matjazp

It's working here. That 0 is error code (and ? is error text) returned from WireHttp::GetHttpCode(). Hard to tell what went wrong. Can you provide original image?

https://github.com/processwire/processwire-issues/issues/1893

Thx Robin! While the Save + Close button is there (as expected), the js part makes the edit links all open home page, eg. they all have the link /processwire/page/edit/?id=1&quick_edit=1 I fixed that with: const $links = $('.PageListActionEdit a:not([data-autoclose])'); $links.each(function() { $link = $(this); $link.attr('data-autoclose', '#save-and-close').attr('href', $link.attr('href') + '&quick_edit=1'); }); Just to mention that using uk-hidden class (as a "belt and suspender" approach) would only work in AdminThemeUIkit, leaving out newer (and older) admin themes.

Now that you explain it, I remember this behavior of modal.js. But, unfortunately, it doesn't work here, only the Save button is cloned. Looking at the modal.js line 417, $icontents.find(buttonSelector) only returns one button, not two as expected. It seems like the Save+Close button is added too late, after the modal has already cloned the Save button. While debugging, I added the button unconditionally and also removed the uk-hidden class, and then edited the page in a new tab with modal=1, and both buttons were there, as expected. But not in modal. I guess if I add the button directly to the ProcessPageEdit.module (or via hook), it would work. But it's working for you, so there must be something with my installation. It's a bit late here, so I'll continue testing tomorrow.

Hey Robin! Thanks for this. I haven't had time to test this, but looking the code I see uk-hidden class on button. Would button even show? (and I know it's there for a reason)

Just to say, that 4.0 beta is working just fine. I had to tweak jquery migrate, but this issue will be fixed soon.

I don't like toggles at all. I find them confusing. I use a checkbox or radio group. It’s clear what you want, accessible and doesn’t make me think.

https://github.com/processwire/processwire-issues/issues/1867 ?

@Lutz I think this is addressed in https://github.com/processwire/processwire/commit/24fc229be563de0157ed18ee199dfab889286169 , but this was already fixed in 3.0.184 so I wonder if this is something new?

Please open an issue report.

I installed SkyScraper profile (last commit from year 2013, not compatible with PW3) on PW 3.0.221 and loaded home page: 1.10 seconds: no cache enabled 1.00 seconds: built-in cache enabled 0.35 seconds: AIOM+ enabled

1.10 seconds: no cache enabled 1.00 seconds: built-in cache enabled 0.095 seconds: AIOM+ enabled I'm not using debug timer, just timings in browser network tab. @bernhardperhaps you could test AIOM+ on the same setup as you tested with ProCache (29 seconds ???).

Nope, as I don't have ProCache, but based on what I've read and info from the forum, ProCache serves static files and there might be directives needed in .htaccess. I'm on IIS so it's not an option for me. AIOM+ leverages the template page cache built in core and uses a small php file to render those files. It's not perfect as modification of index.php is needed for best results (I have a feature request open). It also works if you modify some settings in site config.php, but it's not that effective.

Yes, it even works on PHP 8.3 and it's caching option is almost as fast as ProCache.

You can try my fork of aiom https://github.com/matjazpotocnik/ProcessWire-AIOM-All-In-One-Minify