Issue #3151364 by Charlie ChX Negyesi, amateescu, alexpott, jhodgdon:... (33b050ca) · Commits · project / drupal

core/lib/Drupal/Component/Transliteration/PhpTransliteration.php

+18 −0

Original line number	Diff line number	Diff line
		@@ -58,6 +58,21 @@ class PhpTransliteration implements TransliterationInterface {
		*/
		protected $genericMap = [];

		/**
		* Special characters for ::removeDiacritics().
		*
		* Characters which have accented variants but their base character
		* transliterates to more than one ASCII character require special
		* treatment: we want to remove their accent and use the un-
		* transliterated base character.
		*/
		protected $fixTransliterateForRemoveDiacritics = [
		'AE' => 'Æ',
		'ae' => 'æ',
		'ZH' => 'Ʒ',
		'zh' => 'ʒ',
		];

		/**
		* Constructs a transliteration object.
		*
		@@ -93,6 +108,9 @@ public function removeDiacritics($string) {
		if (strlen($to_add) === 1) {
		$replacement = $to_add;
		}
		elseif (isset($this->fixTransliterateForRemoveDiacritics[$to_add])) {
		$replacement = $this->fixTransliterateForRemoveDiacritics[$to_add];
		}
		}

		$result .= $replacement;

+11 −0

Original line number	Diff line number	Diff line
		@@ -22,3 +22,14 @@ function search_post_update_block_page(&$sandbox = NULL) {
		return $block->getPluginId() === 'search_form_block';
		});
		}

		/**
		* Mark everything for reindexing after diacritics removal rule change.
		*/
		function search_post_update_reindex_after_diacritics_rule_change() {
		$search_page_repository = \Drupal::service('search.search_page_repository');
		foreach ($search_page_repository->getIndexableSearchPages() as $entity) {
		$entity->getPlugin()->markForReindex();
		}
		return t("Content has been marked for re-indexing for all active search pages. Searching will continue to work, but new content won't be indexed until all existing content has been re-indexed.");
		}

+2 −2

Original line number	Diff line number	Diff line
		@@ -58,8 +58,8 @@ public function providerTestPhpTransliterationRemoveDiacritics() {
		// Test all characters in the Unicode range 0x01CD to 0x024F.
		['ǍǎǏ', 'AaI'],
		['ǐǑǒǓǔǕǖǗǘǙǚǛǜǝǞǟ', 'iOoUuUuUuUuUuǝAa'],
		['ǠǡǢǣǤǥǦǧǨǩǪǫǬǭǮǯ', 'AaǢǣGgGgKkOoOoǮǯ'],
		['ǰǱǲǳǴǵǶǷǸǹǺǻǼǽǾǿ', 'jǱǲǳGgǶǷNnAaǼǽOo'],
		['ǠǡǢǣǤǥǦǧǨǩǪǫǬǭǮǯ', 'AaÆæGgGgKkOoOoƷʒ'],
		['ǰǱǲǳǴǵǶǷǸǹǺǻǼǽǾǿ', 'jǱǲǳGgǶǷNnAaÆæOo'],
		['ȀȁȂȃȄȅȆȇȈȉȊȋȌȍȎȏ', 'AaAaEeEeIiIiOoOo'],
		['ȐȑȒȓȔȕȖȗȘșȚțȜȝȞȟ', 'RrRrUuUuSsTtȜȝHh'],
		['ȠȡȢȣȤȥȦȧȨȩȪȫȬȭȮȯ', 'ȠȡȢȣZzAaEeOoOoOo'],