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

Using timezone: Eastern Standard Time
* fcrepo-bot joins04:09
* fcrepo-bot leaves05:26
* dwilcox leaves06:08
* cbeer leaves06:34
* cbeer joins06:39
* pivotal-bot_____ joins06:42
* pivotal-bot____ leaves06:45
* awead joins07:24
* awead leaves07:25
* awead joins07:47
* nikhiltri joins08:00
* nikhiltri leaves08:05
* jcoyne joins08:34
* jonroby joins08:51
<pivotal-bot_____>Jonathan Roby added comment: "Got a reply!08:52
"Hi Jonathan,
Sorry about the length of time it has taken to get back to you. To speed this
p�" https://www.pivotaltracker.com/story/show/79278412
* MohamedAR joins08:53
<pivotal-bot_____>Jonathan Roby added comment: "And a reply back: ""08:59
"Hi Thomas (I assume)
My name is Jon Roby and i'm taking part in an open source project �" https://www.pivotaltracker.com/story/show/79278412
* awead leaves09:00
* awead joins09:02
<jonroby>where would a ns2 namespace come from? I'm seeing it on the oaiprovider output09:19
* awead leaves
* awead joins09:20
<jcoyne>ns2 sounds like a prefix (an abbreviation for a namespace), not a namespace
The prefix should be defined in the root node09:21
xmlns:ns2="..."
<jonroby>jcoyne: Ok thanks, i'll investigate more
jcoyne: you're correct, thanks, just making sure09:24
<pivotal-bot_____>Esme Cowles added "fcr:versions should default to text/turtle like other RDF resources" https://www.pivotaltracker.com/story/show/8158355609:32
Esme Cowles started "fcr:versions should default to text/turtle like other RDF resources" https://www.pivotaltracker.com/story/show/81583556
* github-ff joins09:35
[fcrepo4] escowles created fcr-versions-default-turtle (+1 new commit): http://git.io/hYTb5g
fcrepo4/fcr-versions-default-turtle a61de2c Esmé Cowles: Changing fcr:versions produces declaration to default to text/turtle
* github-ff leaves
<pivotal-bot_____>Esme Cowles added comment: "https://github.com/fcrepo4/fcrepo4/pull/595" https://www.pivotaltracker.com/story/show/81583556
* github-ff joins
[fcrepo4] escowles opened pull request #595: Changing fcr:versions produces declaration to default to text/turtle (master...fcr-versions-default-turtle) http://git.io/yFaxBw
* github-ff leaves
<pivotal-bot_____>Esme Cowles finished "fcr:versions should default to text/turtle like other RDF resources" https://www.pivotaltracker.com/story/show/81583556
<jonroby>Anyone: where would be the best place to put the wiki doc for the Oai Provider?09:37
<awoods>jonroby: Let's start as a child page of: https://wiki.duraspace.org/display/FF/Administrator+Guide09:41
* gregjansen joins
* gregjansen leaves09:42
* travis-ci joins09:47
fcrepo4/fcrepo4#3012 (fcr-versions-default-turtle - a61de2c : Esmé Cowles): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/commit/a61de2c7b2f7
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/39263619
* travis-ci leaves
* osmandin joins09:57
* github-ff joins10:11
[fcrepo4] awoods pushed 1 new commit to master: http://git.io/VaX4ng
fcrepo4/master 9b27b81 Andrew Woods: Merge pull request #595 from fcrepo4/fcr-versions-default-turtle...
* github-ff leaves
<pivotal-bot_____>Andrew Woods delivered "fcr:versions should default to text/turtle like other RDF resources" https://www.pivotaltracker.com/story/show/81583556
Mohamed Mohideen Abdul Rasheed started "Debug UNH issue" https://www.pivotaltracker.com/story/show/8144350610:19
* scossu joins10:25
<awoods>cbeer: with the removal of the namespaces service/endpoint, what is the expectation around creating resources with namespaces in their path/identifier?10:27
cbeer: curl -v -XPUT http://localhost:8080/rest/pid:10010:28
<cbeer>awoods: none. in my email, i called it out as a todo.
awoods: but you can do it, either by pre-registering a namespace in the CND, or by sending in some RDF in a namespace-aware format, that defines your prefix.10:29
<awoods>cbeer: Sure. Have you given thought to the "CRUD-style" interaction? I will create a ticket.10:30
<cbeer>awoods: none beyond what was in my email. we probably need to start with what namespace prefix in a URI even mean.10:31
* longshou joins10:32
<cbeer>(or, if people just like using colons in URIs.)10:33
<pivotal-bot_____>Andrew Woods added "CRUD on namespaces" https://www.pivotaltracker.com/story/show/81587996
Chris Beer finished "Make LDP containers optional" https://www.pivotaltracker.com/story/show/8141141610:34
Andrew Woods edited "Setup and document installing Grinder on AWS" https://www.pivotaltracker.com/story/show/81535724
Andrew Woods edited "Setup and document installing Grinder on AWS" https://www.pivotaltracker.com/story/show/81535724
<cbeer>awoods: wait, CRUD to create namespaces? isn't that what we had, and what was no good?
(and doesn't solve the problem anyway?)
<jcoyne>cbeer++ namespace prefixes in uris is horrid.10:35
<awoods>cbeer: Although there are workarounds, it seems logical to be able to establish a prefix:namespace.10:36
<cbeer>jcoyne: i'm not sure they're horrid. i just don't know what they mean, and they need to be used within the scope of a request, not within some magical new endpoint.
<awoods>cbeer: and to remove them.
cbeer: but, if there is no community need, we can let it ride.10:37
<jcoyne>cbeer: Besides the fact that it would remove a whole range of otherwise valid URIs (ones with colins)
* acoburn joins10:38
<cbeer>awoods: i think there is no community need for a new fcr:namespace endpoint. i'm not arguing about the idea of namespace prefixes in URIs, i'm arguing against your CRUD for namespaces ticket. we had that, it didn't work conceptually.
<awoods>cbeer: I am not suggesting a namespace endpoint10:39
<cbeer>awoods: then https://www.pivotaltracker.com/story/show/81587996 needs some more detail. i understood it to mean CRUD for namespaces.
<pivotal-bot_____>feature: CRUD on namespaces (unscheduled) / owner:
<awoods>cbeer: Here is your quote: "we should probably consider better ways of exposing this in CRUD-style requests instead of offering a separate feature."
