Log of the #fcrepo channel on chat.freenode.net

Using timezone: Eastern Standard Time
* dwilcox joins00:04
* dwilcox leaves00:09
* dwilcox joins00:10
* dwilcox leaves00:26
* jrgriffiniii joins00:32
* dwilcox joins06:55
* dwilcox leaves07:29
* dwilcox joins07:49
* dwilcox leaves07:54
* dwilcox_ joins
* dwilcox_ leaves07:59
* jrgriffiniii leaves08:41
* dwilcox joins08:51
* dwilcox leaves08:59
* kefo joins09:00
* jrgriffiniii joins09:07
* acoburn joins09:14
* kefo leaves09:15
* bseeger joins
* dwilcox joins09:18
* ajs6f joins09:24
* peichman joins09:33
* bseeger leaves09:41
<ajs6f>acoburn: Did you folow this :https://github.com/apache/jena/pull/140 ?10:16
acoburn: Might be interesting for the Camel toolkit.
<acoburn>ajs6f: I didn't see that PR, but it looks great10:17
<ajs6f>acoburn: There's a whole story behind that on the PR to which @stain linked. Some random nice person want to make Jena's OSGi artifact work better and we got into getting a new release of Dexx Collections and stuff, but the improvements came at a price. In the older PR, there are some useful notes about how various versions of Karaf do and do not play well with Jena.10:19
* whikloj joins10:22
<acoburn>ajs6f: "The OSGi users are used to this process of a painfull dependcy resolution" β€” too true
ajs6f: it looks like jpullmann is working on a features.xml provisioning artifact for jena and karaf10:23
<ajs6f>acoburn: Aye. But in hindsight, the fact that so many languages express encapsulation more richly than they express dependency was a time bomb.
acoburn: Yes, shucks! I now remember that I wanted to put you two in contact for that very reason, but it slipped my mind.10:24
* bseeger joins10:25
<ajs6f>acoburn: There must be _dozens_ of people who have built Karaf features, Sling bundlelists, or whatever Websphere or Wildfly or Eclipse Virgo use over the years, for Jena. I don't think that stuff belonds in Jena proper, but it would be nice for there to be a sidecar project maintained by those interested, no?
<acoburn>ajs6f: yes, definitely10:26
<ajs6f>acoburn: Okay, get on that. I'll expect results by COB today.
<acoburn>ajs6f: a lot of projects (camel, activemq, etc) have a separate groupId for this, which provides some separation from the core codebase10:27
<ajs6f>acoburn: Yeah, it's not the mechanical questions. I think AndyS would probably veto it. He's very sparing with committments and he wants efforts to go to functionality.
<acoburn>ajs6f: that's completely fair β€” putting this into a sidecar project would work, too10:28
<ajs6f>acoburn: How do you want to start? I do not know who jpullman isβ€” do you?10:30
<acoburn>ajs6f: I don't know him either10:31
<ajs6f>acoburn: Urb. not much of a lead there. I don't think github can send someone a message.
acoburn: Maybe just users@jena.apache.org or dev@jena.apache.org.10:32
<acoburn>ajs6f: that would work
<ajs6f>acoburn: Do we have anything we could extract from your Camel work to offer as a starting place?10:36
<acoburn>ajs6f: there's this: https://git.io/vrUkY10:37
ajs6f: in my experience, these features.xml files need to be manually curated
ajs6f: I've tried using a maven plugin for this, but that never worked all that well10:38
<ajs6f>acoburn: Yeah, the idea is to pull out the Jena stuff and manually curate that in isolation.
<acoburn>ajs6f: pax exam ends up being a real life-saver
<ajs6f>acoburn: Or at least that was my idea.10:39
acoburn: I wonder if Clerezza has stuff for this.10:40
<acoburn>ajs6f: yes, they do
<ajs6f>acoburn: Maybe that's the starting point? Do we use that?10:41
<acoburn>ajs6f: I used to use clerezza for the camel/rdf stuff, but clerezza had a slightly older version (missing a certain class), and so there were some issues with our existing integration testing framework
<ajs6f>acoburn: Urb. That's going to be an issue. I'm only interested in doing this if it keeps pace with jena.10:42
<acoburn>ajs6f: I know that clerezza updated it's jena facades, but that was after I had migrated off; I don't know if they target 3.x or 2.x
<ajs6f>acoburn: Well, I don't know where to begin. Actually Apache Con is probably the right place ot begin.10:44
<acoburn>ajs6f: probably. I haven't been to that conference since 2013, so maybe it's time to go again10:45
<ajs6f>acoburn: Kind of too late for this year http://www.apachecon.com/10:46
<acoburn>ajs6f: yes, and there are deep discounts for registering early10:47
ajs6f: especially for apache committers
<ajs6f>acoburn: Okay, Plan B. Maybe a cross-list post to dev@jena and dev@clerezza?
acoburn: How does Marmotta wire up?
<acoburn>ajs6f: I know very little about marmotta10:48
<ajs6f>acoburn: https://github.com/apache/marmotta/blob/develop/platform/marmotta-core/src/main/resources/META-INF/beans.xml10:50
acoburn: is a CDI beans file. But I don't know how much that means.
acoburn: Urg, looks like something custom: https://wiki.apache.org/marmotta/Customizing#Modules10:51
* dwilcox leaves10:53
<acoburn>ajs6f: they're using some kind of shim: https://github.com/apache/marmotta/tree/develop/libraries/ldpath/ldpath-backend-jena
* github-ff joins
[fcrepo4-vagrant] whikloj created FCREPO-2020 (+1 new commit): https://git.io/vrUt910:54
fcrepo4-vagrant/FCREPO-2020 ed9b1ab Jared Whiklo: Roll master back to the last release
* github-ff leaves
<acoburn>ajs6f: based on jena 2.7.x
* github-ff joins10:55
[fcrepo4-vagrant] whikloj opened pull request #45: Roll master back to the last release (master...FCREPO-2020) https://git.io/vrUtj
* github-ff leaves
<ajs6f>acoburn: That's for LDPath backends, yes, but they're using it all over.
