[ https://issues.apache.org/jira/browse/OAK-3730?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15045279#comment-15045279 ]
Julian Reschke edited comment on OAK-3730 at 12/7/15 5:23 PM: -------------------------------------------------------------- [~mreutegg]: this is a somewhat hacky (*) approach to implement {{VersionGCSupport.identifyGarbage}} which tries to only SQL-Select IDs that can potentially be split documents: {noformat} List<String> keyPatterns = Arrays.asList(new String[] { "_:p%", "__:p%", "___:p%", "____:p%", "_:h%", "__:h%", "___:h%", "___:h%" }); {noformat} (I believe this includes all split docs and all longpath docs, thus resulting in a new - harmless - false positives) (*) as it depends on assumptions about ID pattern formats was (Author: reschke): [~mreutegg]: this is a somewhat hacky (*) approach to implement {{VersionGCSupport.identifyGarbage}} which tries to only SQL-Select IDs that can potentially be split documents: {{noformat}} List<String> keyPatterns = Arrays.asList(new String[] { "_:p%", "__:p%", "___:p%", "____:p%", "_:h%", "__:h%", "___:h%", "___:h%" }); {{noformat}} (I believe this includes all split docs and all longpath docs, thus resulting in a new - harmless - false positives) (*) as it depends on assumptions about ID pattern formats > RDBDocumentStore: implement RDB-specific VersionGC support for lookup of > split documents > ---------------------------------------------------------------------------------------- > > Key: OAK-3730 > URL: https://issues.apache.org/jira/browse/OAK-3730 > Project: Jackrabbit Oak > Issue Type: Technical task > Components: rdbmk > Reporter: Julian Reschke > Fix For: 1.4 > > Attachments: OAK-3730.diff > > > This requires a change to the table layout (adding SD_TYPE and more), thus > needs thought about how to migrate from older versions of the persistence. -- This message was sent by Atlassian JIRA (v6.3.4#6332)