13:00:13 <gsmet> #startmeeting
13:00:14 <jbott> Meeting started Tue May 23 13:00:13 2017 UTC.  The chair is gsmet. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:00:14 <jbott> Useful Commands: #action #agreed #help #info #idea #link #topic.
13:00:15 <jbossbot> Title: MeetBot - Debian Wiki
13:00:22 <gsmet> hi all!
13:00:44 <sannegrinovero_> hello!
13:00:46 <gmorling> hey
13:01:01 <sannegrinovero_> who is gmorling ?? welcome..
13:01:08 <yrodiere> hi
13:01:08 <gmorling> hey
13:01:22 <gsmet> DavideD: you there?
13:01:39 <sannegrinovero_> he's coming. We're working together today.
13:02:06 <DavideD> Here I am
13:02:21 <gsmet> nice, we're all there, let's get going \o/
13:02:30 <gsmet> #topic Progress Davide
13:02:56 <DavideD> I''ve worked on grouping for Infinispan
13:03:09 <DavideD> It seems to work, I'll send a PR at the end of the meeting
13:03:28 <gsmet> OK cool, did it improve the number of queries significantly?
13:03:29 <sannegrinovero_> nice. Is that Infinispan embedded or remote?
13:03:38 <gmorling> except for @emmanuel if he wants to join, too?
13:03:43 <DavideD> Infinispan remote
13:03:51 <sannegrinovero_> ah even better
13:04:02 <DavideD> I'm not sure gsmet, I'll chek
13:04:15 <gsmet> you have existing tests with Byteman
13:04:31 <DavideD> Well, yes, it passes the tests
13:04:39 <gsmet> but they might be in one of the contrib repo, probably Redis
13:05:05 <DavideD> There are some in core that were disable and check th eexact operations executed
13:05:31 <DavideD> Those work and now infinispan is aliagned with the other dialect supporting the grouping
13:05:37 <gsmet> see https://github.com/hibernate/hibernate-ogm-redis/blob/master/redis/src/test/java/org/hibernate/ogm/datastore/redis/test/performance/RedisPerformanceTest.java
13:06:07 <gsmet> might be a good way to check for the improvement
13:06:11 <DavideD> gsmet, I missed those ones, I'll check after the chat
13:06:33 <gsmet> they are per datastore
13:06:48 <gsmet> it might help measuring the improvements
13:07:00 <DavideD> Sounds good
13:07:32 <DavideD> That's what I've done this week.
13:07:52 <gsmet> #topic Next 2 weeks Davide
13:07:57 <sannegrinovero_> FYI the Infinispan team is working on supporting transactions over Hot Rod, so it will be nice to eventually combine these all.
13:08:31 <DavideD> I've talked with Scott Marlow about adding a property in OGM to pass the connection to the datastore via JNDI
13:09:34 <DavideD> I'm not sure what to focus next, any suggestion? I haven't tested OGM iwth JDK 9
13:09:41 <DavideD> shoudl I have a look at that?
13:10:07 <gsmet> hmm taking a look at if/how grouping can work with neo4j would be the next step I think
13:10:12 <gsmet> then we could release a version
13:10:14 * emmanuel for your information just joined
13:10:30 <DavideD> Ok, sounds good to me. It should be easier with Neo4j
13:10:32 <gsmet> sannegrinovero_: agreed?
13:10:38 <gsmet> DavideD: not so sure :)
13:10:43 <yrodiere> emmanuel: o/
13:10:52 <gsmet> DavideD: well it might be easier because you might be unable to do anything :)
13:10:53 <sannegrinovero_> gsmet: sure I'm looking forward to a release
13:11:02 <DavideD> gsmet, :) I hope not
13:11:03 <sannegrinovero_> gsmet: not sure if that should be a blocker though?
13:11:10 <gsmet> or you might have to use the MongoDB approach
13:11:32 <gsmet> well, if it doesn't take too long, it's probably the best moment to deal with it
13:11:43 <DavideD> I think so as well
13:12:01 <sannegrinovero_> DavideD: about passing the connection. Couldn't Scott inject the "connection object" directly by reference? JNDI seems unnecessary
13:12:02 <gsmet> and we will check the grouping checkbox
13:12:36 <gsmet> so I would do:
13:12:40 <gsmet> grouping for Neo4j
13:12:41 <gsmet> release
13:12:52 <gsmet> move to ORM 5.2 and JDK 8
13:13:10 <gsmet> and Sanne could start working on Infinispan 9 integration
13:13:40 <sannegrinovero_> [on the datasource] Actually injecting it should be trivial, the problem is to agree on a well defined class definition for the interfaces. A module included in WildFly ?
