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

Using timezone: Eastern Standard Time
* kaarefc joins00:55
* kaarefc leaves01:45
* kaarefc joins03:20
* fasseg joins07:06
* kaarefc leaves08:22
* awoods leaves08:26
* awoods joins09:03
<pivotal-bot>A. "Killmation" Soroka added comment: "Um... yeah. org.fcrepo.kernel.translations... I don't get it?" https://www.pivotaltracker.com/story/show/6139121609:09
Andrew Woods added comment: "Sounds good. I thought you were implying moving the new package out of kernel." https://www.pivotaltracker.com/story/show/6139121609:12
* ajs6f joins09:14
* github-ff joins
[fcrepo-jms-indexer-pluggable] escowles pushed 1 new commit to test-query-bugfix2: http://git.io/IA3wkw
fcrepo-jms-indexer-pluggable/test-query-bugfix2 77d7454 Esmé Cowles: Making file-based tests less brittle, introducing new TestIndexer class with associated (non-functional) IndexerGroupIT tests for more direct testing of IndexerGroup functionality
* github-ff leaves
* github-ff joins
[fcrepo4] ajs6f pushed 1 new commit to MoreIteration: http://git.io/GOt6-A
fcrepo4/MoreIteration fb0653b ajs6f: Additional logging
* github-ff leaves
<mikeAtUVa>Was the choice of Turtle as the default response type for fedora 4 web service requests deliberate?
<ajs6f>awoods: ^^^ https://www.pivotaltracker.com/story/show/61166142
<pivotal-bot>feature: Alter FedoraRepositoryNodeTypes, FedoraRepositoryWorkspaces, FedoraIdentifiers and FedoraRepositoryNamespaces to use iterating RDF generation (rejected) / owner: A. "Killmation" Soroka
* gregjansen leaves
<ajs6f>There's yr logging.09:15
<pivotal-bot>A. "Killmation" Soroka finished "Alter FedoraRepositoryNodeTypes, FedoraRepositoryWorkspaces, FedoraIdentifiers and FedoraRepositoryName..." https://www.pivotaltracker.com/story/show/61166142
<ajs6f>mikeAtUva: I would prefer RDF/XML, myself, not because it's _good_ but because librarians find it unthreatening.09:16
<mikeAtUVa>ajs6f: I would prefer something that most web browers choose to render.
<ajs6f>RDF/XML wins there, too.
<escowles>i think RDF/XML would be a better fit for the indexer (and totally agree about browser-renderable being good)
<ajs6f>We could include a stylesheet, for style.
Nuff said. I'll write a ticket for awoods.09:17
<escowles>but i think a lot of RDF people hate RDF/XML, find it unreadable, etc.
<ajs6f>IT is bad.
Sorry _it_ is bad. RDF/XML
But we make choices for more people than RDF people.
<escowles>ajs6f: yes, RDF/XML is terrible, but i still like it more than n3/turtle
<awoods>mikeAtUVa: I thought our content-negotiation sent text/html to the browser?
<ajs6f>If it were up to me, I'd write in Scala and Clojure and publish nothing but OWL to the endpoints.
<mikeAtUVa>ajs6f: people who have opinions about RDF serialization probably have the ability to set request headers.
awoods: perhaps for some service calls...09:18
awoods: mostly I'm looking at the version ones, and others for which we haven't made an HTML template...
<ajs6f>We have two defaults here: RDF or HTML, and from RDF< which format?
<mikeAtUVa>awoods: perhaps I should have made an HTML template;)
<ajs6f>I say RDF?XML, because it's the least bad. Aim high!09:19
<awoods>mikeAtUVa: yes, I think making the template is the pattern used for other endpoints.
<escowles>we should definitely have a template for HTML output of the version info for the browser09:20
<awoods>mikeAtUVa: Are there other endpoints that need templates? If so, feel free to create some tickets.
<mikeAtUVa>awoods, escowles: alrighty... I'll get right on it.09:21
<pivotal-bot>A. "Killmation" Soroka added "Confirm defaulting behavior for conneg at our JAX-RS endpoints" https://www.pivotaltracker.com/story/show/61473832
<ajs6f>afk bbl
<mikeAtUVa>awoods: I haven't noticed any other glaring omissions in the HTML experience...09:22
<awoods>mikeAtUVa: That is a good thing... or else you would be obligated to create tickets.
<mikeAtUVa>awoods; I'm happy enough to write tickets.09:23
<awoods>mikeAtUVa: Thanks. The team spirit is appreciated.09:24
<pivotal-bot>Mike Durbin added "Write a velocity template for the fcr:versions endpoint." https://www.pivotaltracker.com/story/show/6147423609:26
Mike Durbin edited "Write a velocity template for the fcr:versions endpoint." https://www.pivotaltracker.com/story/show/6147423609:27
Mike Durbin estimated "Write a velocity template for the fcr:versions endpoint." as 2 points https://www.pivotaltracker.com/story/show/61474236
* travis-ci joins09:28
[travis-ci] futures/fcrepo4#1272 (MoreIteration - fb0653b : ajs6f): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/43531308f9c9...fb0653b00875
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/14546147
* travis-ci leaves
* ermadmix joins09:30
<bljenkins>Project fcrepo-fixity-corrupter build #492: SUCCESS in 1 min 2 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/492/09:34
Project fcrepo-kitchen-sink build #673: STILL UNSTABLE in 2 min 20 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/673/09:37
* tecoripa joins
<bljenkins>Project fcrepo-jms-indexer-pluggable build #281: UNSTABLE in 4 min 20 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-jms-indexer-pluggable/281/09:38
<pivotal-bot>Esme Cowles added comment: "I have updated https://github.com/futures/fcrepo-jms-indexer-pluggable/pull/14 to make the remaining file-ba..." https://www.pivotaltracker.com/story/show/6135191409:39
<tecoripa>question about pull request protocol: once we submit a pull request and put the URL to the request in the pivotal ticket, should I then finish the ticket?09:42
* github-ff joins
[fcrepo4] awoods pushed 1 new commit to master: http://git.io/FP3NGA
fcrepo4/master 3ac7aab ajs6f: Add streaming RDF to following endpoints...
* github-ff leaves
<pivotal-bot>Andrew Woods added comment: "Resolved with: https://github.com/futures/fcrepo4/commit/3ac7aab8271c26913d6d820eaaced8b249ee2ebf" https://www.pivotaltracker.com/story/show/6116614209:43
Andrew Woods delivered "Alter FedoraRepositoryNodeTypes, FedoraRepositoryWorkspaces, FedoraIdentifiers and FedoraRepositoryNamespaces to..." https://www.pivotaltracker.com/story/show/61166142
* github-ff joins
[fcrepo4] awoods closed pull request #166: More iteration (master...MoreIteration) http://git.io/gYTjrQ
* github-ff leaves
* ksclarke leaves09:54
* ksclarke joins
* ksclarke leaves
* ksclarke joins
* travis-ci joins09:57
[travis-ci] futures/fcrepo4#1274 (master - 3ac7aab : ajs6f): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/49613d5a18d4...3ac7aab8271c
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/14547501
* travis-ci leaves
<pivotal-bot>Andrew Woods accepted "Alter FedoraRepositoryNodeTypes, FedoraRepositoryWorkspaces, FedoraIdentifiers and FedoraRepositoryNamespaces to ..." https://www.pivotaltracker.com/story/show/61166142
<ajs6f>tecoripa: Finishing the ticket tells the reviewer to start review, so yes.