<cbeer>awoods: and that's what you're trying to capture in that ticket?10:40
<awoods>cbeer: yes
<pivotal-bot_____>Chris Beer edited "CRUD requests should support namespace prefixes" https://www.pivotaltracker.com/story/show/81587996
Chris Beer edited "CRUD requests should support namespace prefixes" https://www.pivotaltracker.com/story/show/8158799610:41
<cbeer>awoods: that's what I mean by exposing it in CRUD requests.10:42
<osmandin>afk
<awoods>cbeer: that is much more clear
<cbeer>awoods: but we can also making jcoyne happy and just map colons to nonsense characters and decide that it isn't a feature.10:43
because i still don't know what it means to have URIs with namespaces embedded within them.
<awoods>cbeer: are you planning on removing fcr:sparql?
<cbeer>awoods: i think we should remove it.10:44
<jcoyne>cbeer: I don't have a pressing need for them to be removed, but it is confusing.
<pivotal-bot_____>Andrew Woods added "Remove fcr:sparql" https://www.pivotaltracker.com/story/show/81589024
<awoods>cbeer: Also, I have an N3 file that I am trying to push into F4, and am getting errors "Cannot convert BigDecimal value to long", which makes me wonder what value there is in the type-mapping in ValueConverter.java versus encoding all RDF values and their type info in Strings?10:49
<cbeer>awoods: in master, we're still converting literals to their JCR types when it isn't lossy. encoding them as strings was #588, but that'll need to be redone (if it's what we're going to do)10:50
<awoods>cbeer: I am raising the question, and am interested in the pros/cons of taking the #588 approach.10:51
<cbeer>awoods: then, yes, that is the pro. we get out of the business of type mapping.
<awoods>cbeer: Would the future, optional "validation" on RDF values come from Jena or JCR?10:52
<cbeer>awoods: not JCR. maybe not even fcrepo4.
<awoods>cbeer: as of yesterday, we are accepting RDF values that are not well-formed, no?10:53
<cbeer>awoods: yes
<awoods>cbeer: and there was talk of throwing an error in that case or not.
<cbeer>"that case"?
<awoods>cbeer: in the case where the RDF value is not well-formed10:54
cbeer: the question is, where would that checking/validation come from...
cbeer: JCR type-mapping or RDF type-mapping.
<cbeer>awoods: that is the question. and in #591, we sounded like we were happy to punt on the idea until later.
<awoods>cbeer: s/RDF/Jena/
<cbeer>awoods: it's certainly not JCR.10:55
<awoods>cbeer: agreed, put
cbeer: So it sounds like pushing RDF values into JCR strings would not impact the future ability to throw an exception when RDF values are not well-formed.10:56
s/put/punt/ (two comments up)
<cbeer>awoods: no, it would not. RDF validation and JCR storage are unrelated (other than storing literals in JCR types entails an extremely strict validation)10:57
<awoods>cbeer: Is there benefit in mapping (in the cases where we can) RDF types into JCR types?10:58
<cbeer>awoods: probably. at the very least, we get to work within JCR and take advantage of whatever optimizations they can make, and play nice with others.10:59
<barmintor>awoods: I was corresponding w/ RTC last night re: the search thing11:05
awoods: I get the feeling that question is being received as “No search facility” instead of “The RDF search facility alone, and not the other thing"
awoods: at least, that was the tone of his question to me11:06
<pivotal-bot_____>Chris Beer added "Specify our notion of even-more-basic-than-ldp-BasicContainer" https://www.pivotaltracker.com/story/show/8159183011:07
Chris Beer deleted "Specify our notion of even-more-basic-than-ldp-BasicContainer" https://www.pivotaltracker.com/story/show/8159183011:09
Chris Beer started "Make LDP containers optional" https://www.pivotaltracker.com/story/show/8141141611:15
<awoods>barmintor: just off standup11:16
barmintor: I assume you saw the back-and-forth on fedora-tech about the question of internal-admin-search.11:17
<barmintor>awoods: yes, just want to report how my steering committee rep sees the question11:18
<awoods>cbeer: ok, it sounds like staying with mapping into JCR-types is a good thing.
<cbeer>awoods: does it?
<awoods>cbeer: that is my interpretation of your response above.11:19
<cbeer>awoods: then, no. you asked what the benefits were.
<awoods>barmintor: Thanks. I will be talking with RTC momentarily.11:20
<pivotal-bot_____>Andrew Woods edited "fcr:versions should default to text/turtle like other RDF resources" https://www.pivotaltracker.com/story/show/8158355611:22
* github-ff joins11:28
[fcrepo4] cbeer force-pushed ldp-containers from e785707 to d9d5c5c: http://git.io/KIdBxA
fcrepo4/ldp-containers d9d5c5c Chris Beer: Fedora objects can be any type of LDP container...
* github-ff leaves
<pivotal-bot_____>Chris Beer finished "Make LDP containers optional" https://www.pivotaltracker.com/story/show/8141141611:33
* github-ff joins11:34
[fcrepo4] cbeer force-pushed ldp-containers from d9d5c5c to 49d4004: http://git.io/KIdBxA
fcrepo4/ldp-containers 49d4004 Chris Beer: Fedora objects can be any type of LDP container...
* github-ff leaves
<pivotal-bot_____>Chris Beer edited "Support the full range of LDP containers" https://www.pivotaltracker.com/story/show/81411416
Chris Beer edited "Support the full range of LDP containers" https://www.pivotaltracker.com/story/show/8141141611:35
* travis-ci joins
fcrepo4/fcrepo4#3015 (ldp-containers - d9d5c5c : Chris Beer): The build has errored.
Change view : https://github.com/fcrepo4/fcrepo4/compare/e78570744d94...d9d5c5c49757
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/39277189
* travis-ci leaves
* github-ff joins11:39
[fcrepo4] cbeer force-pushed ldp-containers from 49d4004 to 92733d1: http://git.io/KIdBxA
fcrepo4/ldp-containers 92733d1 Chris Beer: Fedora objects can be any type of LDP container...
* github-ff leaves
<pivotal-bot_____>Longshou Situ added comment: "@awoods With the current status of the fcrepo and message consumer implementation, we only need to fix it in�" https://www.pivotaltracker.com/story/show/7914393011:40
Longshou Situ finished "The sparql recipes no longer pass for 4.0.0-beta-04-SNAPSHOT" https://www.pivotaltracker.com/story/show/79143930
* travis-ci joins
fcrepo4/fcrepo4#3016 (ldp-containers - 49d4004 : Chris Beer): The build has errored.
