Non-Latin File Names Are Kept Apart and Listed
Files named in Hebrew, Arabic, Cyrillic, Chinese and other non-Latin scripts are now stored, listed and counted correctly. Before, two uploads whose names had the same length could be treated as one document, so a collection's file list showed fewer files than its file count, under names made of underscores.
- Every file is listed: a collection's file list and its file count now come from the same files, so the two numbers match.
- Real names: files are shown, and found by name, under the name they were uploaded with.
- One file never replaces another: uploads with different names stay separate documents. Uploading a file again under the same name still replaces the earlier version.
- Affected collections are repaired: files missing from the list are listed again, and files that had dropped out of search were indexed again from their originals. No action is needed.