Here's a simple patch against the top of the 4.6-0.98 branch, which is used for testing.

If the general direction of that patch is good, I'll clean it up a bit and make a patch against master.

Thanks for the patch, Lars Hofhansl. It looks good. We're shooting to roll a new RC for 4.7.0 tomorrow, so let me know if you need this in there.

FYI, we release out of the 4.x and master branches. No plans to release from 4.6 branches.

As far as testing, I'd recommend adding one to QueryPlanTest.testExplainPlan().

Doesn't need to be in 4.7.0, although maybe it's strange to add this in 4.7.1.

I'm a bit confused about the branches. I see 4.x-HBase-1.x. But then the the pom from the master branch has 4.7.0-HBase-1.1.
So do I put this into master and cherry-pick down to the 4.x branches? Or is master not used?

It'd be fine for the next release (4.7.1 or 4.8.0). Since it's 100% client-side, it'd be fine in a 4.7.1 release.

4.x-HBase-0.98 is used to release the 0.98 compatible releases for 4.5, 4.6, 4.7, etc.
4.x-HBase-1.0 is used to release the 1.0 compatible releases for 4.5, 4.6, 4.7, etc.
master is used to release the 1.1+ compatible releases for 4.5, 4.6, 4.7, etc.

We create a 4.6-HBase-0.98, 4.6-HBase-1.0, and 4.6-HBase-1.1 branch upon release of the 4.6 release so that in the event there is user demand for patch releases (and an RM willing to step up) we have a branch on which to commit fixes.

Our typical modus operandi has been to continue active development on the 4.x and master branches to minimize back porting while ensuring that we maintain backward compatibility with previous minor releases such that users can upgrade through a rolling restart. Our contract (http://phoenix.apache.org/upgrading.html) is to support two minor releases back, but for 4.7 we've tested back as far as 4.4.

Thanks I guess what threw me off was the 4.x-HBase-1.x branch.

Here's a patch against master. Also includes a simple test. Should be good to commit.
If I get a +1 I'll commit to master and all three 4.x branches.

+1. Thanks, Lars Hofhansl. I'm confused by the existence of a 4.x-HBase-1.x branch too - I'll ask about that on the dev list.

Actually, I don't see a 4.x-HBase-1.x branch: https://git-wip-us.apache.org/repos/asf?p=phoenix.git;a=summary

Hmm... git branch -a gives me this:
...
remotes/origin/4.x-HBase-0.98
remotes/origin/4.x-HBase-1.0
remotes/origin/4.x-HBase-1.1
...

But when I checkout 4.x-HBase-1.1 I can't pull from it. Weird.

The 4.x-HBase-1.1 was a mistake and was deleted. We release the 1.1 releases from master.

Pushed to master, 4.x-HBase-0.98, and 4.x-HBase-1.0.
(I might have accidentally (re)created the 4.x-HBase-1.1 branch)

Offtopic from this jira... Should I/we delete the 4.x-HBase-1.1?

Should I/we delete the 4.x-HBase-1.1?

Yes, please.

SUCCESS: Integrated in Phoenix-master #1149 (See https://builds.apache.org/job/Phoenix-master/1149/)
PHOENIX-2697 Provide a SERIAL hint. (larsh: rev 0b1a180f1d9acc23cf58ecfc84a67aac110160cd)

• phoenix-core/src/main/java/org/apache/phoenix/execute/ScanPlan.java
• phoenix-core/src/main/java/org/apache/phoenix/iterate/SerialIterators.java
• phoenix-core/src/main/java/org/apache/phoenix/iterate/ExplainTable.java
• phoenix-core/src/main/java/org/apache/phoenix/execute/AggregatePlan.java
• phoenix-core/src/test/java/org/apache/phoenix/compile/StatementHintsCompilationTest.java
• phoenix-core/src/main/java/org/apache/phoenix/parse/HintNode.java

Done. Deleted branch 4.x-HBase-1.1.