Log of the #duraspace-ff channel on chat.freenode.net

Using timezone: Eastern Standard Time
* fasseg leaves03:36
<pivotal-bot____>Andrew Woods added comment: "@osmandin, can you please submit a PR?" https://www.pivotaltracker.com/story/show/6522232004:32
* benpennell2 joins06:24
* benpennell1 leaves
* dwilcox joins08:00
* ksclarke leaves08:02
* dwilcox leaves08:18
* dwilcox joins08:28
* ksclarke joins08:34
* ksclarke leaves08:35
* nbanks joins08:48
* ksclarke joins08:50
* github-ff joins08:57
[fcrepo4] ksclarke opened pull request #256: Updated throws indent formatting to match Java now that new Checkstyle supports this (master...checkstyle-formatting-66144930) http://git.io/zrjeZw
* github-ff leaves
* dwilcox leaves09:08
<pivotal-bot____>Kevin Clarke added comment: "MCHECKSTYLE-169 is fixed but not released yet, so I just updated the separate fcrepo maven-checkstyle-plugi..." https://www.pivotaltracker.com/story/show/6614493009:09
Kevin Clarke finished "Resolve Checkstyle/Eclipse formatting rules" https://www.pivotaltracker.com/story/show/66144930
* ermadmix joins09:10
* dwilcox joins
* tecoripa joins09:11
* dwilcox leaves09:18
* gregjansen joins
* ajs6f joins09:27
* ksclarke leaves09:29
* ksclarke joins
* dwilcox joins09:33
<ajs6f>benpennell1: Is PR/254 up to date and ready to go? 'Cause I'll merge it this morning, if so.09:34
<benpennell2>i've made the suggested changes so ti should be ready to go09:36
<ajs6f>k, thanks!09:37
* github-ff joins09:38
[fcrepo4] ajs6f closed pull request #254: Extend kernel API to include more interfaces (master...kernal-api-interfaces) http://git.io/k94usQ
* github-ff leaves
<pivotal-bot____>A. "Flambulacio" Soroka added comment: "Merged with https://github.com/futures/fcrepo4/commit/826efd9a54ec2e2e25ebac72e23e7f786e6f72d0" https://www.pivotaltracker.com/story/show/64468400
* dwilcox leaves
<pivotal-bot____>A. "Flambulacio" Soroka delivered "Extend kernel API to include more interfaces" https://www.pivotaltracker.com/story/show/6446840009:39
* dwilcox joins
<ajs6f>afk bbs09:43
* ajs6f leaves
* dwilcox leaves09:44
* ajs6f joins09:47
<pivotal-bot____>A. "Flambulacio" Soroka added comment: "@ksclarke : Is it worth doing this move now, or do you think we should wait for MCHECKSTYLE-169,..." https://www.pivotaltracker.com/story/show/6614493009:48
* fasseg joins09:54
* dwilcox joins
* scossu joins09:57
* dwilcox leaves09:59
<pivotal-bot____>A. "Torcrexilly" Soroka estimated "Support for authentication in JMS indexer" as 3 points https://www.pivotaltracker.com/story/show/6492915610:00
A. "Torcrexilly" Soroka added comment: "@ericjames : I don't understand this issue: can you clarify what the behavior here was and what ..." https://www.pivotaltracker.com/story/show/6286760410:01
Kevin Clarke added comment: "I wondered this myself (since a released mvn-checkstyle should have 169 in it in addition to the newer Chec..." https://www.pivotaltracker.com/story/show/6614493010:02
* scossu leaves
<pivotal-bot____>A. "Torcrexilly" Soroka added comment: "@ericjames : Can you break the other issues you mention above off into separate tickets?" https://www.pivotaltracker.com/story/show/6285448810:03
A. "Torcrexilly" Soroka added comment: "@escowles : Are you concerned with the FedoraEvent type here, or the format of the JMS message i..." https://www.pivotaltracker.com/story/show/6285448810:04
Benjamin Pennell started "Modify benchtool to support testing transactions" https://www.pivotaltracker.com/story/show/6634383210:05
* github-ff joins10:07
[fcrepo4] osmandin opened pull request #257: Cache Federated SHA-1 values (master...65222320) http://git.io/3bBmJQ
