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

Using timezone: Eastern Standard Time
* ksclarke leaves01:07
* dwilcox joins08:46
* mikeAtUVa joins09:14
* ksclarke joins09:15
<pivotal-bot>Esme Cowles started "Add support for HTTP OPTIONS verb" https://www.pivotaltracker.com/story/show/6946750009:34
* scossu joins
<mikeAtUVa>Could someone remind me again why we invoke session.logout() after every update to the rest API?10:00
* scossu leaves10:01
* gregjansen joins
<awoods>mikeAtUVa: what is your suggestion?10:02
mikeAtUVa: long-lived sessions create a performance bottleneck.10:03
<mikeAtUVa>awoods: I'm just curious... does session.save() fail to persist the changes or something? Are we worried about inadvertent longstanding sessions?
* scossu joins10:04
<mikeAtUVa>awoods: alright... that makes sense. I'm just trying to find the right model for hanging on to a lock... by default it's session based... there's also a token, but that's pretty cumbersome (like transactions are cumbersome).10:05
<awoods>mikeAtUVa: If you do not logout, what would happen to that session after the request?
* mikeAtUVa shrugs...
awoods: hopefully we'd use it subsequent requests from the same client... in the normal Http Session sense (implemented with cookies I believe).10:06
Having "session == request" is sort of limiting... though perhaps that is the essence of REST.10:08
<awoods>mikeAtUVa: there is the notion of statelessness.10:11
* scossu leaves
* dwilcox leaves10:13
* dwilcox joins10:16
* dwilcox leaves10:20
* dwilcox joins10:44
* longshou joins10:53
* scossu joins11:06
* scossu leaves11:18
* dwilcox leaves11:19
* dwilcox joins11:20
* dwilcox leaves
* dwilcox joins
* dwilcox leaves11:22
* dwilcox joins
* dwilcox_ joins11:23
* dwilcox leaves
* longshou leaves11:24
* github-ff joins11:47
[fcrepo4] escowles pushed 1 new commit to ldp-updates: http://git.io/zLTLhw
fcrepo4/ldp-updates 6153212 Esmé Cowles: Adding support for OPTIONS requests on nodes with appropriate headers specified by LDP (Fixes https://www.pivotaltracker.com/story/show/69467500)
* github-ff leaves
<pivotal-bot>feature: Add support for HTTP OPTIONS verb (started) / owner: Esme Cowles
Esme Cowles added comment: "https://github.com/futures/fcrepo4/pull/285" https://www.pivotaltracker.com/story/show/6946750011:48
Esme Cowles finished "Add support for HTTP OPTIONS verb" https://www.pivotaltracker.com/story/show/69467500
Esme Cowles started "Return timestamp when creating node" https://www.pivotaltracker.com/story/show/6830156211:57
<cbeer>go escowles go12:09
<escowles>trying to sneak one more in before 12:30...
<cbeer>i'm optimistic i'm going to get something working with this Prefer header this morning... although i'm starting to suspect we need to build a better way to inject HTTP headers into the response12:11
* travis-ci joins12:12
[travis-ci] futures/fcrepo4#1721 (ldp-updates - 6153212 : Esmé Cowles): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/6106b93a8cd8...61532121c842
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/23048661
* travis-ci leaves
* scossu joins12:13
<escowles>that sounds like a good idea -- with the different sets of predicates the Prefer header can specify, it'll get a lot more complicated
* longshou joins
* scossu leaves12:16
<pivotal-bot>Andrew Woods added comment: "@ksclarke, is this ticket actually finished? or are you going to continue with larger content sets? (i.e. 6..." https://www.pivotaltracker.com/story/show/6866296412:53
Chris Beer added comment: "I think this may be blocked by #69467742" https://www.pivotaltracker.com/story/show/69469628
<cbeer>escowles: do you know if the order of modeshape's child node iterator is predictable?12:54
<pivotal-bot>Kevin Clarke added comment: "I want to try running a 60k set." https://www.pivotaltracker.com/story/show/6866296412:55
<escowles>cbeer: i don't know, but i would guess not
<pivotal-bot>Kevin Clarke started "Execute cross-backend performance testing regime" https://www.pivotaltracker.com/story/show/68662964
<cbeer>escowles: yeah, that's what i'm afraid of.. to do the Prefer header, i've had to split the getTreeTriples into a thing that does the parent/child triples and the thing that actually inlines the child resources
but i don't have a way to relate the two12:56
<escowles>is the parent/child triple generator still in JcrRdfTools (i.e. in its children like FedoraResource)?13:00
<cbeer>hm, this is interesting: https://github.com/tacitknowledge/jcr-mock13:02
escowles: there's an accessor in JcrRdfTools, but the actual logic is in HierarchyRdfContext13:03
