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

Using timezone: Eastern Standard Time
* kaarefc joins03:04
* kaarefc leaves03:30
* kaarefc joins05:24
* pivotal-bot_____ leaves07:19
* pivotal-bot_____ joins
* nbanks joins08:36
* fasseg joins08:58
<pivotal-bot_____>Frank Asseg started "Add configuration files used to the wiki SCC benchmark results for better reproducability" https://www.pivotaltracker.com/story/show/5941536009:03
Frank Asseg added comment: "Added configuration file links to https://wiki.duraspace.org/display/FF/Performance+evaluation+on+the+SCC+Cl..." https://www.pivotaltracker.com/story/show/5941536009:13
Frank Asseg finished "Add configuration files used to the wiki SCC benchmark results for better reproducability" https://www.pivotaltracker.com/story/show/59415360
* kaarefc leaves09:17
* gregjansen joins09:19
* ksclarke leaves09:48
* ksclarke joins09:50
<pivotal-bot_____>Andrew Woods added "Bump project dependencies" https://www.pivotaltracker.com/story/show/59487140
Andrew Woods edited "Bump project dependencies" https://www.pivotaltracker.com/story/show/59487140
Andrew Woods started "Bump project dependencies" https://www.pivotaltracker.com/story/show/59487140
Andrew Woods edited "Bump project dependencies" https://www.pivotaltracker.com/story/show/59487140
* kaarefc joins09:52
* osmandin joins09:55
<awoods>fasseg: Are you using a load-balancer?09:58
in your SCC tests?
<fasseg>awoods: yeah apache with mod-jk and simple round-robin
<awoods>fasseg: I do not see that mentioned on your wiki page... and therefore, am not sure Osman has put one up.09:59
<fasseg>hmm I see, want me to add a section about loadbalancing?
<awoods>fasseg: That would be great...10:00
<fasseg>will do...
<awoods>fasseg/osmandin: The goal being to have reproducible environment.
* kaarefc leaves10:02
<pivotal-bot_____>Andrew Woods added comment: "Working branch: https://github.com/futures/fcrepo4/compare/dep-update" https://www.pivotaltracker.com/story/show/5948714010:04
* tecoripa joins10:06
* ajs6f joins10:07
<pivotal-bot_____>Osman Din added comment: "That won't help because the particular parser exception seems to be w.r.t an issue with the data being read. I..." https://www.pivotaltracker.com/story/show/58726512
Osman Din finished "Intermittent test errors: AtomJMSIT " https://www.pivotaltracker.com/story/show/5872651210:08
Andrew Woods started "Add configuration files used to the wiki SCC benchmark results for better reproducability" https://www.pivotaltracker.com/story/show/5941536010:09
<awoods>fasseg: I added the load-balancer note to: https://www.pivotaltracker.com/story/show/59415360
<pivotal-bot_____>feature: Add configuration files used to the wiki SCC benchmark results for better reproducability (started) / owner: Frank Asseg
<pivotal-bot_____>Andrew Woods edited "Test different confguration settings of infinispan/jgroups in order to raise performance on the SCC cluster" https://www.pivotaltracker.com/story/show/5940135010:10
Andrew Woods edited "change to modeshape URLBinaryValue and FileSystemConnector to have a option for large files" https://www.pivotaltracker.com/story/show/59285784
<awoods>osmandin: You are done with the following ticket: https://www.pivotaltracker.com/story/show/58726512 ?10:12
<pivotal-bot_____>bug: Intermittent test errors: AtomJMSIT (finished) / owner: Osman Din
<osmandin>awoods: yes, i think the small change referenced in the ticket should fix the current problem. for the other stack trace (that i'm getting) I think another ticket should be created if build fail because of that. Please refer to comments in the ticket.10:13
<awoods>osmandin: thanks, I saw the comment.
<pivotal-bot_____>Frank Asseg added comment: "Added a section about load balancing: https://wiki.duraspace.org/display/FF/Performance+evaluation+on+the+SC..." https://www.pivotaltracker.com/story/show/5941536010:14
<awoods>osmandin: load-balancer?
<fasseg>awoods: I added the section and finished the ticket again then....
