[
https://issues.apache.org/jira/browse/HBASE-12522?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14221793#comment-14221793
]
Sean Busbey commented on HBASE-12522:
-------------------------------------
{quote}
Sean Busbey is the branch-1 patch that different from master? It should not be.
I am going over the master patch in the review board.
{quote}
I don't expect it to be, but I haven't checked how non-clean the pick is yet.
{quote}
I've noticed that you are using yoda conditionals which is not the practice in
hbase. We do not have to fix that for this patch, but it is good to be
consistent throughout the code base.
{code}
if (null == providerId) {
{code}
{quote}
Ah yes. Habit from my C days; hadn't noticed a conscientious choice in HBase.
I'll change my habit and update the code style section of the ref guide.
{quote}
Is this a typo?
{code}
+ public void postSync(final long timeInNanos, final int handlerSyncs) {
+ source.incrementSyncTime(timeInNanos/1000000l);
}
{code}
{quote}
I don't think so? which file is this in? Presuming {{source.incrementSyncTime}}
wants milliseconds that would be the correct conversion from nanos.
> Backport WAL refactoring to branch-1
> ------------------------------------
>
> Key: HBASE-12522
> URL: https://issues.apache.org/jira/browse/HBASE-12522
> Project: HBase
> Issue Type: Task
> Components: wal
> Reporter: Sean Busbey
> Assignee: Sean Busbey
> Fix For: 0.99.2
>
>
> backport HBASE-10378 to branch-1.
> This will let us remove the Deprecated stuff in master, allow some baking
> time within the 1.x line, and will give us the option of pulling back follow
> on performance improvements.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
