Log of the #fcrepo channel on chat.freenode.net

Using timezone: Eastern Standard Time
* dchandekstark joins02:47
* dchandekstark leaves02:51
* thomz joins04:09
* dchandekstark joins04:48
* dchandekstark leaves04:53
* dchandekstark joins05:53
* dchandekstark leaves05:58
* jjtuttle leaves07:17
* jjtuttle joins07:20
* manez joins07:57
* dchandekstark joins08:09
* github-ff joins08:19
[fcrepo-transform] awoods closed pull request #27: remove acoburn from maintainer list (master...update_docs) https://git.io/v6mc0
* github-ff leaves
* ajs6f joins08:30
* github-ff joins
[fcrepo-transform] acoburn reopened pull request #27: remove acoburn from maintainer list (master...update_docs) https://git.io/v6mc0
* github-ff leaves
* dwilcox joins08:36
* manez leaves09:00
* thomz leaves
* manez joins09:07
<ajs6f>barmintor: I just pulled the middle (common) type parameter out of CompositeConverter, and left the type-safety check in the constructors (instead of in the class header). Is that cool with you? I think it simplifies things dramatically down the line.
* bseeger joins09:08
* mikeAtUVa joins
* dhlamb joins09:38
* bseeger leaves09:40
* manez leaves09:49
* manez joins09:51
* manez_ joins09:53
* manez leaves
* bseeger joins09:56
* github-ff joins09:59
[fcrepo-camel-toolbox] acoburn opened pull request #103: Reimplement LDPath transforms as an OSGi service (master...fcrepo-2070) https://git.io/v6YCB
* github-ff leaves
<barmintor>ajs6f: so it becomes <T> CompositeConverter<A, B>, with <A,T> and <T, B> inputs?10:18
<ajs6f>barmintor: Right.10:19
<barmintor>ajs6f: that seems obvious, I was clearly delirious.
* dwilcox leaves
<ajs6f>barmintor: It's the noise. NYC is too noisy.
<barmintor>ajs6f: yes, it is the noise. it was definitely not an issue of my competence.
* dwilcox joins10:22
<barmintor>ruebot: 3pm call today?10:31
<ruebot>barmintor: yessir10:38
* github-ff joins10:44
[fcrepo-webapp-plus] acoburn opened pull request #47: Change configuration to support mode5 (master...fcrepo-1994) https://git.io/v6YRE
* github-ff leaves
* github-ff joins10:45
[fcrepo-transform] acoburn opened pull request #28: Deprecate fcrepo-transform (master...fcrepo-2106) https://git.io/v6YRy
* github-ff leaves
<barmintor>how can I have this much imposter syndrome about a product I work on11:00
am I wrong, or is the Michigan email not how versioning works?11:01
I would not expect to have a representation of nested version histories, nor would I expect the child node of a parent version to correspond to any named version of the canonical child node11:02
<mikeAtUVa>I would describe our versioning behavior for a subgraph as "undefined"... I'm pretty confident that JCR keeps all the version info, but highly skeptical that there aren't scenarios where we can't get the information back given the current REST API.
<barmintor>mikeAtUVa: yeah, but I also don't think JCR considers the version history of a node as a subgraph of that node11:03
mikeAtUVa: it's like a parallel tree
<ajs6f>It's in (to my limited recollection) an entirely separate workspace.11:04
<mikeAtUVa>Last time I worked on it, it was my understanding that "fcr:versions" would only appear once in your URL, and the path to the nodes beneath it are possibly not properly transformed into useful paths.
<ajs6f>fcr:versions is not the Tardis. It does not allow you freely roam time and space.11:05
<mikeAtUVa>barmintor: ajs6f: yeah, when I say subgraph, I was imprecise... it contains all the child nodes (which are ldp:contained)
* mikeAtUVa doesn't know how ashamed he should feel that he wouldn't recommend anyone build anything against the current API for versioning.11:06
<barmintor>well, the JCR versioning approach is very weird IMO. And we bleed it directly into the API11:08
<mikeAtUVa>Maybe for fedora 5 we can decide what it is, then build it.11:10
<barmintor>I mean the JCR appraoch isn't *dumb*. But it's fundamentally different from a FCR3 version.11:11
* mikeAtUVa needs to familiarize himself with how we implement our bidirectional linking... what happens if you link to a resource then revert that resource to a version before the link was made... should that version-reversion update the other object?11:15
* dhlamb leaves11:17
* dhlamb joins11:18
* dhlamb leaves11:21
* github-ff joins11:38
[fcrepo-module-auth-rbacl] acoburn opened pull request #36: Update configuration to work with MODE5 (master...fcrepo-1994) https://git.io/v6YoO
* github-ff leaves
* github-ff joins11:47
[fcrepo-module-auth-webac] acoburn opened pull request #73: Update configuration for MODE5 (master...fcrepo-1994) https://git.io/v6YKy
* github-ff leaves
* github-ff joins11:51
[fcrepo-module-auth-xacml] acoburn opened pull request #55: Update configuration for MODE5 (master...fcrepo-1994) https://git.io/v6Y6B
* github-ff leaves
<ajs6f>mikeAtUVa: "fedora 5"?