<awoods>fasseg: you're the man
osmandin: Is it safe to assume you have not configured the Yale cluster with a load-balancer?10:19
<osmandin>awoods: yes, i haven't used a load balancer.. i think there were only 50 objects in the benchmark10:20
i'll set up one and regenerate the graphs
<fasseg>osmandin: btw: I have the same issue when deleting objects from fcrepo4, the sproadically turn up again, after a while...10:21
I guess will have to create a bug ticket for that
<awoods>fasseg: Are you generating new datastreams for each PUT?10:22
<osmandin>fasseg: ok. i have the same issue
fasseg: are those new 50 mb new objects
<fasseg>awoods: yes from java.util.Random()
<awoods>fasseg: good
fasseg: Maybe add an "issues discovered" section on your wiki page.10:23
<fasseg>I might take a while to generate 50mb of data using Random but still no comparison to the time needed for ingest
thats a couple of ms < 1% of ingest time
<awoods>fasseg: Or you can generate it before running the test... or just add a single character to an existing 50mb file so the SHA-1 is different.10:24
<fasseg>Indeed, it would just be a simple change in the code, but I haven't done it yet ;) .....I'll write it down and commit the change today10:25
<gregjansen>ajs6f: question about graph responses and reconciling ACL endpoint with it's representation in triple land10:29
* tecoripa1 joins
<gregjansen>ajs6f: my child node subject looks like this <http://localhost:57808/rest/workspace:fedora/rolesTest/authz:rbacl>, but the endpoint is at fcr:accessRoles (at the moment)10:30
ajs6f: i.e. that is the suffix on the path
* tecoripa leaves
<gregjansen>ajs6f: https://gist.github.com/gregjan/7138214#file-object-rdf-n3
ajs6f: so, my first question is.. should I make them match?10:31
ajs6f: in gist search for authz:
<ajs6f>Which one is the triple actually describing— the node or the endpoint?10:33
<gregjansen>ajs6f: so currently the endpoint that I've added for acl edits is here: @Path("/{path: .*}/fcr:accessRoles")
ajs6f: I think the triple is describing the node, i.e. our serialization code just grabs all children
<ajs6f>Okay, so then I would stay with the node's subject, not the endpoint's subject.10:34
We're not presenting a list of endpoints, right? We're presenting a set of assertions about the node...
<gregjansen>okay, so leave them separate. I just didn't want to violate some kind of practice there10:35
<fasseg>awoods: heh I actually did it the right way, the time needed for the datastream generation is not included in the duration calculation ;)
<gregjansen>I think so. Seems like we are describing JCR nodes
<awoods>fasseg: nice
<ajs6f>Well, we are desccribbing Fedora resources, which are very often nodes.
Sometimes we might want to describe an endpoint, but that's a different _thing_.
<gregjansen>sure okay.. well then my next question might be, should this ACL node be included or not in the graph and if so, should I be implementing some graph representation other than this default serialization of the node..10:36
<ajs6f>The ACL is a useful thing to know about the node… I suppose we should include it.10:37
Do we have a reaonable predicate to use… e.g. hasACL?
That kind of thing?
<cbeer>http://www.w3.org/wiki/WebAccessControl ?10:38
<awoods>ask, and you shall receive.10:40
<cbeer>ajs6f: what ever happened to injecting the graph subjects object into the jax-rs resources?10:41
<ajs6f>cbeer: {sigh} It's on my list.
<gregjansen>very interesting
i wonder if roles can really map to acl:mode10:42
* ksclarke leaves
<ajs6f>cbeer: You're welcome to move on it. I've got a huge pile of other RDF crap in front of me.
cbeer: It should look much like injectable Sessions.
<cbeer>ajs6f: ok. just checking if you went down that road and found trouble.
<ajs6f>cbeer: No, just made a ticket to remind me to go down that road and find the trouble.
<gregjansen>I'd love to use an existing vocab, but read/write/append/control won't allow extensible roles, unless I am mistaken, which is quite likely10:44
<awoods>cbeer/ajs6f: It is worth noting that InjectableSessions will be fundamentally changed with jersey 2.
cbeer/ajs6f: Therefore, it may be best to waiting on doing something similar, again, with GraphSubjects.10:45
<gregjansen>in any case, I will add a future ticket for describing role-based acls in RDF
<ajs6f>awoods: Yes. But the JAX-RS enhancements in 2 look a lot like the jersey 1 prototypes. I don't think we'd have to redo much. And how soon will we do Jersey 2? We've had the ticket for months.
<pivotal-bot_____>Gregory Jansen added comment: "After consulting with folks in IRC, determined that the description of the node is not the same as the de..." https://www.pivotaltracker.com/story/show/5826343010:46
Gregory Jansen started "Reconcile access roles rest endpoints with resource graphs" https://www.pivotaltracker.com/story/show/58263430
<awoods>ajs6f: The main blocker I see for jersey 2 (besides it being an intensive update) is the dependency Metrics has on Jersey 1.10:47
<pivotal-bot_____>Gregory Jansen added comment: "For example of current serialization, see
Th..." https://www.pivotaltracker.com/story/show/58263430
<ajs6f>awoods: grr. Metrics!