13:14:01 <emmanuel> there is no interface
13:14:06 <emmanuel> it injectst he raw objects
13:14:19 <emmanuel> eg MongoDatasource (fake xample) for Mongo
13:14:23 <DavideD> sannegrinovero_, the idea is to inject the native connector/client
13:14:30 <sannegrinovero_> sure but which *type*, we need to at least make sure we load the MongoDatasource from the same module
13:14:43 <emmanuel> yes good point
13:15:00 <emmanuel> that's where we start to be screwed :D
13:15:13 <emmanuel> it will work for a while ;)
13:15:59 <gsmet> DavideD: sannegrinovero_: OK with the plan ^
13:16:01 <gsmet> ?
13:16:08 <DavideD> It works for me
13:16:13 <sannegrinovero_> gsmet: re OGM release +1
13:16:24 <sannegrinovero_> why are you asking my blessing specifically ? :D
13:16:37 <gsmet> because I know you don't agree :)
13:16:59 <emmanuel> gsmet: sannegrinovero_ disagrees that he disagrees
13:17:01 <sannegrinovero_> ok. No, it looks good. I just regret we're late compared to our plans but that's life.
13:17:24 <gsmet> yes, it's how it is
13:17:39 <gsmet> and having grouping for 2 core dialects will make a better announce
13:17:39 <sannegrinovero_> emmanuel: I'm not sure I can agree to that :P
13:17:49 <gsmet> DavideD: anything else?
13:17:57 <DavideD> no, that's all
13:18:06 <gsmet> thanks
13:18:11 <gsmet> #topic Progress Emmanuel
13:18:16 <gsmet> emmanuel: something to say? :)
13:18:24 <gsmet> no pressure...
13:18:33 <emmanuel> I made some progress on the finding the prod teams
13:18:51 <emmanuel> (the question Gunnar and Sanne asked earlier)
13:18:58 <emmanuel> but it does not look like resolved yet
13:19:28 <sannegrinovero_> emmanuel: the process I explained on hipchat seems to work? Not sure if it's worth bothering further.
13:19:29 <emmanuel> Otherwise on the Hibernate front I'll do a generic what's new in Hibernate next week at Voxxed Singapore
13:19:36 <emmanuel> sannegrinovero_: ?
13:19:39 <emmanuel> David is gone
13:19:45 <emmanuel> so there is no validator prod
13:20:10 <emmanuel> or do you think you shevel it to WF and wait for it is actually working in practice for all the pieces?
13:20:21 <emmanuel> HV HSEARCH specifically
13:20:30 <emmanuel> I'd like it to be checked
13:20:32 <sannegrinovero_> emmanuel: yes, as I mentioned that's how it has been done recently.
13:20:50 <emmanuel> and what will happen witht he prd guys moving to eng teams
13:20:57 <emmanuel> still the same?
13:21:10 <sannegrinovero_> emmanuel: I'd agree with you, would be nice for someone to be proactive, esp like testing the version matrix - like the problems I mentioned in the email with Gail.
13:21:11 <emmanuel> Proibably no because it was the idea to offer flexibility per product
13:21:18 <gsmet> (is it something we should discuss publicly?)
13:21:25 <sannegrinovero_> gsmet: no :)
13:22:04 <gsmet> let's deal with it later in the specific channel then
13:22:17 <sannegrinovero_> +1
13:22:22 <gsmet> emmanuel: anything else?
13:22:49 <emmanuel> NOt much, I'm waiting for the lib side of the vision plan
13:22:57 <gsmet> OK
13:23:01 <emmanuel> and I like it more and mroe to have an explicit lib section there
13:23:10 <emmanuel> as it's a cross team sync tool when you think abouot it
13:23:12 <gsmet> sannegrinovero_: did you start the work on it?
13:23:38 <sannegrinovero_> could we keep the community meeting to community work?
13:23:42 <gsmet> or basically, just ping us when you want us to take a look and add something to it
13:23:46 <gsmet> yeah OK
13:24:02 <sannegrinovero_> gsmet: yes thanks, some brainstorming iterations would be great.
