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

Using timezone: Eastern Standard Time
* thomz joins02:38
* chadmills leaves04:18
* cmmills joins
* esm_ joins08:18
* esm_ leaves08:36
* esm_ joins08:44
* esm_ leaves09:02
* mikeAtUVa joins09:04
* bseeger joins09:18
* peichman joins09:22
* bseeger leaves09:38
* github-ff joins09:41
[fcrepo-camel] ruebot opened pull request #116: Update Danny's GitHub username. (master...danny) https://git.io/v6U0Q
* github-ff leaves
* ajs6f joins09:42
* github-ff joins09:45
[fcrepo-camel-toolbox] ruebot opened pull request #102: Update Danny's GitHub username. (master...danny) https://git.io/v6UEz
* github-ff leaves
* dchandekstark joins09:46
* github-ff joins09:47
[fcrepo-java-client] ruebot opened pull request #15: Update Danny's GitHub username. (master...danny) https://git.io/v6UEM
* github-ff leaves
* github-ff joins10:11
[fcrepo-camel-toolbox] acoburn pushed 1 new commit to master: https://git.io/v6U2X
fcrepo-camel-toolbox/master 089aaf4 Nick Ruest: Update Danny's GitHub username. (#102)
* github-ff leaves
* github-ff joins10:12
[fcrepo-camel] acoburn pushed 2 new commits to master: https://git.io/v6U21
fcrepo-camel/master abb5a57 nruest: Update Danny's GitHub username.
fcrepo-camel/master 0170b79 Aaron Coburn: Merge pull request #116 from ruebot/danny...
* github-ff leaves
* github-ff joins
[fcrepo-java-client] acoburn pushed 2 new commits to master: https://git.io/v6U2S
fcrepo-java-client/master 0a895ea nruest: Update Danny's GitHub username.
fcrepo-java-client/master 4ad944f Aaron Coburn: Merge pull request #15 from ruebot/danny...
* github-ff leaves
<dchandekstark>anybody up for a question? ajs6f ;)10:15
<ajs6f>dchandekstark: Sure, you can always ask. Whether you get a useful answer is another question.10:16
<dchandekstark>ajs6f: fair enough
i think i might be seeing a race condition in the PUTing of a non-RDF source
and trying to track down why it might behappening10:17
my fcrpo log shows two different threads doings a PUT on the same binary resource URI
one sec apart
<ajs6f>dchandekstark: Okay. That would be at the initiative of client, presumably?10:19
<dchandekstark>ajs6f: not intentionally
* dhlamb joins
<dchandekstark>that is, that would not be normal
<ajs6f>dchandekstark: Intentionally or not, can you assure yourself that is it not, in fact, the case?
<dchandekstark>ajs6f: i have not yet so assured myself10:20
<ajs6f>dchandekstark: But you have a resource in hand that seemed to create the problem? Maybe we can use that to set up a test case.
<dchandekstark>ajs6f: i may have to hop over to hydra land to see if anyone knows about this
i suspect if it's at the client level it's in the ruby ldp code10:21
<ajs6f>dchandekstark: If we can isloate to one side of the req-res cycle, that would be a good start.
<dchandekstark>ldp client that is
ajs6f: let me look into ldp first, b/c i think it does have some retry stuff in there
* th5 joins
<dchandekstark>AFAIK there's nothing odd about the object or file10:22
* travis-ci joins
fcrepo4-exts/fcrepo-java-client#46 (master - 4ad944f : Aaron Coburn): The build passed.
Change view : https://github.com/fcrepo4-exts/fcrepo-java-client/compare/03c561e29605...4ad944f8d86b
Build details : https://travis-ci.org/fcrepo4-exts/fcrepo-java-client/builds/149197654
* travis-ci leaves
<ajs6f>dchandekstark: "Retry stuff" might be interesting, yes. PUT is theoretically idempotent, so this may be a situation where not-so-good client code uncovered bad server code. Yay?10:23
* travis-ci joins10:27
fcrepo4-exts/fcrepo-camel#273 (master - 0170b79 : Aaron Coburn): The build passed.
