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

Using timezone: Eastern Standard Time
* ksclarke leaves01:20
* jrgriffiniii joins01:54
* jrgriffiniii leaves01:56
* jrgriffiniii joins07:04
* jrgriffiniii leaves07:05
* dwilcox joins07:26
* ajs6f joins07:40
* dhlamb joins08:21
* mhwood joins08:33
* escowles leaves08:35
* jrgriffiniii joins08:50
* awead joins08:51
* jrgriffiniii leaves08:52
* awead leaves08:59
* scossu joins09:00
<ajs6f>awoods_: Do you need help with FCREPO-1537?09:02
* ksclarke joins09:04
* scossu leaves
* awead joins
* mhwood leaves09:21
<awoods>ajs6f: that would be great.
<ajs6f>awoods: Stop changing your name. It's shifty.09:22
awoods: I can either look at the one you just made for me or help you with 1537.
Which?
* mhwood joins
<awoods>ajs6f: 1537 is in-flight. It would be nice to have it land.
<ajs6f>awoods: What is you problem with it? Do you understand why I don't accept your current solution?09:23
awoods: Do you want me to factor IdentifierConverter into the two JAX-RS classes/
?
<awoods>ajs6f: I have not dug into your latest response yet... working on something else at the moment. But it seems that you have a specific idea of what you would like to happen there... and I support you making it happen.09:24
<ajs6f>awoods: See, that wouldn't be me helping you. That would be me doing your ticket. And I want _you_ to grow, awoods.09:25
<awoods>ajs6f: thanks, buddy.
<ajs6f>awoods: It's got to hurt to heal.09:26
awoods: Seriously, if you don't agree with my comments then I don't want to spend time doing that work. Decide whether you agree with me before I do anything.
<awoods>ajs6f: ok, stay tuned
<ajs6f>awoods: I will knock off the other ticket while you do that.09:27
* yinlin leaves09:49
<ajs6f>awoods: We are apparently adding inbound references to RDF returned for almost any request: ReferencesRdfContext.09:56
awoods: Do you know if that was in response to some genuine user need, or just because we could?
* mhwood leaves09:57
<awoods>ajs6f: There are a few (undocumented) use cases that drove its creation, around providing more context in single responses.10:01
<ajs6f>awoods: Well, it completely destroys the idea of single-subject, for obvious reasons.
<awoods>ajs6f: fyi, we expose it here: https://wiki.duraspace.org/display/FEDORA4x/RESTful+HTTP+API+-+Containers#RESTfulHTTPAPI-Containers-GETRetrievethecontentoftheresource10:02
ajs6f: with http://fedora.info/definitions/v4/repository#InboundReferences
<ajs6f>awoods: Yeah, I realized we had a header for it. Just not whether it was an LDP thing or our thing.
<awoods>ajs6f: our thing
<ajs6f>awoods: If we want to keep that (and I accept that it has its value) then there isn't much point in tests for single-subject. It just becomes a "principle" to which we theoretically adhere as much as we can.10:03
<awoods>ajs6f: The principle can be a little stronger: we give you single subject unless you ask for more.10:04
<ajs6f>awoods: Agreed. And I suppose _that_ can be tested.
<awoods>ajs6f: this also relates to the etag discussions10:05
<ajs6f>awoods: Eh?
awoods: Oh, changes coming from a different resource.
<awoods>ajs6f: i.e. if another inbound reference is created, the etag should change.
<ajs6f>awoods: That way can lead to madness. "If anything anywhere on the Web that refers to this thing changes, its ETag must change."10:06
awoods; We have to have a clear public algorithm.
<awoods>ajs6f: cbeer has his heels firmly planted on this one.
<ajs6f>awoods: About ETags?
<awoods>ajs6f: yes10:07
<ajs6f>awoods: I didn't know he cared.
<awoods>ajs6f: cbeer... very caring
<ajs6f>awoods: Does he like the current behavior?
<awoods>ajs6f: I will let him speak for himself, but my understanding is that he agrees that if a potential representation of the resource changes, its etag should change.10:08
<ajs6f>awoods: Is that the current behavior?
<awoods>ajs6f: "GET with inbound references" is a potential representation
<ajs6f>awoods: For the last time, IS THAT THE CURRENT BEHAVIOR?10:09
<awoods>ajs6f: we do not update etags for all scenarios, no.
ajs6f: due to our etag algorithm, not due to intentional design.