* ajs6f giggles
mikeAtUVa: You have fun with that!
barmintor: I am cleaning up your missing Javadoc params. How come this built for you but not for me?/11:56
<barmintor>that seems unlikely
<ajs6f>param descriptions, that i
Example, FedoraResourceImpl l 812 : * @param node
No description. Doesn't build for me.
<barmintor>worked for me and travis11:57
<ajs6f>I have no interest in Travis' opinions. Travis is flakier than my worst girlfriend in the 90s.
<barmintor>do we have different checkstyle rulesheets or something?11:59
<ajs6f>barmintor: Cripes, I hope not. This is really tedious.12:00
* dchandekstark leaves12:15
* bseeger leaves12:16
<ajs6f>barmintor: Interesting (and nice, for once). SInce we no longer retrieve nodes just to produce identifiers for them, the DefaultIdentifierTranslator that we use for tests no longer needs a JCR Session.12:31
<barmintor>Hmm. Not even for versions?12:32
or maybe it doesn't care about versions
<ajs6f>barmintor: I don't think it would. It doesn't seem, to anyway.
<barmintor>well that's that then
<ajs6f>barmintor: https://www.youtube.com/watch?v=wahLIbwZTLQ&feature=youtu.be&t=2712:35
* dchandekstark joins12:36
<ajs6f>barmintor: One of the reason I don't like inverse() and toDomain is because you may have to override both.12:59
jeez, FedoraLdpIT is a big set of tests.13:03
<barmintor>yes it is13:04
114 right?
it's a bunch
* barmintor has a 10 minute reprieve on a 3 hour block of meetings
<ajs6f>Thousands of lines.
barmintor: Don't forget to pee.13:05
<barmintor>that's good advice
<ajs6f>barmintor: This is annoying. The correct inDomain for the composition is first().inDomain(a) || second().inDomain(first().apply(a));13:07
sorry, first().inDomain(a) && second().inDomain(first().apply(a));
You have to check it all the way through, which means you ahve to do the first step.
Oh, well.
<barmintor>do you want to axe inDomain()? IDK if there's another way to provide it accurately.13:09
<ajs6f>barmintor: At the moment I just want to cut through this last bit like Conan. Not the guy on the TV, the guy from the old pulps. With the axe.
barmintor: Let's land your PR and then we can sit down and think. There are a number of things to think about, the most important of which, for my money, is typing the identifiers for reals.13:10
barmintor: I should have this PR-to-your-PR over to you in just a little bit. It's been a while since I looked at any of this. It's faintly horrifying.13:11
<barmintor>the typing may lessen the need for inDomain anyway- I think the big use case for it is the number of places we're just operating on random strings
<ajs6f>barmintor: Yup. I think you're dead right.13:12
<barmintor>whereas as InternalId(String id) could just tell you no
<ajs6f>Let the computer do the work.
<barmintor>ok, time for meetings.13:13
* barmintor salutes
* dchandekstark leaves13:24
* manez leaves13:25
* github-ff joins13:27
[fcrepo4] ajs6f pushed 1 new commit to SuggestionsForBarmintor: https://git.io/v6YFw
fcrepo4/SuggestionsForBarmintor c74686e ajs6f: More fat for the frying pan
* github-ff leaves
* manez joins
* acoburn joins13:32
* dwilcox leaves13:41
* travis-ci joins13:47
fcrepo4/fcrepo4#4649 (SuggestionsForBarmintor - c74686e : ajs6f): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/compare/c99c1d0c122b...c74686e85a35
Build details : https://travis-ci.org/fcrepo4/fcrepo4/builds/150112413
* travis-ci leaves
* dwilcox joins13:54
<acoburn>ajs6f:barmintor: since the mode5 branch has been merged, is there any need to keep that branch in the public repo?13:55
I'd be happy to delete it
<ajs6f>acoburn: i don't know of any reason.
