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

Using timezone: Eastern Standard Time
* ksclarke leaves02:16
* kaarefc joins03:23
* fasseg joins06:48
* escowles leaves06:51
* escowles joins06:54
* escowles leaves07:04
* escowles joins07:17
* escowles leaves07:19
* escowles joins07:28
* ajs6f joins08:35
awoods: No standup today, right?08:36
* ajs6f leaves09:13
* gregjansen joins09:14
<awoods>ajs6f: correct
* ajs6f joins09:17
<awoods>ajs6f ^^
<ajs6f>thx09:18
* ksclarke joins09:33
* escowles leaves09:48
* github-ff joins09:56
[fcrepo4] ajs6f pushed 1 new commit to KillWarnings: http://git.io/64zlWA
fcrepo4/KillWarnings 8673ded ajs6f: Simpler config for IT HTTP client
* github-ff leaves
<pivotal-bot>Andrew Woods accepted "Javadoc Package Documentation" https://www.pivotaltracker.com/story/show/5495224409:58
Andrew Woods edited "Review/Modify Async Storage API" https://www.pivotaltracker.com/story/show/59855470
Andrew Woods edited "Define needs of FileSystem Federation" https://www.pivotaltracker.com/story/show/60276468
* travis-ci joins10:09
[travis-ci] futures/fcrepo4#1189 (KillWarnings - 8673ded : ajs6f): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/53762ba6e962...8673dedd9c2b
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/14025958
* travis-ci leaves
<bljenkins>Project fcrepo-fixity-corrupter build #435: SUCCESS in 2 min 27 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/435/10:27
Project fcrepo-kitchen-sink build #636: STILL FAILING in 7 min 42 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/636/10:34
* levar1 joins10:57
* nbanks joins11:00
<barmintor>Transaction question: Should GETs in a TX refresh its expiry?11:03
(related: this TX expiration IT is going to be finicky for reasons unrelated to Jersey2)11:04
<ajs6f>barmintor: I vote no. GET shouldn't alter state.11:09
Maybe we need a method for that, tho'?
<barmintor>ajs6f: I dunno. I sympathetic to that stance, but then POST to a resource in a TX is actually updating 2 things, which seems… shifty11:10
<ajs6f>barmintor: agreed. This is out of left field, but how about pATCH?11:11
<barmintor>PATCH?
Let me be clear: I'm not looking for a way to refresh
<ajs6f>Yeah. You're changing an aspect of the resource. (It's expry.)
<barmintor>(that's a sep issue)
I'm trying to figure out how best to test the expiry11:12
<ajs6f>You mean how to get it without changing it?
<barmintor>well, you're not getting *it*, you're getting something else in its context
<ajs6f>Wait, just read your q again, you said GETs -in- a tx, which i read as -of- a tx.
Now I get it.11:13
get it
GET it
See what I did there?
<barmintor>yes. yes, I see it.
<ajs6f>{sigh}
So this is like, does a non-mutating query keep the transaction alive, or does it push the expry forward?11:14
Or both?
Wait, I guess if you do two, you do one.
<barmintor>correct!
<ajs6f>Hm. I'm trying to think of what people tend to find with RDBMSes. Not that we have to do that, but "least surpirse" and all.
<barmintor>but if it does, we have to make the expiry test just sit around for a long time.
OTOH, if it *does*, then we can have session factory take care of the expiry refresh11:15
<ajs6f>Based on the request method?
<barmintor>Based on you asking for a session in a TX at all
<ajs6f>For the test, can't we adjust the timout, run the test, then put he timeout back?11:16
<barmintor>Well, it's hard
<ajs6f>k
You do much better with that async test stuff than I .
<barmintor>You adjust the timeout to be short, and the TX is expired before you see it's there
<ajs6f>Hm. Yeah, absolute times in test just stink.
<barmintor>you make it too long, and you have a 10sec sleep in your IT
I was going to make it do GETs until it failed, but that seems like a petty constraint to put on TX for the sake of one IT11:17
<ajs6f>What if we had a different GC for txes. Like a pool-ish approach...11:18
I wonder if you can get the time of last actiivty from a MODE session.
<barmintor>SoftRefs? Maybe. Would make it more difficult to handle the rollback for expired TXs.11:19
<ajs6f>Urg. True. We could be clever at expiry time and instead of just desotrying the session, put it into a box of death in which rollback would occur but in which it could not be used for new actions.11:20
barmintor: I vote that for the moment, GET _doesn't_ refresh tx.11:22
<barmintor>Ha. I was about to type that I was moving the refresh to SessionFactory, and just sleeping for 10 ro 12 seconds
<ajs6f>Meaning any method _does_ refresh tx?
<barmintor>yeah
I think it's hard to justify some activity not refreshing the tx session11:23
<ajs6f>Ok. I have no problem with that, really. The ten second thing is kind of lame, tho.
I don't think of txes as being containers of activity. I think of them as being containers of mutations.
<barmintor>well, we live in a world replete with lameness, and our lot is to struggle against that lame tide until the dying of the test suite.
<ajs6f>"when the last dingdong of doom has clanged and faded from the last worthless rock hanging tideless in the last red and dying evening"...11:24
<barmintor>is that Jack Vance?
<ajs6f>Close: William Faulkner, from his Nobel acceptance speech.11:25
<barmintor>Oh, ha.
<pivotal-bot>A. "Pholeotic" Soroka accepted "Killing compiler warnings" https://www.pivotaltracker.com/story/show/6081025011:27
A. "Pholeotic" Soroka finished "Killing compiler warnings" https://www.pivotaltracker.com/story/show/60810250
<ajs6f>awoods: ping?11:29
afk bb i 511:32
* ajs6f1 joins11:37
* ajs6f leaves11:38
<barmintor>ajs6f1: what is the subject of the graph returned from nodeTypes supposed to be?
* ajs6f joins11:39
<barmintor>ajs6f: what is the subject of the graph returned from nodeTypes supposed to be?11:40
<ajs6f>eh? what is nodeTypes?
You mean the HTTP methods that cbeer added?
