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

Using timezone: Eastern Standard Time
* scossu leaves01:04
* jonroby leaves02:00
* jgpawletko joins07:55
* dwilcox joins07:59
* nikhiltri joins08:00
* nikhiltri leaves08:01
* scossu joins08:07
* ksclarke joins08:21
* awead joins08:29
* MohamedAR joins08:38
* scossu leaves09:02
* scossu joins09:04
<pivotal-bot>Andrew Woods delivered "Camel: Licensing" https://www.pivotaltracker.com/story/show/8222613609:44
Andrew Woods added comment: "https://wiki.duraspace.org/display/FF/Setup+OAI-PMH+Provider" https://www.pivotaltracker.com/story/show/8148636009:47
Andrew Woods finished "Test OAI-PMH module" https://www.pivotaltracker.com/story/show/81486360
<f4jenkins>Yippee, build fixed!10:08
Project fcrepo4 build #2571: FIXED in 21 min: http://jenkins.fcrepo.org/job/fcrepo4/2571/
Project fcrepo4-release-tests build #8: FAILURE in 2 min 42 sec: http://jenkins.fcrepo.org/job/fcrepo4-release-tests/8/10:26
* jonroby joins10:31
* barmintor joins10:34
<jonroby>awoods: having the same problem with the 400 error code for the integration tests. How do I modify the test harness to recognize the oai prefix?10:36
<awoods>jonroby: You will likely want to include a custom CND in the tests.10:37
<jonroby>awoods: I have new integration tests to insert properties using sparql-update. how do I include a custom CND?10:38
<awoods>jonroby: I posted some links the other day to you in IRC that describe how to specify an alternate CND from your respository.json.
<barmintor>awoods: I’m looking at that test now10:49
<awoods>barmintor: IRC is probably better than email.
* tecoripa joins10:54
* nikhiltri joins11:00
* longshou joins11:02
<pivotal-bot>Longshou Situ started "Remove "Update CND" from HTML UI" https://www.pivotaltracker.com/story/show/8207010611:03
<awoods>MohamedAR: Great work with the ModeShape issue: https://github.com/ModeShape/modeshape/pull/131511:11
<barmintor>awoods: Yep, if TestAPIM.testAddDatastream runs before TestAPIM.testValidate, the latter will fail11:16
sprater ^^11:17
<awoods>barmintor: Is that a new test from sprater?11:18
<sprater>awoods, barmintor: really? doesn't ring a bell
* ajs6f joins11:19
<awoods>barmintor: I assume one or both of those tests needs to be updated in order to be more self-contained?
<barmintor>awoods: Nope. But JUnit tests do not run in a guaranteed order, and FCR3 no longer overwrites existing objects on ingest w/ existing PIDs
awoods: Exactly.
<sprater>barmintor: do you have a commit link?
<barmintor>sprater: There is no commit to link
<sprater>barmintor: I mean -- is there a commit I made (or the merged commit) that shows the change I made?11:20
<barmintor>I’m doing a poor job explaining: For whatever reason JUnit4 has, the tests are running in a different order for config C
it’s nothing you did
<sprater>barmintor: ah. I'll take off my sackcloth, then11:21
barmintor: you'd think we would have run into this before. I know others (you, me, others) did a lot of work to make each test self-sufficient11:22
barmintor: do you have time to look at the test right now? if not, I can take a look at it a little bit later
<barmintor>sprater: this comes up every once in a while. One of the big testing refactors I did fixed it for the demo objects, but these are “special” objects
<pivotal-bot>Andrew Woods delivered "Create initial Grinder test" https://www.pivotaltracker.com/story/show/8213226211:27
Andrew Woods added comment: "@danieldavis1, @ksclarke ...take a look at this work. I think it is a great first step." https://www.pivotaltracker.com/story/show/8213226211:28
<barmintor>I am not going to miss this “intentionally fail to authenticate 50 times and print the stacktraces” test11:34
* barmintor is running configB11:38
<pivotal-bot>Andrew Woods started "Organize and Collapse fcrepo-ontologies" https://www.pivotaltracker.com/story/show/8207535611:43
<f4jenkins>Yippee, build fixed!