<acoburn>ajs6f: thanks, I'll delete it
* github-ff joins13:56
[fcrepo4] acoburn deleted modeshape5 at ca71061: https://git.io/v6YxD
* github-ff leaves
* bseeger joins14:00
* github-ff joins14:06
[fcrepo4] ajs6f pushed 1 new commit to SuggestionsForBarmintor: https://git.io/v6Yhl
fcrepo4/SuggestionsForBarmintor 33ecd15 ajs6f: Minor touches
* github-ff leaves
* github-ff joins14:08
[fcrepo4] ajs6f pushed 1 new commit to SuggestionsForBarmintor: https://git.io/v6Yh9
fcrepo4/SuggestionsForBarmintor 17a1ec6 ajs6f: Name change
* github-ff leaves
* dchandekstark joins14:09
* dchandekstark leaves14:16
<ajs6f>ruebot: is there a meeting page for the im/ex meeting in a bit?14:17
<acoburn>ajs6f: I have some suggestions for barmintor, too, but perhaps I'll wait until your changes have been merged with his branch14:18
<ajs6f>acoburn: Well, did you see what I did? It might affect/conflict with you.
<acoburn>ajs6f: that's exactly what I mean — I've only just begun looking through barmintor's PR14:19
ajs6f: I don't want to get too far along until your changes make it into the PR
<ajs6f>acoburn: If— we'll see what barmintor thinks. But I think he won't find anything too surprising in what I'm suggesting. It's mostly just cleanup and tightening things.
<acoburn>ajs6f: mostly what I'll be suggesting is removing the translator().apply(…) stuff and replacing it with Optional.of(…).map(translator).get()14:20
<ajs6f>acoburn: I really do want to start typing the identifiers. But step-by-step.
acoburn: Why? How is that better?14:21
<acoburn>ajs6f: I think it reads better
ajs6f: and allows for nicer composition
<ajs6f>acoburn: I am not seeing nicerness. The composition, yes, I see that. But I'm not sure that we want the same notion of composition for translation as we do get from java for the Optional "monad".14:22
acoburn: Did you see barmintor/my notes on whether to throw an excpetion for an out-of-bounds input
<acoburn>ajs6f: true, maybe just using andThen14:23
<ajs6f>acoburn: Well, that's what we're doing now, with that custom CompositeConverter type.14:24
<acoburn>ajs6f: cool, like I said, I've only just started looking through the changes
<ajs6f>acoburn: I think the whole thing would be simpler if we just agreed to call every resource "George", like George Foreman did in his family.14:25
* travis-ci joins
fcrepo4/fcrepo4#4650 (SuggestionsForBarmintor - 33ecd15 : ajs6f): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/compare/c74686e85a35...33ecd15fb408
Build details : https://travis-ci.org/fcrepo4/fcrepo4/builds/150120415
* travis-ci leaves
<ajs6f>acoburn: You know, let's land this thing for the week (or ish) and then we really should ask whether the types of the typed identifiers that barmintor and I want turn out to be monadic. That might clarify a lot of things.14:26
Are we really translating across a map, or are we packing and unpacking networks of references?14:27
* travis-ci joins14:28
fcrepo4/fcrepo4#4651 (SuggestionsForBarmintor - 17a1ec6 : ajs6f): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/compare/33ecd15fb408...17a1ec6ec128
Build details : https://travis-ci.org/fcrepo4/fcrepo4/builds/150120928
* travis-ci leaves
<ajs6f>travis, stf up.
I feel like Travis is our version of Donny, from The Big Lebowski. No one wants to hear from him until he falls over dead. Then he's a big pain in the neck even after death.
<acoburn>ajs6f: I agree, getting this into master soon would be a good thing; we can then refine it from there14:29
<ajs6f>acoburn: None of us like having 4-digits-worth-of-line-changes PR flying around.14:30
<acoburn>ajs6f: especially when there are several of such PRs
<ajs6f>acoburn: Are you think of Jena 3?