<barmintor>FedoraRepositoryNodeTypes is an http resource that describes the nodeTypes in the repo11:41
<ajs6f>Right, thats what I meant. cbeer did that (cbeer++)
I don't know, because i haven't looked at what predicates he used.
Presumably… the repo?
* ajs6f1 leaves
<ajs6f>Lemme go look at that.
barmintor: no. what he did was use the actual (namespaced) name of the nodeType as an URI and make that the subject of descriptive triples. IOW, there could be a dozen "subjects" of that graph.11:44
(IOW, it's not likely at all to be a connected graph.)
<barmintor>Ah
Complicated
RdfProvider is displeased
<ajs6f>A little. But we could connect it by adding a couple of repoSubject hasNodeType nodeType triples.11:45
RdfProvider needs to die and be replaced by a streaming equiv. There's a ticket… {sigh}
Do we even have a repoSubject?
<barmintor>fix, then replace11:46
<ajs6f>What is a Fedora repository… _about_...
<barmintor>tests must pass
<ajs6f>Then do that ticket. {grin}
<barmintor>THIS… TEST… SHALL… PASS!
* barmintor gandalfs
<ajs6f>'¡No pasarán!'
*Spanish anarchists*11:47
Oh, wait, no. That was a Communist army. My bad.
That was a whole confusing part of history.
<barmintor>b/c they didn't have git yet11:48
Ah ha. The problem was debug logging.11:53
<ajs6f>What problem?
<barmintor>this "what is the subject" problem
onto the next problem!
<ajs6f>Debug logging was calling a method on an object to log?
<barmintor>sort of11:54
<ajs6f>That's such a bad thing to do, it must have been me.
<barmintor>nope, me
nice try though
<ajs6f>Good save.
* github-ff joins11:59
[fcrepo-aws-puppet] fasseg created slim-setup (+1 new commit): http://git.io/jcdGyg
fcrepo-aws-puppet/slim-setup 4add23d fasseg: removed yourkit config, and updated puppet scripts to use Oracle's JVM
* github-ff leaves
<ajs6f>barmintor: http://stackoverflow.com/questions/3972489/including-git-repo-contents-as-subdir-with-maven
we could do submodules but automate the checkout12:00
* nbanks leaves
<barmintor>http://somethingsinistral.net/blog/git-submodules-are-probably-not-the-answer/12:03
<ajs6f>Urg. That won't load for me, for some reason. But my other reading does point to the idea that submodules are meant for incorporating thrid-party code when you don't have an artifact-based transitive management tool like Maven.12:07
<barmintor>If I seem difficult on this score, it only because my limited experience with submodules in Hydra was, frankly, horrible
<ajs6f>No, I appreciate that submodules may be a bad idea.
But I still feel just as strongly that fcrepo4 is too big, and things like rss and webhooks don't belong there.
Ideally, it's just a parent.
<barmintor>I think we all want a way to pull in optional dependencies that is different than what we've got12:08
<ajs6f>Right.
I dunno… like a modularization framework that is independent of our build framework? {He mutters to himself in the dark closet in which they've locked him to avoid hearing his insane whispers.}12:09
<barmintor>You mean something like FELIX, I assume12:10
<ajs6f>Sure, OSGi could help (with a considerable investment). Or something else.
I'm not doctrinaire about the tool, just that we need one.
<barmintor>PATCHES WELCOME12:16
<ajs6f>No. I don't buy it. PW is a fine response to "I want this feature!" but not to "Our plumbing is leaking! (Concerns.)" This is _our_ deal.
<barmintor>Sure, but I have no ideas on how to do it.12:17
IDEAS WELCOME
<ajs6f>Well, I have ideas, but I can't get any traction, because (very reasonably) our tech leads point out that it's hard to sell modularization as an investment.
eddies and awoods and I have gone around this block many times. We came up with narratives...12:18
like "People will be able to contribute much more easily!"
or "Integrators can assemble the functionality they want without using dev tools!"
<barmintor>there's also a considerable barrier to development that some up when you move from J2EE -> JAXRS -> OSGi
<ajs6f>or "COnfiguration will be sane!"
Yes. True.
Although I think Blueprint (the OSGi flavor of Spring XML) helps immensely.12:19
Almost all of our current wiring would immediately work.
<barmintor>in my experience, "People will ba able to contribute more easily" is just not true. Getting Felix runnig and understanding the contracts beneath it is hard.
<ajs6f>That was the whole point of Blueprint and tools like Karaf— protecting people from having to understand the contracts.12:20
And in my (not huge) experience, they work.
<barmintor>To be fair, i hold myself up as a litmus test of whether it is too hard or not
<ajs6f>All the dynamicity is hidden.
<barmintor>which may not be a fair reflection of the approach
<ajs6f>You are a terrible example. Very few people come in with your experience or acumen.
nbanks might be a better example.12:21
If we try this, and people like nbanks are asking questions about OSGi concepts, we did it wrong.
But we can do it right.
I really just want what eddies wanted— drop an artifact into a container and it works. But I want it for extension stuff _and_ the core itself.12:23
Also I want a pony.
<barmintor>http://www.nps.gov/caha/supportyourpark/adoptapony.htm12:26
<awoods>ajs6f/barmintor: in summary, GET refreshes TXN expiry, git submodules are likely not the answer, we need both build-time and run-time wiring approaches.12:28
* ajs6f leaves
* ajs6f1 joins
afk bbs
<barmintor>awoods: That is a reasonable summary, tho point 1 has not happened yet
* ajs6f1 leaves12:29
<awoods>barmintor/ajs6f: Let's take up this wiring discussion in earnest and come do a decision... before Christmas? before 4.0-Beta?
* github-ff joins12:35
[fcrepo4] barmintor pushed 1 new commit to jersey2: http://git.io/X4FPMA
fcrepo4/jersey2 33056e8 Benjamin Armintor: take out problematic debug logging that assumes non-null graph subject; send Accept header in nodeTypes IT
* github-ff leaves
* ajs6f joins12:44
awoods: sure.12:45
<awoods>ajs6f: decide pre-4.0-Beta?