<tecoripa>ajs6f: ok, thanks09:58
<pivotal-bot>Scott Prater finished "Create repository.json with ServletContainerAuthenticationProvider" https://www.pivotaltracker.com/story/show/61409536
A. "Killmation" Soroka added comment: "@escowles You want me to take this?" https://www.pivotaltracker.com/story/show/6135191409:59
<awoods>tecoripa: As a process note, I typically will "Deliver" and "Accept" tickets that have no review comments. If there are comments for the ticket owner to address, I tend to "Reject" the ticket to provide a visual indicator that the ticket is back in your court (not because the ticket is actually "rejected")10:02
* github-ff joins
[fcrepo4] mikedurbin opened pull request #170: Supressed links to modify nodes on the frozen node HTML view. (master...versioning-3) http://git.io/fMu_FQ
* github-ff leaves
<ajs6f>awoods* is an accepting personality.
<pivotal-bot>Mike Durbin added comment: "https://github.com/futures/fcrepo4/pull/170" https://www.pivotaltracker.com/story/show/6113303010:03
Mike Durbin estimated "Better expose nodes of type nt:frozenNode." as 2 points https://www.pivotaltracker.com/story/show/61133030
Mike Durbin finished "Better expose nodes of type nt:frozenNode." https://www.pivotaltracker.com/story/show/61133030
Mike Durbin started "Write a velocity template for the fcr:versions endpoint." https://www.pivotaltracker.com/story/show/61474236
<ajs6f>afk bbs
<bljenkins>Project fcrepo-fixity-corrupter build #493: SUCCESS in 1 min 53 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/493/10:04
<pivotal-bot>Esme Cowles added comment: "@ajs6f Please do -- I'm not sure what's going wrong with the IT, but I think it's very close to what we were..." https://www.pivotaltracker.com/story/show/6135191410:09
* ajs6f leaves10:11
* osmandin joins10:14
* ajs6f joins
* ajs6f leaves10:22
* ajs6f joins10:23
<pivotal-bot>A. "Killmation" Soroka edited "Build failing: fcrepo-jms-indexer-pluggable" https://www.pivotaltracker.com/story/show/6135191410:26
<awoods>All: Some interesting results are coming out of the single-node testing: https://wiki.duraspace.org/display/FF/Single-Node+Performance+Testing#Single-NodePerformanceTesting-IngestTestMatrix10:32
<bljenkins>Project fcrepo-fixity-corrupter build #494: SUCCESS in 2 min 28 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/494/10:33
<ajs6f>awoods: Trying to use your tip about pulling PRs as well as branches. It worked ysterday, but today I get:10:34
Your configuration specifies to merge with the ref 'pr/14'
from the remote, but no such ref was fetched.
ever seen that?
awoods: So fcrepo4 is both twice as fast and twice as slow as fcrepo3.10:35
<awoods>ajs6f: I do not think I have seen it. One thing to note, I tend to "checkout master" then "branch -D pr-xxx" if the PR has been updated and I want to pull down the new latest.
<ajs6f>k, will try that
<awoods>ajs6f: yes, it is good to know we can spin the F3/4 performance story in any direction.10:36
<ajs6f>awoods: We offer quantum performance: it depends on who's looking at what.
<awoods>apparently10:37
<ajs6f>escowles: where is the new "dummy" test indexer class?
<awoods>I suspect we can do a more clear job of detailing which performance-test setups are in-play for each of the tests.
<bljenkins>Project fcrepo-kitchen-sink build #674: STILL UNSTABLE in 4 min 24 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/674/
<escowles>ajs6f: TestIndexer is in fcrepo-jms-indexer-core/src/main/java/org/fcrepo/indexer10:38
<ajs6f>hm. I guess I still haven't managed to get the PR
<bljenkins>Yippie, build fixed!10:39
Project fcrepo-jms-indexer-pluggable build #283: FIXED in 8 min 36 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-jms-indexer-pluggable/283/
* gregjansen joins
<awoods>git checkout master; git branch -D pr-xxx; git pull; git checkout -b pr-xxx origin/pr/xxx
<ajs6f>ah, finally go oit.
i just did git checkout —track origin/pr/##10:40
<awoods>that is probably easier/better
<ajs6f>it give you a branch name "pr/###" but that's fine with me.
<pivotal-bot>Benjamin Armintor started "Reduce calls to #getMixinNodeTypes in FedoraTypesUtils" https://www.pivotaltracker.com/story/show/6141139610:55
Andrew Woods delivered "Create repository.json with ServletContainerAuthenticationProvider" https://www.pivotaltracker.com/story/show/61409536
* github-ff joins
[fcrepo4] awoods pushed 2 new commits to master: http://git.io/-Grw-Q
fcrepo4/master 9a0815c Scott Prater: Sample repository.json for secured repository
fcrepo4/master c8cee87 Andrew Woods: Merge pull request #168 from sprater/fcrepo-61409536...
* github-ff leaves
<pivotal-bot>Chris Beer added "Fix o.f.k.o.SimpleObserver so it doesn't blow the heap out with large volumes of objects" https://www.pivotaltracker.com/story/show/6148261010:57
<barmintor>cbeer: are we working in the same aea of codebase?10:58
<ajs6f>omw to standup, will be a few minutes
* ajs6f leaves
<cbeer>barmintor: i'm not touching it. i was hoping to convince you to :P
<pivotal-bot>Andrew Woods edited "Fix o.f.k.o.SimpleObserver so it doesn't blow the heap out with large volumes of objects" https://www.pivotaltracker.com/story/show/6148261010:59
<awoods>standup11:01
<cbeer>awoods: link?11:02
<awoods>https://plus.google.com/hangouts/_/calendar/eW91cm1lZGlhc2hlbGYuY29tXzVlYzdpNXQ2Z282dTdidHI4aTVrbGJxOTUwQGdyb3VwLmNhbGVuZGFyLmdvb2dsZS5jb20.opn3ai06k1ge0no877ei3cku4o
https://wiki.duraspace.org/display/FF/2013-11-18+to+11-29+FF+Stand-up+Mtgs
<pivotal-bot>Chris Beer edited "Fix o.f.k.o.SimpleObserver so it doesn't blow the heap out with large volumes of objects" https://www.pivotaltracker.com/story/show/61482610
* ajs6f joins11:03
what's the new standup url?11:04
<mikeAtUVa>https://plus.google.com/hangouts/_/calendar/eW91cm1lZGlhc2hlbGYuY29tXzVlYzdpNXQ2Z282dTdidHI4aTVrbGJxOTUwQGdyb3VwLmNhbGVuZGFyLmdvb2dsZS5jb20.opn3ai06k1ge0no877ei3cku4o
<barmintor>https://plus.google.com/hangouts/_/calendar/eW91cm1lZGlhc2hlbGYuY29tXzVlYzdpNXQ2Z282dTdidHI4aTVrbGJxOTUwQGdyb3VwLmNhbGVuZGFyLmdvb2dsZS5jb20.opn3ai06k1ge0no877ei3cku4o
<escowles>summary of the tests i've run (not yet linked to profiles): https://wiki.duraspace.org/display/FF/Test+Results+Summary11:06
summary of all the tests (with some links to profiles): https://wiki.duraspace.org/display/FF/Single-Node+Performance+Testing
<mikeAtUVa>https://wiki.duraspace.org/display/FF/Versioning#Versioning-TurnofNodeVersioning11:07
<pivotal-bot>Scott Prater accepted "Create repository.json with ServletContainerAuthenticationProvider" https://www.pivotaltracker.com/story/show/6140953611:09
* travis-ci joins11:13
[travis-ci] futures/fcrepo4#1276 (master - c8cee87 : Andrew Woods): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/3ac7aab8271c...c8cee873cfb7
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/14551635
* travis-ci leaves
<bljenkins>Project fcrepo-fixity-corrupter build #495: SUCCESS in 1 min 39 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/495/11:17