<acoburn>ajs6f: yes
<ajs6f>acoburn: Yeah, I would love to get this and that down. MODE5 is already in. That's a much clearer road to 4.7.14:31
<acoburn>ajs6f: have you had a look at some of my follow-on PRs related to the MODE5 transition?
ajs6f: https://jira.duraspace.org/browse/FCREPO-199414:32
<ajs6f>acoburn: No, I saw that but I assumed you were just correcting conf. Why did awoods do that when he issued the main branch?
why didn't, I mean
<acoburn>ajs6f: no idea. I was just trying to fix them so that they can run
ajs6f: as it is, jenkins is in a tizzy so the artifacts aren't going to sonatype
<ajs6f>acoburn: Hm. Well, thanks for cleaning up after awoods.14:33
<barmintor>ajs6f: I don;t think any of this looks controversial, even given my procliivity for disagreeing with you. What are you sneaking by me?14:39
<ajs6f>barmintor: Doesn't it taste a little _too_ savory? The secret is: I used a splash of fish sauce in the dressing!14:40
barmintor: You'd never guess, would you?14:41
<acoburn>ajs6f: my son was using fish sauce on his salad last night
ajs6f: I kept my distance
<barmintor>ajs6f: I think I left some dead methods in HttpResourceConverter, actually. Marked up with a warning suppression14:42
<ajs6f>acoburn: It's actually pretty good in a vinagrette, no joke. Just a shot, like adding an anchovy. Try it before you knock it.
<barmintor>ajs6f: but I'll accept this PR
* ajs6f bursts into tears
You like me! You really like me!
<acoburn>ajs6f: I'll have to try it tonight
<barmintor>ajs6f: isn't it completely absurd how many classes these things touch?14:43
<ajs6f>acoburn: Keep it simple. Good olive oil, maybe some citrus or apple cider vinegar, a little salt and pepper, and a shot of fish sauce. Maybe a pinch of cayenne.14:44
<acoburn>ajs6f: mmmm, tasty
<ajs6f>barmintor: Well, it speaks to the rapidity with which we went from "Jeez, we should make a Fedora 4." to "Jeez, I'm so sick of Fedora 4."
barmintor: Despite acoburn's furious poundings, there is not nearly as much clarity inside the core as we would all like.14:45
barmintor: What the hell is the relationship between the models, the impls, and the *service classes?
<acoburn>ajs6f: my understanding of the service classes is that they are injected into the various non-impl layers (jms, http, auth, etc)14:49
ajs6f: or was that just rhetorical?14:50
<ajs6f>acoburn: I have no *&)*&*) idea.
<barmintor>I like that definition, but I'm not sure the corrollary (no service use in the kernel impl) is true.
Let me just check the design documents.14:51
* mikeAtUVa leaves
<barmintor>btw acoburn I've pulled ajs6f's changes into the PR
<ajs6f>The design docs? Let me just check here...14:52
<acoburn>barmintor: I saw that, but I'm about to have meetings for the rest of the day, so I doubt that I'll have anything substantive to say until next week
* ajs6f looks up his butt
<barmintor>acoburn: no problem.
<ajs6f>No, don't see anything about them there.
barmintor: Can you delete the branch from which I sent that PR, or must I?14:54
<barmintor>ajs6f: from the fcrepo4 org? I can do it
ajs6f: it is done14:55
<ruebot>https://wiki.duraspace.org/display/FF/2016-08-05+-+Import+-+Export+Planning+Meeting -- 5 minutes or so
* github-ff joins
[fcrepo4] barmintor deleted SuggestionsForBarmintor at 17a1ec6: https://git.io/v6OkW
* github-ff leaves
<ajs6f>barmintor: Rock on.
<barmintor>that's right I did, github! and I'll delete future suggestions too!
<ajs6f>Fedora 4: We're Not Interested In Your Oinions.14:57
<bseeger>Fedora 4: We're Not Interested in Your Opinions or Onions.
<ajs6f>Damn, the smooth jazz sounds of Freeconfcall always put me in the mood for the weekend.
Actually, I'm quite interested in onions.
I grow several patches of walking (sometimes called Egyptian) onions myself.14:58
They're a topsetting variety and they're just finishing setting their top bulbs this time of tyear.
* ajs6f is here
<barmintor>I'll never hear the Bangles the same way again14:59
oh crap 3pm
<ajs6f>Get on the beat, barmintor.
<bseeger>*is here *
* jhs joins15:00
<ruebot>barmintor: beeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeen!15:01
<barmintor>let's MEET15:02
<ajs6f>My maybe is the size of all outdoors.15:03
Move the dates?