Change view : https://github.com/fcrepo4-exts/fcrepo-camel/compare/7b115c9094b5...0170b7939be1
Build details : https://travis-ci.org/fcrepo4-exts/fcrepo-camel/builds/149197511
* travis-ci leaves
<dchandekstark>ajs6f: i'm not seeing anything at the client level10:30
ruby ldp - faraday - net/http
* coblej joins10:31
* dhlamb leaves
<dchandekstark>ajs6f: https://gist.github.com/dchandekstark/87699e66d04d8524ca66ff28da10291510:34
* travis-ci joins
fcrepo4-exts/fcrepo-camel-toolbox#261 (master - 089aaf4 : Nick Ruest): The build passed.
Change view : https://github.com/fcrepo4-exts/fcrepo-camel-toolbox/compare/dbe1247946d5...089aaf442ca4
Build details : https://travis-ci.org/fcrepo4-exts/fcrepo-camel-toolbox/builds/149197479
* travis-ci leaves
<ajs6f>dchandekstark: All that segment of log tells me is that something really genuinely is trying to do that PUT twice.10:35
dchandekstark: What version of fcrepo4 are you using?10:36
<dchandekstark>ajs6f: 4.5.1
<ajs6f>dchandekstark: https://github.com/fcrepo4/fcrepo4/blob/fcrepo-4.5.1/fcrepo-http-api/src/main/java/org/fcrepo/http/api/FedoraLdp.java#L36710:37
<dchandekstark>ajs6f: not sure why that isn't there ...
<ajs6f>dchandekstark: I don't see any way that is a server-side issuance of the second request,10:38
dchandekstark: What isn't there?
<dchandekstark>ajs6f: nm
<ajs6f>dchandekstark: Eh?10:39
<dchandekstark>ajs6f: sorry, i meant scratch my "not sure"10:40
<ajs6f>dchandekstark: Ah, okay. The problem is that seeing two log statements like that indicates that two threads (such as are named in the log statements) are coming in with two requests. That's not really something the server can cause to happen.
* diegopino joins10:41
<dchandekstark>ajs6f: i understand - but shouldn't one result in a conflict error?10:42
<ajs6f>dchandekstark: What are you actually getting?
<dchandekstark>ajs6f: no error, just those INFO statements10:43
<ajs6f>dchandekstark: And the resource?
<dchandekstark>ajs6f: the resource exists
<ajs6f>dchandekstark: ; That sounds right. The second PUT just overwrites the first. Unless there is some kind of versioning enabled?10:44
<dchandekstark>ajs6f: i see, ok
<ajs6f>dchandekstark: Are you trying to set up a situation in which the resource is protected in some sense?10:45
<dchandekstark>ajs6f: no, there was an error higher in the stack
which led me to investigate what happened
<ajs6f>dchandekstark: So we're cool?10:46
<dchandekstark>ajs6f: i think so, i'd still like to know what caused the second put, but seems clear that's not fedora
<ajs6f>dchandekstark: So , here: Kyrille DeBose10:48
Let's try that again.
dchandekstark: So , here: https://github.com/fcrepo4/fcrepo4/blob/fcrepo-4.5.1/fcrepo-http-api/src/main/java/org/fcrepo/http/api/FedoraLdp.java#L37310:49
is where the bitstream is getting overwritten, and that's defined here: replaceResourceBinaryWithStream
is where the bitstream is getting overwritten, and that's defined here: https://github.com/fcrepo4/fcrepo4/blob/fcrepo-4.5.1/fcrepo-http-api/src/main/java/org/fcrepo/http/api/ContentExposingResource.java#L583
dchandekstark: And if you want, you can grab the source and put some more logging in there to get as much info as possible about wha the client is doing. Honestly, though, you'll probably do just as well and much quicker to just turn up the logging for your container all the way.10:51
<dchandekstark>ajs6f: ok, thx10:55
* dhlamb joins10:56
<ajs6f>dchandekstark: Sorry we did get to the root of it— let me know if you don't nail it. We can call in some committers who know Hydra (unlike me).10:57
<dchandekstark>ajs6f: we're kind of suspicious now that it might have been a hiccup in tomcat
although that's a sort of wild guess10:58
<ajs6f>dchandekstark: We have five cats and I don't think I've ever heard one hiccup. Lots of sneezing and hairballs and the occasional puking incident, but no hiccups. Never thought about that before. Weird.10:59
<dchandekstark>ajs6f: like i said wild guess
* thomz leaves11:02
* github-ff joins11:22
[fcrepo4-vagrant] awoods pushed 1 new commit to master: https://git.io/v6U1a
fcrepo4-vagrant/master 9a2d1b4 Yinlin Chen: 1. remove unused variables (#51)...