Project fcrepo4-release-tests build #9: FIXED in 1 min 44 sec: http://jenkins.fcrepo.org/job/fcrepo4-release-tests/9/
* barmintor is running configQ11:44
<pivotal-bot>Andrew Woods added comment: "Please document your LevelDB inspection/modification details on the wiki as a child page of: ""11:45
https://wiki.dur�" https://www.pivotaltracker.com/story/show/81443506
<ajs6f>awoods: As you crush the NSes, are you going to update the various clients (the JVM client, the "indexer" client)?
<awoods>ajs6f: I suspect I will... I like for the builds to be working.11:46
<ajs6f>weirdo
<awoods>ajs6f: I was also thinking about moving our bug-tracking to JIRA with the 4.0 release... thoughts?11:47
<ajs6f>awoods: I think I'm probably the only person who will say this, but I like JIRA a lot more than Pivotal.11:48
awoods: One of the big things we need to deal with if we select for JIRA, is customization.
awoods: The basic JIRA setup is pretty bad. But it's enormously customizable.
awoods: From the basic workflow, to integrations, etc.
<barmintor>awoods: duraspace needs to get you a new laptop11:49
<awoods>barmintor: in December this one is 3-years old... which is the replacement cycle.
<barmintor>Merry Christmas!
(the machine I just built FCR3 on is a 2010 Mac Pro)11:50
<awoods>ajs6f: The trouble with pivotal is that it is not particularly easy for the community to add tickets.11:51
<ajs6f>I am seriously starting to think about taking out an AWS instance just as a remote builder. Setting up my own personal CI.
<awoods>ajs6f: JIRA is integrated with the wiki accounts.
<ajs6f>awoods: Yeah. And confluence generally (e.g. lots of macros and things).
Meaning integrated with conf, generally.
<awoods>ajs6f: I have been thinking the same, re:AWS-builder
<ajs6f>awoods: If I had a better-than-DSL connection, I might actually just run my IDE remotely over X.11:52
awoods: Have you ever heard of a convenient setup for "personal CI"? Were you think of just running Jenkins?11:53
<awoods>ajs6f: I think you would needs much, much better bandwidth than DSL for that not to be painful.
<ajs6f>awoods: Yep, that's why I'm going to start by "offloading" the builds themselves.
<awoods>ajs6f: I was thinking less automated builds and more along the lines of a mvn env for running "clean install" on projects.11:54
<ajs6f>awoods: Yeah, that would be utterly simple. I'm just thinking about the "UI" for it. I'd like to not have to edit locally, juggle git stuff, then build remotely. I'd like to have something more automatic. But not too automatic. Just the right amount of magic.11:55
awoods: Can you easily mount a filesystem from an AWS instance on a local box? In a reasonably-performant way?
This:11:58
http://www.oed.com/view/Entry/123113
was the OED "word of the day" yesterday.
I think we should make awoods the Mpret of Fedora.
<awoods>whatever12:00
<barmintor>An Albanian word, for which all usages are quoted refs to Albania… this is a weird inclusion12:03
<ajs6f>yes, I thought that too.12:04
<awoods>ajs6f: There are some FUSE impls for mounting S3, but mounting an EBS volume seems less likely.12:06
<ajs6f>awoods: Right. Oh, well. It would still be nice to just get the big, resource-hungry build processes themselves off of my laptop.12:08
<awoods>ajs6f: you and me both
<ajs6f>I know: I'll buy a hundred Raspberry PI boards and build my own cloud.12:09
<barmintor>awoods, sprater bamboo is into configQ now, looking good12:10
I wonder why we didn’t call it configD12:11
ding ding https://bamboo.duraspace.org/browse/FCREPO-CORE-13512:13
<ajs6f>Q is for quiddity.12:14
awoods/cbeer: Did you see that message to the list from "Dulip Withanage"?12:15
He was PUTing to an fcr:versions endpoint, trying to create a new version, and instead was ccreating a new resource.
A new object, rather.
Which actually makes sense, but maybe we should document that, or let PUT fall through to POST on that endpoint?12:16
I can imagine other people missing that, too.
I think he might have been creating objects called "fcr:versions", or something, which is probably not what anyone would want.12:17
<awoods>barmintor: was Q for "quick"?12:21
* barmintor shrugs
<awoods>ajs6f: I also saw that Dulip's request headers were wrong.12:22
<ajs6f>awoods: Oh, I didn't catch that. You want to send a reply to that effect?
