Enable ENT_SUBSTITUTE flag in Html::escape

If this works reliably, then it would be even be a major reason for bumping PHP requirements to 5.4.

I wouldn't recommend mandating PHP 5.4 be required for D8, but this definitely could be a good optional feature for when PHP 5.4 is available.

Do we have a baked in means of detecting PHP version, or would something like

    if (defined('ENT_SUBSTITUTE')) {
      $replace_flags = ENT_QUOTES & ENT_SUBSTITUTE;
    }
    else {
      $replace_flags = ENT_QUOTES;
    }

be appropriate?

On the assumption the above is ok i've rolled this patch. Tests pass on my local on php 5.3.25 and 5.4.15 - not entirely sure what to do about writing additional unit tests to cater for this. Do we want php version checking logic in our unit tests, or do we just case that the cases are what we expect?

PHP 5.4 is already a requirement and Twig already does this so it would also keep things in line with Twig's escape filter.

https://github.com/twigphp/Twig/blob/1.x/lib/Twig/Extension/Core.php#L1011

+1 to this issue.

Re-rolled.

This issue seems like a really good idea. Postponing on #2550945: Add Html::escape()

FYI twig is using this internally so another reason to have this.
https://github.com/twigphp/Twig/blob/641090378dfff2913929306fbbf227269ba...

+++ b/core/lib/Drupal/Component/Utility/Html.php
@@ -371,7 +371,8 @@ public static function decodeEntities($text) {
+   * including "&eacute;" and "&lt;" to "é" and "<", converting invalid
+   * multi-byte sequences to the Unicode Replacement Character ("�").

Let's not add this documentation to the bit about how this method is different from Html::decodeEntities. It belongs in the list of conversions above.

The list is a list of entity equivalents though, if we want it to be in the list we could drop the " with their HTML entity equivalents" bit?

a newline crept in

+++ b/core/lib/Drupal/Component/Utility/Html.php
@@ -366,7 +366,8 @@ public static function decodeEntities($text) {
+   * (for example, "<" becomes "<"), and invalid UTF-8 encoding
+   * will be converted to the Unicode Replacement Character ("�").

I think it should be "replacement character" - too many capitals :). I like where these docs are now. Thanks.

Fixing some more test fails

Awesome! Thanks for jumping on this after the HTML::escape blocker was in.

clarifying some comments

Thank you, green means go.

Normal tasks need a beta evaluation.

Committed 00360b9 and pushed to 8.0.x. Thanks!

Thanks for adding the beta evaluation to the issue summary.

Wow, this landed fast! Great job, @stefan.r :)