* github-ff leaves
<ajs6f>awoods: Did you see what acoburn did in fcrepo-transform? I still want to merge the functionality in -transform into -camel, but the LDCache thing is going to answer some questions for people confused about how to index linked data.11:24
<awoods>ajs6f: I saw that he did that work, but I have not yet reviewed it.11:25
ajs6f: I assume you like what you see?
<ajs6f>awoods: You won't find anything surprising. Please review if you want to, because I really want to merge that. I have reviewed it.
<awoods>ajs6f: I will give it a look later today... after the next 4hours of meetings.11:26
<ajs6f>awoods: You need one of these: http://www.nydailynews.com/news/national/driver-busted-dummy-passenger-seat-hov-lane-scam-article-1.149556811:27
awoods: Send that to meetings instead.
* github-ff joins
[fcrepo-webapp-plus] awoods pushed 2 new commits to master: https://git.io/v6UMW
fcrepo-webapp-plus/master cac0e02 Aaron Coburn: Properly configure activemq directory...
fcrepo-webapp-plus/master 30d1f15 Andrew Woods: Merge pull request #45 from acoburn/fcrepo-2102...
* github-ff leaves
* bseeger joins11:31
* acoburn joins
* acoburn1 joins11:34
* acoburn leaves11:35
* acoburn2 joins
* acoburn2 leaves
* acoburn joins11:36
* travis-ci joins
fcrepo4-exts/fcrepo-webapp-plus#149 (master - 30d1f15 : Andrew Woods): The build passed.
Change view : https://github.com/fcrepo4-exts/fcrepo-webapp-plus/compare/dea00db472cf...30d1f15cfc7d
Build details : https://travis-ci.org/fcrepo4-exts/fcrepo-webapp-plus/builds/149220423
* travis-ci leaves
<acoburn>awoods: the current README for fcrepo-transform isn't entirely accurate
awoods: I haven't been able to inject marmotta endpoints as is described11:37
<awoods>acoburn: on a call...
<acoburn>awoods: I'll add a comment to the PR
<awoods>acoburn: thanks
<ajs6f>acoburn: Is that something to be fixed in a newar-term future commit?
<acoburn>ajs6f: I can't get a util:list to be injected properly with Spring
<ajs6f>acoburn: By id?
* acoburn1 leaves
<acoburn>ajs6f: yeah, the ID matches and the value-type matches, but it doesn't resolve11:39
ajs6f: my current tack is to simple reimplement this in OSGi
ajs6f: which uses less code and is more flexible
ajs6f: but the current PR does _more_ than the existing fcr-transform does
<ajs6f>acoburn: Wait, inside an fcrepo-transform that you still propose will be distributed as part of the core webapp?11:40
<acoburn>ajs6f: it just doesn't do _everything_
<ajs6f>acoburn: Or you are taking the whole thing outside the core webapp?
<acoburn>ajs6f: I'm taking the entire thing out
<ajs6f>acoburn: In this PR?
<acoburn>ajs6f: no
ajs6f: this PR is simply a way to allow us to move forward with jena 3.x
<ajs6f>acoburn: Okay, now I'm with you, I think. First, this PR, which will do single point LDCache, with the single point being your repo, then the full monte.11:41
<acoburn>ajs6f: yes, exactly
afk bbl
* ajs6f leaves
* acoburn leaves
* coblej leaves12:12
* dwilcox joins
* dhlamb leaves12:13
* peichman leaves12:18
* bseeger leaves12:22
* diegopino leaves12:30
<barmintor>dwilcox: is there an instructor call?12:31
<awoods>barmintor: yes...12:32
barmintor: we are just getting off of a call
* tpendragon joins
awoods dwilcox: Can one of you give me permission to join the call? It's been requesting for a bit12:33
* dchandekstark leaves12:41
* cmmills leaves12:42
<barmintor>awoods: I'm here. I have a call with ruebot at 3pm, but I can talk12:46
awoods: (I'm on the call as well)
<awoods>barmintor: let's talk after the current call... if you are available.12:47
* dhlamb joins12:51
* bseeger joins12:58
* dchandekstark joins13:10
* jdukart joins13:18
hey awoods, dwilcox fyi your RC3 for 4.6.0 has 4.5.2 links: https://github.com/fcrepo4/fcrepo4/releases13:19
<awoods>jdukart: yes, the 4.5.2-SNAPSHOT version will turn into the name of the 4.6.0 release.13:20
jdukart: meaning... its ok.
