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

Using timezone: Eastern Standard Time
* fcrepo-bot leaves01:51
* longshou leaves03:20
* dwilcox joins06:33
* dwilcox leaves07:30
* dwilcox joins09:00
* github-ff joins09:14
[fcrepo4] awoods pushed 1 new commit to master: http://git.io/AwNXkQ
fcrepo4/master c822f0e Mike Durbin: Suspend support for update operations using the FedoraFileSystemConnector....
* github-ff leaves
* ksclarke joins
* github-ff joins09:15
[fcrepo4] awoods closed pull request #399: Suspended support for update operations using the FedoraFileSystemConnec... (master...fs-federation-disallow-writes) http://git.io/Y6nkrQ
* github-ff leaves
<pivotal-bot>Andrew Woods added comment: "Resolved with: https://github.com/fcrepo4/fcrepo4/commit/c822f0e9c5db22fef1eb90e8ceca8ec130890c61" https://www.pivotaltracker.com/story/show/71854364
Andrew Woods added "Document Read-only FileSystemConnector" https://www.pivotaltracker.com/story/show/7225702209:16
Andrew Woods edited "Document Read-only FileSystemConnector" https://www.pivotaltracker.com/story/show/72257022
Andrew Woods delivered "Disable read/write filesystem federation." https://www.pivotaltracker.com/story/show/71854364
* osmandin joins09:22
<pivotal-bot>Gregory Jansen added comment: "Yes, that should work too. However, anything configured on the eclipse project manually will need to be r..." https://www.pivotaltracker.com/story/show/7126229809:24
Gregory Jansen added comment: "@ksclarke see comment above this." https://www.pivotaltracker.com/story/show/71262298
* gregjansen joins09:26
* osmandin leaves09:27
* travis-ci joins09:28
[travis-ci] fcrepo4/fcrepo4#2093 (master - c822f0e : Mike Durbin): The build passed.
[travis-ci] Change view : https://github.com/fcrepo4/fcrepo4/compare/485183693832...c822f0e9c5db
[travis-ci] Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/26298196
* travis-ci leaves
<bljenkins>Project fcrepo4 build #1926: UNSTABLE in 30 min: http://ci.fcrepo.org/jenkins/job/fcrepo4/1926/09:45
awoods: Suspend support for update operations using the FedoraFileSystemConnector.
<pivotal-bot>Andrew Woods accepted "Disable read/write filesystem federation." https://www.pivotaltracker.com/story/show/7185436409:47
Kevin Clarke added comment: "Hi @gregoryjansen I use "Maven > Update Project" quite a bit too but I've never seen it lose the Checkstyle..." https://www.pivotaltracker.com/story/show/71262298
Kevin Clarke added comment: "or, perhaps you were saying we could have the maven dep module and pom configs check out the checkstyle con..." https://www.pivotaltracker.com/story/show/7126229809:51
* tecoripa1 joins10:09
* tecoripa1 leaves10:25
<pivotal-bot>Kevin Clarke edited "Add testing deps to fcrepo4-bom" https://www.pivotaltracker.com/story/show/7185184810:34
Kevin Clarke added comment: "@gregoryjansen But, aside from making the Checkstyle configs more readily available to the eclipse-cs plugi..." https://www.pivotaltracker.com/story/show/7126229810:35
* github-ff joins10:36
[fcrepo4] escowles created sonar-coverage (+2 new commits): http://git.io/nTMwFg
fcrepo4/sonar-coverage 2e046e8 Esmé Cowles: Moving duplicated code to AbstractResource, removing unreachable error handling, adding IT for move/copy to non-existent parent
fcrepo4/sonar-coverage 1f4d140 Esmé Cowles: Removing unused predicate, adding tests
* github-ff leaves
<pivotal-bot>Chris Beer added "fcrepo-jms should send post-translation identifiers, not the raw JCR paths" https://www.pivotaltracker.com/story/show/7226279010:41
* scossu joins10:42
* longshou joins10:51
* travis-ci joins10:54
[travis-ci] fcrepo4/fcrepo4#2094 (sonar-coverage - 1f4d140 : Esmé Cowles): The build passed.
[travis-ci] Change view : https://github.com/fcrepo4/fcrepo4/compare/2e046e875f05^...1f4d1400ac36
[travis-ci] Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/26304592
* travis-ci leaves
* osmandin joins10:59
* ermadmix joins11:05
<pivotal-bot>Chris Beer edited "fcrepo-jms should send post-translation identifiers, not the raw JCR paths" https://www.pivotaltracker.com/story/show/7226279011:10
Chris Beer added "fcrepo-jms should include some information about the session that made the change" https://www.pivotaltracker.com/story/show/7226578211:13
* osmandin leaves11:19
<gregjansen>so what is happening with the committer call?11:24
martin and I are in a hangout alone
<escowles>gregjansen: it was on the google hangout -- did you call readytalk and have to suffer through the hold music?