Change view : https://github.com/fcrepo4/fcrepo4/compare/d9d5c5c49757...49d40044f0de
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/39277936
* travis-ci leaves
* longshou leaves11:44
<f4jenkins>Project fcrepo-module-auth-rbacl build #314: UNSTABLE in 4 min 25 sec: http://jenkins.fcrepo.org/job/fcrepo-module-auth-rbacl/314/11:47
* ksclarke joins11:50
<f4jenkins>Yippee, build fixed!12:00
Project fcrepo-module-auth-rbacl build #315: FIXED in 3 min 56 sec: http://jenkins.fcrepo.org/job/fcrepo-module-auth-rbacl/315/
* github-ff joins12:01
[fcrepo4] nikhiltri opened pull request #596: Applied code changes to fresh clone. (master...aic-velocity-templates) http://git.io/sQ112Q
* github-ff leaves
<pivotal-bot_____>Nikhil Trivedi added comment: "@awoods: Ok, I applied the changes to a fresh fork, hope this works: https://github.com/fcrepo4/fcrepo4/pul�" https://www.pivotaltracker.com/story/show/7920780412:02
<jcoyne>We're talking in #scholarsphere whether the terminology we should use.12:22
Does fcrepo 4 have "NonRdfSources" or does it have "Datastreams"
can we settle on one term?
s/whether/about/
* scossu leaves12:23
* awead leaves12:30
* awead joins
<osmandin>afk12:32
<cbeer>jcoyne: it would be healthy to think about non-rdf sources with an RDF description.12:33
<barmintor>jcoyne: Hope cbeer will chime in if I’m wrong, but I don’t think “datastream” has anything to offer over nonRDFSource except familiarity
<cbeer>and i really hope to s/Object/Container/; s/Datastream/BinaryDescription/ in the codebase.12:34
<barmintor>cbeer++
<jcoyne>cbeer++
* longshou joins
<jcoyne>Does "BinaryDescription" mean "a binary with a description"?
<cbeer>jcoyne: no. it means a description of a binary.12:35
jcoyne: here's the LDP class model: https://dvcs.w3.org/hg/ldpwg/raw-file/default/images/ldpc-hierarchy.png
<jcoyne>Ahh, so presently a Datastream object is the entity accessed via fcr:metadata?12:36
<cbeer>jcoyne: LDP gives you an RDF source to describe non-rdf sources.
jcoyne: pretend you never heard the word datastream.
<jcoyne>cbeer: chime in on https://github.com/projecthydra/active_fedora/issues/508 please
<cbeer>jcoyne: but if you're trying to map fcrepo3 concepts (don't.), yes, the fcrepo3 datastream representation is at fcr:metadata12:37
<barmintor>I think it’s going to be much easier to go the other way12:38
<pivotal-bot_____>Longshou Situ started "Adding datastream without file in HTML UI causes ugly error" https://www.pivotaltracker.com/story/show/78847348
<cbeer>barmintor: agreed. i think there's a lot of de-programming going from 3 => 4.
<barmintor>I know how to fake the RDF describing binary content in an API on FCR3
<jcoyne>cbeer: so it would help if you removed "Datastream" from the fcrepo source & docs12:39
<barmintor>Wouldn’t usually think forward-compat easier than backwards-compat, but there it is
<cbeer>jcoyne: where is it in the wiki? i excised a bunch of references recently.12:40
<jcoyne>cbeer: I don't know that it is
cbeer: awead said "I would stick to whatever terms Fedora is using... if it's still using "datastream" we'd probably want to stick with that, but I agree that "child_resource" is a better fit. I just don't want users getting confused."12:41
* scossu joins12:45
<pivotal-bot_____>Chris Beer added "Remove fedora:object/fedora:datastream references from the HTML UI "Create object" panel (and JAX-RS endpoint?)" https://www.pivotaltracker.com/story/show/81601312
<cbeer>^ that's the only reference I know about that makes me sad.12:46
* awead leaves
* awead joins
<jcoyne>cbeer: do you know of any documentation (other than the LDP spec) about using has/isMemberOfRelation12:47
<cbeer>jcoyne: the LDP primer?12:48
https://dvcs.w3.org/hg/ldpwg/raw-file/tip/ldp-primer/ldp-primer.html
* awead leaves12:49
<barmintor>cbeer: Am I right that org.fcrepo.kernel.services.Service is a dead interface? It looks like all the impls are only called from tests.12:50
<cbeer>barmintor: i believe you are right.12:51
<barmintor>well, this is what branches are for
<cbeer>early and often.12:53
* awead joins12:54
<pivotal-bot_____>Kevin Clarke started "Setup and document installing Grinder on AWS" https://www.pivotaltracker.com/story/show/8153572413:41
Jonathan Roby added comment: "Draft doc at https://wiki.duraspace.org/pages/viewpage.action?pageId=64327748" https://www.pivotaltracker.com/story/show/8148636013:44
* gregjansen joins13:49
<jonroby>awoods: i'm having problems commiting to fcrepo-labs/oaiprovider. I'm not that experienced at git and I don't know if its because I cloned the source as-per the instractions? :(13:53
* gregjansen leaves
<awoods>jonroby: can you provide me a link to your github fork?
<jonroby>awoods: sorry, I have to go afk for a while, i'll be back to sort this out. Using the instructions, it should work if you change the line in oai.xml to org.fcrepo.oai.dublincore.JcrPropertiesGenerator13:57
afk
<escowles>cbeer: ping?13:58
<cbeer>escowles: pong
<escowles>cbeer: how do you feel about returning the updated triples for a node after it's updated (so the client can use them instead of doing a get request)13:59
IIRC, the LDP spec says that clients must not expect it, but doesn't say they can't use it if it's there14:00
<cbeer>escowles: PUT, POST, and PATCH?
<escowles>yes, all three
though jcoyne would know which methods were most important14:01
<cbeer>escowles: and this is just to keep active-fedora in sync with the repository state?
<escowles>cbeer: yes, because right now every update is followed by a get to keep in sync
<cbeer>escowles: and creating non-rdf resources would have a different behavior?
<escowles>cbeer: i hadn't thought about non-rdf resources
<jcoyne>I'm only concerned about RdfSources14:02
I wouldn't want to return a binary resource because Fedora isn't adding properties to that resource itself
<escowles>jcoyne: is AF doing mostly PUTs to update RDF resources?14:03
<cbeer>escowles: i'm not thrilled about the idea. i'm not sure that it's a big burden to make clients update their state, especially when we offer PATCH to avoid just that.
(and now that fcrepo4 is doing a better job round-tripping RDF, I hope it's less of a problem)
<jcoyne>cbeer: this would be a huge performance improvement, especially for ingesting & indexing a large batch14:04
<cbeer>jcoyne: and a huge performance hit when you don't care.
