Re: [DISCUSS] Wrapping log calls in Conditionals like isDebugEnabled, isTraceEnabled, etc.

I doubt that the JIT can optimize the call out. That's what you're asking it to do is optimize a method call away. -- Mike Stolz Principal Engineer, GemFire Product Lead Mobile: +1-631-835-4771 Download the GemFire book here.

Debugging Multiple DUnit VMs with IntelliJ

Hi all, I remembered a couple email threads discussing how it was possible to debug multiple DUnit processes from IntelliJ, but couldn't find a wiki article or a very complete description, so I wrote a wiki article. It's a little barebones right now, but there it is. Please edit or comment with an

Awaitility may time out before evaluating the condition

Hello All, Dale and I encountered an interesting Awaitility bug where an Awaitility using "untilAsserted" timed out before it evaluated the condition. It appears that Awaitility starts its timer before the thread has completely started, and when the duration specified in "atMost" is reached, a Tim

[Spring CI] Spring Data GemFire > Nightly-ApacheGeode > #1036 was SUCCESSFUL (with 2432 tests). Change made by Mark Paluch.

--- Spring Data GemFire > Nightly-ApacheGeode > #1036 was successful. --- Scheduled with changes by Mark Paluch. 2434 tests in total. https://build.spring.io/bro

Re: [DISCUSS] and the NEW Apache Geode 1.7.0 release branch has been created

Unfortunately it would require a fix to the build—it’s not about producing the release candidate. It’s when a user builds from the source release that the version info is ignored. Anthony > On Sep 10, 2018, at 10:02 AM, Nabarun Nag wrote: > > Hello Anthony, > > I plan to do that while creat

Re: [DISCUSS] Wrapping log calls in Conditionals like isDebugEnabled, isTraceEnabled, etc.

Alright I'm more up-to-date on this thread. Some things to consider: The lambdas look great, but we'd have to start splitting the Geode clasess. They're so big right now, that if you change a bunch of log statements to use lambdas you'll hit the max number of lambdas on many of our classes. We hit

Re: [DISCUSS] Wrapping log calls in Conditionals like isDebugEnabled, isTraceEnabled, etc.

The reason we use these guards is that the Log4j2 implementation is more expensive than hitting a volatile. Please don't change anything unless you start writing lots of JMH benchmarks! The existing code went through lots of iterations of perf testing that isn't part of Geode. Every Log4j2 Logger

Re: [DISCUSS] and the NEW Apache Geode 1.7.0 release branch has been created

Hello Anthony, I plan to do that while creating the release candidate. If there are no concerns raised on the release branch, I will start with the process soon. Regards Nabarun Nag On Mon, Sep 10, 2018 at 8:51 AM Anthony Baker wrote: > Looks good Naba! Only thing I see right now is that buil

Re: [DISCUSS] Wrapping log calls in Conditionals like isDebugEnabled, isTraceEnabled, etc.

I think that logging in hot paths/loops is probably something that should be avoided. And if it is necessary, I suspect that the JIT will short-circuit that call since debug levels don't generally change. There probably are a few hot paths that need to optimize logging, but that's not the majority

Re: [DISCUSS] and the NEW Apache Geode 1.7.0 release branch has been created

Looks good Naba! Only thing I see right now is that building from the source distribution does not use the .buildinfo file, leaving the version string empty. Anthony > On Sep 7, 2018, at 9:15 AM, Nabarun Nag wrote: > > CORRECTION: if '*no*' concerns are raised, we will start with the voting