* ajs6f joins13:53
* dhlamb leaves14:30
* dhlamb joins14:34
* manez joins14:51
dwilcox: working on getting a cCLA signed by the IF for Fedora. Is this the one: https://wiki.duraspace.org/display/DSP/Contributor+License+Agreements ?14:52
* peichman joins15:05
* peichman leaves15:06
* peichman joins
* peichman leaves15:09
* dhlamb leaves15:17
* bseeger leaves15:27
* manez leaves
* bseeger joins15:30
* dchandekstark leaves
* jdukart leaves15:33
* bseeger leaves15:43
* bseeger joins15:49
* bseeger leaves15:51
* bseeger joins15:52
* github-ff joins16:02
[fcrepo-transform] ajs6f created ForAcoburn (+7 new commits): https://git.io/v6T0W
fcrepo-transform/ForAcoburn ac89a80 Aaron Coburn: Replace Jena LDCache backend with a linkeddata backend...
fcrepo-transform/ForAcoburn 0d26952 Aaron Coburn: support injection of multiple ldcache endpoints
fcrepo-transform/ForAcoburn 83cce7c Aaron Coburn: update javadocs
* github-ff leaves
* dchandekstark joins16:05
<barmintor>slowly making progress https://docs.google.com/document/d/1ws1x7bry3RqB305XxtqIgru5c8heTDDk03jkbCA7iW4/edit#16:11
* github-ff joins16:17
[fcrepo-transform] ajs6f force-pushed ForAcoburn from 3bd1a7e to 0858854: https://git.io/v6TzU
fcrepo-transform/ForAcoburn 0858854 ajs6f: Well, it may be crappy, but it works for now. Patches to my patch welcome.
* github-ff leaves
* dwilcox leaves16:18
<ajs6f>awoods: I'm going to wait to see acoburn's reaction to my PR-for-his-PR. But other than that, I'm ready to merge his stuff, so let me know when you are comfortable with same.16:20
* travis-ci joins16:22
fcrepo4-exts/fcrepo-transform#98 (ForAcoburn - 3bd1a7e : ajs6f): The build passed.
Change view : https://github.com/fcrepo4-exts/fcrepo-transform/compare/ac89a8006884^...3bd1a7eb9a07
Build details : https://travis-ci.org/fcrepo4-exts/fcrepo-transform/builds/149290367
* travis-ci leaves
<ajs6f>barmintor: ping16:24
<barmintor>ajs6f: hiya16:25
* travis-ci joins
fcrepo4-exts/fcrepo-transform#99 (ForAcoburn - 0858854 : ajs6f): The build passed.
Change view : https://github.com/fcrepo4-exts/fcrepo-transform/compare/3bd1a7eb9a07...08588547a7aa
Build details : https://travis-ci.org/fcrepo4-exts/fcrepo-transform/builds/149294030
* travis-ci leaves
<ajs6f>barmintor: Is it sorta kinda basically like fixity works like versioning, except instead of states, it's fixity checks? But the whole "special kind of container with added magic" is the same?
barmintor: And the added magic is Memento falvor?16:26
<barmintor>ajs6f: that is the gist, yeah
<barmintor>ajs6f: with some spec of behavior around Digest headers
<ruebot>ajs6f: is it making sense? barmintor and i just spent another hour with it.16:27
<barmintor>ajs6f: and leaving it to the impl to decide what a fixity resource serializes as
<ajs6f>barmintor: That's because of the difference between transmission and on-demand, right? The header is for transmission.
barmintor: Oh, hell yes. The first rule of fixity is you don't talk about how you write down fixity.