* mjordan joins
* mikeAtUVa joins15:04
<ajs6f>Ain't Haskell, ain't crap.15:09
* dchandekstark joins15:10
* dchandekstark leaves
* dchandekstark joins
<ajs6f>bseeger+ # gets only one plus for uncertainty
<bseeger>one must feel comfortable with uncertainty. ;)15:13
* yinlin joins
<ajs6f>bseeger: I have children. I am not comfortable with uncertainty, as much as overwhelmed by it.15:18
Papa Bear?!15:22
<bseeger>ajs6f: yeah, that's been my experience as well, though it's getting easier now that my youngest is almost 615:25
<mikeAtUVa>For 5,6,7: are we thinking of exporting a resource from one URL/URI and importing into a different one? Is that really the feature we're talking about?15:26
<ajs6f>Ben "Max Headroom" Armintor15:34
<mikeAtUVa>ajs6f: do you have the same qualms about at post to FCR versions versioning all contained resources?15:35
<ajs6f>mikeAtUVa: good point. I've never really thought about it, because I try not to give myself nightmares by thinking about our versioning stuff
<barmintor>consider... DONE15:39
<ajs6f>barmintor: Consider the tiny ant.15:40
<ajs6f>He never stops to worry.
He's much too busy scurrying about.
<barmintor>yeah, no idea how we would preserve those hash uris15:45
I guess they are nodes
<ajs6f>barmintor: see why I didn't sign up for this?15:46
Our crummy triplestore is all grown up.
<mjordan>please don't be mad
<ajs6f>mjordan: We don't get mad.15:48
<mikeAtUVa>And this client will work against other fedora implementations, right?15:52
<ajs6f>mikeAtUVa: It already does. It works against all of the other impls.
<mikeAtUVa>ajs6f: Does a single resource bag of a binary include the related metadata?15:54
<ajs6f>mikeAtUVa: good question: it depends on what you think a "resource" is.15:55
mikeAtUVa: I think a bitstream and metadata are one resource.
mikeAtUVa: but barmintor _must_ disagree.
* manez leaves15:57
<ajs6f>I wish barmintor would stop plpugging his phone into his Echplex.15:58
p-p-p-p-p-p-plugging his ph-ph-ph-ph-phone.
<barmintor>I I I ddddddon't know know know what at at at yooooooou mean
MODE loads too many nodes15:59
<ajs6f>ruebot is a tornado of doodling
<barmintor>I would still like mjordan and ruebot to work with me on the bagit profile we were talking about16:00
<ajs6f>MODE is made of nodes. In fact, if you ever meet Randall Hauch, he's actually just a graph of JCR nodes floating about like gossamer fishing nets.
<barmintor>since I think it's interesting independent of this sprint :)
<ajs6f>barmintor: What is that profile for?
barmintor: Moving stuff from one LDP instance to another?16:01
<barmintor>ajs6f: while the id stuff will lower the amount of node fetching on objects with a lot of ref properties, MODE still demands that it load the nodes for all of them
ajs6f: no, the bagit think is more a way to introduce context and structure metadata to a bag16:02
<ajs6f>barmintor: WTF? Wait, what's the distinction between "fetching" and "loading"
<barmintor>ajs6f: no difference, using synonmously
barmintor: ^^^
What goes around comes around.
Everything old is new again. Etc.
<barmintor>yes, but this one uses json-ld and not a new spec :)16:03
<ajs6f>barmintor: I'm sure cwilper said "Yes, but this one uses RDF and not a new spec :) ."
<barmintor>cwilper was always so willing to do new things, I prefer to shoehorn crap16:04
<ajs6f>barmintor: So just having a node with prop x => node y means MODE insists on loading node y? Into the same session?
barmintor: Are you thinking of Akubra?
<barmintor>ajs6f: ask me again on Monday, I'm still sorting out stuff.16:05
<ajs6f>barmintor: Fair enough. Good luck.
barmintor: I'll start working on the 3store/filesystem reimpl.16:06
* yinlin leaves16:15
* dwilcox leaves16:16
* mjordan leaves16:33
* ajs6f leaves
* mikeAtUVa leaves16:50
* bseeger leaves17:02
* mikeAtUVa joins17:12
* acoburn leaves17:20
* mikeAtUVa leaves17:29
* jhs leaves17:36
* dchandekstark leaves18:11
* dchandekstark joins19:12
* dchandekstark leaves19:18
* dchandekstark joins20:03
* dchandekstark leaves20:41
* dchandekstark joins21:42
* dchandekstark leaves21:47
* f4jenkins joins23:13

Generated by Sualtam