[LUCENE-8525] throw more specific exception on data corruption

• Type: Improvement
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Lucene Fields: New
• Assignee: Unassigned
• Reporter: Vladimir Dolzhenko
• Created: 05/Oct/18 10:15
• Updated: 28/Aug/22 15:36

Description

DataInput throws generic IOException if data looks odd

DataInput:141

there are other examples like BufferedIndexInput:219, CompressionMode:226 and maybe DocIdsWriter:81

That leads to some difficulties - see elasticsearch #34322

It would be better if it throws more specific exception.

As a consequence SegmentInfos.readCommit violates its own contract

/**
   * @throws CorruptIndexException if the index is corrupt
   * @throws IOException if there is a low-level IO error
   */