* github-ff leaves
<pivotal-bot____>Mike Daines started "Review AuthZ Use Cases (Mike)" https://www.pivotaltracker.com/story/show/66298620
Osman Din added comment: "@awoods, https://github.com/futures/fcrepo4/pull/257
Please let me know if you want me to merge into one comm..." https://www.pivotaltracker.com/story/show/65222320
A. "Torcrexilly" Soroka added comment: "Okay, let's wait until @awoods is back to determine the right rhythm for this. That's just a few..." https://www.pivotaltracker.com/story/show/6614493010:08
A. "Torcrexilly" Soroka added comment: "I will review this today and would very much appreciate a squash beforehand." https://www.pivotaltracker.com/story/show/6522232010:09
A. "Torcrexilly" Soroka added comment: "We fixed this, correct? @barmintor killed all of the long-lived Sessions many weeks ago, IIRC..." https://www.pivotaltracker.com/story/show/6148261010:12
Esme Cowles added comment: "@ajs6f The intention here was to have an event that indicated that some properties had changed on a node. I..." https://www.pivotaltracker.com/story/show/62854488
A. "Torcrexilly" Soroka added comment: "So this would lead to a one-to-one mapping between properties changed and events emitted?" https://www.pivotaltracker.com/story/show/6285448810:13
A. "Torcrexilly" Soroka added comment: "@benjaminpennell : Am I correct in thinking that you "inadvertently" did this by doing the API t..." https://www.pivotaltracker.com/story/show/6317248210:14
Benjamin Pennell added comment: "@ajs6f Yes probably, the only class in org.fcrepo.jcr (FedoraJcrTypes) got moved over into fcrepo-kerne..." https://www.pivotaltracker.com/story/show/6317248210:20
A. "Torcrexilly" Soroka estimated "Separate Java constants now found in fcrepo-jcr into a new artifact" as 1 point https://www.pivotaltracker.com/story/show/63172482
A. "Torcrexilly" Soroka delivered "Separate Java constants now found in fcrepo-jcr into a new artifact" https://www.pivotaltracker.com/story/show/6317248210:21
* ajs6f leaves10:22
<pivotal-bot____>Esme Cowles added comment: "No, I think the ideal situation would be to emit one event for each node that had a property changed. So if..." https://www.pivotaltracker.com/story/show/6285448810:23
<gregjansen>So, is anyone else using TCPPING to setup their cluster?10:25
<tecoripa>gregjansen: you mean, instead of MPING?
<pivotal-bot____>Eric James added comment: "This is just the observation that the first curl POST created "wikicheck" while second one created "indexable..." https://www.pivotaltracker.com/story/show/62867604
<gregjansen>tecoripa: yes10:26
tecoripa: multicast is not enabled on my hosts
<tecoripa>gregjansen: not yet, but I'm about to head down that path. Nothing but pain trying to get discovery to work with MPING.
works for the first three nodes, then the last three fail - can't find the coordinator or other nodes.10:27
<gregjansen>tecoripa: yeah, I never saw a cluster of more than 1 host
does mping has num_initial_members attribute?
<tecoripa>gregjansen: for a few brief and shining moments late last night, I had five of mix six hosts talking to each other10:28
but I couldn't reproduce it again.
let me check, I don't think so...
no10:29
the other think I was going to try was FD_SOCK: failure detection on TCP sockets
http://www.jgroups.org/manual-3.x/html/protlist.html#FD_SOCK10:30
* dwilcox joins
<tecoripa>s/think/thing/10:31
* ajs6f joins
* nbanks leaves10:38
* dwilcox leaves10:39
* nbanks joins10:40
<tecoripa>ajs6f: I won't be at the standup today10:43
ajs6f: but I just sent a somewhat detailed status report to ff-internal
<ajs6f>tecoripa: Okay, thanks!