<ajs6f>is that march?
<awoods>no later than March.12:46
<ajs6f>For a decision, or for a plan, or for the beginnings of impl, or...?
* travis-ci joins12:47
[travis-ci] futures/fcrepo4#1190 (jersey2 - 33056e8 : Benjamin Armintor): The build is still failing.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/38404783e3de...33056e87495a
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/14033753
* travis-ci leaves
<awoods>ajs6f: At a minimum, let's have a decision and a plan in place pre-4.0-Beta. If we can also get going on an impl by then, great.12:48
<ajs6f>Sounds good tome.
<awoods>ajs6f: I think the main thing is putting some deep thought into weighing the options and viability soon.12:49
<ajs6f>Well, I've got my opinions: CDI over OSGi would be best, but extremely expensive. Blueprint (aka SPring DM) over OSGi would be very good and much cheaper. CDI flat (in a vanilla webapp) could be a nice lighterweight alternative, _if_ we can tune injection to be easy to use.12:50
which barmintor is currently doing.12:51
Ish.
<awoods>ajs6f: In the flat CDI case, is it clear to you how a repo-admin can select which external modules to enable at deploy/run-time?12:52
<ajs6f>We have to use the only container available— WEBI-INF/lib12:53
_Or_ we build our own, which is dumb.
We wouldn't do that.
So you drop a JAR in WEB-INF/lib and restart the webapp.
Without a true module container available, that's a given.12:54
We might be able to _disable_ modules at runtime.
And if we were willing to track wiring in the repo, we might be able to do it all, but that amounts to writing our own modulariation framework. See above.12:55
I don't want to invent the Fedora module framework.
<awoods>ajs6f: Is a webapp restart required? or could we just have the CDI tooling (HK2) refresh its ResourceConfig.
<bljenkins>Project fcrepo-fixity-corrupter build #436: SUCCESS in 1 min 50 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/436/12:56
<ajs6f>awoods: If scanning worked just right, maybe we could do that. That is not very diff from an app restart, tho. And HK2 isn't a full CDI solution.
HK2 does JSR-330 only, to my knowledge.12:57
Not that we shouldn't look at it as a useful tool. (That we are already using via Jersey2.)
<awoods>ajs6f: AFAYK, HK2 to roughly equivalent to Guice?12:59
* nbanks joins
<awoods>s/to/is/
<ajs6f>Something like that. It's interested in DI, not in contexts.
CDI is exactly both.
<bljenkins>Project fcrepo-kitchen-sink build #637: STILL FAILING in 3 min 53 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/637/13:00
<ajs6f>I'm guessing that's why barmintor pointed out that HK2 is not a replacement for Spring.
I don't know that much about Guice, tho.
<awoods>ajs6f: Although you approach this on the wiki, would you be willing (not necessarily right now) to add some "scenarios" to this page which detail what we expect the repo-admin and developer to be able to do:https://wiki.duraspace.org/display/FF/Design+-+Wiring+and+configuration13:03
<ajs6f>What's the diff between a scneario and a use case?13:04
awoods: Would you be willing to try to gather some input from repo managers and devs who are not on the project? I know what I want to do. I know what my repo manager wants to do. But I'm pretty strange, and mikeAtUVa is only one example.13:07
<awoods>ajs6f: That is a good suggestion. I can beat the bushes to get stakeholder input on this topic.13:09
<pivotal-bot>Andrew Woods added "Define build-/run-time module wiring use cases" https://www.pivotaltracker.com/story/show/6086334613:10
Andrew Woods edited "Define build-/run-time module wiring use cases" https://www.pivotaltracker.com/story/show/60863346
Andrew Woods edited "Define build-/run-time module wiring use cases" https://www.pivotaltracker.com/story/show/60863346
<ajs6f>Just keep shakin' that bush.13:11
http://www.youtube.com/watch?v=G7QwTTn3FIk
awoods: I just read through some stuff about Guice, and it seems that HK2 is in the same space, although I would say that it seems to be a little more developed in terms of using qualifiers and so forth. But I don't know if that's an accurate observation because i haven't used Guice.13:14
* escowles joins13:15
<cbeer>barmintor: you figured that all out?13:29
<pivotal-bot>Andrew Woods added "Remove statics from NodePropertiesTools" https://www.pivotaltracker.com/story/show/6086561813:42
A. "Pholeotic" Soroka added comment: "Should this wait on a merge of some PR?" https://www.pivotaltracker.com/story/show/6086561813:44
Osman Din added comment: "Will follow up with documentation before next sprint planning call." https://www.pivotaltracker.com/story/show/6061567013:50
Osman Din unstarted "Test Basic AuthZ" https://www.pivotaltracker.com/story/show/60615670
<cbeer>ajs6f: i'm thinking about taking on https://www.pivotaltracker.com/story/show/6086561813:53
<pivotal-bot>feature: Remove statics from NodePropertiesTools (unscheduled) / owner: A. "Pholeotic" Soroka
<cbeer>you think that should wait for barmintor's work?
or your rdf streaming stuff?
<ajs6f>Defintely not waiting for RDF sreaming stuff.
I don't _think_ that would wait on barmintor… barmintor?
<cbeer>hearing no reply, i'm going to spend my day poking at the kernel API.14:06
<ajs6f>no reply to the question about killing statics?
<cbeer>no, from barmintor. i'll just avoid any refactoring that goes above -kernel, i guess14:07
<ajs6f>k14:08
I could spend a whole sprint on kernel, just tightening bolts.
<awoods>ajs6f: Are you on next sprint?14:09
<ajs6f>awoods: Am I not?
I thought so.
<awoods>ajs6f: I see a "?" : https://wiki.duraspace.org/display/FF/Beta+Sprint+Schedule
<ajs6f>awoods: Did you ever hear back from tecoripa about his experience with CDI?
<awoods>nope
<cbeer>ajs6f: if only our ship wasn't half afloat, taking on water, and not really a ship anyway. more of a floating dock.
