Extra </p> inside blockquote

[audwan]

When using blockquote </p> is inserted directly in front of </blockquote>, making the code invalid XHTML.

Example:

<blockquote>This is a blockquote</blockquote>

Gives the following result:

<blockquote>This is a blockquote</p></blockquote>

Seems like ​this forum thread adresses the same issue in the support forum.

Using <blockquote> inside <li> is invalid. Use <q> instead.

For me, <blockquote>Hello, world. Wonder what this will do.</blockquote> puts <p>...</p> inside the <blockquote> fine.

This is with r5142. Closing as worksforme.

On version 2.3.1 this is still an issue.

For example, a post with only <blockquote>some quote</blockquote> will result in <blockquote>some quote</p></blockquote>, which is invalid XHTML.

This seems to me to be a result of:

1. formatting.php:78 $pee = str_replace('</blockquote></p>', '</p></blockquote>', $pee);
2. not having/not adding an opening "<p>" tag at the beginning of the post (even though it does get added for other all-text entries)

ENV: WP trunk r6301

Confirmed bug exists as described.

The bug won't reproduce if there are no blank lines before the blockquotes, or if there is a paragraph before the blockquote. That likely explains it working for rob1n .

Repro having a block quote by the only text, inserted using visual editor.

Removed that part about blockquote in a list from description.

Debuggin wpautop function php file

This bug is still there.
When wpautop deals with that content :

    text before
    <div class="whatever">
      <blockquote>
        text inside the blockquote tag
      </blockquote>
      Some text inside the div embedding the blockquote and after the closing blockquote tag
    </div>

wpautop returns the following output :

<p>    text before</p>
<div class="whatever">
<blockquote><p>
        text inside the blockquote tag
      </p></blockquote>
<p>      Some text inside the div embedding the blockquote and after the closing blockquote tag
    </p></div>

Everything is ok. The output of wpautop is just like we want it to be.
But when we try the following input :

    text before    <div class="whatever">
      <blockquote>
        text inside the blockquote tag
      </blockquote>
      Some text inside the div embedding the blockquote and after the closing blockquote tag
    </div>

wpautop returns :

<p>    text before
<div class="whatever">
<blockquote><p>
        text inside the blockquote tag
      </p></blockquote>
<p>      Some text inside the div embedding the blockquote and after the closing blockquote tag
    </p></div>

Obviously that's not a valid HTML. There's an openning 'p' tag just at the beginning, which one won't be closed.

We can reproduce that bug also using the following HTML code sample :

    text before    <div class="whatever"><blockquote>
        text inside the blockquote tag
      </blockquote>
      Some text inside the div embedding the blockquote and after the closing blockquote tag
    </div>

After passing that content to wpautop, we have, as output, the following content :

<p>    text before
<div class="whatever">
<blockquote>
        text inside the blockquote tag
      </p></blockquote>
<p>      Some text inside the div embedding the blockquote and after the closing blockquote tag
    </p></div>

As you can see there's some orphan tags. The openning p-tag before the beginning text and a closing p-tag just before the closing blockquote-tag.

There's other examples. But the two given are clear enough.

I've spent some hours looking for a solution or a patch. Helping is better that complaining. I've extracted the function, put it in a separated file and give it the real output of 'get_shortcode_regex()', because wpautopo uses it. you can have a look on the page I used here ​http://www.flegme.fr/wpautopdebug.php. And here's the same page with my bug fix applied ​http://www.flegme.fr/wpautopdebugafter.php (the php deugging file is attached as wpautopdebug.php)

So the solution is rather simple. In the wpautop php code at line 37,38 you've got :

	$pee = preg_replace('!(<' . $allblocks . '[^>]*>)!', "\n$1", $pee);
	$pee = preg_replace('!(</' . $allblocks . '>)!', "$1\n\n", $pee);

I've noticed that those two lines add newlines '\n' before and after some special tags (including the 'div' tag and the 'blockquote' tag). But the first regular expression adds one '\n' only where the second one adds two ones.

