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

Using timezone: Eastern Standard Time
* ksclarke leaves01:01
* scossu leaves01:29
* acoburn joins07:38
<pivotal-bot__>Aaron Coburn added "Fix index.html links in message-consumer" https://www.pivotaltracker.com/story/show/8339335607:40
Aaron Coburn started "Fix index.html links in message-consumer" https://www.pivotaltracker.com/story/show/83393356
Aaron Coburn finished "Fix index.html links in message-consumer" https://www.pivotaltracker.com/story/show/83393356
* awead joins08:02
<pivotal-bot__>Aaron Coburn added comment: "A potentially related issue occurs when the following sequence is followed: ""08:04
* Create a node at /parent
* Add �" https://www.pivotaltracker.com/story/show/83268844
* scossu joins08:57
<pivotal-bot__>Stefano Cossu added comment: "I still think that the DELETE and INSERT statements should be carried out in a single query for each node be�" https://www.pivotaltracker.com/story/show/8326884409:03
* scossu leaves09:05
* github-ff joins09:29
[fcrepo-camel] acoburn opened pull request #20: Transform accept header (master...transform_accept_header) http://git.io/RUCaCw
* github-ff leaves
* ksclarke joins09:34
* osmandin joins10:08
* scossu joins10:14
<pivotal-bot__>Andrew Woods added "Transform accept header" https://www.pivotaltracker.com/story/show/8340503610:21
Andrew Woods edited "Transform accept header" https://www.pivotaltracker.com/story/show/83405036
Andrew Woods started "Transform accept header" https://www.pivotaltracker.com/story/show/83405036
Andrew Woods finished "Transform accept header" https://www.pivotaltracker.com/story/show/83405036
Andrew Woods added comment: "https://github.com/fcrepo4-labs/fcrepo-camel/pull/20" https://www.pivotaltracker.com/story/show/8340503610:22
Andrew Woods edited "Transform accept header" https://www.pivotaltracker.com/story/show/83405036
* ajs6f1 joins10:27
<awoods>ajs6f1: ping10:31
<ajs6f1>pung
<awoods>ajs6f1: There is a bug that I wanted to discuss with you
<ajs6f1>You already are.
<awoods>ajs6f1: it is like a dream come true10:32
<ajs6f1>awoods: Do you see Freddy Kruger anywhere?
<awoods>ajs6f1: NamespaceRdfContext...
ajs6f1: is now including two copies of "fedora=http://fedora.info/definitions/v4/repository#"10:33
<ajs6f1>Weird.
<awoods>ajs6f1: which is a problem
<ajs6f1>To my memory, it's just pulling the jJCR namespaces and serializing them. But I could be wrong. Let me go look.
<awoods>ajs6f1: It has to do with the collapsing of namespaces and the magic at line:7410:34
ajs6f1: re line:74, there is no longer an "fcrepo" prefix
ajs6f1: It should be "fedora"
ajs6f1: but the "fedora" prefix is already defined in the CND10:35
<ajs6f1>awoods: I don't undertstand— that line appears to be converting "jcr" to something.
<awoods>ajs6f1: line:74 is converting jcr->fcrepo
<ajs6f1>Yeah.
<awoods>ajs6f1: but, that results in two prefixes that define the same namespace10:36
ajs6f1: fedora=http://fedora.info/definitions/v4/repository# and fcrepo=http://fedora.info/definitions/v4/repository#
<ajs6f1>awoods; That is definitely not an error for any RDF serialization I know of…
Where is the error getting thrown?
Or are you just wanting to get rid of the "fcrepo" prefix because it's obsolete?10:37
<awoods>ajs6f1: The error shows up in the UI, with the pre-populated PREFIXes in the Properties-Update form
<ajs6f1>awoods: Some kind of a problem with a map, or the like?10:38
<awoods>ajs6f1: since the sparql-update machinery chokes when two prefixes define the same namespace
ajs6f1: additionally, as you mentioned, we want to get rid of the fcrepo prefix
<ajs6f1>awoods: Oh, that's weird. It really shouldn't, I don't think. That's ugly, but legit SPARQL, isn't it
<awoods>ajs6f1: it should not be used
<ajs6f1>awoods: Well, what about just blocking the jcr ns?
