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

Using timezone: Eastern Standard Time
* dwilcox joins07:44
<pivotal-bot____>Andrew Woods added comment: "I am getting a build error on the PR-496 branch: ""07:56
2014-10-09 07:54:32.726:WARN:oejuc.AbstractLifeCycle:FAILED �" https://www.pivotaltracker.com/story/show/79973150
* dwilcox leaves08:01
* mohamed joins08:09
* Guest81093 leaves08:13
* dwilcox joins08:39
* awead joins08:40
<pivotal-bot____>Esme Cowles added comment: "@longshous: After testing this more thoroughly, I think this is just another wrinkle on the slowdown with a la�" https://www.pivotaltracker.com/story/show/8021389808:54
* dwilcox leaves08:56
* dwilcox joins09:00
* rtuckey joins09:02
* osmandin joins09:18
* ksclarke joins09:33
* gregjansen joins09:44
* gregjansen leaves10:03
* gregjansen joins10:09
* claussni joins10:22
* longshou joins10:36
* nikhiltri joins10:44
* nikhiltri leaves10:47
* dwilcox leaves10:50
* dwilcox joins10:53
<f4jenkins>Project fcrepo4-T2 build #17: UNSTABLE in 5 min 50 sec: http://jenkins.fcrepo.org/job/fcrepo4-T2/17/
* nikhiltri joins10:55
* acoburn joins10:58
<pivotal-bot____>Chris Beer added comment: "Yes, you need to build fcrepo-module-auth-rbacl from that PR too." https://www.pivotaltracker.com/story/show/7997315011:03
Chris Beer added comment: "(yay circular dependencies!)" https://www.pivotaltracker.com/story/show/79973150
* scossu joins11:04
* gregjansen leaves11:08
* acoburn1 joins11:12
* yc_ joins
* acoburn leaves11:14
* awead leaves
* gregjansen joins11:15
just joined the call, hey guys and gals
3,518 faculty at UNC11:28
oops, wrong channel11:32
* claussni leaves
* awead joins11:35
<pivotal-bot____>Longshou Situ added comment: "@escowles Thank you very much for your efforts to profiling it. I'll close the ticket right the way." https://www.pivotaltracker.com/story/show/8021389811:37
Longshou Situ finished "POSTing RDF to federated filesystem nodes is very slow" https://www.pivotaltracker.com/story/show/80213898
Longshou Situ started "Document available indexers" https://www.pivotaltracker.com/story/show/7652347011:38
* ajs6f joins11:40
* escowles leaves11:42
<cbeer>ajs6f: good news; Jersey 2.x providers call .dispose in a useful place (no more try { } finally { session.logout } for us)11:45
<ajs6f>cbeer: Excellent… {steeples fingers}11:46
cbeer: Is that a JAX-RS thing, or HK2?
Or something else...
<cbeer>dunno. It's the @Inject / @Provider annotations.11:47
and now to ensure anyone touching a JAX-RS resource means a nasty, nasty merge conflict.
* yc_ leaves
<ajs6f>cbeer: That's weird. @Inject/@Provider are JSR 330...
* dwilcox leaves
<ajs6f>I guess HK2 impls them, which isn't surprising.
Even Spring impls them.
* longshou leaves11:48
<ajs6f>Oh, wait. Looking at:11:49
it looks like dispose() is HK@-specific.
I wish we could get to full-on CDI. Then this would all be a bit simpler. But getting toJersey2/HK2 is a big step forward.11:50
<cbeer>ajs6f: hm. this was going swimmingly until i hit transactions.. there's a single request that shouldn't close its injected @Session.11:56
(or @Injected Session)
i think we're setting a flag somewhere in the session, and I guess i could check for it in the provider11:57
although I have the HttpServletRequest in scope too
<ajs6f>cbeer: IIRC, fasseg did that originally by dynamic-proxying the Session.
That may or may not be a good idea any more.
<cbeer>ajs6f: nope. i did it.
<ajs6f>cbeer: With a flag?11:58
<cbeer>ajs6f: the flag is so we know what session goes with what transaction id
<pivotal-bot____>Stefano Cossu added "SPARQL search for non-persisted triples" https://www.pivotaltracker.com/story/show/80389314
<ajs6f>cbeer: So it's a txid, and we check that tx before allowing save or logout?
* dwilcox joins12:04
<cbeer>ajs6f: this is only on the first request to create a transaction; after that, the injected session is wrapped in our proxy that has that logic12:20
ajs6f: but to create the transaction in the first place, the injected session is just a plain session.