<barmintor>osmandin: I think ajs6f is trying to say that there's 2 issues- 1) Any actual parse error you observe, and 2) the intermittent issues caused by waiting for statically defined amounts of clocktime when threads require unpredictable amounts of time to complete11:19
<ajs6f>what barmintor said
<cbeer>here's the weird little squiggle in fcrepo4: https://wiki.duraspace.org/download/attachments/39028088/Screen%20Shot%202013-11-26%20at%208.15.00%20.png?version=1&modificationDate=1385482769179&api=v211:20
<bljenkins>Project fcrepo-kitchen-sink build #675: STILL UNSTABLE in 3 min 59 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/675/11:21
Project fcrepo-jms-indexer-pluggable build #284: UNSTABLE in 6 min 46 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-jms-indexer-pluggable/284/11:22
<barmintor>the technique he mentions lets you handle #2 in some contexts (when you can actually be notified of thread completion), but it's a tougher nut to crack with IT (when you're waiting for requests effects to propagate)
<osmandin>barmintor: thx. my proposed solution was for #1.11:24
<barmintor>osmandin: No problem! Just making sure we're all communicating :)
cbeer: I think the lack of backtraces in that YK profile was b/c we were looking at the CPU snapshot, and not the memory snapshot11:29
(or at least I was)
<cbeer>barmintor: yeah, i think i found the right place finally
let me run this and wait for the heap to explode and i'll post a summary and the dump somewhere11:30
and you can see what you make of it.
<barmintor>but it leads me to basically the same place: SimpleObserver, JMSTopicPublisher, DefaultFilter, and MODE.JcrObservationManager
<tecoripa>escowles: would you mind taking a quick look at my authz test page at https://wiki.duraspace.org/display/FF/AuthZ+-+No+AuthZ+Fedora+4+Comparison+Performance+Testing and see what you think? Now's the time to change the profile, if you think it should match more what you're doing11:31
you'll notice that I'm dealing with quite a few more objects: 10000
and it's slooooow, at least for ingest
<barmintor>the (transaction?) session node lookups and the system node lookups account for at least 47% of the ConcurrentHashMap entries11:32
<tecoripa>17.22% done, and I started close to an hour ago
<pivotal-bot>Chris Beer added comment: "How much hand-holding did CARL have in evaluating fcrepo4?" https://www.pivotaltracker.com/story/show/61357684
Andrew Woods added comment: "Not much at all. Here is the email I sent to them... and they took it from there.11:35
------------------
Hello ..." https://www.pivotaltracker.com/story/show/61357684
Andrew Woods edited "Review CARL UI Feedback" https://www.pivotaltracker.com/story/show/61357684
<tecoripa>escowles: also, did you turn off versioning prior to running your tests?
<mikeAtUVa>tecoripa, awoods, escowles: would it simplify things to turn off versioning by default until (likely tomorrow) I add the versioning policy feature?11:39
<awoods>mikeAtUVa: It sounds like versioning:off is preferred at this point.11:40
<tecoripa>mikeAtUVa: yes, I think it might be good to do that, seeing as we're all still doing very basic performance tests11:41
and our repository profiles to date specify that versioning is off.
<mikeAtUVa>Ok, I'll do that now.11:43
<pivotal-bot>Mike Durbin added "Turn off versioning by default." https://www.pivotaltracker.com/story/show/6148689411:44
Mike Durbin started "Turn off versioning by default." https://www.pivotaltracker.com/story/show/61486894
<mikeAtUVa>So how do people set up or install fedora4 for testing? usting jetty:run? Which configuration should I update to point to an alternate node type definition that doesn't enable versioning?11:57
<awoods>on a call11:58
<pivotal-bot>Chris Beer added comment: "j.u.ConcurrentHashMap12:02
jcrNodes of o.m.j.JcrXaSession
session of o.f.k.o.DefaultFilter
..." https://www.pivotaltracker.com/story/show/61482610
<tecoripa>mikeAtUVa: we run a number of different profiles. Generally, I believe we've been running with mvn jetty:run.12:03
mikeAtUVa: You can see the different repository profiles we use at https://wiki.duraspace.org/display/FF/Test+-+Repository+Profiles+for+Performance+Testing12:04
<pivotal-bot>Chris Beer added comment: "http://cbeer.info/~chris/java_pid42187.prof (170MB)" https://www.pivotaltracker.com/story/show/6148261012:05
<escowles>tecoripa/mikeAtUVa: i didn't turn off versioning -- i think we can probably wait until tomorrow for the versioning policy feature, it seems easier to just add a parameter to turn it off today rather than have to update and rebuild fcrepo4
<mikeAtUVa>tecoripa: I think for this quick disable, I'll disable versioning in all the configurations in fcrepo-http-commons/src/main/resources/config/
<tecoripa>mikeAtUVa: If you do come up with a different file we can somehow point to at runtime (with a system property?), can you document on that wiki page, so we know how to enable/disable versioning when running the tests?12:06
<awoods>tecoripa: that is defined in the CND... which is not currently a config option12:07
<pivotal-bot>Chris Beer added comment: "The o.a.http, Finalizer, and Grizzly classes are artifacts of blowing the heap. " https://www.pivotaltracker.com/story/show/61482610
<mikeAtUVa>escowles: so I shouldn't complete https://www.pivotaltracker.com/story/show/61486894
<pivotal-bot>feature: Turn off versioning by default. (started) / owner: Mike Durbin
<fasseg>Hmmm isn't the size of a datastream displayed anymore in the web frontend? I can't size a fedora:hasSize predicate in the RDF...12:08
*can't see
<escowles>mikeAtUVa: isn't that part of the versioning policy too? i assumed the update was going to be turning versioning off by default, plus policy support for turning it on in particular cases12:09
<tecoripa>awoods: yes, mikeAtUVa and I were discussing that it's in the CND... the right thing to do would be to PUT an update to that CND, correct?
<gregjansen>tecoripa, escowles: I have a script for testing various config file sets. just polishing it a bit...
tecoripa, escowles: trying to isolate tests configs into subdirs https://gist.github.com/gregjan/766215312:12
<tecoripa>gregjansen: that's very useful. is it in a more-or-less completed state?12:14
<gregjansen>I am adding in the jgroups config file. almost done
<pivotal-bot>Esme Cowles added comment: "I have updated the wiki with more complete test results comparing Fedora 3 v. 4 on UCSD VMs: ""12:15
https://wiki.d..." https://www.pivotaltracker.com/story/show/60556948
Frank Asseg added comment: "Projection over a 4GB ramdisk works fine, and the Java heap looks stable at ~270mb on my local machine. The ..." https://www.pivotaltracker.com/story/show/6108774012:26
Frank Asseg added comment: "Gonna start a projection over a 10x30GB dataset on a local hdd next" https://www.pivotaltracker.com/story/show/6108774012:27
<fasseg>ajs6f: Is it possible that the hasSize property got lost somewhere along the way, I can't see anything about the binary size in the RDF triples returned for a datastream node...12:29
<ajs6f>fasseg: Sure, it's possible.