<jcoyne>By default Hydra always cares.14:05
I know we're not the only kids on the block though
<cbeer>jcoyne: let me try to understand the hydra use case.14:07
jcoyne: as a user, doing CRUD against the repository, I'm pretty unaffected either way, right? you're almost always going to go to the repository for each request (unless you make the app stateful)14:08
<escowles>cbeer: wait, i don't think that's true
cbeer: i think a typical scenario is taking user input, updating the repo, and then immediately doing a GET to update the object state to show to the user14:09
cbeer: and/or to update solr
<cbeer>escowles: don't we typically redirect the user from their POST request back to the object?14:10
<jcoyne>cbeer: but another common use case is "import and index" a directory of objects14:11
<cbeer>jcoyne: as a large batch?
<escowles>cbeer: that's one pattern, but just rendering the edit view is the other, which would use the updated object in memory
<cbeer>escowles: so, iterative editing?14:12
<escowles>cbeer: right "save & close" vs. "save & keep editing"
<jcoyne>especially considering that resources that point at a node cause the node to change its properties
Hmm, I think I need to look more into the PATCH method14:14
<cbeer>jcoyne: i think PATCH is a good fit for iterative editing14:16
for the batch import and index, i think we'd separate the two operations, so having a response body wouldn't do us any good... plus, we'd parallelize the indexing.14:18
<jcoyne>\14:22
The inverse relationships are killing me!14:26
<cbeer>jcoyne: why?
<jcoyne>If I don't have a resource up to date I end up overwriting them14:27
<jonroby>back from afk
<jcoyne>I'm making the assumption they are incoming pointers and thus immutable on the target object
But that's not the case
<cbeer>jcoyne: hm. i don't think anything bad would happen if we made them opt-in instead of on-by-default.14:29
<jcoyne>cbeer: we'll it's either that or switch to using PATCH
<cbeer>jcoyne: which you should probably do anyway. Presumably you know what values are dirty.14:30
<jcoyne>cbeer: I'm wondering how this all fits with activetriples?14:31
<cbeer>jcoyne: activetriples should support ActiveSupport::Dirty?
<jcoyne>It doesn't presently, but it wouldn't be too hard to add it.
then update just sends the changed attributes?
<cbeer>in a PATCH request.14:32
<pivotal-bot_____>Chris Beer added "Only include inbound references when explicitly included" https://www.pivotaltracker.com/story/show/8161189614:33
Chris Beer started "Only include inbound references when explicitly included" https://www.pivotaltracker.com/story/show/81611896
Chris Beer finished "Only include inbound references when explicitly included" https://www.pivotaltracker.com/story/show/8161189614:34
<barmintor>can I rename org.fcrepo.http.commons.AbstractResource -> AbstractJaxrsResource ?
* github-ff joins
[fcrepo4] cbeer created inbound-opt-in (+1 new commit): http://git.io/gzpD5Q
fcrepo4/inbound-opt-in 7d6cb0f Chris Beer: Make inbound-references an opt-in preference
* github-ff leaves
<pivotal-bot_____>Chris Beer added comment: "https://github.com/fcrepo4/fcrepo4/pull/597" https://www.pivotaltracker.com/story/show/81611896
* github-ff joins14:35
[fcrepo4] cbeer opened pull request #597: Make inbound-references an opt-in preference (master...inbound-opt-in) http://git.io/o42n0g
* github-ff leaves
<awoods>barmintor: That may not be exactly the case... do you have something particular in mind with that rename?14:36
<barmintor>awoods: To be less confused about what’s a JAX-RS Resource and what’s a LDPResource14:37
<cbeer>barmintor: it would be great to rename AbstractResource, FedoraBaseResource and ContentExposingResource.14:38
barmintor: and probably shuffle some methods between them.
<awoods>barmintor/cbeer: if it helps with clarity, go for it.14:39
<cbeer>barmintor: i don't exactly understand why good AbstractResource is doing for us anyway. i think http-api is its only real consumer, and i don't think we're often using what it's injecting anyawy.
<barmintor>cbeer: well, you are onto why I’m scrutinizing it- it autowires all those service classes14:40
<cbeer>barmintor: i do think you want to distinguish between the base class for all our JAX-RS endpoints and a base class for JAX-RS endpoints that are about resources.
(eg BaseResource vs ContentExposingResource)14:41
<barmintor>cbeer: yes, this is what I mean. Only use Resource in the context of RDF analogs
<cbeer>+1
* ajs6f joins14:42
<pivotal-bot_____>Mohamed Mohideen Abdul Rasheed started "Ensure functionality fcrepo-webapp-plus" https://www.pivotaltracker.com/story/show/8143362814:48
<f4jenkins>Project fcrepo4-client build #274: FAILURE in 1 min 25 sec: http://jenkins.fcrepo.org/job/fcrepo4-client/274/
<pivotal-bot_____>Jonathan Roby added comment: "PR for package path change in oai.xml - https://github.com/fcrepo4-labs/fcrepo4-oaiprovider/commit/74f9fbe31�" https://www.pivotaltracker.com/story/show/8148636014:55
* MohamedAR leaves14:58
<pivotal-bot_____>Longshou Situ added comment: "@awoods Once changing the FedoraFileSystemConnector to read-only, the following test in fcrepo-http-api will�" https://www.pivotaltracker.com/story/show/8077522614:59
* MohamedAR joins
* travis-ci joins15:04
fcrepo4/fcrepo4#3021 (inbound-opt-in - 7d6cb0f : Chris Beer): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/commit/7d6cb0fec2c0
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/39301583
* travis-ci leaves
<jcoyne>cbeer: https://github.com/ruby-rdf/sparql-client looks pretty meager. Are you just crafting all your SPARQL by hand?15:09
<pivotal-bot_____>Esme Cowles added "Restoring a version of non-RDF resources results in error "This operation requires that the node be versionable"" https://www.pivotaltracker.com/story/show/8161520215:10
Esme Cowles started "Restoring a version of non-RDF resources results in error "This operation requires that the node be versionable"" https://www.pivotaltracker.com/story/show/81615202
* dwilcox joins15:15
* scossu leaves15:22
<cbeer>jcoyne: yep: https://github.com/cbeer/activeresource-ldp/blob/772aebba524911f6fdd094c685cdc90112ad682a/lib/active_resource/ldp/base.rb#L12315:25
* scossu joins15:31
<barmintor>Bleeargh, all the *Service impls are in use via injection15:45
<awoods>barmintor: is that a bad thing, or just frustrating because Eclipse sent you down a rabbit hole?
<barmintor>awoods: B)15:46
err
#2
<jcoyne>cbeer: that doesn't support multiple values does it?
<cbeer>jcoyne: what do you mean?15:47
<jcoyne>INSERT {
<http://localhost:8983/fedora/rest/test/39/85/1a/3a/39851a3a-5ac9-4429-8061-560b948faa4f> <http://fedora.info/definitions/v4/rels-ext#hasConstituent> "#<RDF::Queryable::Enumerator:0x007fc72ec05b00>" .
