Wpautop may create extra <br> tags

[azaozz]

Caused by removing the <br> tag normalization from convert_chars() and by the fact that wpautop() uses negative look behind to check for <br /> followed by a line break.

This is also broken in the JS wpautop.

[azaozz]

The best way to restore the behaviour is to normalize the <br>. This has to be added to the JS wpautop as now we use it on loading the content in TinyMCE.

[azaozz]

In 33377.patch:

• PHP wpautop: replace <br> and <br/> with <br /> before converting line breaks.
• JS wpautop: remove spaces and line breaks after existing <br> before converting line breaks.

[valendesigns]

Works as expected for me. +1

[miqrogroove]

Would it be easier to normalize BRs as newlines? If that's done before newline normalization, it might work.

[miqrogroove]

493	        // Change multiple <br>s into two line breaks, which will turn into paragraphs.

Needs tests for that.

[valendesigns]

A lot of tests fail when you move it before line 493.

[valendesigns]

33377.2.diff​ adds a second test and changes the regex pattern on line 493 to account for oddly formatted break tags when making the conversion to paragraphs.

[valendesigns]

alternative regex pattern

[miqrogroove]

33377.3.diff looks better.

[azaozz]

33377.3.diff looks good for now.

Eventually in PHP we should do something similar to the JS -- remove line breaks after any <br>, replace remaining line breaks with <br>, add line breaks after all <br> to make it prettier.

$pee = preg_replace( '|(<br[^>]*>)\s*\n|i', '$1', $pee );
$pee = preg_replace( '|\s*\n|', '<br />', $pee );
$pee = preg_replace( '|(<br[^>]*>)|i', "$1\n", $pee );

This changes the behaviour so a bit reluctant to do it now, so close to release.

[obenland]

Needs a second review.

[wonderboymusic]

This looks good to me

[azaozz]

In 33624:

Fix creating of extra <br /> tags in both PHP and JS variants of wpautop(). Add PHP tests to catch similar problems in the future.
Props valendesigns, azaozz. Fixes #33377.