<gregjansen>nope, I was just late. is it over then?
<escowles>yep, short and sweet: we talked briefly about using jmeter or grinder for load testing because benchtool's producing unbelievable results and we shouldn't be maintaining our own testing framework anyway11:25
<gregjansen>cool, yes I am exploring both for another project11:26
<terrellt>Oh! Speaking of jmeter
<gregjansen>escowles: neat. they are similar. It seems like JMeter is more pluggable in terms of handling arbitrary results11:27
<escowles>gregjansen: good -- i'm esp. interested in your take on grinder, because tecoripia's the only one who's used it
<terrellt>https://github.com/flood-io/ruby-jmeter Lovely ruby DSL for jmeter.
<gregjansen>escowles: I know as much as you about grinder (or less). I just remember Scott's show and tell
<escowles>the one feature i know we wanted that jmeter doesn't support is generating large random data without having to write it to disk (very useful for testing TB files without needing a lot of client disk space)11:28
terrellt: ruby-jmeter looks awesome -- i'll keep that in mind for our perf. testing11:29
<pivotal-bot>Gregory Jansen added comment: "@ksclarke yes, aside from checkstyle config in eclipse, I am fine. Sounds like that belongs in another ti..." https://www.pivotaltracker.com/story/show/7126229811:31
Gregory Jansen added comment: "@ksclarke for the fcrepo-configs approach, you would only need to use the fcrepo as your maven parent pro..." https://www.pivotaltracker.com/story/show/7126229811:32
Kevin Clarke added comment: "@gregoryjansen Maybe I don't understand then re: fcrepo-configs... doesn't Eclipse's eclipse-cs need to be ..." https://www.pivotaltracker.com/story/show/7126229811:35
<terrellt>Is there documentation on workspaces?11:43
Or more to the point, I need a stable way to generate items with a known path and them wipe them all out for tests. I'd really love to be able to generate things in localhost:8983/fedora/rest/test/ for instance.11:46
and then*
<escowles>terrellt: you can just do PUT /fedora/rest/test to create it, and then POST /fedora/rest/test to generate auto-named children under it, and DELETE /fedora/rest/test to wipe it at the end12:00
you could use a workspace or a transaction to do that, but i'm not sure it would get you much12:01
<pivotal-bot>Longshou Situ added "Create a new child datastream got javax.jcr.nodetype.ConstraintViolationException" https://www.pivotaltracker.com/story/show/7227094612:08
* gregjansen leaves12:10
* dwilcox leaves12:14
<pivotal-bot>Longshou Situ added comment: "Okay, the issue is expressed in https://www.pivotaltracker.com/s/projects/684825/stories/72270946.12:15
I've up..." https://www.pivotaltracker.com/story/show/72122678
Longshou Situ edited "Create a new child datastream got javax.jcr.nodetype.ConstraintViolationException" https://www.pivotaltracker.com/story/show/7227094612:20
* nikhiltri joins12:40
* longshou leaves12:49
* dwilcox joins12:52
<terrellt>escowles++ # There, finally have some consistent testing results in AF813:03
<escowles>terrellt: that's great news13:04
* longshou joins13:29
* ermadmix leaves13:35
* dwilcox leaves13:49
<cbeer>escowles: ping?14:13
<escowles>cbeer: pong
<cbeer>escowles: do you remember any of the discussion about fcrepo-jms?
it seems pretty half-baked right now, and i'm trying to figure out how much are features vs bugs vs ??14:14
e.g. https://www.pivotaltracker.com/story/show/72262790
<escowles>i remember the observer/filtering/bundling stuff -- is that what you mean
<pivotal-bot>feature: fcrepo-jms should send post-translation identifiers, not the raw JCR paths (unscheduled) / owner:
<escowles>cbeer: i saw that ticket but haven't looked at what's actually being sent -- is this related to the transparent hierarchy only or to the current non-transparent stuff too?14:15
<cbeer>escowles: i don't know anything about transparent hierarchy. i was just surprised to get raw JCR paths instead of a URI14:16
another thing i didn't add to the ticket is property names get serialized to the JCR property name (e.g. jcr:lastModified) instead of the fcrepo4 way (RDF URI)14:17
<escowles>oh, we should be sending URIs not paths -- that's probably why we needed to setup one indexer for each repo when we did the multiple repo/single triplestore test a while back
<cbeer>i'd also take it in 3 fields... url to the repo, path component, and the changed property uri14:18
<escowles>yes, just adding a baseURI would fix the paths issue -- are the properties also in the same field?14:19
<cbeer>escowles: yep, right now it's just a path e.g. /test/a/b/c/jcr:lastModified, an event type, and a timestamp.14:20
<escowles>that is pretty half-baked
ideally, the property would be the full public URI14:21
<cbeer>escowles: ok, glad we're on the same page14:23
<escowles>cbeer: are you going to fix that? i could probably pick that up tomorrow morning if not14:24
<cbeer>escowles: i might take it. if i do, i'll start that ticket at least.