<barmintor>ajs6f: yeah, header for transmission, GET for most recent, PUT or POST somewhere for a new check16:28
<ajs6f>barmintor: That would be to fix the unfixable.
ruebot: It is, but I am a bad audience. I have history with you two. You need to try it on someone who comes from scratch, or at least is more scratchy. Maybe Felix the Cat.
ruebot:barmintor: One thing I really hope we can do here is cut loose the discussion about fixity. It's not just checksums and hashes. There's a whole world of fixity over motion, fixity over streams, fixity over data too big to process in batch, semantic fixity, etc.16:30
ruebot++ barmintor++ # keep up the good work!
* ruebot braces self
<awoods>ajs6f: I will give acoburn's PR a look today.16:37
<ajs6f>ruebot:barmintor: Two things: I like what you're doing with Memento, but if you can, it's probably worth getting Rob or someone to glance over it. Maybe you're already doing that. It seems to me that y'all are claiming that a fixity check is a representation of the resource in a pretty strong sense (i.e. that it could legitimately be offered at an HTTP identifier for that resource, which I think is legit, but is definitely something to thin
<barmintor>ajs6f: we're trying to say that it's a separate resource that describes the LDP-NR
ajs6f: so we must have some bad language in there
<ajs6f>awoods: https://www.youtube.com/watch?v=8_Imn-LTs7s
<ruebot>ajs6f: yeah, so far it is just the two of us. but, will definitely make sure we pull rob in for peer review at the very least.
<ajs6f>barmintor: I get that, I'm saying that by offering it as a thing that is on the map-to end of the TimeMap, the inference that I described is available. IOW, if you give me X and a TimeMap for X and one of the things to which that TimeMap points is Y, I think I can infer that Y is a representation of X at a given time in the strong sense I described, because under most normal circumstances in which people use memento, Y actually -was- a repr16:40
In that strong sense.16:41
<barmintor>no no
the LDP-Cf is a timemap for the LDP-Rf, not the LDP-NR
<ajs6f>barmintor: Okay, in that case, yes, you do want to wordsmith that a bit. {grin}16:42
<ajs6f>barmintor: "The LDP-Cf is also a Timemap per https://tools.ietf.org/html/rfc7089" => "The LDP-Cf is also a Timemap per https://tools.ietf.org/html/rfc7089 FOR A THING"
barmintor:ruebot: One other thing: inthe versioning discussion, we eventually decided that since the HTTP API describes representations, not resources, we don't decide how you implement history. It might be mutable (e.g. Rob's use case at that time). It might not be. That's not the API's concern. Is that going to be the approach you take here? Is that your 5.3 and 6.5?16:47
<barmintor>ajs6f: yeah, we're porting that accomodation over whole cloth16:51
ajs6f: the whole thing really boils down to "there's a fixity resource for a LDP-NR, and it's versioned"
"see the versioning spec"
<ajs6f>barmintor: Cute. getting the versioing spec to do your work for you! :)16:56
* mikeAtUVa leaves
<ruebot>ajs6f, barmintor: yeah, we still need to sort out the hierarchy/dependency chain for the specs16:57
* ruebot disappears
<ajs6f>barmintor: It's good, tho. I like this way of thinking, and I would be happy to see it used. It almost makes me wonder if there is a way we can talk about atomic operations as ordered containers of tupled representions and URIs, or something.16:58
Maybe containers with subcontainers that project the tupling.
Mrph, got to go. Nice to hear this work is rolling along!16:59
* barmintor waves
* ajs6f leaves
* bseeger leaves17:06
* th5 leaves17:17
* dchandekstark leaves17:20
* dchandekstark joins17:24
* dchandekstark leaves18:16
* dchandekstark joins19:16
* dchandekstark leaves19:22
* dchandekstark joins20:44
* peichman joins22:22
* peichman leaves22:24
* dhlamb joins
* dchandekstark leaves22:29
* dchandekstark joins22:49
* dchandekstark leaves22:54
* dhlamb leaves23:34
* dhlamb joins23:43
* dhlamb leaves00:31
* dchandekstark joins00:51
* dchandekstark leaves00:57

Generated by Sualtam