink-based), a JS polyfill is needed for other browsers which uses loadCSS to load the CSS. Hopefully other browsers catch up on rel="preload" because it is a very elegant solution which allows the CSS to load sooner then with the old code while still being non-render blocking. What more could one which for (“Unicorns” my 10yo daughter might say, but what does she know)?

Anyways; I have integrated this new approach in a separate branch on GitHub, you can download the zip-file here to test this and all the other fixes and improvements since 2.1.0. Let me know what you think. Happy preloading!

No REST for the wicked

After the PR-beating WordPress took with the massive defacements of non-upgraded WordPress installations, it is time to revisit the point-of-view of the core-team that the REST API should be active for all and that no option should be provided to disable it (as per the decisions not options philosophy). I for one installed the “Disable REST API” plugin.

Lykke Li – No Rest For The Wicked

Watch this video on YouTube.

How to make Autoptimize (even) faster

Less blogposts here lately, mostly because I’m doing custom Autoptimize-development for a partner (more on that later) and because I get a lot of support-questions on the wordpress.org support forums (with approx. between 1500-2000 downloads/ weekday that is to be expected). One of the more interesting questions I got there was about Autoptimize being slow when JS optimization was active and what would be the cause of that. The reply is of interest for a larger audience and is equally valid for CSS optimization;

Typically the majority of time spent in Autoptimize is mainly in the actual minification of code that is not minified yet (purely based on filename; if the filename ends in .min.js or -min.js).

So generally speaking, the way to avoid this is;
1. have a page cache to avoid requests triggering autoptimize (as in that case the cached HTML will have links to cached CSS/JS in it)
2. for uncached pages; make sure AO can re-use previously cached CSS/ JS (from another page), in which case no minification needs to be done (for that you will almost always want to NOT aggregate inline JS, as this almost always busts the cache)
3. for uncached CSS/ JS; make sure any minified file is recognizable as such in the filename (e.g. .min.css -min.js), this can lighten the minification-load considerably (I’ll add a filter in the next version of AO so you can tell AO a file is minified even if it does not have that in the name).

So based on this, some tips;
* make sure you’re not aggregating inline JS
* for your own code (CSS/ JS); make sure it is minified and that the filename confirms this. if you can convince the theme’s developer to do so, all the better (esp. the already minified but big wp-content/themes/bridge/js/plugins.js is a waste of precious resources)
* you could try switching to the legacy minifiers (see FAQ) to see if this improves performance
* you can also check if excluding some un-minified files from minification helps performance (e.g. that bridge/js/plugins.js)

WordPress 4.7 custom background image bug & workaround

If you encountered this bug but are not using Autoptimize, leave a comment below or contact me here. Your info can help understand if the regression has an impact outside of Autoptimize as well!


Gotta love Sarah, but there’s a small bug in Vaughan’s (WordPress 4.7) that breaks (part of the CSS) when Autoptimized. If you have a theme that supports custom backgrounds (e.g. TwentySixteen) and you set that custom background in the Customizer, the URL ends up escaped (well, they wp_json_encode() it actually) like this;

body{background-image: url("http:\/\/localhost\/wordpress\/wp-content\/uploads\/layerslider\/Full-width-demo-slider\/left.png");}

Which results in the Autoptimized CSS for the background-image being broken because the URL is not recognized as such. The bug has been confirmed already and the fix should land in WordPress 4.7.1.

If you’re impacted and can’t wait for 4.7.1 to be released, there are 2 workarounds:
1. simple: disable the “Aggregate inline CSS”-option
2. geeky: use this code snippet to fix the URL before AO has a change to misinterpret it;

add_filter('autoptimize_filter_html_before_minify','fix_encoded_urls');
function fix_encoded_urls($htmlIn) {
	if ( strpos($htmlIn,"body.custom-background") !== false ) {
		preg_match_all("#background-image:\s?url\(?[\"|']((http(s)?)?:\\\/\\\/.*)[\"|']\)?#",$htmlIn,$badUrls);
	  	if ($badUrls) {
			foreach ($badUrls[1] as $badUrl) {
				$htmlIn = str_replace($badUrl, stripslashes($badUrl), $htmlIn);
			}
		}
	}
  	return $htmlIn;
}

So how does Autoptimize work anyway?

A question that has come up a couple of times already is how Autoptimize and it’s cache work. So let’s do some copy/pasting of what I replied earlier on the wordpress.org support forum;

  1. AO intercepts the HTML created by WordPress for a request (using the output buffer)
  2. all references to JS (and CSS) are extracted from the HTML
  3. all original references to JS (and CSS) are removed from the HTML (the code in the original files is left as is, AO never changes those files)
  4. all JS (and CSS) is aggregated (JS in one string, CSS in as many strings as there were media types)
  5. the md5-hash (mathematical/ cryptographic function that generates a quasi-unique string based on another string) of the aggregated JS (and CSS) is calculated
  6. using the md5 AO checks if a cached file with that md5 exists and if so continues to step 8
  7. if no cached file is found, the JS (and CSS) is minified and cached in a new file, with the md5 as part of the filename
  8. the links to the autoptimized JS (and CSS) file in cache are injected in the HTML
  9. the HTML is minified (but not cached in Autoptimize)
  10. the HTML is returned to WordPress (where it can be cached by a page cache and sent to the visitor)

This is especially interesting if you want to understand why the cache size can “explode”; if in step 4 the code is even a bit different from previous requests, the md5-hash in step 5 will be different so the file will not be found in cache (step 6) and the code will be re-minified (which is relatively expensive) and cached (step 7).

And that, my friends, is how Autoptimize works.