Adrien Grand created LUCENE-5342:
------------------------------------
Summary: CompressingStoredFieldsFormat's bulk merge should be
disabled when reader and writer don't have the same chunk size
Key: LUCENE-5342
URL: https://issues.apache.org/jira/browse/LUCENE-5342
Project: Lucene - Core
Issue Type: Bug
Affects Versions: 4.5, 4.5.1
Reporter: Adrien Grand
Assignee: Adrien Grand
Priority: Blocker
Fix For: 4.6
LUCENE-5188 changed the way data is compressed when single documents exceed the
chunk size, meaning that it is not legal anymore to bulk merge if writer and
reader have different chunk sizes.
This bug only happens when mixing two instances of
CompressingStoredFieldsFormat that have the same compression mode but different
chunk sizes, and when there are documents that exceed any of the chunk sizes.
So the default codec is not impacted since it has always had the same chunk
size (16KB).
--
This message was sent by Atlassian JIRA
(v6.1#6144)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
