[JCR-1988] AbstractRecord does inefficient List.indexOf()

• Type: Improvement
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: core 1.4.10, 1.6
• Component/s: jackrabbit-core
• Assignee: Unassigned
• Reporter: Marcel Reutegger
• Created: 19/Feb/09 11:11
• Updated: 10/Oct/10 14:42
• Resolved: 23/Feb/09 09:00

Description

AbstractRecord keeps a list of already used UUIDs and references
them by index when used again in a record. Using a List does not
scale well, when the record grows larger. e.g. a transaction of
10k nodes takes more than a minute on my machine when the journal
is enabled. Most of the time is spent doing List.indexOf() in
AbstractRecord.getOrCreateIndex().