That would be somewhere in PropertiesRdfContext, I think.
* ermadmix leaves12:31
<osmandin>afk 3012:33
<tecoripa>bbl12:38
* tecoripa leaves
<gregjansen>fedora start script for testing: https://gist.github.com/gregjan/7662153 takes a config dir which contains repository.json, infinispan.xml and jgroups-fcrepo-tcp.xml.. Shall I add this to the benchtool project?12:40
* ermadmix joins
<awoods>gregjansen: Sounds like a good idea to me to add it to the benchtool project, but it would be better to hear from those actually using that tool.12:41
<gregjansen>I could do a PR and gather feedback. Hoping this can help us test identical and/or well documented setups on different hardware.12:42
<awoods>gregjansen: There is also another PR for the benchtool submitted by escowles. Is someone reviewing/committing that?12:44
<escowles>awoods: it's an important one for read tests (we were reading f3 ds profiles instead of ds content)12:45
<gregjansen>escowles: well that explains my 1005 mb/s read12:46
<escowles>gregjansen: yep, we all know F3's not that fast
<awoods>escowles: Does that imply that previous runs of benchtool are not exactly valid for F3?
sounds like it12:47
<gregjansen>awoods: yes
<escowles>awoods: for read tests, yes -- ingest/update/delete are unaffected
<gregjansen>awoods, escowles: happy to check this out and run a read test to verify behavior and a diff result
<awoods>gregjansen: great, then commit/merge if all is well.12:48
<gregjansen>awoods: will do
awoods, escowles: my new single node f3 read was 128.43 mb/s13:09
<awoods>gregjansen: What was your old rate?13:10
<gregjansen>awoods: 1005
<awoods>I see: 1005 mb/s
gregjansen: that puts f4 in a better light
<escowles>that sounds more reasonable -- what was your f4 read?
<gregjansen>awoods: new rate is still faster than un-tuned f4 minimal config, but perhaps we knew that, 56 mb/s13:11
awoods: that was with versioning, though probably not an impact on reads
<awoods>gregjansen: right13:12
<gregjansen>awoods: it got higher with more threads, up to 210 with 10 threads
<awoods>afk 30
<escowles>gregjansen: i did some math and i think i'm also seeing about 100 MB/s from f313:13
<awoods>escowles: pulling out the math skills!
<gregjansen>escowles: any reason not to test more threads against f3?
<escowles>awoods: i had to use pen & paper!13:14
gregjansen: because benchtool doesn't support it...
<gregjansen>escowles: ok, good answer
so escowles, shall I try to package my scripts into a bin folder in benchtool?
* ermadmix leaves
<escowles>we should probably refactor benchtool so there is one set of timing/logic and then functions to implement that actual REST calls for F3/F413:15
<gregjansen>escowles: makes sense
<escowles>gregjansen: yes please -- that should go a long way towards making our tests more consistent and comparable
<gregjansen>escowles: okay, will get a PR to you.. I am just going to add a unique stop port and key, so I don't have to call killall java.. mwahahaha13:16
<escowles>gregjansen: yes, i do killall java pretty frequently, but it could be a nasty surprise for someone...
<gregjansen>escowles: cya after lunch, bbi4513:17
<escowles>i'm re-running my ingest tests with versioning disabled, and seeing about 10% improvement (1:59 per 100 obj v. 2:11 per)
* gregjansen leaves13:18
<osmandin>afk (meeting)13:27
* ermadmix joins
<pivotal-bot>Chris Beer added comment: "Wow. Then I'd say that's great feedback if they figured out all the obvious and not-so-obvious stuff. There's..." https://www.pivotaltracker.com/story/show/6135768413:39
* ermadmix leaves14:07
<cbeer>ajs6f: it looks like the ns prefix map isn't getting created right in the streaming RDF (for HTML at least).. any immediate thoughts before I go digging?14:13
(and, writing a test for it..)14:14
<ajs6f> Just that it should be happening in NamespaceRdfContext, but if that isn't getting pulled in, the namespace map won't be there.
If you want o build some RDF for serialziation,
it's usually something like:
new NamespaceRdfContext().concat(myAwesomeRdf)14:15
to add the namespaces.
That doesn't mean I actually did it correctly. Anywhere.
I didn't just add it everywhere because you only need it if you intend to serialize.14:16
So I thought it should probably happen in the controllers, or even better, in the MessageBodyWriter.
Does that makes sense?
<cbeer>ajs6f: sure, I'll take a look. FedoraNodes#describe is a candidate for just doing it everywhere, right
?
(even though some serializations don't need namespaces)14:17
<ajs6f>If you're going to do it anywhere, that would be the place.
<cbeer>ajs6f: what do you think about adding a parameter/new signature to FedoraResource#getTriples to toggle namespaces or not?14:18
<ajs6f>Hm. Dunno.
<cbeer>ok. i'll just do it inline in #describe
and we can see
<ajs6f>I mean, we don't want to make people like you type the same thing all over...
You better do what seems right.
<escowles>speaking of what's right... can anybody help me find the right way to connect o.f.h.commons.responses.BaseHtmlProvider (where I'd like to add something to the response context) to o.f.k.services.NodeService (where I can discover appropriate predicates for a node)?14:21
<ajs6f>@Autowire?14:22
@Inject?
@SomeDICrap
<escowles>exactly -- do i just inject the nodeService into BaseHtmlProvider? Or is there some intermediate place in the REST API (FedoraNodes?) where that should happen?14:23
<ajs6f>No, BaseHmlPro is a law unto itself.14:24
<cbeer>escowles: this is for the inline editor?
<ajs6f>Inject it right into the provider.
Cause you don't know _which_controller is going to be producing results that are going to be serialzied throgh the provider.
<escowles>cbeer: yes, i can lookup the predicates in NodeService, now i"m trying to get them down to the template
<cbeer>ajs6f: hm. newNamepaceRdfContext().concat() blows away the stream's subject, it seems.. guess this won't be that easy14:25
escowles: AJAX callback to fcr:nodetypes?
<ajs6f>cbeer: Streams have a subject?
Oh, is it that topic() thing?
<escowles>cbeer: yes, i was thinking of creating a new fcr:properties to list the properties that could be applied to a node
<cbeer>ajs6f: yes, the topic() thing.14:26
<ajs6f>cbeer: That's a bug, and I bet I know where. Hold on.
<cbeer>escowles: +1. i was just too lazy to do a single-node-type endpoint.
or some kind of filtering endpoint for fcr:nodetypes
<ajs6f>(I think it's that I added topic() after I added withContext() and forgot to update the latter.)
<cbeer>ajs6f: sounds good. while you're at it, want to add the namespace context to #describe? :)14:27
<ajs6f>YES.
