[LUCENE-2049] ArabicStemFilter removes numbers

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Invalid
• Affects Version/s: 2.9, 2.9.1
• Component/s: modules/analysis
• Lucene Fields: New
• Assignee: Unassigned
• Reporter: Jesper Badstue
• Created: 10/Nov/09 12:01
• Updated: 28/Aug/22 12:13
• Resolved: 10/Nov/09 14:11

Description

It seems that if a word consist or contains a number the word is removed by the arabic stemmer.

The bug can be reproduced by creating a unit test in TestArabicStemFilter.java

public void testNonArabicNumbers() throws IOException

{ check("123", "123"); }

This test fails as the returned string is empty.