<pivotal-bot_____>Gregory Jansen finished "Reconcile access roles rest endpoints with resource graphs" https://www.pivotaltracker.com/story/show/58263430
Gregory Jansen added "Add roles-based ACL to RDF serialization" https://www.pivotaltracker.com/story/show/5949197610:49
Gregory Jansen estimated "Add roles-based ACL to RDF serialization" as 3 points https://www.pivotaltracker.com/story/show/5949197610:50
* edInCo joins
<awoods>ajs6f/cbeer: In terms of priorities, I would suggest we save an injectable GraphSubjects for after the Jersey upgrade. There is a raft of "core capabilities" that we should nail down as we approach the end of the year.
<ajs6f>awoods: okay. Just put it on the pile of technical debt. {grin}10:51
afk bbi 5
* ajs6f leaves
<pivotal-bot_____>Andrew Woods accepted "Reconcile access roles rest endpoints with resource graphs" https://www.pivotaltracker.com/story/show/5826343010:55
* ajs6f joins
<gregjansen>awoods: I'd like to finish my "create simple acl model" ticket too. or should that wait on the repo move to futures?10:57
<fasseg>osmandin: I updated the bench-tool to include a throughput calculation.
<gregjansen>awoods: bbi3
<awoods>gregjansen: Is there code associated with "create simple acl model"?10:58
<fasseg>osmandin: and the complete duration calculation is more accurate now
<awoods>gregjansen: Or is that just wiki-work?
<osmandin>fasseg: thanks. is your load balancer on the 8th vm?10:59
<fasseg>osmandin: yep
<awoods>866-740-1260, participant code: 225729511:03
<cbeer>thanks. just looking for that
awoods: i'm on too11:05
<cbeer>i miss google hangouts already.11:06
<barmintor>be right there
* github-ff joins11:12
[fcrepo4] ajs6f pushed 2 new commits to FixRDFTypeJCRMixinTranslation: http://git.io/crrq2w
fcrepo4/FixRDFTypeJCRMixinTranslation 15eb166 ajs6f: Fixed reversion: now including mixin info as rdf:type again
fcrepo4/FixRDFTypeJCRMixinTranslation 5eea9be ajs6f: Added unit test for inclusion of RDF type triples from mixins
* github-ff leaves
<cbeer>focused? i never said the meeting was focused!11:17
<ajs6f>No one ever claimed that Hydra is focused.
<cbeer>is this a familiar IT failure? AtomJMSIT.testDatastreamTerm:213 Got wrong datastream ID! expected:<DATASTREAM> but was:<null>11:29
* github-ff joins
[fcrepo4] ajs6f pushed 1 new commit to FixRDFTypeJCRMixinTranslation: http://git.io/SxuG3w
fcrepo4/FixRDFTypeJCRMixinTranslation c825b32 ajs6f: Fixing checkstyle problems
* github-ff leaves
* github-ff joins11:30
[fcrepo4] ajs6f opened pull request #125: Fix rdf type jcr mixin translation (master...FixRDFTypeJCRMixinTranslation) http://git.io/Wek1Dg
* github-ff leaves
<pivotal-bot_____>A. "Senphitic" Soroka added comment: "https://github.com/futures/fcrepo4/pull/125" https://www.pivotaltracker.com/story/show/59441130
A. "Senphitic" Soroka finished "Replace functionality to translate between rdf:type and JCR mixins" https://www.pivotaltracker.com/story/show/59441130
Ye Cao added comment: "I still have following error when integrating solr indexer into webapp:11:33
Error creating bean with name 'solrInde..." https://www.pivotaltracker.com/story/show/57453280
* eddies joins11:36
<cbeer>dropping off.11:37
<pivotal-bot_____>Ye Cao added comment: "There is no error when "mvn clean install " parent project fcrepo-jms-indexer-pluggable OR fcrepo-jms-indexer-pl..." https://www.pivotaltracker.com/story/show/5745328011:38
* github-ff joins11:39
[fcrepo4] cbeer force-pushed move-copy-57813074 from ab2c787 to cf4a5ac: http://git.io/2a-q0w
fcrepo4/move-copy-57813074 cf4a5ac Chris Beer: expose copy and move operations through the REST API
* github-ff leaves
<eddies>anyone else getting audio cutouts?11:47
ah yes =)
<gregjansen>tecoripa: trying typing?
