13:01:15 <gsmet> #startmeeting
13:01:15 <jbott> Meeting started Tue May  9 13:01:15 2017 UTC.  The chair is gsmet. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:01:15 <jbott> Useful Commands: #action #agreed #help #info #idea #link #topic.
13:01:17 <jbossbot> Title: MeetBot - Debian Wiki
13:01:19 <gsmet> (identification issues)
13:01:27 <gsmet> let's start!
13:01:44 <gsmet> #topic Progress Davide
13:02:20 <DavideD> Last week there was an issue related to using HIbernate OGM with WIldfly
13:03:14 <DavideD> In the end it was an error in the documentation
13:03:39 <DavideD> It didn't show the right property to set for WildFLy to use the right ORM and Search combination
13:04:16 <DavideD> I've sent a PR request to update the workding and it seems the user was able to figure out how to fix
13:04:19 <DavideD> the problem
13:04:34 <gsmet> ok cool, nice he validated the PR
13:05:14 <DavideD> sorry
13:05:20 <DavideD> Client crashed
13:05:40 <DavideD> I have to admit that it is a bit confusing having to set that properties
13:06:48 <DavideD> I've also started to have a look at the documentaion and figure out why somebody shoudl use Hibernate OGM with the datastores
13:07:27 <DavideD> I've sent a PR for MongoDB and Infinispank but I'm not sure if there are specific use cases for Neo4j,  so I'm open to suggestion
13:08:33 <gmorling> “infinispank”, LOL
13:08:34 <sannegrinovero_> I like the name: Infinispank .. we should rebrand :)
13:08:34 <DavideD> I also had a quick exchange with Scott Marlow about wildfly-nosql: https://github.com/wildfly/wildfly-nosql
13:09:00 <DavideD> Damn, the new monitor I bought is too big, I need to increase the font :)
13:09:31 <sannegrinovero_> DavideD: TBH it also feels like you're typing very slow.
13:10:14 <DavideD> The project should provide a way for Hibernate OGM to get the client to connect to the datastores
13:10:23 <DavideD> basically it put them in the JNDI
13:10:53 <DavideD> I'm not sure if we should start to make OGM compatible with it
13:11:50 <DavideD> I also tried to have a look at Hibernate ORM 6 to check if there are problems for OGM
13:11:58 <DavideD> since they included the parser
13:12:17 <DavideD> but it's kind of hard because it does not compile
13:12:41 <DavideD> That's all I guess
13:13:07 <gsmet> you mean ORM 6 does not compile?
13:13:17 <DavideD> sannegrinovero_, yes I need to improve my touch typing abilities
13:13:24 <DavideD> Yes, if run the build it does not work
13:13:34 <DavideD> yet, it's still a WIP
13:13:57 <DavideD> Well, it wasn't working when I tried it, maybe now they fixed it
13:14:07 <gsmet> it might still be a WIP
13:14:15 <DavideD> yep
13:14:30 <gsmet> #topic Next 2 weeks Davide
13:14:47 <DavideD> Do you guys know anything about this Wildfly nosql project?
13:15:06 <gsmet> I remember Scott talking about it in Lisbon
13:15:12 <gsmet> but that's basically it
13:15:41 <DavideD> I'm not sure what to focus next in relation to Hibernate OGM, during the meeting we talked about moving the documentation on the hibernate.org website
13:16:09 <DavideD> do we still want to do that? It should improve the SEO but it seems that now the docs
13:16:20 <DavideD> appear correctly when you search them
13:16:21 <gsmet> IIRC, we wanted an ORM 5.2 compatible version
13:16:36 <gsmet> and the upgrade to the new Infinispan
13:16:58 <gsmet> did you move the contrib repos to the hibernate org?
13:17:03 <sannegrinovero_> gsmet: according to our meeting nodes, we actually want Infinispan9 and ORM 5.2 in the next minor of OGM
13:17:24 <sannegrinovero_> so I need you guys to finish this OGM minor so that we can start on those tasks
13:17:27 <sannegrinovero_> RELEASE!
13:17:35 <DavideD> ok
13:17:48 <DavideD> I mean, we release 5.2 the way is it now?
13:18:14 <sannegrinovero_> It seems to be feature complete.
13:18:18 <sannegrinovero_> Any reason to not release it?
13:18:29 <gsmet> IIRC, we wanted to generalize the grouping?