<awoods>ajs6f: sure, I can respond12:23
<ajs6f>awoods: thnx, mpret
* awoods mercy
<barmintor>it would be slow as molasses, but could the fcr3-store util migrate from default PIDs to UUIDs by using the existing namespace as the seed for a pseudo-random generator, and replaying the sequence up to the PID number?12:28
* barmintor scratches his chin12:29
<awoods>barmintor: you are thinking about the F3 -> F4 migration?12:31
<barmintor>awoods: yeah
Eric’s email12:32
<awoods>barmintor: I missed Eric's email12:33
barmintor: was it good?
<barmintor>awoods: Ah, it was to hydra-tech
<awoods>barmintor: are you an admin of hydra-tech?12:35
barmintor: I am "banned" for some reason
barmintor: I do not feel like I deserve to be banned.12:36
<barmintor>awoods: you probably don’t. I can invite, but I can’t add directly
Wow, you are banned12:37
<awoods>it does not feel good
<pivotal-bot>Chris Beer added "Extract storage location from FixityRdfContext into a new stream with opt-in prefer header toggle" https://www.pivotaltracker.com/story/show/8230330212:41
<barmintor>awoods: I sent an email to the owner, fingers crossed for you12:42
awoods: anyway, he wants to see FCR3 PIDs persisted for migrated objects12:43
<awoods>barmintor: It is nice to have friends.
<barmintor>awoods: I can see doing this as metadata, like <uuid> <was:fedora3> <pid>
awoods: or I can see trying to make the resulting UUID predictable from the pid12:44
<awoods>barmintor: does he want to also be able to use the legacy pids for direct access: GET /fcrepo/pid:123 ?
<barmintor>awoods: I assume so, but it’s not completely clear
awoods: but some searchability of them seems important12:45
awoods: and a predictable algorithm would be a boon to migration tools
<awoods>barmintor: there would basically be four identifiers: 1) the resource path, 2) the modeshape UUID, 3) the legacy PID, 4) the newly created UUID.12:47
<ajs6f>awoods/barmintor: We wouldn't want people to end up with a flat F4 repo with a million children of the root. Do we have the "invisible hierarchy" feature available to prevent that?13:06
<barmintor>ajs6f: if you migrate to UUID, you’re in the storage hierarchy13:07
<ajs6f>barmintor: So that's a yes?
<barmintor>No, b/c that’s not invisible
<ajs6f>barmintor: That's a little worrying. Wouldn't we want the "default mapping" of identifiers to produce a performant hierarchy? Or can we not take responsibility for that...?13:08
<cbeer>(or we go and fix modeshape..)13:09
<barmintor>ajs6f: what I’m saying is that, afaik, if you migrate to UUIDs you are in a hierarchy
<ajs6f>:)
<barmintor>ajs6f: but if we allowed the PIDs unadorned, we would not
<ajs6f>barmintor: That's what I don't quite get: UUIDs are flat amongst themselves… whence does the hierarchy arise?
<barmintor>which is why I am inclined to say: you must migrate, but your PID will be findable via SPARQL
ajs6f: we pairtree the UUIDs, don’t we?13:10
<ajs6f>barmintor: I don't know— do we? That's great if we do.
I'm all in favor of discouraging people from expecting to keep their namespaced F3 pids straight into the new repos. Being able to find things by their old PIDs ought to be enough for any sensible integration.13:11
<awoods>cbeer: re:fixing-modeshape: https://issues.jboss.org/browse/MODE-210913:12
<cbeer>awoods: yep. i've been looking at the code.
it's not even a hard problem.. just needs time.13:13
<barmintor>cbeer: isnt it basically just moving stuff you implemented in FCR4 down into MODE?
<cbeer>barmintor: yep.13:14
<ajs6f>barmintor: It's probably moving stuff that _many_ people have impled over MODE (each a little differently) down into MODE.
<barmintor>ajs6f: yeah, but cbeer is way better
<ajs6f>I will back cbeer to the hilt.