<cbeer>ajs6f++
<ajs6f>Just as long as you take this search stuff.
K?
<cbeer>the MQ-driven search?
<ajs6f>{sigh} Yes.
<cbeer>git rm -f * # done
all the tests pass14:28
<ajs6f>That's what I told awoods. But he's got a bug in his hair about something...
<cbeer>(he's probably on a call)
* awoods not on call
<pivotal-bot>A. "Killmation" Soroka added "Add topic to defined context of RdfStream" https://www.pivotaltracker.com/story/show/6150024814:29
A. "Killmation" Soroka started "Add topic to defined context of RdfStream" https://www.pivotaltracker.com/story/show/61500248
Chris Beer started "Review CARL UI Feedback" https://www.pivotaltracker.com/story/show/61357684
* ermadmix joins
<pivotal-bot>Chris Beer added "Update HTML AJAX call for ingesting objects to raise modal exceptions when ingest fails" https://www.pivotaltracker.com/story/show/6150033014:30
Chris Beer edited "Update HTML AJAX call for ingesting objects to raise modal exceptions when ingest fails" https://www.pivotaltracker.com/story/show/61500330
Chris Beer added "If someone tries to create a new node with an unregistered namespace in the HTML UI, capture that error and give them ..." https://www.pivotaltracker.com/story/show/6150042614:31
<ajs6f>hang on. cbeer?
Did you write new NamespaceRdfContext().concat(goodStuff)?
Because that is going to (correctly) assume you meant to take the topic of the NaRdfCon, not the concat'd stuff.14:32
<pivotal-bot>Chris Beer added "Provide some example SPARQL queries / help text for updating properties in the HTML UI" https://www.pivotaltracker.com/story/show/61500476
Chris Beer added "Detect unregistered namespace prefixes in the update properties section of the HTML UI and gently remind the user tha..." https://www.pivotaltracker.com/story/show/61500542
* gregjansen joins14:33
<cbeer>ajs6f: yes, that's what i did. i guess i could have wrapped the rdf stream in the namespace context later in the stream
<ajs6f>I don't think I offered good tools in RdfStream. I'm fixing that now.14:34
cbeer: What you can do immediately is:14:35
<pivotal-bot>Chris Beer added "Preserve uploaded file name as a node property (at least from the HTML UI; maybe more globally)" https://www.pivotaltracker.com/story/show/61500736
<ajs6f>new NamespacerdfContext.withThisContext(myNewTriples)
That will take the namespaces from the NRDFC (and as soon as I get this PR in, the topic as well) and wrap them around myNewTriples.14:36
<pivotal-bot>Andrew Woods added comment: "Currently, the UI uses dc:title to display resource titles. Do you plan on auto-populating dc:title?" https://www.pivotaltracker.com/story/show/6150073614:37
Chris Beer added "Provide the already-registered namespaces in the fcr:sparql HTML form " https://www.pivotaltracker.com/story/show/61500850
<ajs6f>Does that makes sense? withThisContext() means, "Take everything that isn't the triples themselves from this guy and wrap them around the triples I'm giving you.
<awoods>All: Since futures6 seems to be down for the count, do we have any other services that could host a Tomcat for remotely deploying kitchen-sink on successful builds?14:39
s/services/servers/
<barmintor>if a Node was removed, how does DefaultFilter know if it was a FedoraResource?14:40
* github-ff joins
[fcrepo4] ajs6f created FixRdfStream (+1 new commit): http://git.io/s8ox0Q
fcrepo4/FixRdfStream aa2e2e8 ajs6f: Adding topic() to the context for RdfStream
* github-ff leaves
<awoods>barmintor: nbanks was working the issue, I do not recall where he landed.
<pivotal-bot>A. "Killmation" Soroka added comment: "https://github.com/futures/fcrepo4/pull/171" https://www.pivotaltracker.com/story/show/61500248
* github-ff joins
[fcrepo4] ajs6f opened pull request #171: Adding topic() to the context for RdfStream (master...FixRdfStream) http://git.io/M6Gwdw
* github-ff leaves
<pivotal-bot>A. "Killmation" Soroka finished "Add topic to defined context of RdfStream" https://www.pivotaltracker.com/story/show/61500248
<ajs6f>cbeer: ^^^ That's for you.14:41
<awoods>ajs6f: why did you decide to remove the logback-test.xml?14:42
<ajs6f>I didn't. Maybe a git screw up?
Weird.
Ill add it back and it can be smushed.
* github-ff joins14:43
[fcrepo4] ajs6f pushed 1 new commit to FixRdfStream: http://git.io/g5antw
fcrepo4/FixRdfStream 2ef9343 ajs6f: Adding mysteriously disappearing logback file
* github-ff leaves
<pivotal-bot>Chris Beer added comment: "I hadn't considered it, but that's probably as good a place as any, huh? Or rdfs:label14:45
" https://www.pivotaltracker.com/story/show/61500736
<awoods>ajs6f: smushed. Are you happy cbeer? https://github.com/futures/fcrepo4/pull/171/files14:46
<ajs6f>cbeer: Happy to add other conveniences, as you discover the needs.
<pivotal-bot>Chris Beer added "Question: should we disguise UUID path intermediate nodes, and hide them (at least from the HTML UI)?" https://www.pivotaltracker.com/story/show/61501636
<cbeer>ajs6f: works for me, thanks.14:47
i'll let awoods merge it
* github-ff joins
[fcrepo4] awoods pushed 1 new commit to master: http://git.io/X-6YMA
fcrepo4/master 436ff13 ajs6f: Adding topic() to the context for RdfStream...
* github-ff leaves
* github-ff joins14:48
[fcrepo4] awoods closed pull request #171: Adding topic() to the context for RdfStream (master...FixRdfStream) http://git.io/M6Gwdw
* github-ff leaves
<ajs6f>cbeer: Do you wnat to cut me a ticket for putting the namespaces into RdfStreamProvider, so that it doesn't have to happen in any controller?
<cbeer>ajs6f: if that's a good place for it, sure.14:49
<ajs6f>Or are you working on something that could include that?