13:18:46 <gsmet> it might be a good idea to check what we decided in Lisbon
13:18:56 <gsmet> but I'm pretty sure there are some things missing for this version
13:19:18 <gsmet> note that we should only take care of the core dialects
13:19:22 <sannegrinovero_> right it's missing some things like grouping and batching improvements
13:19:45 <gsmet> it means Neo4j and Infinispan
13:19:48 <sannegrinovero_> however this is now a problem as I can't postpone much longer me working on Infinispan 9
13:20:08 <DavideD> OK, I'll release the alpha and work on the grouping this week
13:20:09 <gsmet> hmm ok
13:20:12 <sannegrinovero_> so maybe we need to revisit the plan..
13:20:49 <DavideD> but sannegrinovero_ what do you need to work on infinisspan? I mean we can create a branch 5.3
13:20:50 <sannegrinovero_> Not that I have nothing to do on Search, but the Infinispan 9 upgrade is getting late for our purposes.
13:20:54 <gsmet> so maybe we could take a look at Neo4j + grouping
13:21:07 <gsmet> it might be a no no due to the referential integrity
13:21:19 <gsmet> and move Infinispan + grouping to next release?
13:21:52 <gsmet> we would release (Infinispan 9 + grouping) + ORM 5.2 in the next release
13:22:06 <sannegrinovero_> sure I could work on Search first and do Infinispan 9 later, I see no point in just hacking on a 5.3 branch if we don't release it soon after though.
13:22:56 <gsmet> DavideD: what do we have in 5.2 right now?
13:23:08 <sannegrinovero_> we had good reasons to want to have an OGM version which would work with Infinispan 8, ORM 5.1 and also have better grouping.
13:23:47 <sannegrinovero_> if this is no longer possible, I'd rather remove grouping.. hard choices.
13:24:08 <gsmet> let's say we release 5.2 next week?
13:24:21 <gsmet> and we see if we can include grouping for Neo4j + Infinispan or not?
13:24:38 <sannegrinovero_> I'd like to consider that yes. Maybe let's hear what Emmanuel has to add in terms of big picture strategy?
13:24:51 <DavideD> There are no new feature currently because the amin focus was to remove the existing repositories, I have a branch where I've upgraded eNeo4j to the latest version that will need some rebase done
13:25:04 <gsmet> OK
13:25:09 <gsmet> that's what I feared
13:25:49 <gsmet> sannegrinovero_: 5.1 already had Infinispan + ORM 5.1 support
13:25:58 <gsmet> 5.2 is basically an empty shell for now
13:26:30 <sannegrinovero_> gsmet: we wanted to have the combination - quoting myself - "Infinispan 8, ORM 5.1 and also have better grouping"
13:26:51 <sannegrinovero_> so my problem is that if we postpone grouping to next minor
13:27:16 <sannegrinovero_> we'll no longer have the grouping feature backwards compatible with Ispn8 and WildFly's ORM version.
13:27:53 <sannegrinovero_> but fair enough, we didn't pour energy on OGM.
13:28:13 <sannegrinovero_> so maybe we want to keep the plan and just shift it forward for some months..
13:28:25 <sannegrinovero_> Or keep the schedule and cancel grouping.
13:28:48 <sannegrinovero_> We just need to decide on this with emmanuel_off
13:29:08 <gsmet> sannegrinovero_: as mentioned, 5.1 already has Infinispan 8 and ORM 5.1
13:29:34 <gsmet> so 5.2 would basically be better grouping (if we work on it) for Neo4j and Infinispan
13:29:36 <sannegrinovero_> gsmet: do I need to re-quote myself again ? :) I'm listing a combination of 3 features, not two.
13:30:12 <sannegrinovero_> "Infinispan 8, ORM 5.1 and ALSO have better grouping" -> we don't have this combination
13:30:23 <gsmet> yeah, well
13:30:25 <gsmet> :)
13:30:51 <gsmet> so basically, Davide starts on grouping and we see where we are next week?
13:30:57 <gsmet> and decide what to do from there
13:31:07 <DavideD> ok, sounds good to me
13:31:31 <gsmet> what to do might be skip 5.2 altogether
13:31:59 <gsmet> but I suspect the Infinispan 9 work might take a while
13:32:16 <sannegrinovero_> we can of course do grouping later if we agree on that. I'm merely stressing that in that case, we won't have the combination of (those 3) and AFAIR we considered that important to have for potentially promoting it as a supportable combination.