* ksclarke leaves
<tecoripa>ajs6f: it seems like gregjansen is in somewhat the same boat I'm in (for different reasons)10:44
<ajs6f>tecoripa: Meaning unable to get to standup?
* ksclarke joins
<tecoripa>ajs6f: so I think we're both going to try all TCP, all the time (for both discovery and state transfer). It's slower, but hopefully more robust.
ajs6f: no, meaning not able to use UDP for discovery.
<ajs6f>tecoripa: Oh, that's annoying. That means people hav to maintain lists of IPs.10:45
Oh, well.
<tecoripa>ajs6f: not necessarily. They do have some good macros: SITE_LOCAL, for example, which means bind to your private 192.168.0.0 address10:46
<gregjansen>ajs6f, tecoripa: think I found success
<tecoripa>gregjansen: do tell10:47
<ajs6f>gregjansen: Yay!
<gregjansen>the jgroups file I was using was pretty different from the stock one in fcrepo, so I went back to stock.. I think Frank had made a number of big changes to it. Anyway, after plugging in my host details it seems to be working better.. cofirming10:48
confirming it
* dwilcox joins10:49
<tecoripa>gregjansen: if you confirm it, can you template it and check it into the puppet-fcrepo git project, under templates/?
then I'll grab it and try it out, too
<gregjansen>yep, cluster of 7 is working
yep
<tecoripa>yay!
+5 for gregajansen10:50
* dwilcox leaves
<gregjansen>tecoripa: don't get too excited, it has a list of my local hostnames in initial hosts under TCPPING, however you know how to parameterize that10:51
* dwilcox joins
<gregjansen>better than I
<ajs6f>gregjansen: Wait, so you are or are not using UDP multicast?10:52
<gregjansen>I am using TCP for both
<tecoripa>gregjansen: I may try the SITE_LOCAL macro for the bind_addr in TCPPING.10:53
<gregjansen>tecoripa: https://github.com/gregjan/puppet-fcrepo/compare/package-tomcat?expand=1
tecoripa: you mean initial hosts? bind_addr is in TCP element.. I used the puppet function called <%= ipaddress_eth0 %> for that10:54
<tecoripa>are your addresses on eth0 public? looks like they might be10:55
* dwilcox leaves
<tecoripa>ah, I see. initial_hosts doesn't support the macros.10:56
sigh, okay
but yes, puppet will fill those in.
<gregjansen>tecoripa: I mean, if you bind to localhost, then you aren't going to talk to any other nodes, am I right?
<tecoripa>can you push your branch to my repo? you should have permissions now.
<gregjansen>tecoripa: or at least they cannot talk to you..
tecoripa: i will try it10:57
<tecoripa>right. SITE_LOCAL is a macros for private networks, not localhost.
<gregjansen>tecoripa: yeah it worked (push)
<tecoripa>but don't bother: TCPPING initial_hosts doesn't seem to support the macro
<gregjansen>oh
<tecoripa>gregjansen: thanks
<gregjansen>stand up..10:58
<tecoripa>I see that you have FD_SOCK in your config, too.
okay, I gotta bail to another meeting. thanks, gregjansen.
<gregjansen>Confluence seems to be down..
anybody have the hangout url?10:59
* dwilcox joins
<ajs6f>Yup.11:00
Hang on.
<tecoripa>https://plus.google.com/hangouts/_/event/c1glu6soq43r1rr6ou17qtobug8
<gregjansen>thx!