<ajs6f>cbeer; I think of it as more of a Flying Dutchman than a Flying Scotsman.14:10
awoods: Perhaps that question mark just indicates my fundamentally questionable value.
<awoods>ajs6f: In which case I would advocate for making it a solid "x"14:11
<ajs6f>awoods: what is the diff between x and #?
<awoods>ajs6f: # stands for "scrum dog"
<ajs6f>What on Earth is a scrum dog? Never mind. I don't even want to know.14:12
afk bb in 514:14
* ajs6f leaves
<awoods>afk - lunch14:17
* ajs6f joins14:18
<cbeer>ajs6f: it's been way too long since i looked at -kernel14:22
was there some clever injection reason we have every method on NodeService take a Session, instead of creating a NodeService-per-session?14:23
(or, even NodeService-per-node..)
* ajs6f1 joins
<cbeer>11:22 ] @cbeer> ajs6f: it's been way too long since i looked at -kernel
11:23 ] @cbeer> was there some clever injection reason we have every method on NodeService take a Session, instead of creating a NodeService-per-session?
11:23 ] @cbeer> (or, even NodeService-per-node..)
<ajs6f1>I don't even remember NodeService. Did I write that?14:24
<cbeer>no, it's probably my fault.
<ajs6f1>Was that supposed to be like DatastreamService, ObjectService, etc.
<cbeer>https://github.com/futures/fcrepo4/blob/master/fcrepo-kernel/src/main/java/org/fcrepo/kernel/services/NodeService.java
yeah, i think DS + Object extend NodeService, maybe.
nope, guess not.14:25
I think the reason I made it was there are lots of times now we don't care if it's a datastream or an object
* ajs6f leaves
<cbeer>it just seems weird we don't create one per session or something.
<ajs6f1>I would suggest ResourceService, then.
There's another thing— nothing that takes a JCR Item should need to take a session.14:26
Items come attached to the repo through a Session.
<cbeer>ajs6f1: this takes a session + path
<ajs6f1>Is it creating a node?
Sounds like inherited static stuff.14:27
<cbeer>yeah. that's my thinking too. or stuff that hoped it was static.
so, no objection to making a new node service for each session?
(i guess if it was for some clever wiring or injection thing, we could just make a ResourceServiceFactory)14:28
<ajs6f1>no objection, just that we should watch and make sure that we are really saving time/effort by using it. Sometimes it's clearer to just use the JCR API.
<cbeer>you and awoods can fight to the death.14:29
<ajs6f1>I'm talking about our own impls, not what we expose.
<cbeer>ajs6f1: yeah. i think this is getting exposed to -http-api14:30
* ajs6f joins14:32
Oh. Do we need to expose something that carries no abstraction? It seems like just a bunch of shortcuts for JCR calls. The type involved are all JCR types.
* ajs6f1 leaves14:34
* github-ff joins14:37
[fcrepo4] ajs6f force-pushed KillWarnings from 8673ded to 280e503: http://git.io/l_ehEg
fcrepo4/KillWarnings 9b421a7 ajs6f: Simple fixes for type warnings
fcrepo4/KillWarnings 23b949f ajs6f: Using non-deprecated HTTP client machinery
fcrepo4/KillWarnings 9fe1e47 ajs6f: Using non-deprecated HTTP multipart machinery
* github-ff leaves
<ajs6f>awoods: ^^^ is a little cleanup ready for merge14:39
cbeer: Looking again, at least all the interesting return types are Fedora stuff. So I guess that does make sense to expose.14:41
<escowles>awoods: just got an email from declan: i'm on for sprint 814:45
<ajs6f>afk
* ajs6f leaves
* travis-ci joins14:49
[travis-ci] futures/fcrepo4#1191 (KillWarnings - 280e503 : ajs6f): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/8673dedd9c2b...280e5033d451
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/14039516
* travis-ci leaves
<escowles>cbeer: are we using http://purl.org/dc/terms/ or http://purl.org/dc/elements/1.1/ for e.g. dc:title these days?14:58
<cbeer>escowles: we're using /dc/terms in master, and /dc/elements/1.1 in the sparql branch
we should use the element set, i think. and maybe apply dcterms in some optional extension14:59
<escowles>ok, i'll stick with /dc/elements/1.1 -- i was just worried b/c /dc/terms will set the page title, but /dc/elements/1.1/ won't
<bljenkins>Project fcrepo-fixity-corrupter build #437: SUCCESS in 1 min 22 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/437/15:01
Project fcrepo-kitchen-sink build #638: STILL FAILING in 4 min 48 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/638/15:06
<awoods>escowles: great news about your addition of sprint-8!15:15
cbeer: Are ajs6f and I supposed to fight to the death?15:16
<cbeer>awoods: maybe no any more. just a disagreement about how much JCR machinery we should be expose. some of our *Services are just really thin proxies around JCR stuff15:17
but i think we're all in agreement again, because our methods (should) return fcrepo4 concepts, not JCR stuff
(which some of the DatastreamService methods don't.)15:18
<awoods>cbeer: Are folks in disagreement, or has the code just not yet caught up with the agreement?15:19
<cbeer>awoods: nope. i think we're all in agreement again.15:20
<awoods>cbeer: Good, that sounds much better than a blood bath.
<cbeer>confluence--15:21
awoods, escowles: what do you guys think of the approach i'm taking on https://wiki.duraspace.org/display/FF/REST+API+Redesign+Stub?15:23
and by approach, i mean layout and the like, not the copy.15:24
<awoods>cbeer: I will give a look in just a moment.15:26
<pivotal-bot>A. "Pholeotic" Soroka added comment: "I believe that even one core procs can use pipelining to help run multiple threads more efficiently." https://www.pivotaltracker.com/story/show/59603190
<escowles>cbeer: looks pretty
<awoods>cbeer: RE: fcrepo-transform vs. fcrepo-http-api...15:27
cbeer: I am fine with have the admin-search somewhere besides fcrepo-http-api, but I was under the impression that fcrepo-transform was establishing a pattern around storing/retrieving transforms in the repo that were then used in various contexts to dynamically transform repo responses.15:28
cbeer: I am just not seeing that the sparql/admin-search is following that pattern.15:29
cbeer: Would a new fcrepo-admin-search module make more sense?15:30
or fcrepo-search-admin
<cbeer>awoods: maybe.
