Quotes Are Messing Up

[miqrogroove]

Paste this into the body of a new post:

A sentence.  "A quote after 2 spaces."

Preview that, and be amazed. The first quote is backwards. :(

Due to font differences, this is easiest to see on the Twenty Ten theme.

Isn't this caused by the wptexturize filter?
Try to put

remove_filter('the_content', 'wptexturize');

in the theme's functions.php file.

Isn't this caused by the wptexturize filter?

Yes, It's a bug in the texturize filter, this ticket is to fix it.

Related: #16957 ?

Cannot reproduce in the editor.

wptexturize( 'A sentence. "A quote after 2 spaces."' ) returns proper quotes.

Replying to nacin:

Cannot reproduce in the editor.

Sometimes pasting doesn't preserve both spaces. Try removing the space and hitting spacebar twice from Visual editor. That'll do it.

Here's another bug:

"<code>blah</code>"

The trailing quote is backwards when output. 3.4.2 and 3.5 both affected.

Replying to miqrogroove:

"<code>blah</code>"

The trailing quote is backwards when output.

Related: #18549

Any chance of getting a milestone for the 2-spaces bug? I'm in the habit of typing with 2 spaces so this bothers me a bit.

Could not reproduce either:

1. Pasted the example into editor (tried both Text and Visual).
2. Made sure there are two spaces before the quote.
3. The post contains proper quotes:

   A sentence.  “A quote after 2 spaces.”
   

Tested 3.5.1:

<p>A sentence.  &#8221;A quote after 2 spaces.&#8221;</p>

Please always test trunk.

I can reproduce in 3.5.1 with WP Editor plugin activated, but not on a clean install. Could you try on a clean install?

Replying to SergeyBiryukov:

Could not reproduce either:

1. Pasted the example into editor (tried both Text and Visual).

OK couple things to clarify here. I can't reproduce this bug by pasting or by using the Text tab. Go to the Visual tab, type in the spaces, and then see what happens in preview.

I've also confirmed this is reproducible in 3.6-beta1. Any chance this can get on a milestone now?

Replying to miqrogroove:

I've also confirmed this is reproducible in 3.6-beta1. Any chance this can get on a milestone now?

Would need a patch and unit tests for that.

OK, I'm in an optimistic mood today :) The problem is likely in a filter somewhere and I'll track it down.

*sigh*

Not a filter issue. I'm getting different results with identical inputs on two servers now. Starting to suspect an mb string issue. I'll boil this down to a test case for you to look at.

$text = "A sentence. \xC2\xA0\"A quote after 2 spaces.\"";

$opening_quote = '“';
$closing_quote = '”';

$dynamic = array();
$dynamic[ '/(\s|\A|[([{<])"(?!\s)/'    ] = '$1' . $opening_quote . '$2'; 
$dynamic[ '/"(\s|\S|\Z)/'              ] = $closing_quote . '$1'; 

$dynamic_characters = array_keys( $dynamic );
$dynamic_replacements = array_values( $dynamic );

echo preg_replace($dynamic_characters, $dynamic_replacements, $text);

The above test case produces bad output on my BlueHost account, but works perfectly on my home server. Configurations are quite different; unsure the culprit. Suspect mb string.

Oops. Fixed it. The PCRE_UTF8 modifier forces identical output on both servers.

$dynamic[ '/(\s|\A|[([{<])"(?!\s)/u'   ] = '$1' . $opening_quote . '$2'; 
$dynamic[ '/"(\s|\S|\Z)/u'             ] = $closing_quote . '$1'; 

I would think that's safe to do as long as everyone is using UTF-8.

Patch added.

Btw, I don't do unit testing. What can we do to get this on a milestone now?

Thank you Sergey

Anyone have a few minutes to test my patch?