<ajs6f>cbeer: So… wouldn't that pattern continue to work with HK2 injection?12:22
<cbeer>ajs6f: injection yes, but not the dispose. when setting up a transaction, it injected a plain session, it tries to dispose the plain session... i'm just adding some logic to check if it has that namespace prefix hack set up12:23
* acoburn1 leaves12:24
<ajs6f>cbeer: Could the proxy InvocationHandler do the right thing on dispose()?
<cbeer>ajs6f: we're not in a proxy at that point. unless you're suggesting we always proxy our injected sessions?
<ajs6f>cbeer: If it's transactional, aren't we in a proxy? Or is dispose somehow called after the proxy "collapses"?12:25
is dispose called after save/logout?12:26
<cbeer>this is the method: https://github.com/fcrepo4/fcrepo4/blob/master/fcrepo-http-api/src/main/java/org/fcrepo/http/api/FedoraTransactions.java#L73
<ajs6f>Or just save, rather.
<cbeer>we're not in a transaction for that method
<ajs6f>Oh, this is about not disposing the Session that has just become part of a tx?
<ajs6f>Even though we sally want to dispose it at the end of a controller method.
Not Sally. Perfectly nice person though I'm sure she is.12:27
<cbeer>i'll push what i've done once the tests finish, maybe you'll have a better idea
<ajs6f>Can the injector (HK2) take any advice in this situation? Could we set that flag in the session and let the dispose method examine it to determine wether to logout or not?
* github-ff joins12:30
[fcrepo4] cbeer pushed 1 new commit to ldp-resource: http://git.io/x-wYRw
fcrepo4/ldp-resource e96995d Chris Beer: Use SessionProvider#dispose to logout injected sessions...
* github-ff leaves
is the best I have.
only 700 more lines to go before this PR breaks 10000 lines!12:31
* longshou joins12:32
<ajs6f>cbeer: I probably can't look at that before tomorrow. I'm not blocking you, am I?12:36
<cbeer>ajs6f: nope, awoods is.
* travis-ci joins12:44
fcrepo4/fcrepo4#2551 (ldp-resource - e96995d : Chris Beer): The build has errored.
Change view : https://github.com/fcrepo4/fcrepo4/compare/4f36eaa439eb...e96995d8ff07
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/37518566
* travis-ci leaves
* acoburn joins12:52
* dwilcox leaves13:09
* gregjansen leaves13:35
* acoburn leaves13:49
* nikhiltri leaves13:58
* gregjansen joins14:12
<pivotal-bot____>Chris Beer unstarted "Support "soft" deletes" https://www.pivotaltracker.com/story/show/63458666
* osmandin leaves14:13
* osmandin joins14:22
* nikhiltri joins14:23
* gregjansen leaves14:24
* dwilcox joins14:29
* gregjansen joins14:33
* dwilcox leaves14:36
* dwilcox joins14:42
* escowles joins14:46
<ksclarke>osmandin available for a sprint wrap up?15:01
<pivotal-bot____>Esme Cowles added comment: "https://github.com/fcrepo4/ontology/pull/7" https://www.pivotaltracker.com/story/show/7953401615:02
Esme Cowles finished "OAI-Provider predicates" https://www.pivotaltracker.com/story/show/79534016
<osmandin>ksclarke: isn't it tomorrow?
right now, sorry
<osmandin>oh...i'll be there in a minute or two
<pivotal-bot____>Esme Cowles unstarted "Large number of files in Federation failures" https://www.pivotaltracker.com/story/show/7965211815:11
* dwilcox leaves15:24
* ajs6f leaves15:25
* acoburn joins15:26
* ajs6f joins15:32
* github-ff joins15:36
[fcrepo4] osmandin opened pull request #506: Add remove namespace button (master...79359308) http://git.io/wxH5pA
* github-ff leaves
* scossu leaves15:39
<pivotal-bot____>Osman Din added comment: "https://github.com/fcrepo4/fcrepo4/pull/50615:40
having a drop down might another approach, but seems it'd be easier�" https://www.pivotaltracker.com/story/show/79359308
Osman Din finished "Add "Remove Namespace" to UI" https://www.pivotaltracker.com/story/show/79359308
* dwilcox joins15:42
* scossu joins15:43
* dwilcox leaves15:46
* scossu1 joins15:49
* scossu1 leaves
* scossu1 joins15:50
* scossu leaves15:51
<scossu1>Did someone take notes for today's committers' meeting?16:02
* dwilcox joins16:03
<ksclarke>scossu1 esme did; may not be up yet though
<scossu1>ksclarke: Thanks.16:05
<escowles>scossu1: yes, i did take notes, but haven't posted them yet -- will post them later this evening or tomorrow morning16:06
<scossu1>escowles: no problem. I just wanted to add some comments about ordering when they're up.16:07
* ajs6f1 joins16:09
* ajs6f leaves
* dwilcox leaves16:18
* osmandin leaves16:45
* github-ff joins16:50
[fcrepo4] cbeer pushed 1 new commit to ldp-resource: http://git.io/sNueXA
fcrepo4/ldp-resource 6a47770 Chris Beer: When persisting a graph, use exceptions for invalid operations...