<ajs6f>tecoripa++
* tecoripa leaves11:01
<pivotal-bot____>Eric James added comment: "Seems there's just 1 issue here, that for property changes, 1 event is being emitted as path/property no matt..." https://www.pivotaltracker.com/story/show/6285448811:04
* dwilcox leaves11:10
* edInCo joins11:13
<pivotal-bot____>A. "Torcrexilly" Soroka added comment: "I don't know why they are behaving differently-- can you check the logic and see if there's a pr..." https://www.pivotaltracker.com/story/show/6286760411:15
<gregjansen>tecoripa: do you have the workload handy for my benchtool test (wiki down)11:16
<pivotal-bot____>A. "Torcrexilly" Soroka added comment: "The additional hierarchy is available for when users don't particularly care about the identifie..." https://www.pivotaltracker.com/story/show/62867604
* dwilcox joins11:19
<gregjansen>fasseg: what does your benchtool run look like? (no auth to worry about on my cluster)11:21
<ajs6f>all: Wiki should be coing back up.11:22
<pivotal-bot____>A. "Torcrexilly" Soroka added comment: "I don't see this as one issue: there's the issue of having the property identifier separated fro..." https://www.pivotaltracker.com/story/show/6285448811:23
* nbanks leaves
* ksclarke leaves11:27
* ksclarke joins11:31
* dwilcox leaves11:37
* dwilcox joins11:38
* dwilcox leaves11:42
* dwilcox_ joins
* ksclarke leaves11:43
* ksclarke joins11:44
<gregjansen>hmm, I have a question about benchtool and the throughput it reports.. It is reporting bytes over durations logged on threads. I would not expect that number to improve when you add threads. Shouldn't we report to cluster throughput as bytes over clock time, minus data prep time?11:46
<pivotal-bot____>Esme Cowles added comment: "I've added a section to the wiki outlining the single-event for a batch of property updates: ""11:47
https://wiki.d..." https://www.pivotaltracker.com/story/show/62854488
<ajs6f>Throughput is definitely quantity over clock time, to my mind, but the problem with that measurement is that it has just has much to do with your network closet as your code, so we have to be very careful when we make comparisons using it.
<pivotal-bot____>A. "Torcrexilly" Soroka added comment: "That looks nice, @escowles . I'll bring that up on Thursday at the committers' call (or you can,..." https://www.pivotaltracker.com/story/show/6285448811:48
* ksclarke leaves11:53
* ksclarke joins11:54
* github-ff joins
[fcrepo4] osmandin opened pull request #258: Use the cached "CONTENT_DIGEST" property (if it exists) instead of recal... (master...fcrepo-connector-file/sha1) http://git.io/L_jvBA
* github-ff leaves
* ksclarke leaves
<pivotal-bot____>Osman Din added comment: "@ajs6f https://github.com/futures/fcrepo4/pull/258" https://www.pivotaltracker.com/story/show/65222320
A. "Torcrexilly" Soroka added comment: "Thanks @osmandin !" https://www.pivotaltracker.com/story/show/6522232011:55
* ajs6f leaves11:58
* escowles joins12:05
<gregjansen>ajs6f: yes, but when I run 10 threads to really use the cluster, it will make a total throughput calc based on ten, often simultaneous, durations.12:11
ajs6f: oh wait, it is multiplied again by the number of threads.. totally legit12:12
* dwilcox_ leaves12:35
* ksclarke joins12:36
* escowles_ joins13:04
* escowles leaves13:05
* gregjansen leaves13:07
* dwilcox joins13:15
* gregjansen joins13:23
* dwilcox leaves13:44
* dwilcox_ joins13:45
<pivotal-bot____>Gregory Jansen added comment: "Some related edits went on this page: https://wiki.duraspace.org/display/FF/ModeShape+Clustering My pupp..." https://www.pivotaltracker.com/story/show/6609683214:00
Gregory Jansen finished "Install puppet and load balancer on UNC cluster." https://www.pivotaltracker.com/story/show/66096832
* scossu joins14:01
* ajs6f joins14:02
* github-ff joins14:11
[fcrepo4] ajs6f pushed 2 new commits to master: http://git.io/2Jzqlw
fcrepo4/master b08f4a9 osmandin: Use the cached "CONTENT_DIGEST" property (if it exists) instead of recalculating the value.
fcrepo4/master 7e768dd ajs6f: Merge pull request #258 from osmandin/fcrepo-connector-file/sha1...