* github-ff joins15:32
[fcrepo4] escowles pushed 3 new commits to editable-views: http://git.io/SsoAnw
fcrepo4/editable-views ab71c1b Esmé Cowles: Basic add/delete property functionality
fcrepo4/editable-views ad5ee8f Esmé Cowles: Merge branch 'editable-views' of github.com:futures/fcrepo4 into editable-views...
fcrepo4/editable-views 898adcd Esmé Cowles: Adding custom predicate URI field, handling quoted or unquoted URIs and literals, suppressing deletion widget for JCR-managed properties
* github-ff leaves
* github-ff joins15:35
[fcrepo4] escowles opened pull request #153: Making HTML views editable (master...editable-views) http://git.io/AZX8uw
* github-ff leaves
<awoods>cbeer: https://wiki.duraspace.org/display/FF/REST+API+Redesign+Stub looks good.15:36
<cbeer>awoods: ok. i'll press forward.15:37
<pivotal-bot>Andrew Woods added comment: "This command fails on my machine: ""
mvn clean install -Djacoco.skip -T 2.0C" https://www.pivotaltracker.com/story/show/59603190
Esme Cowles added comment: "I've added basic editing functionality: an X link next to properties to delete them, and a form in the sideb..." https://www.pivotaltracker.com/story/show/60781280
Esme Cowles finished "Make property values in HTML UI editable" https://www.pivotaltracker.com/story/show/6078128015:38
<barmintor>cbeer: I get intermittnent failures on the htmlunit tests, and am considering ignoring that suite for now.15:46
* travis-ci joins15:48
[travis-ci] futures/fcrepo4#1192 (editable-views - 898adcd : Esmé Cowles): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/7e50933cd263...898adcd17f8d
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/14041679
* travis-ci leaves
<cbeer>barmintor: +1. i'd rather have jersey2 than html integration tests.
<barmintor>okey doke
<pivotal-bot>Esme Cowles added comment: "I have updated https://wiki.duraspace.org/display/FF/Comparing+Fedora+3+and+Fedora+4 with additional results..." https://www.pivotaltracker.com/story/show/6055694815:50
Esme Cowles finished "Single-node F4 Ingest Benchmarks" https://www.pivotaltracker.com/story/show/60556948
Esme Cowles finished "Single-node F3 Ingest Benchmarks" https://www.pivotaltracker.com/story/show/6055702415:51
<bljenkins>Project fcrepo-fixity-corrupter build #438: SUCCESS in 1 min 23 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/438/15:55
Project fcrepo-kitchen-sink build #639: STILL FAILING in 3 min 54 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/639/15:59
<pivotal-bot>Chris Beer added "FedoraContent PUT should accept checksum header" https://www.pivotaltracker.com/story/show/6087591616:11
Andrew Woods added comment: "Looks good by me, pending comments. @escowles?16:25
@cbeer, would you rather leave the new classes in fcrepo-tra..." https://www.pivotaltracker.com/story/show/60713796
Chris Beer added "Fix up fcr:datastreams for a path-based world" https://www.pivotaltracker.com/story/show/6087699016:31
Andrew Woods added comment: "Have you been able to diagnose this behavior? Or are you still in the phase of documenting it?" https://www.pivotaltracker.com/story/show/6055694816:32
Esme Cowles added comment: "Yes, I'm still documenting what the issue is: I'm currently running a test with 50MB datastreams, and I'm no..." https://www.pivotaltracker.com/story/show/6055694816:37
Chris Beer added "Remove numPids reference from fcr:identifier API" https://www.pivotaltracker.com/story/show/60877478
* github-ff joins16:40
[fcrepo4] awoods pushed 1 new commit to master: http://git.io/C9N4UA
fcrepo4/master 26c4fbe ajs6f: Simple fixes for type warnings...
* github-ff leaves
<pivotal-bot>Andrew Woods delivered "Killing compiler warnings" https://www.pivotaltracker.com/story/show/60810250
Andrew Woods added comment: "This looks good. Although I would suggest adding a section to the wiki which makes it clear to the user whi..." https://www.pivotaltracker.com/story/show/6071383816:45
* travis-ci joins16:54
[travis-ci] futures/fcrepo4#1194 (master - 26c4fbe : ajs6f): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/1c94c0c47474...26c4fbe6fc38
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/14044494
* travis-ci leaves
<pivotal-bot>Chris Beer added "Fix content returned in fcr:workspaces" https://www.pivotaltracker.com/story/show/6087883816:58
* gregjansen leaves17:00
<pivotal-bot>Andrew Woods accepted "Killing compiler warnings" https://www.pivotaltracker.com/story/show/60810250
Andrew Woods added comment: "So is this ticket "Finished"?" https://www.pivotaltracker.com/story/show/6055694817:01
<barmintor>Conclusion: HK2 is like Spring, but worse in every way.17:11
<bljenkins>Project fcrepo-fixity-corrupter build #439: SUCCESS in 1 min 56 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/439/
<awoods>barmintor: That might be better than HK2 being compelling in some ways.17:12
<barmintor>awoods: ...