<awoods>jcr-mock could come in handy, cbeer.13:04
<escowles>cbeer: but JcrRdfTools could save a reference to the HierarchyRdfContext and then you'd be able to access it from the part doing the inlining (if that part had a reference to the FedoraResource or whatever) -- seems a little convoluted, but maybe simpler than trying to work it all out at the JCR level
<cbeer>escowles: i was also thinking about adding or changing the interface to let me pass in a node iterator to the HierarchyRdfContext13:07
or teaching FedoraResource how to give us a consistently ordered child node iterator13:08
<escowles>i just found some references to sling:orderedFolder -- a JCR type (or extension) to order the children and get a predictable iteration order13:10
* scossu joins13:16
* scossu1 joins13:20
* scossu leaves
* edInCo joins13:29
* github-ff joins13:37
[fcrepo4] escowles pushed 1 new commit to ldp-updates: http://git.io/qsxOeg
fcrepo4/ldp-updates c871354 Esmé Cowles: Adding Last-Modified header with create/update timestamp when nodes and datastreams are created/updated (https://www.pivotaltracker.com/story/show/68301562)
* github-ff leaves
<pivotal-bot>feature: Return timestamp when creating node (started) / owner: Esme Cowles
Esme Cowles added comment: "Last-Modified seems perfectly appropriate to me -- I've added it for nodes and datastreams on both POST and ..." https://www.pivotaltracker.com/story/show/6830156213:38
Esme Cowles finished "Return timestamp when creating node" https://www.pivotaltracker.com/story/show/68301562
* travis-ci joins14:03
[travis-ci] futures/fcrepo4#1722 (ldp-updates - c871354 : Esmé Cowles): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/61532121c842...c871354cc57c
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/23056587
* travis-ci leaves
* nbanks joins14:06
* gregjansen leaves14:10
<ksclarke>awoods: you were right about the roles; the problem of one role seems to be on the testing side of things rather than the main code side of things14:13
* ksclarke changes rabbit holes
<pivotal-bot>Esme Cowles added comment: "I think the user would need to do two things for "<> a <http://www.w3.org/ns/ldp#Container" to work: ""14:22
1. Reg..." https://www.pivotaltracker.com/story/show/66168308
Chris Beer added comment: "Do we want to persist the LDP triples, though? I assumed we wouldn't, because they are server-managed and imm..." https://www.pivotaltracker.com/story/show/6616830814:52
* gregjansen joins15:16
<cbeer>escowles: ok, i think i have something working with Prefer: by changing the HierarchyRdfContext to:15:27
HierarchyRdfContext(final javax.jcr.Node node, final IdentifierTranslator graphSubjects, final LowLevelStorageService lowLevelStorageService, final Map<String, String> options)
Think I should change that Map<> to a proper class?
<escowles>cbeer: yeah, probably better to have a real class15:28
<cbeer>ok. figured as much.
<escowles>HierarchyRdfContextOptions #barf
<cbeer>yep. and we probably need a HierarchyRdfContextOptionsFactory15:29
maybe even a HierarchyRdfContextOptionsFactoryBean15:30
* escowles lol15:31
* nbanks leaves15:33
<pivotal-bot>Nikhil Trivedi added comment: "Issues addressed: https://github.com/futures/fcrepo-jms-indexer-pluggable/pull/3615:43
My apologies for creat..." https://www.pivotaltracker.com/story/show/64929156
Nikhil Trivedi finished "Support for authentication in JMS indexer" https://www.pivotaltracker.com/story/show/6492915615:45
* dwilcox_ leaves15:58
* dwilcox joins
<pivotal-bot>Esme Cowles started "Support LDP Paging spec" https://www.pivotaltracker.com/story/show/6946962815:59
<cbeer>escowles: you might want to hold off on that16:00
<pivotal-bot>Esme Cowles added comment: "I see some machinery for paging in FedoraNodes.describe() -- but it doesn't appear to be having any effect (..." https://www.pivotaltracker.com/story/show/6946962816:01
<escowles>cbeer: the upshot is that paging doesn't seem to be working, so that's the main thing to do for that ticket (adding the headers will be very easy)
<cbeer>escowles: yeah, i think i've fixed it in https://www.pivotaltracker.com/story/show/69467742 with the serialization options i'm doing
<pivotal-bot>feature: Support the HTTP `prefer` header (started) / owner: Chris Beer
<cbeer>i think the switch to rdf streaming broke it, and we must not have any itests for it16:02
* dwilcox leaves16:03
<escowles>i'm actually done for the day, so if your update is ready tomorrow am, i can add the LDP paging headers with some ITs so we notice if it breaks again
<cbeer>escowles: sounds good, thanks