* ksclarke leaves13:16
* ksclarke joins13:18
* ksclarke leaves13:23
* sprater leaves13:29
<awoods>afk13:35
* awead leaves13:38
* ksclarke joins13:43
* dwilcox leaves13:44
<barmintor>So right now our whole versioning API involves leaking JCR versions into the REST api13:46
<awoods>barmintor: you are talking about what shows up in the triples of specific REST requests?13:47
<barmintor>awoods: yeah
* awead joins
<barmintor>awoods: and also probably what we mean conceptually13:48
<awoods>barmintor: Maybe there is an opportunity to filter out some of those triples?
barmintor: really, you think the REST API contract exposes leakage?
<barmintor>well, I mean: the way that you make something versionable is to add a <http://www.jcp.org/jcr/mix/1.0versionable> type assertion via PATCH13:49
<awoods>barmintor: actually, that is no longer necessary. If you make a request to create a version of a resource, that mixin is automatically applied.13:51
barmintor: That said, we could probably remove any awareness of that mixin from the user.13:52
now I am really "afk"
be back shortly
<ajs6f>barmintor: I've thrown out in the past the idea that we could take over the Memento model for versioning and expose that.13:54
barmintor: not saying that M is perfect, but it's not JCR.
<barmintor>ajs6f: I’m sympathetic, though I’m trying to work out what it means LDPC wise
<ajs6f>barmintor: Fair question, and I don't pretend to know myself. Maybe we could get RS to raise it as a point of discussion there.13:55
<barmintor>guessing it means that the containers properties are in a container of containers **head explodes**13:56
<ajs6f>barmintor: It's okay. You can put all that confusion in a container and then put that into itself. Just don't put a bag of holding into another one.
* nikhiltri leaves13:57
* dwilcox joins
* nikhiltri joins14:09
* scossu leaves14:24
<jonroby>CAn anyone think of a reason why the rest harness would say a sparql-update to add a property works but then reading that property fails?14:44
*test not rest
* awead leaves14:52
* awead joins14:54
* scossu joins15:01
<awoods>jonroby: not really, unless you were managing the Session and it was not being saved.15:15
<ajs6f>jonroby: But this is in the integtration tests?15:16
<jonroby>awoods: its ok, i found the problem and i'm about finished
awoods: but thank you for responding
<awoods>np15:17
<pivotal-bot>Andrew Woods accepted "Update ValueConverter to exclusively produce JCR String values" https://www.pivotaltracker.com/story/show/8222845215:46
Andrew Woods accepted "Verify/Enable events on federated content" https://www.pivotaltracker.com/story/show/82224996
Andrew Woods edited "Verify/Deny events on federated content" https://www.pivotaltracker.com/story/show/82224996
Andrew Woods accepted "Camel: Licensing" https://www.pivotaltracker.com/story/show/82226136
<jonroby>awoods: i've commited my code to my repo and still have yesterdays PR. would the new changes be part of that PR or do I need to submit a new one?15:54
<awoods>jonroby: If you push your commits on the same branch to your github repo, they should automatically show up in your previous PR.15:55
<jonroby>awoods: I believe they are, I just wasn't sure as i'm not a big git user
<awoods>jonroby: looks like it worked: https://github.com/fcrepo4-labs/fcrepo4-oaiprovider/pull/6/commits15:56
<jonroby>awoods: ok15:57
<awoods>jonroby: apparently you are a bigger git user than you think.15:58
<jonroby>awoods: for this project at least15:59
<awoods>jonroby: what is your version control system of choice?