<awoods>barmintor: At least HK2 is categorically worse.17:13
<barmintor>this really seems to be a library written by sadists at Oracle
like, in the name of making your code less dependent on Spring, it has to know MUCH MUCH MORE about its dependency injection framework17:14
<awoods>barmintor: Is there a way to minimize the HK2 madness?17:15
<barmintor>I have no idea
I assume so, because no one has firebombed Larry Ellison's house yet
<pivotal-bot>Andrew Woods added comment: "@cbeer / @escowles, where do we stand on this one?" https://www.pivotaltracker.com/story/show/6071379617:18
<bljenkins>Project fcrepo-kitchen-sink build #640: STILL FAILING in 8 min 4 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/640/17:20
* nbanks leaves
<cbeer>ok, promoted the REST API docs to https://wiki.duraspace.org/display/FF/REST+API
bbl17:21
<pivotal-bot>Esme Cowles added comment: "This looks good to me -- similar approach to what we're doing at UCSD for our SQL-based triplestore (same Je..." https://www.pivotaltracker.com/story/show/6071379617:27
Esme Cowles added comment: "Well, it's "finished" in the sense that I have an initial set of benchmarks. But really I'd call it "ongoin..." https://www.pivotaltracker.com/story/show/6055694817:28
Andrew Woods added comment: "I guess the response to this comment will only impact your work (#60781280) @escowles, not this ticket.17:30
htt..." https://www.pivotaltracker.com/story/show/60713796
<awoods>escowles: Should we deliver https://www.pivotaltracker.com/story/show/60556948 , and create another ticket with more details of next steps?17:32
<pivotal-bot>feature: Single-node F4 Ingest Benchmarks (finished) / owner: Esme Cowles
Esme Cowles added comment: "@awoods : i'm fine holding off on #60781280 -- there are some pretty clear next steps for improving that wor..." https://www.pivotaltracker.com/story/show/60713796
Andrew Woods added comment: "I am going to review this one and push/comment as appropriate. Feel free to go ahead and create additional ..." https://www.pivotaltracker.com/story/show/6078128017:34
<awoods>escowles: Could you please rebase PR-153 on master. The merge has conflicts.17:37
* levar1 leaves
<awoods>escowles: ping17:38
<escowles>awoods: yes?
<awoods>escowles: two comments above17:39
escowles: re:60556948 and rebasing PR-153
<escowles>yep -- i'll check it out (though i'm leaving for dinner in 10 mins, so won't get it done today)17:40
<awoods>escowles: kk, just let me know.
escowles: mikeAtUVa will be around for testing tomorrow, so ideally we can get this in by then.17:43
* github-ff joins17:44
[fcrepo4] escowles pushed 4 new commits to editable-views: http://git.io/jTlhCA
fcrepo4/editable-views e99a7ef Esmé Cowles: Basic add/delete property functionality
fcrepo4/editable-views 6fbfe23 Esmé Cowles: Basic add/delete property functionality
fcrepo4/editable-views 9bfe2c7 Esmé Cowles: Adding custom predicate URI field, handling quoted or unquoted URIs and literals, suppressing deletion widget for JCR-managed properties
* github-ff leaves
<escowles>awoods: is this better now? https://github.com/futures/fcrepo4/pull/15317:58
awoods: babysitter's here -- so i'll check in tomorrow17:59
<awoods>escowles: I will give it a look after finishing with Mike's PR... unless you want to hear now.
thanks
<pivotal-bot>Andrew Woods added "Optional Versioning" https://www.pivotaltracker.com/story/show/6088232818:03
<bljenkins>Project fcrepo-fixity-corrupter build #440: SUCCESS in 1 min 20 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/440/18:09
* ksclarke leaves18:15
<bljenkins>Project fcrepo-kitchen-sink build #641: STILL FAILING in 6 min 32 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/641/18:16
<pivotal-bot>Andrew Woods added "Version Management" https://www.pivotaltracker.com/story/show/6088334618:20
Chris Beer added "Fix export URL" https://www.pivotaltracker.com/story/show/6088375218:30
Chris Beer added "Normalize responses for Nodes with a trailing slash" https://www.pivotaltracker.com/story/show/6088381418:32
* github-ff joins18:36
[fcrepo4] barmintor pushed 1 new commit to jersey2: http://git.io/mVJj1Q
fcrepo4/jersey2 0e33a71 Benjamin Armintor: http-api passing
* github-ff leaves
<pivotal-bot>Andrew Woods delivered "Simple Versioning" https://www.pivotaltracker.com/story/show/5961426818:47
* github-ff joins
[fcrepo4] awoods pushed 1 new commit to master: http://git.io/otquQQ
fcrepo4/master ece5e32 Michael Durbin: Added basic version creation on all REST API updates commands....
* github-ff leaves
<pivotal-bot>Andrew Woods added comment: "@md, I added the attached minor updates." https://www.pivotaltracker.com/story/show/5961426818:48
* travis-ci joins
[travis-ci] futures/fcrepo4#1196 (jersey2 - 0e33a71 : Benjamin Armintor): The build is still failing.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/33056e87495a...0e33a71fabb4
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/14048986
* travis-ci leaves
* github-ff joins18:50
[fcrepo4] awoods closed pull request #151: Added basic version creation on all REST API updates commands. (master...master) http://git.io/DVG27w
* github-ff leaves
<pivotal-bot>Chris Beer added "Support preconditions on FedoraNodes#delete" https://www.pivotaltracker.com/story/show/6088450818:54
<bljenkins>Project fcrepo-fixity-corrupter build #441: SUCCESS in 57 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/441/
* travis-ci joins19:00
[travis-ci] futures/fcrepo4#1197 (master - ece5e32 : Michael Durbin): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/26c4fbe6fc38...ece5e3264363
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/14049426
* travis-ci leaves
<bljenkins>Project fcrepo-fixity-corrupter build #442: SUCCESS in 56 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/442/19:11
Project fcrepo-kitchen-sink build #642: STILL FAILING in 3 min 1 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/642/19:14
Project fcrepo-jms-indexer-pluggable build #225: UNSTABLE in 4 min 41 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-jms-indexer-pluggable/225/19:15
* ksclarke joins19:16
<pivotal-bot>Chris Beer added "Support pre-conditions on fcr:datastreams" https://www.pivotaltracker.com/story/show/6088547219:17
Andrew Woods added comment: "The rebase in https://github.com/futures/fcrepo4/pull/153 seems junked. I tried to cherry-pick the appropri..." https://www.pivotaltracker.com/story/show/6078128019:22
Andrew Woods rejected "Make property values in HTML UI editable" https://www.pivotaltracker.com/story/show/60781280
Chris Beer added "Provide a way to retrieve backups over the REST API" https://www.pivotaltracker.com/story/show/6088571419:27
Chris Beer added "Build out asynchronous API support using fcr:backup as an example" https://www.pivotaltracker.com/story/show/60885728
Chris Beer added "Should POST fcr:identifier send back status 200 or 201?" https://www.pivotaltracker.com/story/show/6088574819:28
Chris Beer added "If fcr:identifier returns only a single ID, should we send it in the Location header too?" https://www.pivotaltracker.com/story/show/60885756
Chris Beer added "Remove double Expires header from fcr:tx response" https://www.pivotaltracker.com/story/show/6088580819:30
Chris Beer added "fcr:fixity status code should indicate.. something.." https://www.pivotaltracker.com/story/show/6088587019:33
Andrew Woods accepted "Review/Modify Async Storage API" https://www.pivotaltracker.com/story/show/59855470
Andrew Woods edited "Provide a way to retrieve backups over the REST API" https://www.pivotaltracker.com/story/show/6088571419:35
Chris Beer started "review fcrepo4 RESTful error codes" https://www.pivotaltracker.com/story/show/4891859519:36
Andrew Woods added this label to 1 story https://www.pivotaltracker.com/story/show/60885728
Chris Beer started "Use nouns in the fcrepo-http-api sub-resources" https://www.pivotaltracker.com/story/show/49062417
Chris Beer added "Fix created workspace Location" https://www.pivotaltracker.com/story/show/6088600019:37
Chris Beer added "Add preconditions to fcr:nodetypes, if possible" https://www.pivotaltracker.com/story/show/6088602019:38
<awoods>cbeer: Any word on Stanford availability of test servers (for performance testing)?19:45
<cbeer>awoods: sysadmins are on it.