}
WHERE { }
F
That's not quite right.
<cbeer>jcoyne: yes, you'd need to enumerate that.15:48
* dwilcox leaves15:49
<pivotal-bot_____>Andrew Woods delivered "Support the full range of LDP containers" https://www.pivotaltracker.com/story/show/8141141615:52
* 17SAAQSXW joins
[fcrepo4] awoods deleted ldp-containers at 92733d1: http://git.io/MBlE2g
* 17SAAQSXW leaves
* 1JTAATR4I joins
[fcrepo4] awoods closed pull request #594: Support all LDP container types (master...ldp-containers) http://git.io/Fwj9SQ
* 1JTAATR4I leaves
<cbeer>ajs6f: ping?15:58
<ajs6f>pung
<barmintor>Ah, OutOfMemoryError, my old friend.
<cbeer>ajs6f: i'm thinking about LDP paging (https://dvcs.w3.org/hg/ldpwg/raw-file/default/ldp-paging.html), and really just max-member-count.15:59
ajs6f: and how to wedge it into our triple generation code: https://github.com/fcrepo4/fcrepo4/blob/master/fcrepo-http-api/src/main/java/org/fcrepo/http/api/ContentExposingResource.java#L179
<ajs6f>Do we want to spend the time if its "at-risk"?
<cbeer>ajs6f: i think we will want pagination.16:00
<ajs6f>cbeer: Sure, but do we want to do that particular impl at this particular time?
<barmintor>ajs6f, cbeer: the conversation around it has not sounded very at-risk16:01
<cbeer>ajs6f: no time like the present. max-triple-count isn't at-risk, and would probably be the same implementation.
<jcoyne>cbeer: patterns = changes.map do |k, (was,is)|
RDF::Query::Pattern.new(subject, schema[k]['predicate'], k.to_sym).to_s
end
is that k.to_sym in the object position correct?
<ajs6f>cbeer/barmintor: Okay, fari engouh.16:02
<jcoyne>https://github.com/cbeer/activeresource-ldp/blob/772aebba524911f6fdd094c685cdc90112ad682a/lib/active_resource/ldp/base.rb#L127
<cbeer>ajs6f: obviously the naive thing to do is just wrap our RdfStream in something that skips N (triples/subjects) and then reads M (triples/subjects).
ajs6f: but that sounds.. naive.
<ajs6f>cbeer: And really expensive.
<barmintor>cbeer: although it gets to my point that the whole concept of paging implies some order AAAAAARRRRRGGGGGGH
<ajs6f>cbeer: This is where the fat that the JVM is not a stack-based machine bites us.
<cbeer>ajs6f: the only thing i've thought of, yet, is an object we pass into the *RdfContexts that tracks some kind of pagination state.
<ajs6f>cbeer: Yes. What we're talking about are a limited form of continuations.16:03
* dwilcox joins
<ajs6f>cbeer: We have something on which to build— the stack of iterators and transforms that produce triples.
<barmintor>ajs6f: “fari engouh” is a Hungarian spec engineer
<ajs6f>barmintor: Fedo's brother-in-law?16:04
<barmintor>I think so
<cbeer>ajs6f: what's the not-so-bad naive implementation look like?
<ajs6f>cbeer: That stack of iterators. Holding it around in case you get a request for the next page.
<cbeer>barmintor: i think Modeshape is giving us some sort of order. it's not promising what, but i think it's consistent.
ajs6f: implying holding open a JCR session, and having somewhere to stash it?16:05
jcoyne: i don't know offhand.
<barmintor>cbeer: yeah, I meant that it’s an internal contradiction on the LDP-Paging convo
<ajs6f>cbeer: Either in the session, or perhaps better, in a common cache.
cbeer: And if you don't ask for the next page within a certain time, we have to rebuild the stack and start from the beginning to get where you want to be.16:06
cbeer: But if you do, we can find your iterator stack in the cache and keep peeling you off triples.
<barmintor>I think to be complient w/ spec the paging has to be reproducible by URI unless the underlying resource changes16:07
<ajs6f>barmintor: That seems reasonabe.
<f4jenkins>Yippee, build fixed!16:08
Project fcrepo4-client build #275: FIXED in 2 min 28 sec: http://jenkins.fcrepo.org/job/fcrepo4-client/275/
<ajs6f>afk for a smoke, bbs
* ajs6f leaves
* osmandin leaves16:11
* travis-ci joins
fcrepo4/fcrepo4#3022 (master - a3bb336 : Andrew Woods): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/compare/9b27b817b54c...a3bb33680350
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/39309070
* travis-ci leaves
<pivotal-bot_____>Andrew Woods delivered "Only include inbound references when explicitly included" https://www.pivotaltracker.com/story/show/8161189616:12
* github-ff joins
[fcrepo4] awoods closed pull request #597: Make inbound-references an opt-in preference (master...inbound-opt-in) http://git.io/o42n0g
* github-ff leaves
<cbeer>jcoyne: ^16:16
* ajs6f1 joins
<barmintor>FedoraNodes and FedoraLDP have the same path spec16:17
<cbeer>barmintor: yep.
<ajs6f1>barmintor: That's called "belt-and-suspenders".
<barmintor>… is this a trap?
<cbeer>barmintor: and i think ajs6f1 has plans to introduce a 3rd one
<ajs6f1>Same spec, _all the time_.16:18
<barmintor>I don get it
<cbeer>barmintor: FedoraLdp has the LDP bits. FedoraNodes has the non-LDP bits.
barmintor: and JAX-RS 2 is smart enough to deal with that.
<barmintor>Oh, geesh
ok
so it’s just the method endpoints that are different16:19
<cbeer>yep
* ajs6f joins
(In fact, you're in charge of that effort.)
<pivotal-bot_____>Chris Beer added "getResourceTriples should avoid switching subjects too much." https://www.pivotaltracker.com/story/show/8162091616:20
Chris Beer started "getResourceTriples should avoid switching subjects too much." https://www.pivotaltracker.com/story/show/81620916
* ajs6f1 leaves16:22
<ajs6f>barmintor: I think I got cut off by crappy wifi. What I meant to say (about FedoraLdp and FedoraNodes) is that we plan to break up the API anyway, so there will be a refactoring across not just types, but modules. (In fact, you're in charge of that effort.)16:25
* travis-ci joins16:29
fcrepo4/fcrepo4#3023 (master - feb74ec : Andrew Woods): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/compare/a3bb33680350...feb74ecc9583
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/39310773
* travis-ci leaves
<ajs6f>cbeer: Did what I said about paging make any sense? Are you going to try that, or instead try passing the "state of paging" down through the iterator chain?16:30
<cbeer>ajs6f: dunno.
to all counts.