* dwilcox leaves16:01
<jonroby>awoods: we're trying to use git here but not that much. I've used cvs and subversion and a little bit of mercurial. I don't have a favourite. I think git seesm most fault tolerant
awoods: "user tolerant"
<awoods>;)16:02
<jonroby>awoods: it looks like it errored on compile but not on my stuff? on JQL?16:04
* nikhiltri leaves16:05
<pivotal-bot>Jonathan Roby added comment: "PR as above, integration tests added for property injection (description test didn't work)" https://www.pivotaltracker.com/story/show/8148290416:14
Jonathan Roby added comment: "Added integration tests for properties pulled from pom.xml and root node properties" https://www.pivotaltracker.com/story/show/8185050416:15
Jonathan Roby added comment: "After a few emails back and forth, Thomas assures me that possibly next week #fcrepo may be back. will conti�" https://www.pivotaltracker.com/story/show/7927841216:16
Mohamed Mohideen Abdul Rasheed added comment: "Added documentation at: https://wiki.duraspace.org/display/FF/How+to+inspect+LevelDB16:19
Updat�" https://www.pivotaltracker.com/story/show/81443506
Mohamed Mohideen Abdul Rasheed finished "Debug UNH issue" https://www.pivotaltracker.com/story/show/81443506
Andrew Woods added comment: "That is great. Let's keep tabs on this." https://www.pivotaltracker.com/story/show/7927841216:20
Andrew Woods delivered "Debug UNH issue" https://www.pivotaltracker.com/story/show/8144350616:21
Mohamed Mohideen Abdul Rasheed added comment: "Added integration test: https://github.com/ModeShape/modeshape/pull/131516:23
@hchiorean's comm�" https://www.pivotaltracker.com/story/show/72982948
Andrew Woods added comment: "Related commit in fcrepo4-oaiprovider: ""16:25
https://github.com/fcrepo4-labs/fcrepo4-oaiprovider/commit/efbb32c4136�" https://www.pivotaltracker.com/story/show/82228452
<awoods>jonroby: fcrepo4-oaiprovider should be back in business.
<jonroby>awoods: seems to be good16:30
<pivotal-bot>Jonathan Roby added comment: "Documentation updated to state that the version is pulled from the software and not added by the admin." https://www.pivotaltracker.com/story/show/8148636016:31
* ajs6f leaves16:34
<jonroby>awoods: so that PR was ok?16:37
* awead leaves
<awoods>jonroby: I have not looked closely at it. If you are done, please change the ticket state to "Finished"16:38
<jonroby>awoods: will do
<pivotal-bot>Jonathan Roby finished "Integration test for OAI#Identify" https://www.pivotaltracker.com/story/show/8185050416:42
Jonathan Roby finished "Allow admins to manage descriptive content about the repository and make the content available" https://www.pivotaltracker.com/story/show/81482904
Longshou Situ started "Fix oaiprovider ListSetsIT" https://www.pivotaltracker.com/story/show/8214139416:56
* MohamedAR leaves17:04
<pivotal-bot>Andrew Woods finished "Normal user can only change properties but not add child nodes" https://www.pivotaltracker.com/story/show/7298294817:16
* jonroby leaves17:29
<awoods>cbeer: ping17:41
<cbeer>awoods: pong17:48
<awoods>cbeer: I am collapsing the three fedora namespaces...17:50
cbeer: and hit the minor hiccup of mushing rels-ext:isPartOf into fedora:isPartOf17:51
cbeer: rels-ext is not a "managed" namespace (so updates involving rels-ext predicates are allowed), but the fedora namespace is managed.17:52
<cbeer>awoods: why, and, why do we need to touch the rels-ext namespace at all?
<awoods>cbeer: there are some tests that update rels-ext:isPartOf properties
cbeer: Now with LDP containment predicates, rels-ext:isPartOf seems arbitrary17:53
cbeer: I updated those tests with dcterms:isPartOf, but wanted to pass it by you for any red flags.
<cbeer>awoods: ok, lots of red flags.
<awoods>cbeer: like what?17:54
<cbeer>awoods: in fact, i don't know why we don't go back to using info:fedora/fedora-system:def/relations-external#isPartOf and calling it good.
http://www.fedora.info/definitions/1/0/fedora-relsext-ontology.rdfs
<awoods>cbeer: I don't think we want to propagate that legacy namespace
cbeer: <whatever>:isPartOf does not have any special meaning in F4.17:55
<cbeer>awoods: i don't think we need to migrate it into the future. but, if people are already using that, why not let them keep using it?17:56
<awoods>cbeer: sure
cbeer: is there any reason to bake the legacy namespace into our tests?
<cbeer>awoods: oh, but if you were just asking "can i change the tests to use dcterms", i don't have a problem with that.