Instead of converting it to something.10:39
<awoods>ajs6f1: there are a number of ways we can address the problem, yes...
* scossu1 joins
<ajs6f1>awoods: Actually, that whole loop could be factored into a Guava filter and transform.
<awoods>ajs6f1: Would you be interested in putting in a patch?10:40
<ajs6f1>awoods: Not today, but sure, I could do that.
<awoods>ajs6f1: it has to happen today
ajs6f1: I will do it
<ajs6f1>Why?
<awoods>ajs6f1: it is a blocker for the release, which is Thurs10:41
* scossu leaves
<ajs6f1>awoods: Do you accept that blocking the JCR ns would be a fix?
<awoods>ajs6f1: I would rather use a Map that overwrites duplicate keys
ajs6f1: changing line:75 to: "prefix = "fedora""10:42
<ajs6f1>awoods: Why would we want to put the fedora ns in twice at all?
<awoods>ajs6f1: but blocking the jcr ns is reasonable... probably.
ajs6f1: it seems slightly more robust to handle duplicate namespaces if they happen10:43
ajs6f1: but blocking the jcr ns is reasonable... probably.
<ajs6f1>awoods: A duplicate ns for you is a duplicated ns/prefix pair?10:44
<awoods>ajs6f1: a duplicate ns for me is a duplicate key
ajs6f1: key = prefix
<ajs6f1>awoods: That means that the last ns/prefix pair in wins. That's not deterministic.10:45
<awoods>ajs6f1: true
<ajs6f1>and it doesn't allow aliasing, which I admit is a small use case.
<awoods>ajs6f1: do you like the jcr-blocking approach?10:46
<ajs6f1>awoods: I like the idea of not exposing the JCR ns at all. Am I not right in thinking that if that causes problems, we've turned up another bug (one where we expose JCR guts)?
<awoods>ajs6f1: we hide the guts by changing the prefix to "fcrepo" (or rather "fedora")10:47
<ajs6f1>awoods: but the ns URI is still there.
<awoods>ajs6f1: But I am happy to block the guts
ajs6f1: no, actually line:80 then swaps out the ns URI10:48
<ajs6f1>awoods: So right now, we're exposing JCR stuff, but swapping the ns....
So it's not "visibly" JCR.
Doesn't that mean that we're implicitly importing the entire JCR ns (all those predicates)?10:49
And re-publishing it under our ns?
<awoods>ajs6f1: quite possibly10:51
<ajs6f1>awoods: Urg. That seems like a tremendous gap. Well, I don't know how much we can do about that today. But we should think about a much tighter interface there.
awoods: It makes the idea of an ontology-driven API fail down badly.10:52
awoods: Anyway, do you want to do the filtering you described (only one prefix allowed per URI)?
<awoods>ajs6f1: I think one of the primary goals in the new year should be flushing out our ontology-driven API10:53
<ajs6f1>awoods: I think we should flush it, indeed.
Replace it with SOAP
And then we'll have a really clean repo.
<awoods>ajs6f1: It is not that I want to limit one prefix per URI, but rather the Jena machinery requires it.10:54
* awead leaves
<ajs6f1>awoods: If I'm right in remembering that multiple prefixes-per-URI is legit SPARQL, we should file a bug on that. But that's a side note.
<awoods>ajs6f1: We have an immediate bug, I think our options are to block the jcr namespace in NamespaceRdfContext, or use a Map that overwrites duplicate entries (based on key)10:55
<ajs6f1>awoods: If we block the JCR ns, then we're relying on the CND-sourced fedora ns to be there, right? I don't have a problem with that, but we should remember that we now have that assumption.10:56
It doesn't seem unreasonable.
<awoods>ajs6f1: Everything that is defined in our provided CND is required10:57
ajs6f1: Fedora would not work without those definitions.
<ajs6f1>awoods: Okay. Then I think blocking the JCR ns is a better move. It allows people to alias, keeps things deterministic, and fixes the bug.
<awoods>ajs6f1: ok, I will put in a patch against the fcrepo-4.0.0rc branch10:58
<ajs6f1>Wait, it doesn't allow people to alias, because they'll run into the same weakness in Jena.
Oh, well.
awoods: I can do that, but it won't be for a little while (hours).