* nikhiltri leaves
<cbeer>escowles: it's blocking my spotlight solr indexing work right now, but i'll see if there are other spotlight things i should tackle first
<escowles>cbeer: ok, sounds good -- either way this is a good thing to clean up before the beta14:25
<pivotal-bot>Chris Beer added comment: "What HTTP code does that map to currently? " https://www.pivotaltracker.com/story/show/72270946
* scossu leaves14:26
<cbeer>yup. maybe some day i'll get to work on spotlight on top of fcrepo4 again.
<pivotal-bot>Longshou Situ added comment: "@cbeer curl -i -X POST "http://localhost:8080/rest?mixin=fedora:datastream"14:41
HTTP/1.1 405 Method Not Allowe..." https://www.pivotaltracker.com/story/show/72270946
* scossu joins14:42
<pivotal-bot>Chris Beer estimated "fcrepo-jms should send post-translation identifiers, not the raw JCR paths" as 5 points https://www.pivotaltracker.com/story/show/7226279014:43
Chris Beer started "fcrepo-jms should send post-translation identifiers, not the raw JCR paths" https://www.pivotaltracker.com/story/show/72262790
* dwilcox joins14:45
* ermadmix joins14:55
* dwilcox leaves
* dwilcox joins14:56
* nikhiltri joins14:59
<cbeer>escowles: what level of granularity do you think the FedoraEvents should be? (FedoraEvent is the intermediary between the JCR event and the JMSEvent)15:00
escowles: it looks like right now it's half JcrEvent (every property change gets a separate event) and half at the node level (so multiple property changes on the same node get rolled up to a single node event)15:01
<escowles>cbeer: i don't think every property update should get its own event, because that would be a ton of events
<cbeer>escowles: ok, so at the node level?
<escowles>cbeer: yes, that would be good15:02
<cbeer>escowles: what about changes to jcr:data? do those roll up to the fcrepo4 node, or should we distinguish between property changes and content changes?
<escowles>cbeer: seems like it could be nice to have those be separate15:03
* dwilcox leaves
<cbeer>escowles: works for me. the other thing i was thinking was just keeping track of the properties (or jcr:data) that change and including that in the FedoraEvent15:04
<escowles>so it would be an array of properties that were changes? seems like that could be useful -- i think ermadmix had some interest in more granularity like that15:05
<cbeer>escowles: yep. the JCR event has that already, but we're throwing it away
<escowles>cbeer: oh, i wish i'd noticed that when i was updating the filtering logic...
<cbeer>escowles: well, i guess it's not that we're throwing it away.. we're just randomly picking one event to use and throwing the rest away15:06
<escowles>cbeer: yes, i remember doing that to allow at least one property update to go through, but not generate dozens of events for each update15:07
longshou: are we doing sprint wrapup? https://plus.google.com/hangouts/_/event/c1glu6soq43r1rr6ou17qtobug815:13
<longshou>Sure. @4:00pm EST time?15:14
<escowles>longshou: it's supposed to be at 3:00pm EST15:15
so, whenever you're ready
<longshou>Sorry I thought it’s 4:00pm. I am joing it now.15:16
<escowles>longshou: no problem, should be pretty quick
<ermadmix>escowles: cbeer: see "A New Eventing System" in https://wiki.duraspace.org/display/FF/Design+-+PREMIS+Event+Service
I was considering pulling the whole eventing system from the jcr level to the fedora-service level.15:17
<cbeer>ermadmix: nice, thanks.15:18
* dwilcox joins15:19
<longshou>Esme: I am in https://plus.google.com/hangouts/_/event/c1glu6soq43r1rr6ou17qtobug8 now15:22
* dwilcox leaves15:29
<cbeer>escowles: huh. for some reason, i thought the identifier translator was spring-injected15:32
<escowles>cbeer: i'm on a call
<pivotal-bot>Esme Cowles started "Unit and Integration tests to 80%" https://www.pivotaltracker.com/story/show/7067884215:34
Esme Cowles edited "Enable AuthZ on Filesystem Federation" https://www.pivotaltracker.com/story/show/6801672815:39
Esme Cowles edited "Sequencer: MODS.xml extraction" https://www.pivotaltracker.com/story/show/71400142
Esme Cowles edited "Concurrent PUT requests that create hierarchy can result in unexpected results" https://www.pivotaltracker.com/story/show/7163222415:40