13:32:16 <gsmet> so Davide, maybe start on grouping for Infinispan
13:32:38 <gsmet> if we can have this in, Sanne can start working on Infinispan 9 without future merge issues
13:33:13 <DavideD> ok, I'll start with that
13:33:40 <gsmet> it should have more benefits than for Neo4j
13:34:41 <sannegrinovero_> cool
13:35:03 <DavideD> ok, I think that's all then from me
13:35:09 <gsmet> DavideD: first things to decide is to see which model is the best to copy
13:35:32 <DavideD> model?
13:35:34 <gsmet> from batch (MongoDB) or grouping by entity (the others)
13:36:09 <gsmet> anyway, we can discuss it together later
13:36:15 <sannegrinovero_> let's also consider that some of these will support both transactions & batches
13:36:22 <gsmet> let's continue, we're very late
13:36:23 <sannegrinovero_> surely gets complex :)
13:36:30 <sannegrinovero_> +1
13:36:35 <gsmet> #topic Progress Guillaume
13:36:57 <gsmet> so, I worked a bit on merging the core/contrib PR for OGM
13:37:13 <gsmet> I wrote a blog post on simple query string
13:37:23 <gsmet> no real feedback on it (not really surprised)
13:37:33 <gsmet> I think I should probably include part of it in our documentation
13:38:15 <gsmet> next I worked on having HV on part with the last BV spec
13:38:22 <gsmet> working on group conversions lately
13:38:42 <gsmet> I thought it would be easy
13:38:45 <gsmet> but it's not
13:39:04 <gmorling> a dev’s famous last words
13:39:04 <gsmet> and as it was a subject I didn't touch at all before that, it's full of surprises
13:39:16 <gsmet> mostly bad surprises or it wouldn't be fun
13:39:30 <gmorling> groups and sequences are the #1 sore point of BV imho
13:39:48 <gsmet> yeah, was happy so far to have avoided them
13:40:16 <gsmet> I also reviewed a couple of Yoann's PR last week (the easiest ones)
13:40:25 <gsmet> that's pretty much it
13:40:32 <gsmet> #topic Next 2 weeks Guillaume
13:40:58 <gsmet> so, I hope to finish this convert things today, at least getting the existing tests work
13:41:07 <gmorling> cool
13:41:13 <gsmet> I still need to add tests to exercise the new features
13:41:38 <gsmet> I expect quite some review feedback as I'm not very happy about one thing and I suppose Gunnar won't be either :)
13:41:44 <gmorling> oha
13:41:50 <gsmet> we'll see
13:41:53 <gmorling> you made me curious
13:41:59 <gsmet> :)
13:42:02 <gmorling> anyways, i don’t care too much about impl beauty at this point
13:42:31 <gsmet> yeah, let's get it working first
13:42:42 <gsmet> we will be able to beautify it later
13:42:45 <gmorling> +1
13:42:58 <gsmet> next I'll work on the metadata support for container elements
13:43:11 <gsmet> and we should be able to release the next HV version
13:43:30 <gmorling> that’d be great
13:43:42 <gsmet> hope I won't have bad surprises there too, we'll see
13:44:08 <gsmet> and after that, I'll focus on the TCK for a while I think
13:44:23 <gsmet> that's pretty much it
13:44:29 <gsmet> #topic Progress Gunnar
13:44:47 <gmorling> ok, so i’ve been spending last week mostly on debezium and getting into it
13:45:07 <gmorling> fixed some small bugs, merged some PRs, added comments (via disqus) to the blog
13:45:16 <gmorling> everything to get into it
13:45:40 <gmorling> as of this week i’m back to BV now
13:45:48 <gmorling> (that context switch hurts :( )
13:46:01 <gmorling> as you know we are in the public review phase atm.
13:46:24 <gmorling> i’m planning to do an update on the public review draft, likely in exactly two weeks from now
13:46:40 <gmorling> that’s essentially the time we got for ironing out any substantial issues
13:46:48 <gmorling> after that, the ballot will begin
13:47:06 <gmorling> i think we can do some smaller changes after public review, but nothing dramatic
13:47:37 <gmorling> on thursday i’ll have a conf talk on BV
13:47:55 <gmorling> but i’ll meet with emmanuel, so we hopefully can discuss some of the loose ends
13:48:19 <gmorling> and once done with the public draft update i reckon i’ll swing back to DBZ
13:48:30 <gmorling> that’s really it from me
13:48:37 <gmorling> (in terms of looking back and forward)
13:48:42 <gsmet> ok, thanks
13:48:49 <gsmet> #topic Progress Sanne
13:48:52 <sannegrinovero_> so, my last two weeks have been quite exciting
13:48:58 <sannegrinovero_> and busy..