<awoods>cbeer: That is good... but I take it they did not offer an expected date of availability.19:46
<cbeer>nope19:47
awoods: or, if i had looked in my email, they've made all 4 boxes.19:49
<awoods>cbeer: nice
cbeer: It would be great if you could help get to the bottom of single and clustered performance issues (if at the end of the next two sprints we do not have the performance issues nailed, bad news)19:50
<pivotal-bot>Chris Beer edited "review fcrepo4 RESTful status codes" https://www.pivotaltracker.com/story/show/4891859520:09
Chris Beer added comment: "Filed new tickets to fix defects as needed.20:10
https://www.pivotaltracker.com/story/show/60885748
https://www.p..." https://www.pivotaltracker.com/story/show/48918595
Chris Beer finished "review fcrepo4 RESTful status codes" https://www.pivotaltracker.com/story/show/48918595
Chris Beer added comment: "Seems like we've done better than expected. There's some possibly questionable endpoints with no obvious easy..." https://www.pivotaltracker.com/story/show/4906241720:12
Chris Beer finished "Use nouns in the fcrepo-http-api sub-resources" https://www.pivotaltracker.com/story/show/49062417
Chris Beer added "Could/should fcr:export/fcr:import and fcr:backup/fcr:restore get grouped under a single endpoint?" https://www.pivotaltracker.com/story/show/60887190
Chris Beer edited "Could/should fcr:export/fcr:import and fcr:backup/fcr:restore get grouped under a single endpoint?" https://www.pivotaltracker.com/story/show/6088719020:13
Chris Beer added "Find out the behavior of fcr:restore" https://www.pivotaltracker.com/story/show/60887228
Andrew Woods added comment: "This wiki page needs some clean-up based on ModeShape developments. See inline comments: ""
https://wiki.duras..." https://www.pivotaltracker.com/story/show/60276468
Andrew Woods accepted "Define needs of FileSystem Federation" https://www.pivotaltracker.com/story/show/6027646820:14
Andrew Woods accepted "Simple Versioning" https://www.pivotaltracker.com/story/show/59614268
Chris Beer added "Fix response body from creating a node" https://www.pivotaltracker.com/story/show/60887246
Chris Beer edited "Find out the behavior of fcr:restore" https://www.pivotaltracker.com/story/show/6088722820:15
Chris Beer estimated "Fix response body from creating a node" as 1 point https://www.pivotaltracker.com/story/show/60887246
Chris Beer estimated "Could/should fcr:export/fcr:import and fcr:backup/fcr:restore get grouped under a single endpoint?" as 1 point https://www.pivotaltracker.com/story/show/60887190
<awoods>cbeer: There are some outstanding comments on https://github.com/futures/fcrepo4/pull/147
cbeer: Otherwise, it can be pushed.
<cbeer>yep. i saw them. i'll get to it early next week i guess
<pivotal-bot>Chris Beer edited "Add preconditions to fcr:nodetypes, if possible" https://www.pivotaltracker.com/story/show/6088602020:16
Chris Beer estimated "Add preconditions to fcr:nodetypes, if possible" as 2 points https://www.pivotaltracker.com/story/show/60886020
Chris Beer estimated "Fix created workspace Location" as 2 points https://www.pivotaltracker.com/story/show/60886000
Chris Beer edited "Fix created workspace Location" https://www.pivotaltracker.com/story/show/6088600020:17
Andrew Woods accepted "Use nouns in the fcrepo-http-api sub-resources" https://www.pivotaltracker.com/story/show/49062417
Chris Beer added "Update node COPY/MOVE to be cross-workspace ready" https://www.pivotaltracker.com/story/show/6088732220:18
Chris Beer added "Add DELETE endpoint for workspaces" https://www.pivotaltracker.com/story/show/60887336
Chris Beer edited "fcr:fixity status code should indicate.. something.." https://www.pivotaltracker.com/story/show/6088587020:19
Chris Beer estimated "fcr:fixity status code should indicate.. something.." as 1 point https://www.pivotaltracker.com/story/show/60885870
Chris Beer edited "Remove double Expires header from fcr:tx response" https://www.pivotaltracker.com/story/show/6088580820:20
Chris Beer edited "Remove double Expires header from fcr:tx response" https://www.pivotaltracker.com/story/show/60885808
Andrew Woods edited "Should POST fcr:identifier send back status 200 or 201?" https://www.pivotaltracker.com/story/show/6088574820:22
Chris Beer edited "If fcr:identifier returns only a single ID, should we send it in the Location header too?" https://www.pivotaltracker.com/story/show/60885756
Chris Beer estimated "If fcr:identifier returns only a single ID, should we send it in the Location header too?" as 1 point https://www.pivotaltracker.com/story/show/60885756
Andrew Woods edited "fcr:fixity status code should indicate.. something.." https://www.pivotaltracker.com/story/show/60885870
Andrew Woods accepted "review fcrepo4 RESTful status codes" https://www.pivotaltracker.com/story/show/48918595
Chris Beer edited "Should POST fcr:identifier send back status 200 or 201?" https://www.pivotaltracker.com/story/show/6088574820:23
Chris Beer estimated "Should POST fcr:identifier send back status 200 or 201?" as 0 points https://www.pivotaltracker.com/story/show/60885748
Chris Beer edited "Build out asynchronous API support using fcr:backup as an example" https://www.pivotaltracker.com/story/show/6088572820:24
Chris Beer estimated "Build out asynchronous API support using fcr:backup as an example" as 3 points https://www.pivotaltracker.com/story/show/60885728
Chris Beer deleted "Provide a way to retrieve backups over the REST API" https://www.pivotaltracker.com/story/show/60885714
Chris Beer edited "Build out asynchronous API support using fcr:backup as an example" https://www.pivotaltracker.com/story/show/6088572820:25
<awoods>cbeer: Are you viewing the async implementation work as falling into the Christmas Release (i.e. 4.0 Key Capabilities), or are you just documenting the tickets while they are fresh in your mind?