* kefo joins
<acoburn>whikloj: for vagrant, I wonder if a git-flow style of development would be beneficial β€” master tracks the current release while development happen in the dev branch10:58
<ajs6f>whikloj:acoburn: I think y'all need to figure out what the vagrant artifact is _for_.
<whikloj>acoburn: The question is, who generates the WAR artifacts and how often?
ajs6f/acoburn: Or do we change vagrant to instead clone and compile?
<acoburn>ajs6f: I think awoods mostly uses it for demos and workshops10:59
ajs6f: I certainly don't use it
<ajs6f>whikloj:acoburn: Is it for dev? Demos? An example of config?
whikloj:acoburn: If it is for demos, then stability matters most.
<whikloj>ajs6f/acoburn: I'd say it is mostly a demo machine, so stability seems most beneficial11:00
<acoburn>ajs6f: this is precisely why I urge sidecar projects *not* to depend on snapshot artifacts
<whikloj>ajs6f/acoburn: I use sometimes for dev, but I bring it up and replace the fcrepo.war with one I compiled from fcrepo-webapp-plus
acoburn/ajs6f: Maybe a discussion for Thursday's call?11:01
<ajs6f>whikloj: That sounds like an edge case. Commtters often do weird things.
<whikloj>ajs6f: true true
<ajs6f>whikloj: Fine, but I have no real opinion. I'm just trying to push for the idea that you can't make the right decision until you understand the purpose of the thing.
* Iome leaves11:02
* dwilcox joins
<whikloj>ajs6f: ok, but either I can declare its purpose or we'll need a shared understanding
Fedora 4: The fifedom you've always wanted11:03
* dwilcox leaves
* Iome joins11:08
* jrgriffiniii leaves11:33
* jrgriffiniii joins
* bseeger leaves12:03
* jrgriffiniii_ joins12:21
* jrgriffiniii leaves12:22
* dwilcox joins12:24
* bseeger joins12:27
<ruebot>acoburn: i think git-flow would be good for vagrant12:37
acoburn: ...and i'll be on thursday's call!
<acoburn>ruebot: sounds good12:38
* acoburn1 joins12:45
* acoburn leaves12:46
<ruebot>whikloj: awoods is gone this week, right?12:48
<whikloj>ruebot: I believe so13:07
<ruebot>whikloj: should i merge your pr, or wait until we talk about it in a larger context on thursday?13:08
<whikloj>ruebot: Wait, it assumes a course of action that might not be what people want.13:10
* dwilcox leaves
<ruebot>whikloj: @_@13:17
* jrgriffiniii leaves13:29
* jrgriffiniii joins13:31
<osmandin>whikloj: ruebot: Thanks. The PR seems to have fixed the issue. I just tested the PR.13:34
* bseeger leaves13:35
<whikloj>osmandin: np, how do you use the vagrant setup? If we lock it to the last released version is that better or worse for you?
<osmandin>whiklog: I'm using it for the first time for a demo.13:36
<whikloj>osmandin: ok, so last released version is probably preferable for that type of use13:37
* acoburn1 leaves13:38
<osmandin>whikloj: Yes. Also, some people here use it to just try out F4 for themselves, so it's a very useful thing.
* bseeger joins13:48
* dwilcox joins13:51
<ruebot>osmandin: cool. good to hear.13:56
whikloj: i suppose that means i should merge :-)
<whikloj>ruebot: no, we should wait for Thursday13:57
<ruebot>osmandin: i saw your question about vagrant vs base box from yesterday, did that get answered?
whikloj: aye
<osmandin>ruebot: So, in the basebox F4 is absent. Right?13:59
* dwilcox leaves14:01
* acoburn joins
* mikeAtUVa joins14:02
<ruebot>osmandin: yep. it's everything you need for f4, so, then you can spin up a vagrant instance quicker.14:04
osmandin: the key is here: https://github.com/fcrepo4-exts/fcrepo4-vagrant-base-box/blob/master/Vagrantfile#L11 -- so, we use a vanilla ubuntu image there.14:05
...then we export it once it is built, and add it to atlas.hashicorp -- https://atlas.hashicorp.com/fcrepo/boxes/fcrepo4-base14:06
<osmandin>ruebot: Great, thanks.
<ruebot>...which then allows us to use it as a basebox for fcrepo4-vagrant https://github.com/fcrepo4-exts/fcrepo4-vagrant/blob/master/Vagrantfile#L11
* dwilcox joins14:11
* jrgriffiniii leaves14:12
* jrgriffiniii joins14:18
* bseeger leaves14:35
* bseeger joins14:48
* acoburn leaves14:59
* dwilcox leaves15:05
* mikeAtUVa leaves15:18
* dwilcox joins15:30
* dwilcox leaves15:36
* mikeAtUVa joins15:40
* osmandin leaves15:49
* dwilcox joins16:16
* dwilcox leaves16:36
* bseeger leaves16:37
* whikloj leaves17:00
* jrgriffiniii leaves17:11
* dwilcox joins17:33
* dwilcox leaves17:37
* peichman leaves17:52
* kefo leaves18:29
* jrgriffiniii joins18:30
* dwilcox joins18:34
* dwilcox leaves18:39
* mikeAtUVa leaves20:38
* dwilcox joins20:54
* dwilcox leaves21:46
* dwilcox joins22:14
* dwilcox_ joins22:34
* dwilcox leaves22:36
* dwilcox_ leaves23:11