<tecoripa1>I'll call baclk on regular land line in a sec continue
<barmintor>tecoripa1: Are you standing in a solar flare?11:48
<osmandin>it's like the signal from the movie Contact
<gregjansen>or the movie Sunshine
<tecoripa1>i'm back11:49
* ksclarke joins11:56
<osmandin>cbeer: I've seen that AtomJMSIT.testDatastreamTerm:213 . Do you have the full log trace?11:58
* github-ff joins12:02
[fcrepo4] ajs6f created RDFIterationForMutation (+1 new commit): http://git.io/7Qd1tw
fcrepo4/RDFIterationForMutation dd2ae9a ajs6f: Distinguishing between managed and unmanaged RDF
* github-ff leaves
* github-ff joins12:03
[fcrepo4] ajs6f deleted ExposeRdfStream at 72aa4d8: http://git.io/EMHPXA
* github-ff leaves
<pivotal-bot_____>Osman Din started "Run OS level performance tests on Yale cluster" https://www.pivotaltracker.com/story/show/59251828
Osman Din finished "Run OS level performance tests on Yale cluster" https://www.pivotaltracker.com/story/show/5925182812:04
* edInCo leaves12:11
<pivotal-bot_____>Andrew Woods added comment: "Sounds like you just need to add httpmine to fcrepo-jms-indexer-webapp/pom.xml" https://www.pivotaltracker.com/story/show/5745328012:15
<ajs6f>afk bbl12:24
<cbeer>osmandin: i haven't been able to recreate that error.. but i've gotten a bunch of other bad errors out of -jms too12:26
seems to be intermittent
awoods: i'm getting some errors in fcrepo-auth-common with modeshape 3.612:32
<awoods>cbeer: but you are fine with mode-5?12:33
i'm not sure which of those authz errors are expected or not
but the NPEs probably aren't :P
<awoods>cbeer: are you digging in on this?12:34
cbeer: otherwise, I will... unless gregjansen beats me to it.12:35
<cbeer>awoods: probably not very deep
<awoods>cbeer/gregjansen: I will be back in ~20min
* github-ff joins12:40
[fcrepo4] awoods pushed 1 new commit to master: http://git.io/TzDxVA
fcrepo4/master a561e1f Chris Beer: Add ability to COPY and MOVE nodes (and their children)...
* github-ff leaves
<pivotal-bot_____>Andrew Woods added comment: "Resolved with: https://github.com/futures/fcrepo4/commit/a561e1f9082c8feda3ba9df202d7503231349135" https://www.pivotaltracker.com/story/show/57813074
Andrew Woods delivered "Implement node/tree copy." https://www.pivotaltracker.com/story/show/57813074
<cbeer>awoods: how are you merging the pull requests? i'm confused what that commit is12:50
it seems like you squashed the commits and ignored the pull request?12:51
is that what we're doing now?
* github-ff joins
[fcrepo4] cbeer closed pull request #123: Add support for moving and copying nodes via the java and REST APIs (master...move-copy-57813074) http://git.io/hMUCtg
* github-ff leaves
<cbeer>(and, i would have suggested keeping those as 3 separate commits.. certainly 2 separate ones if you wanted to lump the -kernel and -webapp commits together)12:52
* github-ff joins12:56
[fcrepo4] cbeer deleted move-copy-57813074 at cf4a5ac: http://git.io/vwyFLA
* github-ff leaves
* github-ff joins
[fcrepo4] cbeer deleted new-node-location-fcr-content at d9f5b5e: http://git.io/kKlTGw
* github-ff leaves
* travis-ci joins
[travis-ci] futures/fcrepo4#1050 (master - a561e1f : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/d08914b0aa5f...a561e1f9082c