<ajs6f>cbeer: Right on, brother.
<barmintor>these services are all the same16:31
<cbeer>barmintor: all the NodeServices?16:32
<barmintor>yeah
they’re all in use, but they’re all the same thing with tiny differences
* ajs6f leaves
* ajs6f joins
Well, nodes have a lot in common.
<cbeer>barmintor: yep. i'm really hoping someone comes along and makes a ResourceService<T> class.
<barmintor>someone is doing that right now16:33
<cbeer>so we can actually have the same signatures for the same things.
barmintor++
<ajs6f>cbeer: Didn't you do something like it in the move to factor out Fedora-specific types? I'm thinking it might be ResourceService<T extends ACBeerType>.16:34
<cbeer>ajs6f: i think that was in a branch that got left behind.
<ajs6f>cbeer: Oh, well, what the hell.
<cbeer>ajs6f: but, still, T extends FedoraResource isn't too bad.
<ajs6f>cbeer: Yes, that's what I meant.16:35
The meaning of FedoraResource may change, and that's fine.
barmintor: Are you ready to report on finalizing the efforts for API vetting tomorrow?
<cbeer>ajs6f: yep. someone just has to come along and rename it.
<ajs6f>cbeer: I think what barmintor is doing is just plan good.
plain16:36
<cbeer>ajs6f: shame about all the merge conflicts he's going to have.
<ajs6f>sh...
<barmintor>ajs6f: Umm…. no. I’m ready to report on the Columbia site that went live yesterday, and the bugs that kept me at work until 8pm.
<ajs6f>barmintor: Sounds good. I'm looking forward to it.
<barmintor>Wait a minute, I’m pretty nervous about refactoring fcrepo.kernel.services.Service16:38
<ajs6f>Because it's underneath so much?16:39
<barmintor>and so complex
<ajs6f>Sounds like a good reason to refactor it.
But maybe not right before release.
<barmintor>(Service is an empty interface)16:40
<ajs6f>What are its semantics? What does it mark?
<barmintor>That a thing is either a NodeService, and ObjectService,a RepositoryService, a TransactionService, or a VersionService16:41
<ajs6f>Sounds like we might have been reaching for an abstract class with helpers in it.
<barmintor>AbstractService does, in fact, implement Service16:42
it is also the base class of all those other classes
<ajs6f>Better leave it in for a couple of years to make sure that we don't leave anyone behind.
<barmintor>ajs6f++ // brb adding some marker interfaces
<ajs6f>Mark it deprecated so that years from now, when barmintor is closing down the 4-series, he will have something to curse about.16:43
<barmintor>that seems about right
<cbeer>ajs6f: we used to have an abstract class with helpers in it.. but never re-used those helpers in more than one service.
<ajs6f>cbeer: That's the logical dual of DRY.
* MohamedAR leaves16:44
<cbeer>ajs6f: huh. i never noticed this before now, but when we RdfStream#concat, we place the concatenated content before the existing content.16:45
<ajs6f>cbeer: We do that, or Guava does that? It wouldn't surprise me if we absentmindedly reversed them, anyway.16:46
* dwilcox leaves
<cbeer>ajs6f: we do that. i don't think there's any reason for it though.
guess we'll find out what tests break
<ajs6f>cbeer: No. the whole thing is kind of a half-assed Lisp-style list, anyway.16:47
<cbeer>ajs6f: can you think of a good way to test that some RDF is expressed as tersely as possible?16:48
<ajs6f>cbeer: That is an _extremely_ hard problem.16:49
<cbeer>ajs6f: ok. so i shouldn't feel bad solving it by counting how many lines start with non-whitespace?
in turtle.
* jcoyne demands one letter prefixes!
<ajs6f>cbeer: Are you trying to deal with two graphs that you _know_ are the "same"?
cbeer: And meansure which is "bigger"?16:50
cbeer: Or one graph and you're trying to "compress" it?
<cbeer>ajs6f: i have one graph. and i want to make sure it's sufficiently terse.
<ajs6f>cbeer: And "sufficient" means...?
<cbeer>ajs6f: no worse than how it's going to be after this refactor.16:51
<barmintor>ExternalContentService is not a Service
<cbeer>ajs6f: i guess here are my metrics:
1) doesn't switch back and forth between subjects more than necessary16:52
2) doesn't contain duplicate triples
3) assertions with the same subject and predicate should be grouped together
<ajs6f>cbeer: Keep in mind that without some ontological background (which allows us to make judgments about whether two graphs "mean the same thing"), we can do nothing but measure either entailment or syntatic equivalence.
2) That you can do pretty easily, if you're willing to cache triples.
3) That's not really "terseness" at all, is it?
<cbeer>4) if an rdf:type assertion is made, it should come first.
<ajs6f>1) How much is "necessary"?16:53
cbeer: This sounds more and more like order, not "terseness".
<cbeer>ajs6f: no, you're right, it's not really terseness. terseness is probably just a way of measuring aesthetics.
<ajs6f>cbeer: Okay, so we want to order triples?
(WITHOUT making guarantees about them….)
<cbeer>ajs6f: i think we want some more order than we have now (which is pure chaos).16:54
<ajs6f>cbeer: This is to give "nice" human-readable outputs at endpoints
?
Right, okay.
The big problem here will be that you're going to have to sacrifice some laziness.
<cbeer>ajs6f: and i'm making it look right by hand.. and want to leave a test to make sure it only gets better.
ajs6f: maybe my bar is lower than yours... and maybe that's where "necessary" and "sufficient" come into play.16:55
<ajs6f>Well, for testing purposes only, how about pulling them into an in-memory representation, serializing into some form that has the syntactic properties you want, and comparing that/
?
I mean, RDF/XML will group things by subject.
(Or anyway, every serliazer I ever saw does.)