* github-ff leaves
<pivotal-bot____>A. "Torcrexilly" Soroka added comment: "Merged with https://github.com/futures/fcrepo4/commit/7e768ddb3f29b89ddb6daa6a629ee21075b0517d" https://www.pivotaltracker.com/story/show/6522232014:13
A. "Torcrexilly" Soroka delivered "Cache Federated SHA-1 values" https://www.pivotaltracker.com/story/show/65222320
<ajs6f>fasseg: Did we ever try this:14:14
https://wiki.duraspace.org/display/FF/Using+Teamspeak+3
?
<pivotal-bot____>A. "Torcrexilly" Soroka added "Determine whether org.fcrepo.kernel.services.TransactionService can be replaced by javax.transaction.Tran..." https://www.pivotaltracker.com/story/show/6644886614:20
* ajs6f1 joins14:22
* ajs6f leaves
* ajs6f1 leaves14:30
* ajs6f1 joins
* ajs6f1 leaves14:33
* ajs6f1 joins
* ajs6f1 leaves14:46
* ajs6f joins
<scossu>Hi, I have been dealing for several days with an issue in my primary-type branch. Tests keep failing when a SPARQL query is made on a custom property. See https://groups.google.com/d/msg/fedora-tech/wqdXp8XPxcY/NoGXqdraJw4J
Does someone have any hint about where to look?14:47
<ajs6f>scossu: Is this a test you introduced?14:48
<scossu>ajs6f: No, I haven't changed the test, just introduced a restrictive fedora:resource
<ajs6f>So that test (which introduces a property you are disallowing, _should_ fail, no?14:49
<scossu>No, the property is allowed because I add a "anyproperty" mixin before hand, which allows any property.
<ajs6f>Okay, but should the SPARQL query retrieve that property?14:50
<scossu>ajs6f: I'm just not sure if it's indexed.
ajs6f: Does the internal indexer look for a specific mixin or node type?
<ajs6f>Does your "anyproperty" mixin specifically indicate index status?14:51
scossu: the internal indexer belongs to MODE, not Fedora. It looks in the mixin specification to determine indexability, as per JCR.
Let me see if I can find that.
<scossu>ajs6f: Do you mean indexing:indexable?14:52
Or is it for the JMS indexer only?
<ajs6f>scossu: No, that is itself a mixin. We're talking about the mixin definitions, such as the one you made.
The indexing: namespace is Fedoras. this is all happening outside of Fedora, in MODE.
See:14:53
https://docs.jboss.org/author/display/MODE/Defining+custom+node+types#Definingcustomnodetypes-Declaringnodetypes
<scossu>ajs6f: The only thing that changed in the node itself is that it is a fedora:baseObject intead of an nt:folder.
<ajs6f>and notice the "query" aspect for the definition.14:54
I _think_ that's what's governing this functionality, but I could be wrong
Can you tell me what test is failing?
<pivotal-bot____>Eric James added "separate the property identifier from the node identifier in event creation and messaging for property modified events" https://www.pivotaltracker.com/story/show/6645234414:55
<scossu>ajs6f: Ok, I'll try adding the 'query' option.
ajs6f: https://groups.google.com/group/fedora-tech/attach/e27dadaa9978136/TEST-org.fcrepo.integration.transform.http.FedoraSparqlIT.xml?part=4&authuser=0
<ajs6f>Okay. That may help, but if not, we'll dig a little deeper.