<awoods>cbeer: what problem (red flag) do you have?17:57
<cbeer>in fact, i'd just change 'em to some:property or whatever.
awoods: i thought you were suggesting we convert everyone using fedora-rels-ext:isPartOf to use dcterms equivalents.
awoods: but if it is just tests, i don't care.. the predicates are arbitrary (and maybe we should be very clear about that)17:58
<awoods>cbeer: not at all. I am suggesting people use whatever vocab they want...
cbeer: the only real change here is that we used to support a fedora-defined namespace (rels-ext) that users could update (meaning users could update properties from that namespace), and now we do not have a fedora-defined namespace that users can "use".17:59
...and double-verifying that there is nothing special about rels-ext:isPartOf.18:00
<cbeer>awoods: there's nothing special about the predicates in the tests.
awoods: there is something special about info:fedora/fedora-system:def/relations-external#isPartOf
<awoods>cbeer: and how would you define the specialness of "info:fedora/fedora-system:def/relations-external#isPartOf"18:01
<cbeer>awoods: from the RDFS: "A definition of the generic part/whole relationship between fedora objects. The subject is a fedora object representing a part and the predicate is a fedora object representing a whole."
it's a structural relation. dcterms:isPartOf doesn't give you that.
there's nothing special as far as fcrepo is concerned, but it is special for consumers of fcrepo.18:02
<awoods>cbeer: dcterms:isPartOf: "A related resource in which the described resource is physically or logically included."
cbeer: anyways, I think we are good.
* ksclarke leaves18:19
* jonroby joins18:23
* scossu leaves18:34
* MohamedAR joins18:50
* scossu joins19:05
* scossu leaves19:14
* jonroby leaves19:24
* jonroby joins19:29
* MohamedAR leaves19:32
* github-ff joins19:38
[fcrepo4] lsitu opened pull request #625: Removed CND update form. (master...feature/cnd-form) http://git.io/5mnZqA
* github-ff leaves
<pivotal-bot>Longshou Situ added comment: "https://github.com/fcrepo4/fcrepo4/pull/625" https://www.pivotaltracker.com/story/show/8207010619:40
Longshou Situ finished "Remove "Update CND" from HTML UI" https://www.pivotaltracker.com/story/show/82070106
* scossu joins20:00
* scossu leaves
* jonroby leaves20:02
* ajwagner leaves20:21
* ajwagner joins20:22
* dwilcox joins
<pivotal-bot>Aaron Coburn started "Camel: Add maven-managed integration test environment" https://www.pivotaltracker.com/story/show/8222571820:23
Aaron Coburn started "Camel: Triplestore integration" https://www.pivotaltracker.com/story/show/8222691420:24
Aaron Coburn started "Camel: Add checkstyle and enforcer" https://www.pivotaltracker.com/story/show/82226486
Aaron Coburn finished "Camel: Add maven-managed integration test environment" https://www.pivotaltracker.com/story/show/82225718
Aaron Coburn added comment: "Added support for maven-managed integration tests: https://github.com/acoburn/fcrepo-camel/commit/20196280723�" https://www.pivotaltracker.com/story/show/8222571820:26
Andrew Woods added comment: "@aaroncoburn, can you also please add a link in this ticket to your PR?" https://www.pivotaltracker.com/story/show/82225718
Aaron Coburn added comment: "See this commit: https://github.com/acoburn/fcrepo-camel/commit/9a70318d5db1be6ce7654c602e152356b4e4e120" https://www.pivotaltracker.com/story/show/8222648620:29
Aaron Coburn finished "Camel: Add checkstyle and enforcer" https://www.pivotaltracker.com/story/show/82226486
* scossu joins20:30
<pivotal-bot>Aaron Coburn started "Camel: Switch in fcrepo4-client for custom code" https://www.pivotaltracker.com/story/show/8222593420:33
* scossu leaves20:39
<pivotal-bot>Aaron Coburn started "JMS Message property list" https://www.pivotaltracker.com/story/show/82135176
* ksclarke joins20:40
* dwilcox leaves20:57
<pivotal-bot>Longshou Situ added comment: "https://github.com/fcrepo4-labs/fcrepo4-oaiprovider/pull/7" https://www.pivotaltracker.com/story/show/8214139421:05
Longshou Situ finished "Fix oaiprovider ListSetsIT" https://www.pivotaltracker.com/story/show/82141394
* longshou leaves21:06
* jonroby joins21:14
* scossu joins22:35
* scossu leaves22:44
* MohamedAR joins22:46
* MohamedAR leaves23:04
* jonroby leaves23:55
* scossu joins00:40
* scossu leaves00:49
* ksclarke leaves01:06

Generated by Sualtam