Anyway, de-duping alone will require caching every triple but the very last one, right?16:56
Might as well go the whole hog and take the triples into heap all at once for ease-of-use.16:57
cbeer: on a related note, if we're going to have a "house style" on how we offer triples, we should document that, I think.
cbeer: Tests are necessary, but some note about what we expect would be good (motivation for passing the tests).16:58
<pivotal-bot_____>Andrew Woods added comment: "@longshous, yes @Ignore them with a comment indicating that they should be enabled once the FedoraFileSystemC�" https://www.pivotaltracker.com/story/show/8077522617:00
* github-ff joins
[fcrepo4] cbeer created better-output (+2 new commits): http://git.io/Fpvogg
fcrepo4/better-output 173ecad Chris Beer: Re-organize the generated triples to optimize changing subjects
fcrepo4/better-output e9c4d7d Chris Beer: Refactor triple generation to produce aesthetically pleasing graphs
* github-ff leaves
<pivotal-bot_____>Chris Beer finished "getResourceTriples should avoid switching subjects too much." https://www.pivotaltracker.com/story/show/81620916
* github-ff joins17:01
[fcrepo4] cbeer force-pushed better-output from e9c4d7d to 87c7479: http://git.io/ovU8Cw
fcrepo4/better-output 87c7479 Chris Beer: Refactor triple generation to produce aesthetically pleasing graphs
* github-ff leaves
<pivotal-bot_____>Chris Beer edited "getResourceTriples should avoid switching subjects too much." https://www.pivotaltracker.com/story/show/81620916
<cbeer>ajs6f: maybe, yeah, we just need a style guide... maybe someone already has one for RDF (or just turtle), even.17:03
<ajs6f>cbeer: I'm sure many people do. Well, that's why I suggest using a prebuilt serializer— they have that kind of assumption built-in.17:04
<cbeer>ajs6f: RIO seems to do a pretty good job, if it wasn't hobbled by our streams.
* dwilcox joins17:05
<ajs6f>cbeer: But for a test, would it be so bad to roll up the stream?
<cbeer>ajs6f: probably
<ajs6f>cbeer: Really? The streams I've seen going to through tests are a couple of dozen triples at most, maybe a hundred...17:06
* github-ff joins
[fcrepo4] cbeer force-pushed better-output from 87c7479 to 420a80a: http://git.io/ovU8Cw
fcrepo4/better-output 420a80a Chris Beer: Refactor triple generation to produce aesthetically pleasing graphs
* github-ff leaves
* travis-ci joins
fcrepo4/fcrepo4#3024 (better-output - e9c4d7d : Chris Beer): The build has errored.
Change view : https://github.com/fcrepo4/fcrepo4/compare/173ecadaac73^...e9c4d7df1af1
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/39315044
* travis-ci leaves
* travis-ci joins17:07
fcrepo4/fcrepo4#3025 (better-output - 87c7479 : Chris Beer): The build has errored.
Change view : https://github.com/fcrepo4/fcrepo4/compare/e9c4d7df1af1...87c7479f43af
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/39315077
* travis-ci leaves
<ajs6f>out for the day17:09
* ajs6f leaves
* dwilcox leaves17:10
<f4jenkins>Project fcrepo-module-auth-rbacl build #319: UNSTABLE in 4 min 26 sec: http://jenkins.fcrepo.org/job/fcrepo-module-auth-rbacl/319/17:17
Project fcrepo-message-consumer build #826: UNSTABLE in 5 min 25 sec: http://jenkins.fcrepo.org/job/fcrepo-message-consumer/826/17:18
* acoburn leaves17:19
* travis-ci joins17:23
fcrepo4/fcrepo4#3027 (better-output - 420a80a : Chris Beer): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/compare/87c7479f43af...420a80a97451
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/39315526
* travis-ci leaves
<pivotal-bot_____>Andrew Woods added comment: "The above commit is very close to being a Pull-Request. You should be able to navigate to: ""17:27
https://github.com�" https://www.pivotaltracker.com/story/show/81486360
<f4jenkins>Yippee, build fixed!17:30
Project fcrepo-module-auth-rbacl build #320: FIXED in 3 min 46 sec: http://jenkins.fcrepo.org/job/fcrepo-module-auth-rbacl/320/
* awead leaves17:31
<f4jenkins>Yippee, build fixed!17:32
Project fcrepo-message-consumer build #827: FIXED in 5 min 34 sec: http://jenkins.fcrepo.org/job/fcrepo-message-consumer/827/
<pivotal-bot_____>Andrew Woods added comment: "@longshous, nicely done. Pending one minor code review comment: ""
Also, I assume the wiki documentation needs t�" https://www.pivotaltracker.com/story/show/79143930
Andrew Woods rejected "The sparql recipes no longer pass for 4.0.0-beta-04-SNAPSHOT" https://www.pivotaltracker.com/story/show/79143930
Jonathan Roby added comment: "there doesn't seem to be any changes needed for the READ.md. Apart from the package path, it worked out of t�" https://www.pivotaltracker.com/story/show/8148636017:39
* dwilcox joins17:43
* dwilcox leaves17:48
* jcoyne leaves17:50
<pivotal-bot_____>Longshou Situ started "The sparql recipes no longer pass for 4.0.0-beta-04-SNAPSHOT" https://www.pivotaltracker.com/story/show/7914393017:55
* jonroby leaves17:59
* dwilcox joins18:02
* jcoyne joins18:06
* jcoyne leaves18:10
* github-ff joins18:13
[fcrepo4] barmintor created 80959252 (+3 new commits): http://git.io/7w2y6A
fcrepo4/80959252 bb8f7a7 Benjamin Armintor: refactor getClusterExecutor into Function subclass
fcrepo4/80959252 4400630 Benjamin Armintor: push rarely-used repositoryService injection down inheritance tree
fcrepo4/80959252 f357c54 Benjamin Armintor: refactor Service into a genric interface
* github-ff leaves
* barmintor leaves
* jcoyne joins18:18
<pivotal-bot_____>Longshou Situ added comment: "PR https://github.com/fcrepo4-labs/fcrepo4-release-tests/pull/2 again, commit https://github.com/lsitu/fcrep�" https://www.pivotaltracker.com/story/show/7914393018:21
Longshou Situ finished "The sparql recipes no longer pass for 4.0.0-beta-04-SNAPSHOT" https://www.pivotaltracker.com/story/show/7914393018:22
* ksclarke leaves18:27
* travis-ci joins18:35
fcrepo4/fcrepo4#3029 (80959252 - f357c54 : Benjamin Armintor): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/compare/bb8f7a724bab^...f357c548db27
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/39322749
* travis-ci leaves
<pivotal-bot_____>Chris Beer estimated "Remove fedora:object/fedora:datastream references from the HTML UI "Create object" panel (and JAX-RS endpoint?)" as �" https://www.pivotaltracker.com/story/show/8160131218:38
