From fb12a6a8350fd74bac77bb5a90bf1b2038c71881 Mon Sep 17 00:00:00 2001 From: "github-actions[bot]" Date: Wed, 8 Apr 2026 03:54:21 +0000 Subject: [PATCH] Deploy 0a550ca from main --- includes/class-peptide-news-rest-api.php | 19 ++++--------------- 1 file changed, 4 insertions(+), 15 deletions(-) diff --git a/includes/class-peptide-news-rest-api.php b/includes/class-peptide-news-rest-api.php index 70c444d..367bf7c 100644 --- a/includes/class-peptide-news-rest-api.php +++ b/includes/class-peptide-news-rest-api.php @@ -248,21 +248,10 @@ class Peptide_News_Rest_API { // Fallback 2: strip trailing nbsp-separated publisher names. // Google News RSS excerpts use double non-breaking space + publisher name // when the source field doesn't match (e.g., source = "news.google.com"). - // Uses string search instead of regex to avoid UTF-8/PCRE edge cases. - $double_nbsp = $nbsp . $nbsp; - $nbsp_pos = strrpos( $text, $double_nbsp ); - if ( false !== $nbsp_pos ) { - $after = substr( $text, $nbsp_pos + strlen( $double_nbsp ) ); - } else { - $nbsp_pos = strrpos( $text, $nbsp ); - if ( false !== $nbsp_pos ) { - $after = substr( $text, $nbsp_pos + strlen( $nbsp ) ); - } - } - if ( false !== $nbsp_pos && $nbsp_pos > strlen( $text ) * 0.4 ) { - // Only strip if what follows looks like a publisher name (< 60 chars, starts uppercase). - if ( ! empty( $after ) && strlen( $after ) < 60 && preg_match( '/^[A-Z]/u', $after ) ) { - return rtrim( substr( $text, 0, $nbsp_pos ) ); + // Uses mb_ereg for reliable multibyte matching. + if ( preg_match( '/^(.{10,}?)\x{00A0}{1,2}([A-Z][^.!?\x{00A0}]{0,55})$/u', $text, $m ) ) { + if ( mb_strlen( $m[1], 'UTF-8' ) > mb_strlen( $text, 'UTF-8' ) * 0.4 ) { + return rtrim( $m[1] ); } }