<awoods>ajs6f1: It needs to be done today, I will let you know before I start working on it, and you please do the same.10:59
<ajs6f1>kk
awoods: I just git fetch'd, but I don't see a fcrepo-4.0.0rc branch or tag…11:04
awoods: I don't see it from github's ui, either. Where is it?11:05
<awoods>ajs6f1: https://github.com/fcrepo4/fcrepo4/tree/fcrepo-4.0.0-RC11:07
<acoburn>awoods: related to this, the predicates passed to the JMS producer all use the JCR namespace: http://www.jcp.org/jcr/1.0lastModified,http://www.jcp.org/jcr/1.0createdBy
<ajs6f1>weird
<awoods>acoburn: yes, we will need to clean that up... in 4.0.111:08
ajs6f1: let me push the proper branch up11:09
<ajs6f1>awoods: Thnx.
afk bbs
* ajs6f1 leaves
* github-ff joins11:11
[fcrepo4] awoods created fcrepo-4.0.0rc (+1 new commit): http://git.io/VhoetQ
fcrepo4/fcrepo-4.0.0rc c9bf0a8 Andrew Woods: [maven-release-plugin] prepare release fcrepo-4.0.0-RC
* github-ff leaves
* ajs6f joins11:20
<pivotal-bot__>Andrew Woods added "Remove JCR references in JMS messages" https://www.pivotaltracker.com/story/show/8341133411:21
Andrew Woods edited "Remove JCR references in JMS messages" https://www.pivotaltracker.com/story/show/83411334
<awoods>ajs6f: https://github.com/fcrepo4/fcrepo4/tree/fcrepo-4.0.0rc
<ajs6f>awoods: what acoburn said makes me think that we have to think of the ns question differently in future. The case of mapping nses is part of a larger question of mapping between JCR model and Fedora model. There are various places we need to do that. RDF prodution is one, JMS production another. The nses are not conveniences for serialization. They are pointers to the different models.11:22
<awoods>ajs6f: I feel like it fits into the API planning11:23
<ajs6f>awoods: yes, but also into other things. Any functional surface for the repo. E.g. the JMS publshing, or maybe resourcesync. Depending on how we do that.11:24
<awoods>ajs6f: agreed, we need to be intentional about our descriptions of resources.11:25
<ajs6f>awoods: We also need to be extensional about them. IOW, we need to give explicit ontology and not implicit "ontology-by-inadvertent-reference".11:26
* awoods going onto a call11:27
* awead joins11:28
* travis-ci joins11:33
fcrepo4/fcrepo4#3240 (fcrepo-4.0.0rc - c9bf0a8 : Andrew Woods): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/commit/c9bf0a88ce9c
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/42095482
* travis-ci leaves
<ajs6f>is anyone else seeing massive compilation problems in -kernel-inpl for the RC branch? I can't get it to build at all...11:34
awoods: I can't help you with the NamespaceRdfzcontext issue. I can't even get the RC to build.11:39
<osmandin>ajs6f: I also got a build error: "expected to find namepsace..fedora but was fcrepo"11:48
<ajs6f>osmandin: that's different from my problems. I was getting class not found errors.11:49
is anyone getting this to build cleanly?
<osmandin>ajs6f: yeah, mine is a test error from -webapp11:50
<acoburn>I was able to get the fcrepo-4.0.0rc branch to compile, but only on the second attempt. Something about a zip file being closed?11:51
<osmandin>no compilation error on my side11:55
* ajs6f leaves11:57
* ajs6f joins12:02
* ajs6f leaves12:10
* ajs6f joins12:13
afk bbl12:20
* ajs6f leaves
<awoods>awead: ping12:40
<awead>awoods: pong
<awoods>awead: I was mentioning to jcoyne on the projecthydra list that I was interested in you fcrepo-4.0.0-RC testing...12:41
awead: I was mentioning to jcoyne on the projecthydra list that I was interested in your fcrepo-4.0.0-RC testing...
<awead>awoods: yeah, I tried it out yesterday. seems to be working fine.
<awoods>awead: also, there was a minor bugfix that will also need to go in...12:42
<awead>awoods: I had to make a minor patch to ActiveFedora to get it working, but other than that, no problems
<awoods>awead: the bugfix should have no effect on your work, but I wanted to raise it to your attention.
<awead>awoods: ah, so the bugfix is post RC release?
