[BUG] utf8_encode() has been deprecated.

[rosiel]

PHP Note: https://php.watch/versions/8.2/utf8_encode-utf8_decode-deprecated

What steps does it take to reproduce the issue?

• Caught by CI.

• When does this issue occur?
  When running CI on PHP 8.2.

• Where does it occur?
  https://github.com/Islandora/islandora/blob/2.x/modules/islandora_text_extraction/src/Plugin/Field/FieldFormatter/OcrTextFormatter.php#L135-L137

Resolution

mbstring, iconv and intl are suggested as replacements, but I don't know enough to suggest which one (or combination) to use, or whether we have access to these extensions in the default deployment.

[adam-vessey]

Given the utf8_encode() call is predicated on a call to mb_detect_encoding(), which seems to be from mbstring, would seem to point towards mbstring as the shortest path, already expected to be present?

Instead of the current

    if (mb_detect_encoding($contents) != 'UTF-8') {
      $contents = utf8_encode($contents);
    }

Something like:

    $detected_encoding = mb_detect_encoding($contents);
    if ($detected_encoding != 'UTF-8') {
      $contents = mb_convert_encoding($contents, 'UTF-8', $detected_encoding);
    }

?

---

Also, revisiting Drupal's requirements proper, it requires mbstring already: https://www.drupal.org/docs/getting-started/system-requirements/php-requirements#s-mbstring-

[rosiel]

fixed as part of #987