<pivotal-bot>Chris Beer edited "Support pre-conditions on fcr:datastreams" https://www.pivotaltracker.com/story/show/60885472
<cbeer>awoods: i think a limited async API, at least for backup operations, would be very desirable.20:26
<awoods>cbeer: for the Christmas Release?20:27
cbeer: Should we add it to 4.0 Key Capabilities?
<cbeer>stretch goal.
i don't think it's a key capability.
but we do need a way to retrieve backups that isn't log into the server and find it, i think
<pivotal-bot>Chris Beer estimated "Support pre-conditions on fcr:datastreams" as 2 points https://www.pivotaltracker.com/story/show/6088547220:28
<awoods>cbeer: I like stretch goals, especially if the expected goals are "in the bag"
<cbeer>awoods: this might also be low-hanging fruit if the jersey2 stuff pans out.
<pivotal-bot>Chris Beer edited "Support preconditions on FedoraNodes#delete" https://www.pivotaltracker.com/story/show/6088450820:29
Chris Beer edited "Normalize responses for Nodes with a trailing slash" https://www.pivotaltracker.com/story/show/6088381420:30
Chris Beer edited "Fix export URL" https://www.pivotaltracker.com/story/show/60883752
<awoods>cbeer: I agree. But you see my point: The steering team has made a push to get enough developers on the next two sprints to ensure that the agreed upon 4.0 Key Capabilities are achieved. If we do other things instead and fall short on the Key Capabilities that will be a significant failure.20:31
<cbeer>awoods: sure, but doesn't backups fall under the durable storage capability?
<awoods>cbeer: not key: https://wiki.duraspace.org/display/FF/Fedora+4.0+Key+Capabilities20:32
<cbeer>oh. i missed when that got dropped
carry on then,
<awoods>cbeer: I am delighted for you to push some of these other capabilities forward. You do it well. I just want to make sure this short list of Key Capabilities (most notably *performance*) gets wrapped up nicely.20:33
<cbeer>wait. are there multiple versions of this page floating around?20:34
<awoods>cbeer: There is the original page, which is now a sub-page of the link I just sent.
<cbeer>ah ha. i thought i was going crazy.20:35
<awoods>cbeer: In Austin, during the Thurs steering meeting, the original list was trimmed.
<cbeer>that's the page i've been referring to. maybe it needs a big disclaimer on it?
<awoods>cbeer: I will do that now
<cbeer>thanks
<pivotal-bot>Chris Beer added "Improve REST API document for external, technically savvy consumers" https://www.pivotaltracker.com/story/show/6088782620:36
Chris Beer estimated "Improve REST API document for external, technically savvy consumers" as 2 points https://www.pivotaltracker.com/story/show/60887826
Chris Beer started "Improve REST API document for external, technically savvy consumers" https://www.pivotaltracker.com/story/show/60887826
Chris Beer finished "Improve REST API document for external, technically savvy consumers" https://www.pivotaltracker.com/story/show/60887826
<cbeer>just inflating our numbers.20:37
<awoods>https://wiki.duraspace.org/display/FF/Fedora+4.0+Key+Capability+Candidates
cbeer: Sorry for the confusion
cbeer: I also dropped Hydra and Islandora from the list based on your and nbank's feedback.20:38
<pivotal-bot>Chris Beer added comment: "https://wiki.duraspace.org/display/FF/REST+API" https://www.pivotaltracker.com/story/show/6088782620:39
Chris Beer added comment: "There's probably some outstanding TODOs about improving copy and describing our data model, but I think we sh..." https://www.pivotaltracker.com/story/show/6088782620:40
Andrew Woods added "Decide on Wiring and Configuration Approaches" https://www.pivotaltracker.com/story/show/6088798220:44
Andrew Woods added comment: "Follow-on task #60887982" https://www.pivotaltracker.com/story/show/56894946
Andrew Woods accepted "Document runtime wiring options for F4 modules" https://www.pivotaltracker.com/story/show/5689494620:45
Andrew Woods added comment: "Pending code-review comments." https://www.pivotaltracker.com/story/show/60713796
Andrew Woods started "Single-node F3 Ingest Benchmarks" https://www.pivotaltracker.com/story/show/6055702420:48
Andrew Woods started "Single-node F4 Ingest Benchmarks" https://www.pivotaltracker.com/story/show/60556948
Andrew Woods accepted "Improve REST API document for external, technically savvy consumers" https://www.pivotaltracker.com/story/show/6088782620:50
<awoods>out for now20:51
* fcrepo-bot joins22:48
* fcrepo-bot leaves23:55

Generated by Sualtam