<awoods>awead: yes, but it has not been committed to master
awead: would you like to test an RC-2...12:43
awead: or do something else?
awead: I would not expect it to have any effect on you.
<awead>awoods: nah, I can always build against master12:44
awoods: are you still expecting to have 4.0 out by T-giving?
<awoods>awead: yes.
awead: but I did not want to have you test an RC, then release a slightly updated version.
<awead>awoods: k. letm e know if you need anything from me prior to that. I’m out tomorrow traveling, but will check email. I have to run right now, but will be back in about 30 mins.12:45
<awoods>awead: ok, thanks
<awead>awoods: no worries about the updated version...
* awead leaves
* awead joins13:06
<osmandin>afk13:08
<pivotal-bot__>Andrew Woods added "Duplicate namespace prefixes: fedora and fcrepo" https://www.pivotaltracker.com/story/show/8342281613:32
Andrew Woods started "Duplicate namespace prefixes: fedora and fcrepo" https://www.pivotaltracker.com/story/show/8342281613:33
* ruebot_ joins13:43
* ruebot_ leaves
* github-ff joins13:50
[fcrepo4] awoods pushed 1 new commit to fcrepo-4.0.0rc: http://git.io/vBjKfA
fcrepo4/fcrepo-4.0.0rc e33f20c Andrew Woods: Eliminate duplicate namespace prefixes: fedora and fcrepo...
* github-ff leaves
<pivotal-bot__>Andrew Woods added comment: "Resolved with: https://github.com/fcrepo4/fcrepo4/commit/e33f20c5ead4526e95ae8bd34b4f4cc5fdb77d1213:51
on branch -�" https://www.pivotaltracker.com/story/show/83422816
Andrew Woods finished "Duplicate namespace prefixes: fedora and fcrepo" https://www.pivotaltracker.com/story/show/83422816
* ajs6f joins14:09
<awoods>ajs6f: fyi: https://github.com/fcrepo4/fcrepo4/commit/e33f20c5ead4526e95ae8bd34b4f4cc5fdb77d1214:10
ajs6f: note, that is on the fcrepo-4.0.0rc branch
* travis-ci joins14:11
fcrepo4/fcrepo4#3241 (fcrepo-4.0.0rc - e33f20c : Andrew Woods): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/compare/c9bf0a88ce9c...e33f20c5ead4
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/42113878
* travis-ci leaves
<f4jenkins>Project fcrepo4-release-tests build #53: UNSTABLE in 1 min 18 sec: http://jenkins.fcrepo.org/job/fcrepo4-release-tests/53/
<ajs6f>awoods: Fine by me. I'lll let you know Monday if I ever get it to build.
<awoods>ajs6f: if you are trying to build the RC branch, you should probably build the whole baseline, since the pom.xml version number is different.14:12
ajs6f: you would not be able to simply build fcrepo-kernel-impl without having build the other modules first.
<ajs6f>awoods: That's what I was trying.
(Start from scratch)
* github-ff joins14:26
[fcrepo4] acoburn opened pull request #669: aliasing any jcr-namespaced properties to fedora: (master...jms-jcr-properties) http://git.io/H1GHOg
* github-ff leaves
<pivotal-bot__>Aaron Coburn added comment: "https://github.com/fcrepo4/fcrepo4/pull/669" https://www.pivotaltracker.com/story/show/8341133414:27
Aaron Coburn started "Remove JCR references in JMS messages" https://www.pivotaltracker.com/story/show/83411334
Aaron Coburn finished "Remove JCR references in JMS messages" https://www.pivotaltracker.com/story/show/83411334
* github-ff joins14:36
[fcrepo-camel] awoods pushed 1 new commit to master: http://git.io/UIkgAg
fcrepo-camel/master 04247cf Aaron Coburn: Set the Accept header to application/json when using the transform option...
* github-ff leaves
* github-ff joins14:37
[fcrepo-camel] awoods closed pull request #20: Transform accept header (master...transform_accept_header) http://git.io/RUCaCw
* github-ff leaves
<pivotal-bot__>Andrew Woods added comment: "Resolved with: https://github.com/fcrepo4-labs/fcrepo-camel/commit/04247cf0d606568b64ffd324dec29dc2e2307aae" https://www.pivotaltracker.com/story/show/83405036
Andrew Woods delivered "Transform accept header" https://www.pivotaltracker.com/story/show/83405036
* osmandin leaves14:46
* travis-ci joins
fcrepo4-labs/fcrepo-camel#37 (master - 04247cf : Aaron Coburn): The build passed.