<scossu>Thanks. This thing's been haunting me for too long.14:56
<ajs6f>benpennell2: I just noticed something— when you moved some classes into the API module (e.g. NodePropertiesTools) you seem to have left the tests in the kernel module. Did you remember to move the tests with the classes?15:01
afk bbs
<pivotal-bot____>Eric James added "parse property changed events in the indexer-pluggable so as to update indexers (solr, triplestore,??) with the approp..." https://www.pivotaltracker.com/story/show/6645288215:02
Eric James edited "separate the property identifier from the node identifier in event creation and messaging for property modified events" https://www.pivotaltracker.com/story/show/66452344
<scossu>ajs6f: no luck.15:03
I kind of expected it, because 'query' should be on by default.
<pivotal-bot____>Mike Daines added comment: "For example, IP address is mentioned as an access condition in this use case: https://wiki.duraspace.org/dis..." https://www.pivotaltracker.com/story/show/66375794
<scossu>I'll try to add 'queryops' to the * properties, just to doubke checks.15:04
check
<benpennell2>ajs6f: ah no, i did not move the tests over. I can make another commit for that. Also, there are tests for classes that are now named Impl, should I change the names of the tests to match?15:06
* dwilcox_ leaves15:07
<pivotal-bot____>Mike Daines edited "Use Case: Streamlined and secure way of distinguishing open from closed " https://www.pivotaltracker.com/story/show/6629891215:08
<gregjansen>has anyone looked into clean shutdown of modeshape before? I notice that we have no calls to the JcrRepositoryFactory.shutdown methods. Was thinking that a @PreDestroy call to shutdown might help our clusters run more smoothly.
<ajs6f>benpennell2: Don't add a commit, because I already merged the PR. Open a new little bitty ticket for this (resulting in a new PR). And yes, I would change the names of test to match the exact type they are testing.15:09
gregjansen: I don't think anyone has done that, and what you suggest makes a lot of sense. You want to tackle that (especially since the cluster not shutting down rightly thing is going to annoy you a lot).15:10
?
* dwilcox joins15:11
<gregjansen>ajs6f: thought you would say that. work is in progress but I will make a ticket for this. Yeah, I am seeing slews of transactions fail when a node leaves the cluster.15:12
* dwilcox leaves15:13
* tecoripa joins15:15
* escowles leaves
* dwilcox joins
<pivotal-bot____>Gregory Jansen added "Ensure a clean shutdown of the repository on container shutdown." https://www.pivotaltracker.com/story/show/66454096
<ajs6f>gregjansen: Making a ticket is enough. Working it is golden, and if you do, awoods will buy you a beer in Raleigh.15:16
<pivotal-bot____>Gregory Jansen started "Ensure a clean shutdown of the repository on container shutdown." https://www.pivotaltracker.com/story/show/66454096
<tecoripa>gregjansen: coming in at the end of the conversation here...
is it related to the ticket I opened earlier on unclean tomcat shutdown?15:17
https://www.pivotaltracker.com/story/show/65988656
<pivotal-bot____>bug: Tomcat process persists after shutdown in cluster environment (unscheduled) / owner: Scott Prater
<gregjansen>tecoripa: noticed lots of stack traces during node shutdown.. googled them to find they may be dangling due to unclean modeshape shutdowns.. c.f. https://docs.jboss.org/author/display/MODE/Getting+Started#GettingStarted-Stoppingtherepositoryandengine15:18
tecoripa: let me read your ticket ..
<tecoripa>gregjansen: ok. more than likely, they're a lot of the same ones I saw with the tomcat shutdown
I pasted a gist -- link to it in my ticket15:19
* fasseg leaves
<gregjansen>tecoripa: right, I was seeing more than that, but also that. I tried to shutdown a node while test cleanup was running. Saw a bunch of traces from the infinispan transactions, which had been aborted.15:21
I will try and dig up my trace
<tecoripa>gregjansen: ok. I didn't have stack traces, but lots of spinning threads left running, as you see. This was just a plain shutdown after a startup, no transactions work done in between.15:22
<pivotal-bot____>Benjamin Pennell added "Move tests of classes which were moved kernel-api" https://www.pivotaltracker.com/story/show/66454712
<gregjansen>tecoripa: that makes sense. I figure eventually those threads get the axe after which any pending transactions (coming from other hosts) would throw exceptions related to the suspect node.15:23
* dwilcox leaves15:24
<tecoripa>gregjansen: right, I imagine that's what would happen. though the odd thing is the webapp that owns those threads is destroyed -- are those dangling threads basically zombie processes, orphans with no home?