<pivotal-bot>Chris Beer added comment: "Created new tickets for actionable items, tagged with 'user-feedback'. There's some remaining comments that ..." https://www.pivotaltracker.com/story/show/61357684
<cbeer>ajs6f: nope. go for i
t
<pivotal-bot>A. "Killmation" Soroka added "put namespaces into serialized streamed RDF" https://www.pivotaltracker.com/story/show/6150190414:50
Chris Beer added comment: "I'll also summarize some of the positive feedback and open questions for fedora-tech and finish this ticket" https://www.pivotaltracker.com/story/show/61357684
* travis-ci joins14:55
[travis-ci] futures/fcrepo4#1277 (FixRdfStream - aa2e2e8 : ajs6f): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/commit/aa2e2e8824d6
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/14563588
* travis-ci leaves
* travis-ci joins14:59
[travis-ci] futures/fcrepo4#1279 (FixRdfStream - 2ef9343 : ajs6f): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/aa2e2e8824d6...2ef93433e1d8
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/14563763
* travis-ci leaves
<bljenkins>Project fcrepo-fixity-corrupter build #496: SUCCESS in 1 min 22 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/496/15:02
* travis-ci joins15:03
[travis-ci] futures/fcrepo4#1281 (master - 436ff13 : ajs6f): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/c8cee873cfb7...436ff13c431f
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/14564019
* travis-ci leaves
<ajs6f>We sure do log a good deal of loggage.15:08
<bljenkins>Yippie, build fixed!15:09
Project fcrepo-jms-indexer-pluggable build #285: FIXED in 7 min 50 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-jms-indexer-pluggable/285/
* github-ff joins15:21
[fcrepo4] ajs6f created NamespacesForStreamedRDF (+1 new commit): http://git.io/GAwEGA
fcrepo4/NamespacesForStreamedRDF 1e018b8 ajs6f: Added namespace-related triple production to JAX-RS producers
* github-ff leaves
* github-ff joins
[fcrepo4] ajs6f opened pull request #172: Added namespace-related triple production to JAX-RS producers (master...NamespacesForStreamedRDF) http://git.io/pRJiig
* github-ff leaves
<pivotal-bot>A. "Killmation" Soroka added comment: "https://github.com/futures/fcrepo4/pull/172" https://www.pivotaltracker.com/story/show/61501904
A. "Killmation" Soroka finished "put namespaces into serialized streamed RDF" https://www.pivotaltracker.com/story/show/61501904
<ajs6f>cbeer: with ^^^ merged, you shouldn't have to do anything in the controllers. The namespaces and triples that describe them will just dhow up in serialization.15:22
If they can't be serialzied as namespaces, at least there will be triples describing how they _should_ have been serialized.
<bljenkins>Project fcrepo-fixity-corrupter build #497: SUCCESS in 1 min 12 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/497/15:24
Project fcrepo-jms-indexer-pluggable build #286: UNSTABLE in 6 min 3 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-jms-indexer-pluggable/286/15:29
* tecoripa joins15:32
* travis-ci joins15:35
[travis-ci] futures/fcrepo4#1282 (NamespacesForStreamedRDF - 1e018b8 : ajs6f): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/commit/1e018b851cc9
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/14565928
* travis-ci leaves
* ermadmix leaves15:37
* tecoripa leaves15:39
<barmintor>I'll say this: avoiding a couple of long-lived sessions has resulted in a much spikier heap usage graph15:40
<ajs6f>Long lived?
<barmintor>they stay alive a long time
<ajs6f>Associated to managed beans?
<barmintor>Observers and filters15:41
<ajs6f>Right.
<barmintor>things I learned today: JcrSession keeps a ConcurrentHashMap of all the nodes it has read
<ajs6f>My guess: they have to be able to see every mutation that is committed from every (short-lived) session. That may build up in their caches.
Wait, what?! Why?15:42
<barmintor>so it doesn't have to look them up again, silly!
it's a cache!
<ajs6f>I think long lived Sessions are a very bad idea.
<barmintor>so does MODE :P15:43
<ajs6f>(Having introduced them to several of those beans myself.)
<gregjansen>note to self: get a new session
<barmintor>gregjansen++
<ajs6f>"Refresh your Sess"
See, Sess would rhyme with refresh.
See what I did there?15:44
<gregjansen>this will pin an upper limit on our transaction implementation as well
<ajs6f>You mean the lifetime of a xaction?15:45
<gregjansen>yeah
<barmintor>well, it imposes a performance penalty on very large transactions. Thems the breaks.
<ajs6f>Yeah. But then, for the same reasons, maybe that's a good thing.
Long-lived transactions create fragile architecture.
<gregjansen>better to fail faster15:46
<ajs6f>Which we should have thought of in connexion with sessions, come to think of it.
barmintor: Do you think the penalty will really be oppressive to any reasonable use case?
* gregjansen leaves15:47
<barmintor>ajs6f: unclear to me
<ajs6f>"Answer unclear. Ask again later."
<bljenkins>Project fcrepo-fixity-corrupter build #498: SUCCESS in 1 min 23 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/498/
* ksclarke leaves15:51
<bljenkins>Project fcrepo-kitchen-sink build #676: STILL UNSTABLE in 4 min 9 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/676/
* ajs6f leaves15:56
* ajs6f joins15:57
* ajs6f1 joins16:02
* ajs6f leaves
* ajs6f joins16:03
* ajs6f1 leaves
<ajs6f>escowles: ping?16:04
<escowles>ajs6f: yep?
<ajs6f>Knee-deep in the indexer stuff
<escowles>fun fun fun
<ajs6f>DId you ever see something like?
<escowles>ajs6f: ?16:08
* ermadmix joins16:09
* ajs6f leaves
<barmintor>I have never seen anything like16:10
<escowles>i don't even know what like -- it appears to have killed ajs6f mid-thought16:11
<barmintor>it is the most dangerous comparison16:12
<pivotal-bot>Osman Din added comment: "Unable to reproduce (fcrepo-authz builds fine even when F4 is running using jetty:8080). Does this happen only..." https://www.pivotaltracker.com/story/show/6115531216:15
Osman Din started "Fix Checksum Mismatch Message" https://www.pivotaltracker.com/story/show/6100798816:18
<awoods>ajs6f just keeled16:19
afk
<pivotal-bot>Esme Cowles added comment: "I have not yet found a process of looking up the predicates that can be applied to a node.16:24
I was considerin..." https://www.pivotaltracker.com/story/show/60781280
<escowles>afk # going to start some chicken soup...16:25
* ajs6f joins
escowles: network is crappy, so typing uickly
I think I foundthe root of our problems in the indexer tests. We are creating mutiiplle IndexerGroup beans16:26
we create one expliscitly
but jersey autocreates another
and which ne gets injected at any given moment is a matter of DI roulette.
I HATE the way Spring fights with every other framework to own everything.16:28
* ajs6f leaves16:29
* ajs6f joins16:31
<pivotal-bot>A. "Killmation" Soroka added comment: "Can you not just @Autowired the NodeService directly into the HTML provider?" https://www.pivotaltracker.com/story/show/6078128016:33
Osman Din added comment: "https://github.com/osmandin/fcrepo4/commit/af772ef42486eb3d8dfb0642a9439d881f09753f" https://www.pivotaltracker.com/story/show/61007988
Osman Din finished "Fix Checksum Mismatch Message" https://www.pivotaltracker.com/story/show/61007988
<escowles>ajs6f: i tried getting the list of indexers from the indexerGroup, and it was always the same as the injected on16:41
<ajs6f>Yes, but it wasn't the same _IndexerGroup_.
I checked via hashCodes.
I jus fixed it.16:42
<escowles>oh, right -- the IndexerGroup is duplicated -- why? i have never seen anything like
<ajs6f>Because we conflated the Spring contexts for the indexer stuff and the repos stuff.
<escowles>well that sort of makes sense then
<ajs6f>And because Jersey will happily autocreate anything in the packages we tell it to scan.
I fixed that, but the tests are now faiing, but for better reason.16:43
<escowles>so jersey is autogenerating an IndexerGroup and injecting it, and the repo is using the explicitly-created on in the spring config
<ajs6f>Something like that.
We need to keep a clear spearation between the wiring of what we are tesing (indexer stuff) and the wiring of what it uses (repo stuff).16:44
I cannot imagine what some poor bastard who has never seen this code but wants to make an extension module would go through to figure this out.
It's bad enough for us.