Esme Cowles edited "Concurrent REST requests that mint new mixins (e.g. PATCH with rdf:types) can cause RepositoryExceptions" https://www.pivotaltracker.com/story/show/71598760
Esme Cowles edited "Document ContainerRolesPrincipalProvider and HeaderPrincipalProvider" https://www.pivotaltracker.com/story/show/70710348
Esme Cowles edited "Add testing deps to fcrepo4-bom" https://www.pivotaltracker.com/story/show/7185184815:41
Longshou Situ started "Review/Update Configuration Options" https://www.pivotaltracker.com/story/show/72122826
Esme Cowles edited "Determine potential of horizontal scaling" https://www.pivotaltracker.com/story/show/69178852
<escowles>longshou: fcrepo-kernel-impl/target/site/jacoco-it/org.fcrepo.kernel.impl.services/index.html15:43
<pivotal-bot>Chris Beer added "SimpleObserver only works for the default workspace" https://www.pivotaltracker.com/story/show/7229109215:44
<escowles>longshou: sonar dashboard: http://sonar.fcrepo.org/dashboard/index/1
<longshou>Thanks, Esme.15:46
<cbeer>it's kinda weird, but i wonder if fcrepo-jms needs a dependency on -http-commons15:48
i can't think of why you'd listen to JMS if you weren't doing something with the HTTP API too15:49
<escowles>cbeer: yep, it does sound a little strange, but i agree that JMS events do strongly imply you're going ot use the HTTP API (since the main purpose is offloading stuff to an external worker that can only use the HTTP API)15:50
<cbeer>escowles: UriInfo is only available within JAX-RS resources, though, isn't it?
* dwilcox joins15:52
<escowles>cbeer: yes, definitely an injected request-scoped thing
<cbeer>hm.. and the jms producer can't be request-scoped.
<escowles>cbeer: i think we will need to uncouple HttpIdentifierTranslator from the request scope anyway -- eventually that should all be hidden inside FedoraResource anyway15:55
<pivotal-bot>Chris Beer unstarted "fcrepo-jms should send post-translation identifiers, not the raw JCR paths" https://www.pivotaltracker.com/story/show/7226279016:01
* github-ff joins16:13
[fcrepo4] escowles pushed 1 new commit to sonar-coverage: http://git.io/SxhD5Q
fcrepo4/sonar-coverage 587cea3 Esmé Cowles: Adding ITs for TransactionServiceImpl and VersionServiceImpl
* github-ff leaves
<pivotal-bot>Longshou Situ added comment: "@cbeer I am seeing the ConstraintViolationException again after rebuilding master with the latest codes fr..." https://www.pivotaltracker.com/story/show/7227094616:34
* travis-ci joins
[travis-ci] fcrepo4/fcrepo4#2095 (sonar-coverage - 587cea3 : Esmé Cowles): The build passed.
[travis-ci] Change view : https://github.com/fcrepo4/fcrepo4/compare/1f4d1400ac36...587cea394498
[travis-ci] Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/26331838
* travis-ci leaves
<bljenkins>Yippee, build fixed!16:46
Project fcrepo4 build #1928: FIXED in 32 min: http://ci.fcrepo.org/jenkins/job/fcrepo4/1928/
<pivotal-bot>Esme Cowles added "SPARQL Update fails with PathNotFoundException when node path contains spaces" https://www.pivotaltracker.com/story/show/7229680416:56
Esme Cowles edited "SPARQL Update fails with PathNotFoundException when node path contains spaces" https://www.pivotaltracker.com/story/show/7229680416:58
* jcoyne joins17:12
Is there going to be an official alpha 6 release?
It's been 6 weeks since a517:13
<escowles>jcoyne: i think the plan is to release b1 early next week17:17
* dwilcox leaves17:23
* dwilcox joins17:25
* dwilcox leaves17:31
* dwilcox_ joins17:32
<cbeer>escowles: sounds like this JMS thing ought to be a blocker for that17:33
* dwilcox_ leaves17:39
<jcoyne>cbeer: JMS thing?
<cbeer>jcoyne: JMS messages should use URIs, not internal paths and things.17:41
<jcoyne>Ahh.17:42
* ermadmix leaves17:55
* ksclarke leaves18:33
* jcoyne leaves18:46
* scossu leaves19:10
* longshou leaves20:04
* jcoyne joins20:19
* nikhiltri leaves20:55
* ksclarke joins22:04
* jcoyne leaves22:49
<pivotal-bot>Kevin Clarke finished "Update Eclipse/IntelliJ formatting settings files" https://www.pivotaltracker.com/story/show/7126229823:34
* jcoyne joins23:40

Generated by Sualtam