in which case, I suppose they just die a natural thread death, but nodata goes or in or out?15:25
<ajs6f>tecoripa: When you say "the webapp", what do you mean? Do you mean that it no longer responds at endpoints/
?
<tecoripa>I mean that the fcrepo webapp context is destroyed in Tomcat:
<ajs6f>That doesn't mean that no threads from that app are running.15:26
That just means Tomcat no longer maintains an app context for that app.
<tecoripa>right, that's my point. the webapp is gone, but threads are still running.
<gregjansen>tecoripa: after adding calls to shutdown modeshape repo, the thread warnings are gone
<ajs6f>Right, I'm just saying that the Tomcat manager is not a good way to determine shutdown-ness/
<tecoripa>ajs6f: ah, okay15:27
* dwilcox joins
<ajs6f>Shutdownity. Shutdownaiae.
<tecoripa>ajs6f: although we would like our webapp who asked been asked to remove itself to clean up after itself, correct?
gregjansen: nice. is that going into a PR?15:28
<gregjansen>yes15:29
<ajs6f>tecoripa: Oh, yes. Our uses shouldn't care about the subtlties.
users.
<tecoripa>gregjansen. I'll test and mark that ticket as finished, and pat myself on the back for a job well done. ;)
<ajs6f>Be it a web container, an app server, or some weird OSGi thing, whatever it is into which Fedora is deployed shold be able to rely on it shutting down in a reasonable way.15:30
<tecoripa>ajs6f: which I, I mean gregjansen, has apparently just done.
<ajs6f>tecoripa: For that remark, you just earnt the right to review gregjansen's PR.
* dwilcox leaves
<gregjansen>all 4 lines15:31
<ajs6f>tecoripa: Each and every one.
<pivotal-bot____>A. "Torcrexilly" Soroka edited "Ensure a clean shutdown of the repository on container shutdown." https://www.pivotaltracker.com/story/show/66454096
<gregjansen>I am still double checking it on the cluster, have it in today
<tecoripa>gregjansen: what's the word count? how many characters?
<pivotal-bot____>Mike Daines added "HTTPHeaderPrincipalFactory#getGroupPrincipals always returns the empty set" https://www.pivotaltracker.com/story/show/6645551015:32
<ajs6f>Fedora 4: Quality, not quantity.
<gregjansen>1 word
<tecoripa>gregjansen: I just saw it. PreDestoy.
I Approve Your Pull Request.15:33
<gregjansen>tecoripa: wait, where did you see that?
<tecoripa>in your comment on your ticket
<ajs6f>He hasn't sent it yet.
<gregjansen>oh, you are jk
<scossu>ajs6f: I tested a few combinations with the CND, and it looks like the error only goes away if I add the wildcard child definitions directly in fedora:resource.15:35
- * (undefined)15:36
- * (undefined) multiple
So I guess it should be a Fedora function that is in charge of indexing the new nodes…15:37
<ajs6f>fedora:resource the mixin?
No, it's not a Fedora function.