* nbanks joins16:07
* github-ff joins16:08
[fcrepo4] ksclarke opened pull request #287: resolves issue where user with 'writer' role can't list ACL (master...acl-writer-68547530) http://git.io/AaGKUQ
* github-ff leaves
* nbanks leaves
<pivotal-bot>Kevin Clarke added comment: "Resolution: https://github.com/futures/fcrepo4/pull/287" https://www.pivotaltracker.com/story/show/68547530
Kevin Clarke finished "Add node-level or graph-level locking." https://www.pivotaltracker.com/story/show/6609378816:09
* gregjansen leaves
<pivotal-bot>Kevin Clarke started "Add node-level or graph-level locking." https://www.pivotaltracker.com/story/show/66093788
<ksclarke>sorry mikeAtUVa, just accidentally 'finished' your ticket
(it was above the one I was trying to 'finish'16:10
<pivotal-bot>Kevin Clarke finished "Unable to list ACL if user includes "writer" role" https://www.pivotaltracker.com/story/show/68547530
<mikeAtUVa>ksclarke: too bad you didn't actually finish it;)16:11
* gregjansen joins
* scossu joins16:12
* scossu1 leaves
<pivotal-bot>Esme Cowles unstarted "Support LDP Paging spec" https://www.pivotaltracker.com/story/show/6946962816:14
Mike Durbin added "Modeshape doesn't release locks held by sessions when logout() is invoked." https://www.pivotaltracker.com/story/show/6957487816:27
Mike Durbin started "Modeshape doesn't release locks held by sessions when logout() is invoked." https://www.pivotaltracker.com/story/show/69574878
* gregjansen leaves16:32
* gregjansen joins16:36
<pivotal-bot>Kevin Clarke started "FedoraFileSystemConnectorIT test fails when running mvn install without running a clean first" https://www.pivotaltracker.com/story/show/6901380216:59
* gregjansen leaves17:00
* gregjansen joins17:03
* gregjansen leaves17:26
* github-ff joins17:44
[fcrepo4] cbeer created prefer (+1 new commit): http://git.io/Bv5rkw
fcrepo4/prefer aab525c Chris Beer: Implement Prefer: HTTP headers to control LDPC hierarchy serialization
* github-ff leaves
<pivotal-bot>Chris Beer added comment: "https://github.com/futures/fcrepo4/pull/288" https://www.pivotaltracker.com/story/show/69467742
Chris Beer finished "Support the HTTP `prefer` header" https://www.pivotaltracker.com/story/show/69467742
* github-ff joins
[fcrepo4] cbeer opened pull request #288: Implement Prefer: HTTP headers to control LDPC hierarchy serialization (master...prefer) http://git.io/QR6AaA
* github-ff leaves
* dwilcox joins17:59
<pivotal-bot>Chris Beer edited "Add fcrepo4 integration tests based on LDP use cases and requirements" https://www.pivotaltracker.com/story/show/6947140618:03
* dwilcox leaves
* gregjansen joins18:25
* gregjansen leaves
<pivotal-bot>Andrew Woods delivered "Unable to list ACL if user includes "writer" role" https://www.pivotaltracker.com/story/show/6854753018:33
* github-ff joins
[fcrepo4] awoods pushed 2 new commits to master: http://git.io/KtTCbw
fcrepo4/master 8802dd6 Kevin S. Clarke: resolve acl issue where user with reader and writer roles can't read
fcrepo4/master 0540f6a Andrew Woods: Merge pull request #287 from ksclarke/acl-writer-68547530...
* github-ff leaves
* travis-ci joins18:35
[travis-ci] futures/fcrepo4#1724 (prefer - aab525c : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/commit/aab525cb0c5b
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/23075056
* travis-ci leaves
* scossu leaves18:42
<pivotal-bot>Andrew Woods accepted "Multiple Fedoras Feeding Single Triplestore" https://www.pivotaltracker.com/story/show/6885335418:45
* travis-ci joins18:52
[travis-ci] futures/fcrepo4#1726 (master - 0540f6a : Andrew Woods): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/6c9cdcaba6c4...0540f6aaf5f8
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/23078601
* travis-ci leaves
<pivotal-bot>Andrew Woods accepted "Unable to list ACL if user includes "writer" role" https://www.pivotaltracker.com/story/show/6854753018:59
* dwilcox joins
* dwilcox leaves19:03
* edInCo leaves19:50
<pivotal-bot>Longshou Situ added comment: "https://github.com/futures/fcrepo-sample-dataset/pull/1" https://www.pivotaltracker.com/story/show/6919058820:20
* longshou leaves21:42
* scossu joins22:18
* scossu leaves22:47
* scossu joins22:53
* scossu leaves22:57
* scossu joins
* scossu leaves
<pivotal-bot>Kevin Clarke added comment: "60k runs ended with same error that the 500k runs did (see above ticket). " https://www.pivotaltracker.com/story/show/6866296423:31
Kevin Clarke finished "Execute cross-backend performance testing regime" https://www.pivotaltracker.com/story/show/6866296423:33

Generated by Sualtam