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/12995116
* travis-ci leaves
* ksclarke leaves12:58
* ksclarke joins13:06
* edInCo joins13:13
<awoods>cbeer: The process is:13:21
- submit PR
- click "finish" in pivotal and include PR link
- I review the code and push to master
** If the PR has multiple commits, I squash them so one commit is associated with the pivotal ticket
** You are welcome to pre-squash in the PR, or indicate that you want specific commits to be un-squashed
** It seems like one final commit per ticket is clean, however.
<cbeer>awoods: ok, thanks. i'll be sure to do that next time.13:22
<pivotal-bot_____>Andrew Woods accepted "Implement node/tree copy." https://www.pivotaltracker.com/story/show/5781307413:27
Andrew Woods estimated "Replace functionality to translate between rdf:type and JCR mixins" as 3 points https://www.pivotaltracker.com/story/show/59441130
Andrew Woods added comment: "@osmandin, please include a comment or wiki link that documents the completion of this ticket." https://www.pivotaltracker.com/story/show/5925182813:29
Andrew Woods added "Upgrade Modeshape Version to: 3.6.0.Final" https://www.pivotaltracker.com/story/show/5950483213:31
Andrew Woods edited "Upgrade Modeshape Version to: 3.6.0.Final" https://www.pivotaltracker.com/story/show/59504832
* ajs6f1 joins13:33
* ajs6f leaves
<pivotal-bot_____>Osman Din added comment: "Test results are at : https://wiki.duraspace.org/display/FF/Test+-+Cluster+Setup+-+Yale" https://www.pivotaltracker.com/story/show/5925182813:40
<barmintor>awoods: since FCR3 is ending, I'm just removing final markers on methods when it interferes with testing. ok?13:43
(Also, I hate public, final instance methods)
<awoods>barmintor: agreed, public final stinks. But presumably these were there for a reason... but as you note, 3.7 is the end of the line. Have at it.13:44
* barmintor cackles malevolently
<gregjansen>barmintor: wouldn't that make them *more* final? <jk>13:45
* minellium joins13:57
* ajs6f1 leaves13:58
* minellium leaves
* ajs6f joins13:59
<awoods>Is the pivotal site down? https://www.pivotaltracker.com/14:01
* ajs6f leaves14:12
<pivotal-bot_____>Gregory Jansen edited "Extend PEP to support enforcement of access policies for graph representations. (filtering of child nodes)" https://www.pivotaltracker.com/story/show/5505351014:16
* tecoripa1 leaves14:20
* tecoripa joins14:22
<awoods>It looks like the FileSystemConnector work that ermadmix has been working got a little shout-out in the Modeshape 3.6.0 release: https://community.jboss.org/message/843008?et=watches.email.thread#843008
<tecoripa>cbeer: did you reslove your AtomJMSIT integration test problem?14:29
* github-ff joins14:32
[fcrepo4] ajs6f pushed 1 new commit to RDFIterationForMutation: http://git.io/YRRY8A
fcrepo4/RDFIterationForMutation aaca4e0 ajs6f: Developing machinery to compare new and old material for updating
* github-ff leaves
* ajs6f joins
<awoods>tecoripa: There is this fix from osmandin, that has not been pushed to master yet: https://www.pivotaltracker.com/story/show/5872651214:43
<pivotal-bot_____>bug: Intermittent test errors: AtomJMSIT (finished) / owner: Osman Din
<tecoripa>awoods: ok, thanks. I did get it fixed, earlier on... but that tests seems to be a bit twitchy. I haven't verified this, but my guess is if a build failed earlier, an activemq daemon listener process may be left running somewhere14:49
<ajs6f>tecoripa: nbanks once spotted something like that, but I don't think he ever resolved it.14:59
<tecoripa>logging out, logging back in and re-running the build seems to resolve it, every time. :(15:00
hence my suspicion about some process running, bound to a port...
again, could be easily confirmed/denied next time the error crops up.
* nbanks leaves15:04
* osmandin leaves15:18
* osmandin joins15:19
<pivotal-bot_____>Andrew Woods edited "Upgrade Modeshape Version to: 3.6.0.Final" https://www.pivotaltracker.com/story/show/5950483215:20
Andrew Woods edited "Upgrade Modeshape Version to: 3.6.0.Final" https://www.pivotaltracker.com/story/show/59504832
<awoods>Team: I am going to be out tomorrow morning, but ajs6f will run the stand-up.15:21
<ajs6f>right into the ground15:22
<awoods>You did not mention that, ajs6f :0
* kaarefc joins
<ajs6f>You really should have known.