* github-ff leaves
* acoburn leaves16:52
* gregjansen leaves16:54
* travis-ci joins17:01
fcrepo4/fcrepo4#2553 (ldp-resource - 6a47770 : Chris Beer): The build has errored.
Change view : https://github.com/fcrepo4/fcrepo4/compare/e96995d8ff07...6a47770bdf28
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/37541330
* travis-ci leaves
* ajs6f1 leaves17:05
* scossu1 leaves17:39
* github-ff joins17:40
[fcrepo4] cbeer pushed 1 new commit to ldp-resource: https://github.com/fcrepo4/fcrepo4/commit/2e0e3cd5919f8d6c955ff6e97721adb4b9cf5e89
fcrepo4/ldp-resource 2e0e3cd Chris Beer: When PersistingRdfStreamConsumer#consume fails, wrap any exception in a MalformedRdfException
* github-ff leaves
<cbeer>escowles: ping?17:45
* travis-ci joins17:49
fcrepo4/fcrepo4#2555 (ldp-resource - 2e0e3cd : Chris Beer): The build has errored.
Change view : https://github.com/fcrepo4/fcrepo4/compare/6a47770bdf28...2e0e3cd5919f
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/37545657
* travis-ci leaves
* ksclarke leaves18:18
* ruebot_away leaves18:25
* ruebot joins18:26
* ruebot leaves
* ruebot joins
* nikhiltri leaves18:36
* nikhiltri joins18:37
* nikhiltri leaves18:38
<escowles>cbeer: pong?
<cbeer>escowles: i've noticed the fcrepo-http-api (and probably others) are using a no-op slf4j logger18:40
which is unfortunate for any logging that happens in fcrepo-*18:41
i'd guess we used to get logback as a transitive dependency of something somewhere.
i think we should make logback a test-jar dependency in -kernel-impl, maybe?18:42
* scossu joins18:46
<escowles>i don't know why we are using the no-op slf4j logger -- is there a reason why we can't just use logback directly in fcrepo-http-api?18:47
<cbeer>escowles: well, we're using the no-op slf4j logger because we haven't provided an impl (which i assume we were getting transitively before..)18:48
and i don't know enough about java logging to know what the right thing to do is
<escowles>cbeer: i hate java logging with a fiery passion and usually just try to make it stop -- maybe awoods or somebody else knows what we should do?18:50
* github-ff joins
[fcrepo4] cbeer pushed 1 new commit to ldp-resource: http://git.io/51q_eg
fcrepo4/ldp-resource 633f72c Chris Beer: Update error conditions when content-type (and/or an empty body) are provided
* github-ff leaves
<escowles>not that they're on irc at the moment...
<cbeer>escowles: yeah, maybe i should ping ajs6f on an issue somewhere.
i'm sure he's highly opinionated.
<escowles>ajs6f is reliably opinionated
* travis-ci joins19:03
fcrepo4/fcrepo4#2557 (ldp-resource - 633f72c : Chris Beer): The build has errored.
Change view : https://github.com/fcrepo4/fcrepo4/compare/2e0e3cd5919f...633f72ce48b7
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/37551379
* travis-ci leaves
* scossu leaves19:12
* scossu joins20:00
* scossu leaves
* awead leaves
* awead joins20:02
* awead leaves20:03
* longshou leaves20:16
* ksclarke joins21:24
* github-ff joins21:50
[fcrepo4] cbeer created ldp-resource-hash (+1 new commit): http://git.io/RecQvQ
fcrepo4/ldp-resource-hash 4e83a74 Chris Beer: Support Hash URIs within resources
* github-ff leaves
* github-ff joins21:52
[fcrepo4] cbeer opened pull request #508: Support Hash URIs within resources (ldp-resource...ldp-resource-hash) http://git.io/q6Xglw
* github-ff leaves
* travis-ci joins21:54
fcrepo4/fcrepo4#2558 (ldp-resource-hash - 4e83a74 : Chris Beer): The build has errored.
Change view : https://github.com/fcrepo4/fcrepo4/commit/4e83a74dbd60
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/37559985
* travis-ci leaves
* scossu joins22:29
* scossu leaves22:45
* scossu joins22:46
* ksclarke leaves23:16

Generated by Sualtam