13:49:03 <sannegrinovero_> first: preparing for Summit and actually deep-dive learning about some of our other products.
13:49:11 <sannegrinovero_> I also spent some time learning about the transaction logs within Elasticsearch;
13:49:18 <sannegrinovero_> specifically the time it takes to run the Search testsuite is killing us so I want to do something about it.
13:49:28 <sannegrinovero_> Then actually getting to Summit - working a bit in the 5am timezone - but mostly to catch up with complex questions via email
13:49:32 <sannegrinovero_> While helping out people interested in our projects for the rest of the day
13:49:43 <sannegrinovero_> on the Search side I've been working on some very large but simple refactorings, again on decupling from Class as a type, to pave the road for free-form.
13:49:51 <sannegrinovero_> I started over and it's a race to finish it now soon or it will be another huge conflict which I'll not be able to rebase
13:50:00 <sannegrinovero_> (the old branch was having too many conflicts)
13:50:06 <sannegrinovero_> this is the 4th attempt btw..
13:50:13 <sannegrinovero_> I had also worked on JGroups 4 upgrade but Yoann decided to refactor it all :D
13:50:19 <sannegrinovero_> So that work is doomed already as well..
13:50:24 <sannegrinovero_> On my near future:
13:50:30 <sannegrinovero_> I need to work on OGM now as Devoxx UK is in 2 days
13:50:36 <sannegrinovero_> I also have two half days off this week.
13:50:44 <sannegrinovero_> It's a problem as I won't be able to finalize the Search refactoring but I can't move Devoxx ;)
13:50:50 <sannegrinovero_> So maybe we'll need to integrate a milestone of my refactoring.. finishing the whole thing is not doable
13:50:55 <sannegrinovero_> after devoxx though ;)
13:51:08 <sannegrinovero_> After Devoxx I'm not sure if I should spend a week on Search or OGM.
13:51:16 <sannegrinovero_> OGM seems overdue, but we're also getting close to important milestones for Search.. would be good to push it to the finish line.
13:51:29 <sannegrinovero_> I also have 2 more Search presentations locally
13:51:36 <sannegrinovero_> It looks like we need to talk about the OGM roadmap, and decide based on that.
13:51:41 <sannegrinovero_> Especially now that Debezium joined the group, there's an impact and ignoring that is probabably the worst strategy.
13:51:50 <sannegrinovero_> that's it for me ;)
13:52:06 <gsmet> ok, thanks
13:52:21 <gsmet> Yoann is on PTO today so we're right on time.
13:52:25 <gsmet> thanks all
13:52:27 <sannegrinovero_> questions?
13:53:10 <gsmet> did we start work on HS 6?
13:53:20 <sannegrinovero_> gsmet: in some sense..
13:53:26 <gsmet> (I suppose your freeform refactoring is part of that)
13:53:41 <sannegrinovero_> gsmet: most of the internal changes we're doing are "nice to have" for 5.8
13:53:45 <sannegrinovero_> but would be blockers for 6
13:53:56 <sannegrinovero_> e.g. https://github.com/hibernate/hibernate-search/pull/1388
13:54:06 <sannegrinovero_> Filter literally doesn't exist as a type on newer Lucene versions
13:54:20 <gsmet> yes, OK
13:54:27 <gsmet> so some preliminary work is ongoing
13:54:31 <gsmet> nice to know
13:54:32 <sannegrinovero_> In my experience it's easier to migrate by getting rid of the deprecated types one by one
13:54:40 <sannegrinovero_> rather than having a massive 20,000 compile errors ;)
13:54:52 <sannegrinovero_> at least we have testsuites to keep us on track
13:55:15 <sannegrinovero_> also, it gives us the opportunity to put nice "Deprecations" in place to help end users  to move on
13:55:26 <sannegrinovero_> but that's not our primary goal..
13:55:39 <sannegrinovero_> I'm taking your objections from the meeting in consideration ;)
13:56:14 <gsmet> OK, thanks :)
13:56:21 <gsmet> let's close this meeting?
13:56:24 <sannegrinovero_> ok
13:56:29 <sannegrinovero_> good timing!
13:56:42 <gsmet> #endmeeting