You can ask people on the last spring how well I did with that.15:23
<fasseg>im off for today, ill see you guys tomorrow15:24
<pivotal-bot_____>Andrew Woods added "ModeShape 3.6 FileSystemConnector performance" https://www.pivotaltracker.com/story/show/5951141015:28
* fasseg leaves
<pivotal-bot_____>Andrew Woods edited "ModeShape 3.6 FileSystemConnector performance" https://www.pivotaltracker.com/story/show/59511410
* ajs6f leaves15:30
* ajs6f joins15:32
<pivotal-bot_____>Andrew Woods added comment: "Pending code-review comments." https://www.pivotaltracker.com/story/show/5872651215:33
* ermadmix joins15:34
<pivotal-bot_____>Andrew Woods accepted "Run OS level performance tests on Yale cluster" https://www.pivotaltracker.com/story/show/5925182815:35
Gregory Jansen added comment: "Not attachments, but I am halfway through that build now anyway.." https://www.pivotaltracker.com/story/show/5950483215:44
<awoods>gregjansen: attached.15:45
<pivotal-bot_____>Andrew Woods added comment:s https://www.pivotaltracker.com/story/show/59504832
Osman Din added comment: "https://github.com/osmandin/fcrepo4/commit/5eb8a53558387bc2f42fd112bee443c2a6ef2a19" https://www.pivotaltracker.com/story/show/5872651215:47
<gregjansen>awoods: thanks, got it15:59
<awoods>gregjansen: This is not super-high priority, but it seemed like you would be the one to tap the issue most quickly.
<gregjansen>It looks as though ModeShape has fixed the copy projected nodes problem where binaries were not copied (became references to external binaries)
<awoods>gregjansen: That is a treat.16:00
<gregjansen>awoods: there is a lot to look at in mode 3.6 as well
yeah, Mike D. will be happy to see that
<pivotal-bot_____>Andrew Woods added comment: "It looks like the formatting issue has to do with tabs instead of spaces. See line 131.16:08
I will take care of..." https://www.pivotaltracker.com/story/show/58726512
<barmintor>awoods: testing the changes mikeAtUVa requested - https://github.com/duraspace/fcrepo/compare/af4d1f62b2103cc6207bc21d7d5785d7904f5ad5...3.7.x-checkStore16:12
<barmintor>those are the tests that were outstanding16:14
<awoods>barmintor: So you are done?16:15
<barmintor>awoods: AFAIK
* github-ff joins
[fcrepo4] ajs6f pushed 1 new commit to RDFIterationForMutation: http://git.io/1nkOOA
fcrepo4/RDFIterationForMutation 75410ce ajs6f: Building new iterator-driven persistence machinery
* github-ff leaves
<barmintor>awoods: mike asked me to email him when I thought it was wrapped up, so there's that16:16
<awoods>barmintor: Mike is out, but maybe tecoripa can give it a final review.
* github-ff joins
[fcrepo4] ajs6f pushed 1 new commit to FixRDFTypeJCRMixinTranslation: http://git.io/5h8uSA
fcrepo4/FixRDFTypeJCRMixinTranslation cbad0c2 ajs6f: Removing accidental inclusion of experimental class
* github-ff leaves
<gregjansen>ajs6f: this explains why I haven't seen access denied exceptions for graph serializations so far: https://issues.jboss.org/browse/MODE-203616:17
<ajs6f>awoods: ^^^ see last commit
<awoods>ajs6f: got it.
<ajs6f>gregjansen: WTF?
seems like kind of a hole.16:18
<gregjansen>fixed in 3.6
<awoods>gregjansen: And that fix is what then broke your integration test?
<ajs6f>3.6 makes everything better. fcrepo3.6 modoe3.6, everything.
<gregjansen>nope, integration test was something else. still working it
actually, that could be related..16:19
<ajs6f>"When we try to pick out anything by itself, we find it hitched to everything else in the Universe."16:20
— John Muir
<awoods>barmintor: What is your timeline? It sounded like you wanted to pull the trigger on 3.7.1 tomorrow (or Saturday). Is it basically in the reviewers' court(s) now?
<barmintor>awoods: yes16:21
(writing the email to committers now)
<awoods>barmintor: It would be great if tecoripa and mikeAtUVa could give a look.16:22
* ajs6f leaves16:28
* github-ff joins16:29
[fcrepo4] awoods pushed 1 new commit to master: http://git.io/zJmyVw
fcrepo4/master 69ffe69 osmandin: Fix AtomJMSIT test failures...
* github-ff leaves
<pivotal-bot_____>Andrew Woods added comment: "Resolved with: https://github.com/futures/fcrepo4/commit/69ffe6958b68c89ad38560a2585f2b32f32931e7" https://www.pivotaltracker.com/story/show/5872651216:31
Andrew Woods delivered "Intermittent test errors: AtomJMSIT " https://www.pivotaltracker.com/story/show/58726512
* ajs6f joins16:34
<gregjansen>FYI, it *was* MODE-2036 that broke my tests.. also it is graph serialization that is calling Node.getNode().. So we have managed to chase our tails back to that issue yet again. Putting child node URIs into graph output will require that we construct the URI differently than this:16:35
<ajs6f>We can't call getNode()?
