Re: [hibernate-dev] Pooled Optimiser Improvements

With my original patch the intention was that that the thread local blocks were smaller than the incrementSize, so not every thread local allocation would require a DB call. Your patch changes that approach but I don't think it actually matters that much, the overall performance should still be

Re: [hibernate-dev] Pooled Optimiser Improvements

https://github.com/scottmarlow/hibernate-orm/commits/pooledOptimizer_5.x is looking more correct now, if others want to look at that. On 12/15/2015 07:58 PM, Scott Marlow wrote: > > > On 12/15/2015 05:58 PM, Scott Marlow wrote: >> >> >> On 12/15/2015 05:40 PM, Scott Marlow wrote: >>> I changed th

Re: [hibernate-dev] Pooled Optimiser Improvements

On 12/15/2015 05:58 PM, Scott Marlow wrote: > > > On 12/15/2015 05:40 PM, Scott Marlow wrote: >> I changed the new test methods a bit. [2] seems to be passed the tests >> but I am not understanding how PooledThreadLocalLoOptimizer should >> coordinate with the AccessCallback to allocate the next

Re: [hibernate-dev] Pooled Optimiser Improvements

On 12/15/2015 05:40 PM, Scott Marlow wrote: > I changed the new test methods a bit. [2] seems to be passed the tests > but I am not understanding how PooledThreadLocalLoOptimizer should > coordinate with the AccessCallback to allocate the next chunk of > sequence numbers. > > We seem to be able

Re: [hibernate-dev] [ORM] Synchronization on AbstractLoadPlanBasedLoader

Looks like this is related to HHH-8679 and HHH-8939: IIRC, the code in AbstractLoadPlanBasedLoader was copied from Loader. Here's the commit that removed the comment you mention from Loader# wrapResultSetIfEnabled and added a synchronized block to retreiveColumnNameToIndexCache (HHH-8679): https:

Re: [hibernate-dev] Pooled Optimiser Improvements

I changed the new test methods a bit. [2] seems to be passed the tests but I am not understanding how PooledThreadLocalLoOptimizer should coordinate with the AccessCallback to allocate the next chunk of sequence numbers. We seem to be able to call AccessCallback.getNextValue() to get the next

Re: [hibernate-dev] refresher needed on running Hibernate unit tests in Intellij?

On 12/15/2015 03:51 PM, andrea boriero wrote: > i think you are using java 7 instead of java 8.. Not really sure but commenting out the below line seemed to help me. As far as I can tell, I have jdk 1.8 selected for the Hibernate project. > > On 15 December 2015 at 20:32, Scott Marlow

Re: [hibernate-dev] refresher needed on running Hibernate unit tests in Intellij?

Yes, this points back to a discussion we had here wrt allowing Java 8 for tests and the problems that causes for IDE imports. On Tue, Dec 15, 2015 at 2:52 PM andrea boriero wrote: > i think you are using java 7 instead of java 8.. > > On 15 December 2015 at 20:32, Scott Marlow wrote: > > > Hi,

Re: [hibernate-dev] refresher needed on running Hibernate unit tests in Intellij?

i think you are using java 7 instead of java 8.. On 15 December 2015 at 20:32, Scott Marlow wrote: > Hi, > > I created a new Java project for the folder containing Hibernate master > and imported the gradle build file (accepted defaults). When I tried > running a unit test, I get a compile erro

[hibernate-dev] refresher needed on running Hibernate unit tests in Intellij?

Hi, I created a new Java project for the folder containing Hibernate master and imported the gradle build file (accepted defaults). When I tried running a unit test, I get a compile error in org.hibernate.test.schemaupdate.SchemaUpdateDelimiterTest on the call to "Files.readAllLines( output.t

Re: [hibernate-dev] Pooled Optimiser Improvements

Those tests tend to assert the increments. We seem to agree that this ThreadLocal one can skip gaps of values. I'd look there first. On Tue, Dec 15, 2015 at 11:34 AM Scott Marlow wrote: > I'm trying to move the optimizer to PooledThreadLocalLoOptimizer [1]. > We are currently failing some new

Re: [hibernate-dev] Pooled Optimiser Improvements

I'm trying to move the optimizer to PooledThreadLocalLoOptimizer [1]. We are currently failing some new unit tests, which are cloned from existing PooledLoOptimizer tests which might be part of the problem. Scott [1] https://github.com/scottmarlow/hibernate-orm/tree/pooled-optimiser-hack On 12

Re: [hibernate-dev] [ORM] Synchronization on AbstractLoadPlanBasedLoader

I did not add those comments; they were just in some code I copied over into that class. On Tue, Dec 15, 2015, 4:02 AM Sanne Grinovero wrote: > On 15 December 2015 at 01:46, Steve Ebersole wrote: > > It's very possible that code comments may no longer be pertinent. > > Right, that's what I'm tr

Re: [hibernate-dev] [CI] MariDB size on jenkins slaves

It seems that the latest version of MariaDB has a feature that simplify the process. I'll try On Tue, Dec 15, 2015 at 10:42 AM, Sanne Grinovero wrote: > On 15 December 2015 at 10:23, Davide D'Alto wrote: > > Hi, > > the file ibdata1 created by MariaDB has reached the size of 1G. > > This happ

Re: [hibernate-dev] [CI] MariDB size on jenkins slaves

On 15 December 2015 at 10:23, Davide D'Alto wrote: > Hi, > the file ibdata1 created by MariaDB has reached the size of 1G. > This happens because when deleting the tables the size of the db does not > decrease. > > There are some possible solutions I can try, but I need to change the > configurat

[hibernate-dev] [CI] MariDB size on jenkins slaves

Hi, the file ibdata1 created by MariaDB has reached the size of 1G. This happens because when deleting the tables the size of the db does not decrease. There are some possible solutions I can try, but I need to change the configuration of the db and probably upgrade it. Would that be ok? I've a

Re: [hibernate-dev] [ORM] Synchronization on AbstractLoadPlanBasedLoader

On 15 December 2015 at 01:46, Steve Ebersole wrote: > It's very possible that code comments may no longer be pertinent. Right, that's what I'm trying to figure out. Do you remember which possible deadlock it might have referred to? > > On Mon, Dec 14, 2015 at 10:26 AM Sanne Grinovero > wrote: >