[
https://issues.apache.org/jira/browse/OAK-763?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13681140#comment-13681140
]
Marcel Reutegger commented on OAK-763:
--------------------------------------
I like the general idea to use a branch for the checkpoint feature. Some
questions:
Do we need two new methods for something very similar to the existing branches?
Wouldn't it be better to change the existing branch method or introduce an
overloaded version of NodeStore.branch() with a lifetime? I'd probably even
favor a single branch method with a lifetime parameter. Wouldn't it be
beneficial for the implementation to always have a hint on how long it must
keep an un-merged branch?
Ah, I think now I understand why those two steps/methods are there. We need a
marker that we can store in the index and later resume.
Alternatively we could move the checkpoint reference to NodeStoreBranch and
have a single NodeStore.branch(@Nullable String branchRef, long lifetime).
> Asynchronous indexing
> ---------------------
>
> Key: OAK-763
> URL: https://issues.apache.org/jira/browse/OAK-763
> Project: Jackrabbit Oak
> Issue Type: New Feature
> Components: core
> Reporter: Alex Parvulescu
> Assignee: Alex Parvulescu
> Attachments: OAK-763-mv-scheduler.patch, OAK-763.patch,
> OAK-763-v2.patch, OAK-763-v4.patch
>
>
> It would be good to have the possibility to update indexes asynchronously,
> outside the current commit hook (or editor).
> This would be used for the index impls which can be a bit behind, like the
> ones that do text extraction from binaries.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
