Improve sanitize_title_with_dashes % removal

[duck_]

The current method of % removal involves placeholders to prevent stomping on escape sequences:

// Preserve escaped octets.
$title = preg_replace('|%([a-fA-F0-9][a-fA-F0-9])|', '---$1---', $title);
// Remove percent signs that are not part of an octet.
$title = str_replace('%', '', $title);
// Restore octets.
$title = preg_replace('|---([a-fA-F0-9][a-fA-F0-9])---|', '%$1', $title);

This leads to sanitize_title_with_dashes('---aa---') producing %aa instead of just aa.

[2189]

[duck_]

25021.negative-lookahead.diff​ replaces the placeholder system with a single regex with a negative lookahead. This fixes the issue, but is very slightly slower in my very limited testing.

[SergeyBiryukov]

Related: #3329

[nacin]

I'm going to guess "very slightly slower" is insignificant enough to warrant this change? We've always suggested that sanitize_title() and friends are not light and thus should be avoided where possible anyway.

How might this affect sanitize_title_for_query() (as in, working with older slugs that don't sanitize the same way)?

[SergeyBiryukov]

Related: ticket:14773:14

[nacin]

Replying to nacin:

How might this affect sanitize_title_for_query() (as in, working with older slugs that don't sanitize the same way)?

As long as this doesn't break anything, I'm game.