<gregjansen>well, actually, this PEP is pretty weird.. it probably needs an special condition for jcr:content16:36
<awoods>gregjansen: It looks to be a PEP that is only used in integration tests.
<gregjansen>this PEP is saying no for that particular node, so this is the behavior we will see when a child node is protected from current user. In this case the behavior of the PEP can be adjusted to make the test sane.16:37
yes, but I am sure to see a few more like this when I test the roles stuff against mode 3.6, since it includes protected children.16:38
I will give it a shot actually, brb
<awoods>gregjansen: Will the flow breakdown in the case of protected children?
<ajs6f>Won't someone think of the children?!? Oh, good, Greg is.16:39
gregjansen: Do we generally need a way to get child _URIs_ without getting the children themselves?
<gregjansen>If we want to point at them in the parent's graph, I'd say yes.. Seems like we will need that to take a middle road on the 403 children question16:40
<ajs6f>That, and also to impl LDP.16:41
I think. Maybe we can just not do that part.
<gregjansen>hmm, we may only be generating the jcr:content URI that way.. my other tests passed.
<ajs6f>No, I'm pretty sure the others get pulled… hold on, let me check.16:42
Interesting… any other child comes from getNodes(), not getNode(name).16:43
maybe that matters?
jcr:content is the only one we pull by name, which stands to reason. It's the only one we _know_ is special.16:44
<gregjansen>the aggregate one could skip protected children .. SILENTLY!
make sense re:jcr:content
<ajs6f>Returns all child nodes of this node accessible through the current
* <code>Session</code>. Does <i>not</i> include properties of this
* <code>Node</code>. The same reacquisition semantics apply as with {@link
* #getNode(String)}. If this node has no accessible child nodes, then an
* empty iterator is returned.
I think you're right, gregjansen. It's silently ignoring protected kids.
* travis-ci joins
[travis-ci] futures/fcrepo4#1052 (master - 69ffe69 : osmandin): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/a561e1f9082c...69ffe6958b68
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/13005883
* travis-ci leaves
<gregjansen>ignorance is bliss, right ajs6f?16:46
<ajs6f>I couldn't say— I've never studied the matter.
<tecoripa>ajs6f: so that makes you happy, right?16:47
<gregjansen>okay, so normally we are not allowing separate ACLs on the jcr:content node.. so this fail is by virtue of this dummy PEP, I will fix the build first..16:48
<bljenkins>Yippie, build fixed!
Project fcrepo4 build #1233: FIXED in 16 min: http://ci.fcrepo.org/jenkins/job/fcrepo4/1233/
awoods: Fix AtomJMSIT test failures16:49
<pivotal-bot_____>Andrew Woods accepted "Intermittent test errors: AtomJMSIT " https://www.pivotaltracker.com/story/show/58726512
<bljenkins>Project fcrepo-fixity-corrupter build #298: SUCCESS in 1 min 15 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/298/16:50
* kaarefc leaves16:51
<bljenkins>Project fcrepo-kitchen-sink build #549: STILL UNSTABLE in 2 min 34 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/549/16:53
<ajs6f>gregjansen: For future reference, wouldn't we want to allow separate ACLs for jcr:content? Like when I want to let you see the metadata but not the content?
<gregjansen>you're right, probably depends on type of PEP.. in a role-based PEP you probably would not.. but in a permissions based ACL system, yeah16:54
<ajs6f>Just something to keep in the back of our collective mind.16:55
* ermadmix leaves
<gregjansen>ajs6f: in rbac I have doubts about using datastream level roles actually.. folks will want it, but it seems like a bad idea if you've got roles16:56
<tecoripa>gregjansen: why is that? don't you want roles to be able to apply down to the datastream level?16:58
<gregjansen>they should definitely be applied at the ds level, but should they be assigned at the ds level?
<tecoripa>they should be assigned at the node level, right? but refer to a ds within the node?17:00
<ajs6f>If we don't do that, we better supply some good ways to make assertions about content with fine granularity, and be able to connect roles and types.
<gregjansen>tecoripa: this is more a comment on manageability than anything else. Why not encapsulate which datastreams are protected in the role definition, instead of in an ACL, where you have to tweak it. (This assumes that this particular datastream is not special)
* osmandin leaves17:01
<gregjansen>ajs6f, tecoripa: I would think that you want as few ACLs as possible to get the job done. So granularity is nice if you really need it, but a pain otherwise. If you are protecting the same datastream name/types on many objects, that sounds like a role thing. But if this datastream is special, that seems like an ACL thing.17:02
<tecoripa>gregjansen: I get it. what you said17:03
<gregjansen>tecoripa: it helps me to think about each ACL as something a human being would decide upon17:04
<tecoripa>we would prefer the former, but do we have to allow for the latter?