13:24:20 <gsmet> emmanuel: thanks for the disruption :)
13:24:27 <gsmet> #topic Progress Guillaume
13:24:37 <emmanuel> you asked for my work related to Hibernate
13:24:40 <emmanuel> I gave you that
13:24:50 <gsmet> so I worked on automating the last bits of BV release process
13:24:55 <gsmet> so we are now fully automated
13:25:03 <gsmet> then I worked on the RI mostly
13:25:20 <gsmet> implementing what's needed to be on par with BV 2.0.0.Beta2
13:25:31 <gsmet> and changing a few things here and there
13:25:50 <gsmet> I also worked on the TCK to remove some testing utils that shouldn't have been there
13:25:57 <gsmet> and make the tests more consistent
13:26:08 <gsmet> it helped found a couple of bugs/mistakes
13:26:14 <gsmet> find*
13:26:29 <gsmet> #topic Next 2 weeks Guillaume
13:26:45 <gsmet> so I plan to release HV 6 Beta2 tomorrow morning
13:26:57 <gsmet> if Gunnar merges my PRs :)
13:27:17 <gsmet> next I will work on 2 things:
13:27:21 <gmorling> yes, next thing after the meeting :)
13:27:29 <gsmet> - adding the TCK testable markers to the spec
13:27:49 <gsmet> - starting moving some tests from HV to the TCK
13:28:04 <gmorling> it’s exactly one month until proposed final draft
13:28:06 <gsmet> (the testing utils are now consistent between the 2 projects so that should be easier)
13:28:14 <gmorling> do you think you’ll get along with the tCK?
13:28:21 <gmorling> or should someone else help out?
13:28:39 <gsmet> the plan is to know at the end of the week
13:29:01 <gsmet> and decide then
13:29:12 <gmorling> ok
13:29:22 <gsmet> we need a final of the RI for the final draft too?
13:29:38 <gmorling> i need to look how it was done last time
13:29:43 <gmorling> either final or a very good CR
13:29:47 <gmorling> i think the latter
13:30:03 <gsmet> I think we're not that far from CR
13:30:10 <gsmet> the remaining bits are not that important I think
13:30:11 <gmorling> yeah, same here
13:30:47 <gsmet> what I'm not sure about is the time I should spend on the RI
13:30:59 <gsmet> that's the big unknown right now
13:31:06 <gmorling> depends on what it is
13:31:12 <gmorling> if it’s missing features we need it
13:31:21 <gmorling> if it’s more polishing, then it can wait a bit
13:31:35 <gsmet> a bit meaning after the final draft?
13:31:50 <gmorling> tck and missing features are vital for the final
13:31:52 <gsmet> anyway, let's be clear on that before the end of the week
13:32:05 <gmorling> polishing in the RI can be done in a 6.0.1, too, if needed
13:32:28 <gsmet> and see with the team if we reorganize a bit for June or not
13:32:50 <gmorling> sounds good
13:32:56 <gsmet> that's pretty much it for me
13:33:09 <gsmet> ah, except end of the week is tomorrow :/
13:33:25 <gmorling> hehe, yes, worst case we’ll have to see on monday
13:33:26 <gsmet> so early next week it is
13:33:53 <gsmet> yeah I need some time to start the actual work on the spec/TCK
13:33:59 <gsmet> to see how it goes and plan from there
13:34:13 <gsmet> #topic Progress Gunnar
13:34:49 <gmorling> so i’ve been working on an update to the Public Review draft for BV until last week
13:34:58 <gmorling> i pushed it to the jcp folks, but they haven’t published it yet
13:35:08 <gmorling> which concerns me a bit, as the ballot is about to begin in one week
13:35:24 <gmorling> so i’ve asked heather (jcp chief) today about it
13:35:34 <gmorling> i hope they’ll do it quickly now
13:35:52 <gmorling> after that i mostly focused on debezium
13:36:04 <gmorling> i’d like to do “my” first release soon
13:36:20 <gmorling> it’s mostly bug fixes for this time around
13:36:46 <emmanuel> (ah, I've some hopes to give feedback on BV 1.1-2.0 diff review in my next two weeks)
13:36:59 <gmorling> or smaller feature requests; e.g. typically people will use a new column type not supported yet (say geometry columns in dbz) so we are adding that
13:37:20 <gmorling> of course i’m still fighting my ways through the dbz code base
13:37:36 <gmorling> there’s a parser for mysql ddl statements which causes some constant trouble it seems
13:37:48 <gmorling> not sure what to do about it
13:37:58 <gmorling> as pe randall there’s no authoritative and complete grammar for that
13:38:02 <gmorling> so the parser is hand-written
13:38:12 <gmorling> and often needs love to support further types of statements
13:38:32 <gmorling> so that’s that
13:38:37 <gmorling> i’ll get back to BV next week
13:38:46 <gmorling> and pick up the loose ends for the proposed final draft
13:39:12 <gmorling> we’ve filed some issues needing clarifcation, so that’s what i’ll mostly do
13:39:23 <gmorling> ha, i guess that already was my plan for the next two weeks
13:39:26 <gmorling> :)
13:39:34 <gsmet> yup :)
13:39:37 <gsmet> anything else?