Chris Beer started "Remove fedora:object/fedora:datastream references from the HTML UI "Create object" panel (and JAX-RS endpoint?)" https://www.pivotaltracker.com/story/show/81601312
* github-ff joins
[fcrepo4] cbeer created remove-html-datastream (+1 new commit): http://git.io/uq92bw
fcrepo4/remove-html-datastream 7880667 Chris Beer: Remove datastream references from HTML UI controls
* github-ff leaves
<pivotal-bot_____>Chris Beer added comment: "https://github.com/fcrepo4/fcrepo4/pull/600" https://www.pivotaltracker.com/story/show/81601312
Chris Beer edited "Remove fedora:object/fedora:datastream references from the HTML UI "Create object" panel (and JAX-RS endpoint?)" https://www.pivotaltracker.com/story/show/81601312
Chris Beer finished "Remove fedora:object/fedora:datastream references from the HTML UI "Create object" panel (and JAX-RS endpoint?)" https://www.pivotaltracker.com/story/show/81601312
* github-ff joins
[fcrepo4] cbeer opened pull request #600: Remove datastream references from HTML UI controls (master...remove-html-datastream) http://git.io/FunutQ
* github-ff leaves
* scossu leaves18:41
* jcoyne leaves18:47
<f4jenkins>Project fcrepo4-client build #280: FAILURE in 1 min 34 sec: http://jenkins.fcrepo.org/job/fcrepo4-client/280/18:52
* travis-ci joins18:53
fcrepo4/fcrepo4#3030 (remove-html-datastream - 7880667 : Chris Beer): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/commit/788066741f27
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/39325146
* travis-ci leaves
<f4jenkins>Project fcrepo-module-auth-rbacl build #322: UNSTABLE in 3 min 17 sec: http://jenkins.fcrepo.org/job/fcrepo-module-auth-rbacl/322/
* dwilcox leaves19:23
* jcoyne joins19:30
<pivotal-bot_____>Andrew Woods added comment: "Based on your first comment in this ticket, I thought you wanted to change the README.md from /oai to /rest/o�" https://www.pivotaltracker.com/story/show/81486360
* github-ff joins19:46
[fcrepo4] escowles created non-rdf-restore (+1 new commit): http://git.io/L_iKjg
fcrepo4/non-rdf-restore 1718401 Esmé Cowles: Fixing versioning check when reverting versions of non-RDF resources
* github-ff leaves
<pivotal-bot_____>Esme Cowles added comment: "https://github.com/fcrepo4/fcrepo4/pull/601" https://www.pivotaltracker.com/story/show/8161520219:47
Esme Cowles finished "Restoring a version of non-RDF resources results in error "This operation requires that the node be versionable"" https://www.pivotaltracker.com/story/show/81615202
* github-ff joins
[fcrepo4] escowles opened pull request #601: Fixing versioning check when reverting versions of non-RDF resources (master...non-rdf-restore) http://git.io/0LnLrA
* github-ff leaves
* dwilcox joins19:49
* jcoyne leaves19:50
* dwilcox leaves19:56
* travis-ci joins
fcrepo4/fcrepo4#3032 (non-rdf-restore - 1718401 : Esmé Cowles): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/commit/171840199dc4
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/39331641
* travis-ci leaves
* jcoyne joins19:59
<f4jenkins>Yippee, build fixed!20:02
Project fcrepo-module-auth-rbacl build #323: FIXED in 4 min 35 sec: http://jenkins.fcrepo.org/job/fcrepo-module-auth-rbacl/323/
* github-ff joins20:03
[fcrepo4] escowles pushed 1 new commit to non-rdf-restore: http://git.io/WR895Q
fcrepo4/non-rdf-restore f49c3fd Esmé Cowles: Using FedoraBinary.getDescription() instead of node-munging
* github-ff leaves
<pivotal-bot_____>Andrew Woods delivered "The sparql recipes no longer pass for 4.0.0-beta-04-SNAPSHOT" https://www.pivotaltracker.com/story/show/7914393020:10
* github-ff joins20:13
[fcrepo4] cbeer deleted merge-for-ajs6f at ba20064: http://git.io/qQ5AkQ
* github-ff leaves
* github-ff joins
[fcrepo4] cbeer deleted lang-tag-literals at aa20cf1: http://git.io/JxIB4g
* github-ff leaves
* travis-ci joins20:14
fcrepo4/fcrepo4#3034 (non-rdf-restore - f49c3fd : Esmé Cowles): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/compare/171840199dc4...f49c3fdefe7f
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/39332900
* travis-ci leaves
<f4jenkins>Project fcrepo4-release-tests build #1: UNSTABLE in 2 min 4 sec: http://jenkins.fcrepo.org/job/fcrepo4-release-tests/1/20:16
Project fcrepo4 build #2509: FAILURE in 2 min 48 sec: http://jenkins.fcrepo.org/job/fcrepo4/2509/20:18
* jcoyne leaves20:19
<f4jenkins>Project fcrepo-module-auth-rbacl build #324: UNSTABLE in 4 min 27 sec: http://jenkins.fcrepo.org/job/fcrepo-module-auth-rbacl/324/20:20
* jcoyne joins20:28
<pivotal-bot_____>Andrew Woods added "Enable random ports on fcrepo4-release-tests" https://www.pivotaltracker.com/story/show/8164041020:29
Andrew Woods edited "Enable random ports on fcrepo4-release-tests" https://www.pivotaltracker.com/story/show/81640410
* dwilcox joins20:36
* github-ff joins20:43
[fcrepo4] lsitu opened pull request #602: Enforced the FedoraFileSystemConnector to be read-only. (master...feature/readonly-connector) http://git.io/KgQehw
* github-ff leaves
<pivotal-bot_____>Longshou Situ added comment: "https://github.com/fcrepo4/fcrepo4/pull/602" https://www.pivotaltracker.com/story/show/80775226
Longshou Situ finished "Make FedoraFileSystemConnector Read-Only" https://www.pivotaltracker.com/story/show/8077522620:44
Longshou Situ started "Enable random ports on fcrepo4-release-tests" https://www.pivotaltracker.com/story/show/81640410
* longshou leaves20:45
<pivotal-bot_____>Andrew Woods finished "Velocity templates should be called for ascenedant primary types" https://www.pivotaltracker.com/story/show/7920780420:51
* awead joins20:58
* dwilcox leaves20:59
* awead leaves21:05
* awead joins21:14
* ksclarke joins21:19
<f4jenkins>Project fcrepo-message-consumer build #831: UNSTABLE in 1 hr 5 min: http://jenkins.fcrepo.org/job/fcrepo-message-consumer/831/21:21
* awead leaves21:22
* dwilcox joins21:33
* jcoyne leaves21:37
* dwilcox leaves21:41
* jcoyne joins22:21
* awead joins22:25
* dwilcox joins22:58
* awead leaves23:07
* awead joins23:08
* dwilcox leaves23:11
* dwilcox joins23:13
* dwilcox leaves23:14
* awead leaves23:20
* jcoyne leaves23:32
* dwilcox joins00:25

Generated by Sualtam