<ajs6f>Sounds like a question for the product owner.
<awoods>It would be a shame to not allow setting an ACL on a datastream.17:06
<gregjansen>In a role-based environment, I would protect it by setting datastream metadata.. But in a permissions-based environment (like the ModeShape access controls are) then I would put an ACL on the datastream content node I think.
awoods: I'll not argue with that
I think some specific datastreams need special protection, it's just unusual is all.17:08
awoods: I can make you a pull request that includes the fix and the pom change, you want that?17:09
<awoods>gregjansen: You know I do.
<tecoripa>gregjansen: "setting datastream metadata". Do you have an example of this? I'm thinking that the jcr:content just contains a blob...
<gregjansen>awoods: coming right up
<awoods>gregjansen: You can just use the normal pivotal/git flow
gregjansen: i.e. start the ticket17:10
finish the ticket
add a link to the PR in the ticket
<tecoripa>gregjansen: so where would you put the ds metadata, and how would it link to the jcr:content blob?
<ajs6f>tecoripa: Nothing prevents you from setting properties on jcr:content nodes in the usual way.
They're full-fledge nodes.17:11
It ain't like fcrepo3.
<gregjansen>they have a jcr:data property which holds the blob
<ajs6f>(Although even there, you could use RELS-INT).
out for the day17:12
* ajs6f leaves
<tecoripa>gregjansen, ajs6f: okay, got it.17:13
* github-ff joins17:14
[fcrepo4] gregjan opened pull request #126: ModeShape upgrade to 3.6, including necessary change to test PEP (master...pivotal-59504832) http://git.io/rzLlGw
* github-ff leaves
<pivotal-bot_____>Gregory Jansen added comment: "pull request is here https://github.com/futures/fcrepo4/pull/126" https://www.pivotaltracker.com/story/show/5950483217:15
Andrew Woods started "Upgrade Modeshape Version to: 3.6.0.Final" https://www.pivotaltracker.com/story/show/5950483217:17
Andrew Woods estimated "Upgrade Modeshape Version to: 3.6.0.Final" as 1 point https://www.pivotaltracker.com/story/show/59504832
Andrew Woods finished "Upgrade Modeshape Version to: 3.6.0.Final" https://www.pivotaltracker.com/story/show/59504832
<tecoripa>hey greg...17:19
<awoods>you may want to use his irc handle, tecoripa.17:20
gregjansen ^^^17:21
<gregjansen>thx awoods, we've gotten back in our own room.
<tecoripa>awoods: yep, sorry about that... typing in wrong window17:23
<awoods>no sweat.17:25
* gregjansen leaves
* barmintor leaves18:04
* edInCo leaves18:16
* tecoripa leaves19:12
* ksclarke leaves20:36
* ksclarke joins20:39
* github-ff joins22:11
[fcrepo4] ajs6f pushed 1 new commit to RDFIterationForMutation: http://git.io/4uQT9A
fcrepo4/RDFIterationForMutation f540613 ajs6f: Continuing to build iterating persistance machinery
* github-ff leaves
* github-ff joins22:24
[fcrepo4] ajs6f pushed 1 new commit to RDFIterationForMutation: http://git.io/-uUhnw
fcrepo4/RDFIterationForMutation 6dfd2e4 ajs6f: More unit tests
* github-ff leaves
* github-ff joins22:44
[fcrepo4] ajs6f pushed 1 new commit to RDFIterationForMutation: http://git.io/12hBCg
fcrepo4/RDFIterationForMutation a623168 ajs6f: More unit tests
* github-ff leaves
* github-ff joins22:49
[fcrepo4] ajs6f pushed 1 new commit to RDFIterationForMutation: http://git.io/M2L5PA
fcrepo4/RDFIterationForMutation 6813bdd ajs6f: More unit tests
* github-ff leaves
<bljenkins>Project fcrepo-fixity-corrupter build #299: SUCCESS in 4 min 40 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/299/23:25
Project fcrepo-fixity-corrupter build #300: SUCCESS in 3 min 24 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/300/00:09
Project fcrepo-kitchen-sink build #550: STILL UNSTABLE in 9 min 52 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/550/00:19
* kaarefc joins00:22
* kaarefc leaves00:23
* kaarefc joins01:14
* kaarefc leaves01:29
* kaarefc joins01:39
* kaarefc leaves01:56
* ksclarke leaves02:10

Generated by Sualtam