Change view : https://github.com/fcrepo4-labs/fcrepo-camel/compare/5bf5a3502005...04247cf0d606
Build details : http://travis-ci.org/fcrepo4-labs/fcrepo-camel/builds/42118549
* travis-ci leaves
<awoods>afk14:48
* jgpawletko joins15:20
* osmandin joins15:21
* osmandin leaves15:26
<awoods>awead: If you are interested, and have the time, please feel free to test the war file built from the fcrepo-4.0.0rc branch: https://s3.amazonaws.com/f4-artifacts/fcrepo-webapp-4.0.0-RC.war15:56
<awead>awoods: ok16:00
<awoods>awead: thanks, and let me know if anything comes up.
<awead>awoods: will do… may not get to it before the day is up, but I doubt there would be any problems16:01
<awoods>awead: np. I do not expect there to be any issues. The difference is one minor bugfix: https://github.com/fcrepo4/fcrepo4/commit/e33f20c5ead4526e95ae8bd34b4f4cc5fdb77d1216:02
<awead>awoods: yeah, I don’t see anything that looks like it was cause problems16:03
<awoods>awead: I will be continuing with testing tomorrow... followed by Thurs release, if all is well.16:04
<awead>awoods: ok. I’m away stating tomorrow until Monday or Tuesday
<awoods>awead: be safe16:05
<awead>awoods: thanks!
* acoburn leaves16:29
* scossu joins16:40
* scossu1 leaves16:41
* github-ff joins17:07
[fcrepo-message-consumer] awoods pushed 1 new commit to fcrepo-4.0.0rc: http://git.io/H-Cakg
fcrepo-message-consumer/fcrepo-4.0.0rc 6d8c640 Aaron Coburn: Fix links in webapp index page...
* github-ff leaves
<pivotal-bot__>Andrew Woods added comment: "Resolved with: https://github.com/fcrepo4/fcrepo-message-consumer/commit/6d8c640b934e521a756b48441f4e5f04813a�" https://www.pivotaltracker.com/story/show/8339335617:08
Andrew Woods edited "Duplicate namespace prefixes: fedora and fcrepo" https://www.pivotaltracker.com/story/show/83422816
Andrew Woods edited "Remove JCR references in JMS messages" https://www.pivotaltracker.com/story/show/83411334
Andrew Woods accepted "Transform accept header" https://www.pivotaltracker.com/story/show/8340503617:09
* ajs6f leaves
<pivotal-bot__>Andrew Woods edited "Remove JCR references in JMS messages" https://www.pivotaltracker.com/story/show/83411334
Andrew Woods edited "Fix index.html links in message-consumer" https://www.pivotaltracker.com/story/show/83393356
<f4jenkins>Yippee, build fixed!17:12
Project fcrepo4-release-tests build #54: FIXED in 1 min 14 sec: http://jenkins.fcrepo.org/job/fcrepo4-release-tests/54/
* scossu1 joins
* scossu2 joins17:15
* scossu leaves17:16
* scossu1 leaves
<pivotal-bot__>Andrew Woods added comment: "Partially addressed with: ""17:17
https://github.com/fcrepo4/fcrepo-message-consumer/commit/760b3ba8e5f222c574af0b1d7�" https://www.pivotaltracker.com/story/show/83268844
Andrew Woods edited "Triples removed from triplestore index when a related resource is created" https://www.pivotaltracker.com/story/show/83268844
Andrew Woods edited "Use JUnit @Test 'timeout' attribute instead of our own code" https://www.pivotaltracker.com/story/show/8333062817:18
Andrew Woods added comment: "Pending further discussion." https://www.pivotaltracker.com/story/show/8341133417:19
Andrew Woods rejected "Remove JCR references in JMS messages" https://www.pivotaltracker.com/story/show/83411334
* ksclarke leaves18:13
* tecoripa joins18:24
* sprater leaves18:37
* scossu2 leaves18:53
* awead leaves21:42
* ksclarke joins21:44
* jgpawletko leaves23:55
* ksclarke leaves00:49

Generated by Sualtam