<ajs6f>awoods: _Then I look forward to cbeer's PR._
<awoods>ajs6f: https://jira.duraspace.org/browse/FCREPO-149810:10
<ajs6f>awoods: No PR there.
<awoods>ajs6f: nothing
<ajs6f>awoods: A void, dark and windy, like my guts after too many bean tacos.
awoods: Do you have any sense that anyone (other than me) is interested in reworking the RDF generation framework for Java 8.10:27
?
<awoods>ajs6f: I suspect there is much interest... acoburn is probably the closest to turning that interest into implementation.10:29
<ajs6f>awoods: Yeah, I suspected as much. Then I will leave it be.
awoods: It's kind of a mess right now. It's very hard to figure out where any given triple got generated.10:30
<awoods>ajs6f: is that something that rewriting with Java8 structures will help?
<ajs6f>awoods: Depends on how you do it, but it certainly could. Just the concision alone would be good. If you look at what I did in the original Java 8 branch, we saved four digits of linecount in the RDF generation stuff.10:31
awoods: And I was nowhere near taking full advantage.
awoods: RdfStream was a special-purpose impl of the current Java 8 Streams API, because cbeer and I didn't have the Streams API available to us.10:32
<awoods>ajs6f: I agree, the RDF generation logic is a bit tricky to trace. I usually fall back on: https://github.com/fcrepo4/fcrepo4/blob/master/fcrepo-http-api/src/main/java/org/fcrepo/http/api/ContentExposingResource.java#L19110:35
<ajs6f>awods: Yes, that's where I'm wallowing now, because I'm trying to kill fcr:metadata-subject triples. That says something really bad about us, when in order to understand some business logic in a kernel impl, you have to go to an HTTP API module and trace around, hoping to find some clue as to what's going on underneath.10:36
* dwilcox leaves11:20
* escowles joins11:26
* dwilcox joins11:32
* awoods_ joins12:03
* awoods leaves12:06
* awoods__ joins12:24
* awoods_ leaves12:27
* jrgriffiniii joins12:28
* jrgriffiniii leaves12:29
* github-ff joins13:08
[fcrepo4] ajs6f created FCREPO-1640 (+1 new commit): http://git.io/vmK5E
fcrepo4/FCREPO-1640 8b27b56 ajs6f: Fix for FCREPO-1640
* github-ff leaves
* github-ff joins13:10
[fcrepo4] ajs6f opened pull request #844: Fix for FCREPO-1640 (master...FCREPO-1640) http://git.io/vmKdD
* github-ff leaves
<ajs6f>awoods: ^^^ a little rough, but tak a look13:11
<awoods>ajs6f: thanks...13:13
* travis-ci joins13:21
fcrepo4/fcrepo4#3850 (FCREPO-1640 - 8b27b56 : ajs6f): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/commit/8b27b561b56b
Build details : https://travis-ci.org/fcrepo4/fcrepo4/builds/71449972
* travis-ci leaves
<ajs6f>escowles: Do you get what I'm saying on that PR? Use a function from property -> iterator<triple> and you'll be okay.13:42
<escowles>ajs6f: yep, that makes sense
<ajs6f>escowles; Rock on like https://en.wikipedia.org/wiki/Uluru.13:43
* awead leaves15:03
* awead joins15:04
* dwilcox leaves15:20
* github-ff joins16:01
[fcrepo4] escowles pushed 1 new commit to fcrepo-1497-indirect-container: http://git.io/vm6xG
fcrepo4/fcrepo-1497-indirect-container b0c99a7 Esmé Cowles: Addressing test and session comments
* github-ff leaves
* dhlamb leaves16:12
* travis-ci joins16:15
fcrepo4/fcrepo4#3852 (fcrepo-1497-indirect-container - b0c99a7 : Esmé Cowles): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/compare/ed08d37a4ed1...b0c99a781a5e
Build details : https://travis-ci.org/fcrepo4/fcrepo4/builds/71474045
* travis-ci leaves
* jgpawletko leaves16:23
* JPD_ joins16:29
* JPD_ leaves16:30
* jgpawletko joins16:42
* the_mgt joins16:44
good evening16:45
<ajs6f>Good night.16:58
* ajs6f leaves
* jgpawletko leaves17:43
* jgpawletko joins18:22
* awoods_ joins19:12
* awoods leaves19:16
* the_mgt_ joins19:49
* the_mgt leaves19:52
* escowles leaves20:02
* ksclarke leaves20:05