<escowles>on the editable-views front -- i'll try autowiring the nodeservice in the HtmlBaseProvider so it can lookup properties and stuff them in the velocity context16:45
mmmm... stuffing
<ajs6f>yeah. An @autowired thing is cheap, just a reference.
But this whole HTML saga makes me even more sure that we ARE building our own views framework.16:46
And we should probably adopt someone else's.
<escowles>yes, i agree that we are
<osmandin>afk
<ajs6f>barmintor: Write a views framework so we can adopt yours!
<barmintor>...
<ajs6f>DONE!
<escowles>but it's at least a somewhat-limited-scope views framework16:47
<ajs6f>escowles: That's been the hanging question since this srping.
<escowles>elipsis would actually be a good name for a views framework, i bet there's already one named that
<ajs6f>Can we keep it limited enough to avoid the large footprint of another dependency.
"Elipsis: The views framework that lets everything remain unsaid."
<escowles>http://www.infini-software.com/Encyclopedie/Developpement/Infini/Ellipse/English/Index.wp16:48
<barmintor>well, that's something:
<ajs6f>I got to run so I can bike home before the roads start to ice up. I'll bang more on this indexing stuff when I get tired of stuffing my big mouth.
With turkey.
<escowles>afk # chopping up chicken...16:49
<barmintor>LinkedBlockingQueue.take() calls were: 18k calls, 9.6M ms
went to: 18k calls, 321 ms16:50
* github-ff joins
[fcrepo-jms-indexer-pluggable] ajs6f created ajs6fThrashing (+2 new commits): http://git.io/wYcJAQ
fcrepo-jms-indexer-pluggable/ajs6fThrashing 6f53762 ajs6f: Thrashing away, on a rainy day...
fcrepo-jms-indexer-pluggable/ajs6fThrashing bde01a2 ajs6f: Thrashing all the way to sunlight
* github-ff leaves
<ajs6f>Who owns those Queue? MODE?16:51
<barmintor>YK suggests all threads CPU from 25.7M ms -> 19.9M
ajs6f: yes
that's the map underneath session
<ajs6f>barmintor++
If we can get reasonable close to bound to IO or disk, that's all we can do.16:52
We can't make anyone's pipes any fatter.
Happy Thanksgiving to anyone to whom that means anything!16:53
see y'all on the other side of a large dead bird.
* ajs6f leaves
<bljenkins>Yippie, build fixed!16:56
Project fcrepo-jms-indexer-pluggable build #288: FIXED in 5 min 51 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-jms-indexer-pluggable/288/
<pivotal-bot>Chris Beer finished "Review CARL UI Feedback" https://www.pivotaltracker.com/story/show/6135768417:08
Chris Beer estimated "Update HTML AJAX call for ingesting objects to raise modal exceptions when ingest fails" as 1 point https://www.pivotaltracker.com/story/show/61500330
Chris Beer estimated "If someone tries to create a new node with an unregistered namespace in the HTML UI, capture that error and give t..." https://www.pivotaltracker.com/story/show/61500426
Chris Beer estimated "Provide some example SPARQL queries / help text for updating properties in the HTML UI" as 1 point https://www.pivotaltracker.com/story/show/61500476
Chris Beer estimated "Detect unregistered namespace prefixes in the update properties section of the HTML UI and gently remind the user..." https://www.pivotaltracker.com/story/show/61500542
Chris Beer estimated "Preserve uploaded file name as a node property (at least from the HTML UI; maybe more globally)" as 1 point https://www.pivotaltracker.com/story/show/61500736
Chris Beer estimated "Provide the already-registered namespaces in the fcr:sparql HTML form " as 0 points https://www.pivotaltracker.com/story/show/61500850
Chris Beer estimated "Question: should we disguise UUID path intermediate nodes, and hide them (at least from the HTML UI)?" as 0 points https://www.pivotaltracker.com/story/show/61501636
* ermadmix leaves17:13
<barmintor>cbeer: is that a question of URL shortening?
<cbeer>barmintor: maybe. the CARL feedback just said they were confused by the intermediate children17:14
* mikeAtUVa leaves17:15
<cbeer>and there may be other reasons we'd like to introduce artificial hierarchy (especially for people who want to pretend fcrepo4 is just an fcrepo3 and stick a bunch of named objects off the root node)
* ermadmix joins17:16
* github-ff joins17:22
[fcrepo4] barmintor opened pull request #173: reduce node lookups in event processing (master...reduceLookups) http://git.io/Ldz6cA
* github-ff leaves
* osmandin leaves
<barmintor>Who wants to review a PR?17:23
<awoods>barmintor: who doesn't?17:24
<cbeer>barmintor++17:25
<awoods>barmintor: which ticket are we talking about?
<barmintor>awoods: it's linked in the comments for the PR
<awoods>convenient
<barmintor>I know!17:26
<cbeer>awoods: can you also review and merge http://github.com/futures/fcrepo4/pull/172
<awoods>well thank you very much, cbeer.
cbeer: You have this one on your plate: https://www.pivotaltracker.com/story/show/61446628
<pivotal-bot>feature: Custom Property IT (finished) / owner: Andrew Woods
* github-ff joins17:28
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/KYCJ7A
fcrepo4/master 1df8cea Chris Beer: Merge pull request #169 from futures/custom-properties...
* github-ff leaves
<cbeer>oops. i think i delivered the wrong ticket
<pivotal-bot>Chris Beer delivered "Custom Property IT" https://www.pivotaltracker.com/story/show/61446628
<cbeer>oh, no. i didn't save it. that's what happened17:29
i'm still surprised that actually works. not complaining, though.
<awoods>cbeer: you saw something different before?17:31
barmintor: Is the ticket associated with pr-173 finished? or is this PR an interim push? https://www.pivotaltracker.com/s/projects/684825/stories/6141139617:32
<pivotal-bot>Andrew Woods added comment: "https://github.com/futures/fcrepo4/pull/173" https://www.pivotaltracker.com/story/show/6141139617:33
Benjamin Armintor finished "Reduce calls to #getMixinNodeTypes in FedoraTypesUtils" https://www.pivotaltracker.com/story/show/61411396
<barmintor>awoods: yes.
<awoods>that answers that
<barmintor>FOR NOW.
<awoods>that is all we have
<barmintor>ok, I have to run. If no one else can, I nominate escowles to review that PR, since he's running throughput tests and will be able to crush my dreams of improvement.17:35
<awoods>barmintor: I am looking at it now...17:36
barmintor: But having escowles eyes on it as well would be great.
<escowles>barmintor/awoods: i can check it out after i'm done with my no-versioning tests
<barmintor>escowles++ // be gentle17:38
* ermadmix leaves
* fasseg leaves17:39
<pivotal-bot>Andrew Woods added "Fix fcrepo-auth-common package name" https://www.pivotaltracker.com/story/show/6151470617:40
* travis-ci joins17:43
[travis-ci] futures/fcrepo4#1285 (master - 1df8cea : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/436ff13c431f...1df8cea1ebb5
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/14572744
* travis-ci leaves
<bljenkins>Project fcrepo-fixity-corrupter build #499: SUCCESS in 1 min 40 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/499/17:59
Project fcrepo-kitchen-sink build #677: STILL UNSTABLE in 2 min 58 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/677/18:02
Project fcrepo-jms-indexer-pluggable build #289: UNSTABLE in 5 min 38 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-jms-indexer-pluggable/289/18:03
<pivotal-bot>Andrew Woods added comment: "Did you run benchtool against this PR? Does it show an improvement over the previous 9% CPU?" https://www.pivotaltracker.com/story/show/6141139618:06
Andrew Woods added comment: "Some minor comments, but otherwise looks good. @escowles?" https://www.pivotaltracker.com/story/show/6141139618:07
Andrew Woods accepted "Custom Property IT" https://www.pivotaltracker.com/story/show/6144662818:22
* github-ff joins18:46
[fcrepo4] awoods pushed 1 new commit to master: http://git.io/WBQDBA
fcrepo4/master 42f3f2d Michael Durbin: Supressed links to modify nodes on the frozen node HTML view.
