[OJS 3.1.1.0] TinyMCE calls for google font even when enable_cdn is disabled

[MarHerUMR]

In to share as little data as possible we disabled the enable_cdn option in the config. But it turns out that the backend still makes calls to //fonts.googleapis.com/css?family=Noto+Sans:400,400italic,700,700italic.

As far as I can see the call is made because the plugins/generic/tinymce/styles/content.css has the import:

@import url('//fonts.googleapis.com/css?family=Noto+Sans:400,400italic,700,700italic');

I would have assumed that the google fonts would be disabled in the front- and backend when disabling enable_cdn (especially because of the advise to disable it because of GDPR). But I'm not quite shure how this can be disabled here.

[asmecher]

@MarHerUMR, at a glance it looks like there's a separate Noto Sans import (in lib/pkp/classes/template/PKPTemplateManager.inc.php) that is dependent on the CDN setting being enabled -- you should be able to remove the reference in plugins/generic/tinymce/styles/content.css without any negative effects. Could you try this and confirm?

[NateWr]

plugins/generic/tinymce/styles/content.css loads styles that are only applied in the iframe of the tinymce editor. That means that the noto font isn't loaded separately, and removing it from content.css will cause the content inside the editor to not be styled properly.

However, @MarHerUMR, if you are already disabling the CDN, then you won't have this font anyway, so you're probably safe removing that line in content.css for the time being.

We'll probably need to figure out a solution that takes enable_cdn into account when loading the TinyMCE content CSS file. I'll take this into account with the TinyMCE work I'm doing in #3594.

[NateWr]

Ooh, it supports multiple stylesheets. 🙏 https://www.tinymce.com/docs-3x/reference/configuration/Configuration3x@content_css/ Ok, let me address this real quick.

[NateWr]

PR:
#3767

Tests:
pkp/ojs#1999

[NateWr]

This needs a commit to the tinymce plugin to add the new/modified CSS.

[NateWr]

This was addressed as part of #3594