When I modify the code that it inserts two newlines before and after the special tags, the bug disapears.

So those two lines become :

	$pee = preg_replace('!(<' . $allblocks . '[^>]*>)!', "\n\n$1", $pee);
	$pee = preg_replace('!(</' . $allblocks . '>)!', "$1\n\n", $pee);

Maybe i've missed something and my solution is not the good one... But that worked fine for me.

Regards

Archibald Leaurees

see also: #4298, #6041, #6984, #2833, #9437, #7988

in 11256:

<blockquote>some quote</blockquote>

works fine

    text before
    <div class="whatever">
      <blockquote>
        text inside the blockquote tag
      </blockquote>
      Some text inside the div embedding the blockquote and after the closing blockquote tag
    </div>

clunky but works too

    text before    <div class="whatever">
      <blockquote>
        text inside the blockquote tag
      </blockquote>
      Some text inside the div embedding the blockquote and after the closing blockquote tag
    </div>

fails

    text before    <div class="whatever"><blockquote>
        text inside the blockquote tag
      </blockquote>
      Some text inside the div embedding the blockquote and after the closing blockquote tag
    </div>

fails

and the patch fixes everything indeed.

add an extra \n before block elts

better patch, that fixes #6041 as well

just closed #6401 as dup. second patch fixes this:

<dd>paragraph 1

paragraph 2

paragraph 3</dd> 

second patch also fixes #7988:

<div><a href="xx"> <img src="yy" /> </a> <p>text</p> </div>

was meant #6041, even. it fixes #6376 as well:

<dl> <dt>term</dt> <dd>paragraph

paragraph2</dd> </dl> 

there is one side effect, however. it turns:

<ul>
<li>foobar</li>
</ul>

into:

<ul>
<li>
<p>foobar</p>
</li>
</ul>

fix <li>\n<p>foo bar</p>\n</li>

3rd patch keeps everything fixed and sanitizes list items as needed.

Don't think it's a good idea to preg_split() on a \n followed by a white space. That would affect a lot of the internal workings of wpautop. It is affected by line breaks in the html formatting and the workaround is to keep that formatting to a minimum. It works when using the visual editor as all html formatting is removed there when saving.

IMHO to go ahead with this and all of the wpautop related patches, we will need comprehensive unit tests first.

it's not too urgent, as these tickets have waited for years anyway.

the preg_split() is actually mostly equivalent. this:

$pee = preg_replace("/\n\n+/", "\n\n", $pee); // take care of duplicates 
$pees = preg_split('/\n\s*\n/', $pee, -1, PREG_SPLIT_NO_EMPTY); 

is more or less the same as this, in a less optimized/generalized manner:

$pees = preg_split('/\n(?:\s*\n)+/', $pee, -1, PREG_SPLIT_NO_EMPTY); 

semantically, it means the same: find double \n ignoring lines that have space chars. the suggested regex merely does a better job at it, and it's an optimization that can readily be removed if it's not wanted, since it has no effect on the ticket.

anyway, I'm not too familiar with the unit tests api, so moving this over to needs-unit-tests... I totally agree a few are needed on this front.

the patch is an improvement to the current behavior, however.

this is the one that needs unit-tests.

punting pending unit tests

Another long-player in the formatting bug league, +1 for getting a definition, unit-tests and merging with related ones.

Agreed, we should get all related tickets for autop (PHP and JS) together and come up with good unit tests for them. Best time to do that is early in the dev. cycle, like couple of weeks after 2.9 is out.

No longer early in a dev cycle.. flicking to 3.1

Note, in trunk currently, the result of wpautop('<blockquote>This is a blockquote</blockquote>) is:

<blockquote><p>This is a blockquote</p></blockquote>

UPDATE::
after looking into the problem further I was able to reproduce the above behavior.

wpautop('<blockquote>This is a blockquote</blockquote>');

returns

<blockquote><p>This is a blockquote</p></blockquote>

and
wpautop('<p><blockquote>This is a blockquote</blockquote></p>');

<p>
<blockquote>This is a blockquote</p></blockquote>