* github-ff leaves
* github-ff joins18:47
[fcrepo4] awoods closed pull request #170: Supressed links to modify nodes on the frozen node HTML view. (master...versioning-3) http://git.io/fMu_FQ
* github-ff leaves
<pivotal-bot>Andrew Woods added comment: "Resolved with: https://github.com/futures/fcrepo4/commit/42f3f2d124fd2405ccd304da7a4ee40690bf397c" https://www.pivotaltracker.com/story/show/61133030
Andrew Woods delivered "Better expose nodes of type nt:frozenNode." https://www.pivotaltracker.com/story/show/61133030
<bljenkins>Project fcrepo4 build #1453: UNSTABLE in 11 min: http://ci.fcrepo.org/jenkins/job/fcrepo4/1453/18:57
awoods: Supressed links to modify nodes on the frozen node HTML view.
* travis-ci joins18:59
[travis-ci] futures/fcrepo4#1286 (master - 42f3f2d : Michael Durbin): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/1df8cea1ebb5...42f3f2d124fd
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/14576350
* travis-ci leaves
* github-ff joins19:14
[fcrepo4] awoods pushed 1 new commit to master: http://git.io/Eptn_w
fcrepo4/master a4d2ffc osmandin: Updated checksum exception message....
* github-ff leaves
<pivotal-bot>Andrew Woods added comment: "Resolved with: https://github.com/futures/fcrepo4/commit/a4d2ffc8b9ab46b93b17dc436774ae83b57d96c4" https://www.pivotaltracker.com/story/show/61007988
Andrew Woods delivered "Fix Checksum Mismatch Message" https://www.pivotaltracker.com/story/show/61007988
Chris Beer edited "Develop paging mechanism for fcrepo-http-api methods that report RDF." https://www.pivotaltracker.com/story/show/5924016019:16
Chris Beer added comment: "This paging mechanism should also apply and be exposed to the HTML serialization." https://www.pivotaltracker.com/story/show/59240160
Andrew Woods added comment: "Resolved with: https://github.com/futures/fcrepo4/commit/436ff13c431faeedd53f86de9c58858ce272fd2d" https://www.pivotaltracker.com/story/show/6150024819:23
Andrew Woods accepted "Add topic to defined context of RdfStream" https://www.pivotaltracker.com/story/show/61500248
Andrew Woods edited "Review CARL UI Feedback" https://www.pivotaltracker.com/story/show/6135768419:24
Andrew Woods delivered "Review CARL UI Feedback" https://www.pivotaltracker.com/story/show/6135768419:25
Andrew Woods accepted "Review CARL UI Feedback" https://www.pivotaltracker.com/story/show/61357684
Chris Beer added comment: "Summarized other points and discussion items to https://groups.google.com/d/msg/fedora-tech/8RDSWQeGdL4/MfDw4..." https://www.pivotaltracker.com/story/show/61357684
<bljenkins>Yippie, build fixed!19:28
Project fcrepo4 build #1454: FIXED in 14 min: http://ci.fcrepo.org/jenkins/job/fcrepo4/1454/
awoods: Updated checksum exception message.
<pivotal-bot>Andrew Woods accepted "Better expose nodes of type nt:frozenNode." https://www.pivotaltracker.com/story/show/6113303019:29
Andrew Woods accepted "Fix Checksum Mismatch Message" https://www.pivotaltracker.com/story/show/61007988
<bljenkins>Project fcrepo-fixity-corrupter build #500: SUCCESS in 1 min 1 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/500/
Project fcrepo-kitchen-sink build #678: STILL UNSTABLE in 2 min 10 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/678/19:32
<cbeer>awoods: sorry, is https://github.com/futures/fcrepo4/pull/172 on your plate or mine?
<awoods>cbeer: I am reviewing it now... but more eyes are better.19:33
<bljenkins>Yippie, build fixed!
Project fcrepo-jms-indexer-pluggable build #290: FIXED in 4 min 31 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-jms-indexer-pluggable/290/
<awoods>cbeer: Thanks for our review of the CARL feedback... I forwarded your email to Robin and Ed.19:34
s/our/your/
cbeer: Would it be useful to have a discussion of some of the points that came out of the review with more of the dev team? Or do you hope/expect that to come via responses to your email?19:35
<cbeer>awoods: i don't think it warrants a committers call agenda item.19:36
i think we should address the tickets that came out of it, and make sure to get their second impressions after the documentation effort.
<awoods>cbeer: sounds good. Robin and Ed also volunteered to help with wiki documentation.19:37
<pivotal-bot>Andrew Woods added comment: "Looks good to me. Does it solve your concern @cbeer? If so, push it." https://www.pivotaltracker.com/story/show/6150190419:45
Andrew Woods edited "put namespaces into serialized streamed RDF" https://www.pivotaltracker.com/story/show/61501904
* github-ff joins19:46
[fcrepo4] cbeer closed pull request #172: Added namespace-related triple production to JAX-RS producers (master...NamespacesForStreamedRDF) http://git.io/pRJiig
* github-ff leaves
<pivotal-bot>Chris Beer added comment: "Shipped it. I owe it some integration tests, but I'll do that in another pass." https://www.pivotaltracker.com/story/show/61501904
Chris Beer delivered "put namespaces into serialized streamed RDF" https://www.pivotaltracker.com/story/show/61501904
* github-ff joins19:47
[fcrepo4] cbeer deleted NamespacesForStreamedRDF at 1e018b8: http://git.io/Zi2usg
* github-ff leaves
* travis-ci joins19:58
[travis-ci] futures/fcrepo4#1288 (master - 9d8a5ad : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/a4d2ffc8b9ab...9d8a5ad9a920
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/14578391
* travis-ci leaves
<pivotal-bot>Andrew Woods added comment: "Thanks, @osmandin. I just doubled-checked... and you are absolutely correct. I am closing as a non-issue." https://www.pivotaltracker.com/story/show/6115531219:59
Andrew Woods accepted "Parameterize JMS port in fcrepo-authz ITs" https://www.pivotaltracker.com/story/show/6115531220:00
<bljenkins>Project fcrepo-fixity-corrupter build #501: SUCCESS in 1 min 53 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/501/20:09
* ksclarke joins20:36
<bljenkins>Project fcrepo-fixity-corrupter build #502: SUCCESS in 2 min 6 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/502/20:37
Project fcrepo-kitchen-sink build #679: STILL UNSTABLE in 5 min 27 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/679/20:43
Project fcrepo-jms-indexer-pluggable build #292: UNSTABLE in 8 min 42 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-jms-indexer-pluggable/292/20:44
* awoods leaves22:04