13:39:43 <gmorling> thanks for the feedback, @emmanuel, that’ll be highly welcomed
13:39:47 <gmorling> not really, i think
13:40:07 <gsmet> #topic Progress Sanne
13:40:25 <sannegrinovero_> so, my 2 weeks period started with Devoxx UK
13:40:46 <sannegrinovero_> I had a lunch-time quicky slot but still was able to meet some interested people
13:41:06 <sannegrinovero_> after that I've been quite immersed in Search
13:41:42 <sannegrinovero_> the overall progress is great: https://hibernate.atlassian.net/issues/?filter=12266
13:41:43 <jbossbot> Title: Issue Navigator - Hibernate JIRA
13:42:06 <sannegrinovero_> and I hope to speed up more now that I'm getting (again) more familiar with the internals
13:42:27 <sannegrinovero_> I have a huge patch ready, I'm essentially "self-reviewing" it right now
13:42:30 <sannegrinovero_> as it impacts many areas
13:42:48 <sannegrinovero_> then we "just" have to burn down through all the trivial ones still open
13:43:13 <sannegrinovero_> In parallel on the Infinispan side we've been debating some things
13:43:24 <sannegrinovero_> like how to provide XML mapping for Search
13:43:33 <yrodiere> sannegrinovero_: your filter is missing "OR fixVersion = 5.8.next"
13:43:38 <sannegrinovero_> but it mostly boils down to separation of other concerns
13:43:44 <sannegrinovero_> yrodiere: oh, right thanks
13:44:14 <sannegrinovero_> so in parallel with 5.8 we're also keeping an eye on requirements coming from:
13:44:16 <sannegrinovero_> Lucene 7
13:44:20 <sannegrinovero_> Elasticsearch 6
13:44:22 <sannegrinovero_> ORM 6
13:44:24 <sannegrinovero_> Java 9
13:44:38 <sannegrinovero_> and Infinispan, e.g. free-form needs to be pushed
13:44:49 <sannegrinovero_> that's it for me.
13:44:57 <gsmet> OK, thanks!
13:45:15 <sannegrinovero_> ah I'm also planning some more chats with the ORM team about Java9 milestones
13:45:19 <gsmet> I suppose your next 2 weeks are just more progress on this?
13:45:29 <sannegrinovero_> gsmet: sure
13:45:43 <sannegrinovero_> BTW coming monday is bank holiday for those in UK
13:46:00 <gsmet> OK, let's deal with the last one then!
13:46:06 <gsmet> #topic Progress Yoann
13:46:33 <yrodiere> Something Sanne didn't mention is we released 5.8.Beta2
13:46:49 <sannegrinovero_> yrodiere: you did it all :)
13:47:21 <yrodiere> There wasn't any major change in it, mostly bugfixes IIRC
13:47:29 <sannegrinovero_> [fixed the JIRA filter, we've "just" 120 issues on our plate]
13:47:46 <yrodiere> anyway, after that I focused on some minor tickets
13:47:55 <yrodiere> and then I started to work on Elasticsearch normalizers :D
13:48:16 <yrodiere> we decided to introduce those in the embedded Lucene mode too
13:48:33 <yrodiere> (FYI a "normalizer" is an analyzer without a tokenizer)
13:48:51 <yrodiere> that's almost done, I should submit PRs (plural) today
13:49:02 <gsmet> cool!
13:49:47 <yrodiere> I hope to spend less time on the remaining tickets for 5.8 (about 30 of them, but they keep piling up faster than we fix them :/)
13:50:00 <yrodiere> next two weeks
13:50:10 <gsmet> #topic Next 2 weeks Yoann
13:50:17 <yrodiere> Well, those will be about addressing the remaining tickets for 5.8, obviously
13:51:02 <yrodiere> I'll have to take some time to fix stuff on the JSR-352 implementation too, because Mincong is starting to have trouble finding work (there are things that block him and that I should fix myself)
13:51:17 <yrodiere> And that should be it
13:51:25 <gsmet> ok, thanks!
13:51:37 <gsmet> thanks all, we're right on time
13:52:03 <gsmet> #endmeeting