It's definitely MODE. We would have noticed if we wrote that code. :)
<tecoripa>afk bbs15:38
* tecoripa leaves
<ajs6f>Hm. We need to understand how MODE selects what it will add to the Lucene indexes it maintains (that's what supports, ultimately, the SPARQL query endpoint).15:39
* dwilcox joins15:43
<scossu>Yes, the mixin. I'm pretty positive that the test is assigning a fedora:baseObject as a primary type, and a fedora:anyproperty mixin (because I can set zz:name), along with fedora:resource of course
<ajs6f>scossu: I believe you. :) Is this code in your fork? Can you give me a branch ref? I'd like to take a close look at what you are doing as I read up on MODE's query apparatus.15:44
* github-ff joins15:45
[fcrepo4] ajs6f pushed 1 new commit to master: http://git.io/_wkb_A
fcrepo4/master 440a2bc ajs6f: Nonfunctional code cleanup
* github-ff leaves
<scossu>ajs6f: I have to clean up the branch before I commit to my remote. Here is a pate-up of my test: http://pastebin.com/3iYcZB38
<ajs6f>Thanks.15:46
<scossu>See lines 73-74. Test fails on itShouldIndexCustomProperties
My CND file: http://pastebin.com/uCpTJH4Q15:47
See lines 167 and 19015:48
<ajs6f>And amongst your combinations you tried:15:49
[fedora:anyproperty] mixin query 
- * (undefined)
- * (undefined) multiple
?
<scossu>Yes, the paste-up is reverted to before my query option tests.15:52
<ajs6f>Ok.15:53
<pivotal-bot____>Eric James added comment: "Related lickets:
https://www.pivotaltracker.com/s/projects/684825/stories/66452882
https://www.pivotaltracke..." https://www.pivotaltracker.com/story/show/62854488
Gregory Jansen added "Remove duplicate logback.xml shipped in http-api jar" https://www.pivotaltracker.com/story/show/6645747415:58
<scossu>Eric James just coined a wonderful term: lickets.
ajs6f: https://github.com/scossu/fcrepo4/tree/primary-type16:04
* dwilcox leaves
<ajs6f>Thank, scossu.
* dwilcox joins16:05
* dwilcox leaves
* dwilcox joins16:07
* dwilcox_ joins16:11
* dwilcox leaves16:12
* tecoripa joins16:15
* gregjansen leaves16:17
<pivotal-bot____>Eric James added comment: "Could not recreate with latest build. Maybe it was fixed?
curl -X POST "http://localhost:8080/rest/test2/pa..." https://www.pivotaltracker.com/story/show/62867604
<ajs6f>scossu: I'm not going to get to look at this today, but I will look at it first thing tomorrow.16:18
* dwilcox_ leaves
<scossu>ajs6f: No problem, thanks a lot.
* dwilcox joins
* dwilcox leaves
* github-ff joins16:19
[fcrepo4] ajs6f pushed 1 new commit to master: http://git.io/jr74MQ
fcrepo4/master 306b021 ajs6f: Nonfunctional code cleanup
* github-ff leaves
<pivotal-bot____>A. "Torcrexilly" Soroka added comment: "If we can't replicate it, we can't fix it. {grin}16:20
Sure, close the ticket but keep your eyes pee..." https://www.pivotaltracker.com/story/show/62867604
* dwilcox joins
* gregjansen joins
* dwilcox leaves16:37
* dwilcox joins16:39
<pivotal-bot____>Mike Daines added "Improve documentation of PEPs" https://www.pivotaltracker.com/story/show/6646096216:40
<ajs6f>Out for the day— have a good evening everyone...16:52
* ajs6f leaves
* gregjansen leaves17:03
* dwilcox leaves17:05
* dwilcox joins17:08
* dwilcox leaves17:12
* dwilcox joins17:15
* edInCo leaves17:19
* dwilcox leaves17:24
<ksclarke>just curious has the JMS interface been more popular than the sequencer one with fcrepo4 (among the folks developing connectors)?18:01
* ksclarke realizes he should ask that earlier on in the day
* tecoripa leaves18:02
* scossu leaves18:07
* scossu joins18:10
* ksclarke leaves18:16
* ermadmix leaves
* scossu leaves19:14
* ksclarke joins21:22